Thesis and Dissertation Abstracts 1999-2003 (last update: 04/30/2004) For any information, please contact bib-di@inf.puc-rio.br 1999 99_MSc_cusatisjunior Affonso de CUSATIS JUNIOR. Raycasting intervalar de superficies implicitas com aritmetica afim. M.Sc. Diss. Port. Presentation: 23/04/99 72 p. Advisors: Luiz Henrique de Figueiredo (LNCC) and Marcelo Gattass Abstract: This work investigates ray casting - the most natural visualization technique for implicit surfaces - using interval methods. Some robust algorithms are implemented for ray casting generic surfaces, using interval methods and conventional numerical methods in the determination of ray-surface intersections - a central operation in ray casting - and a new interval calculation model is tested: affine arithmetic (AA), proposed as a alternative for standard interval arthmetic (IA). AA was designed in order to avoid the IA's error explosion problem at long interval calculation sequences, and takes into account correlations between terms in a interval expression. So, AA defines more expensive operations but yields more precise results, and this feature may accelerate several interval algorithms. 99_MSc_ferreira Alexandre Guimaraes FERREIRA. Uma arquitetura para visualizacao distribuida de ambientes virtuais. M.Sc. Diss. Port. Presentation: 22/12/99 83 p. Advisor: Marcelo Gattass Abstract: A number of applications require a visualization system that provides multiple visions of an animated virtual environment. The disposal of multiple visualization devices around a user consists in a natural solution to compose a panoramic view of the scene. Virtual environment panoramic views, simultaneously supplied to different users, can be obtained from different points of view, in a more sophisticated arrangement. Each panoramic scene presented to a user is formed by the composition of images generated in each visualization surface. This way, it is possible to imagine each visualization device as a window to the virtual environment. The use of this technology seeks to enhance human perception and to increase user's immersion. This is reached through a quantitative increase of the visual stimuli provided, due to the use of multiple windows. A traditional approach for the development of such system is based on specialized systems that use dedicated hardware, which is responsible for the visualization devices' centralized processing and control. This work proposes a distributed architecture to be used in heterogeneous networks with low-cost graphic workstations, seeking to provide greater scalability, portability and flexibility to the visualization system. The proposed architecture seeks to support the development of applications that present a consistent visual outcome for the virtual environment. To accomplish that, techniques are presented to assure synchronism and integrity among several workstations. 99_MSc_silva Alexandre Reis e SILVA. Infra-estruturas de chaves publicas - caminhos de certificacao e repositorios distribuidos. M.Sc. Diss. Port. Presentation: 12/03/99 114 p. Advisor: Michael Anthony Stanton Abstract: The primary goal of this dissertation is to present a practical study of aspects related to the security applications of telecooperation in geographically distributed intra and intercorporate environments. Issues of public key distribution are studied, and we discuss the use of digital certificates, issued by certification authorities, to distribute keys and their storage in distributed repositories. A comparative study of public key infrastructure standards and implementation is presented and special emphasis is given to certification path processing methods. This study results in a new certificate path processing method, and to show its viability a prototype is implemented, allowing telecooperation applications to be interoperable, even in large corporate environments. This tool accesses distributed repositories in order to retrieve certificates, to perform automatic certificate path construction in a simple and dynamic way, and to validate the resulting paths. 99_PhD_dias Ana Isabel de Azevedo Spinola DIAS. Sistemas reativos: uma abordagem geometrica. Ph.D. Thesis. Port. Presentation: 06/05/99 132 p. Advisor: Edward Hermann Haeusler Abstract: This thesis describes a way to model reactive systems that is based on Sheaf Theory. The starting point for the construction of this model is the interpretation of a reactive system as a transition system with special characteristics, which is associated to a semi-sheaf - an object of a category defined in this dissertation. This category is a topos, and this fact is proved in detail. The categorial relation of semi-sheaves with sheaves and with pre-sheaves is discussed. The notion of bisimulation in the category of semi-sheaves is analyzed 99_MSc_silva Anderson Oliveira da SILVA. Modelo para o desenvolvimento de ferramentas para gerencia de redes combinando SNMP, HTTP, HTML e Java. M.Sc. Diss. Port. Presentation: 23/04/99 71 p. Advisor: Michael Anthony Stanton Abstract: As computer networks continuously grow, automatic network management seems to be required in order to guarantee confiability and performance. SNMP (Simple Network Management Protocol) is the principal protocol for TCP/IP network management, and version 1 is a standard, established and used in the great majority of manageable network equipment. WWW (World Wide Web) and the established user interface of the browsers bring together many Internet services in the same tool of use to network administrators. This dissertation introduces a model for developing network management tools combining SNMP, HTTP, HTML and Java, in order to make tools available to facilitate the management of network equipment from different vendors. This is achieved by combining management information from different sources in order to obtain results of significance to network administrators, as well as using uniform interfaces between different systems and running multiplatform. 99_MSc_bastos Andre Renato Gualter BASTOS. Comunicacao de voz sobre ATM: utilizacao na rede corporativa ATM da Petrobras. M.Sc. Diss. Port. Presentation: 27/04/99 133 p. Advisor: Luiz Fernando Gomes Soares Abstract: Recently, it has been having a strong movement towards the integration of the network technologies, like voice, data and video, in only one network infrastructure. Looking for this target, many companies have decided to use ATM at their network backbones. Petrobras, in the same way, has been installing his new ATM WAN backbone since 1998. In this work we are presenting a migration proposal of the voice service, from the Petrobras' telephony network to a new ATM WAN network. Besides, we also developed a new service, the Desktop Voice over ATM, to provide voice communication between ATM workstations. 99_PhD_ciarlini Angelo Ernani Maia CIARLINI. Geracao interativa de enredos. Ph.D. Thesis. Port. Presentation: 31/04/99 167 p. Advisor: Antonio Luz Furtado Abstract: Within a narrative, be it fiction or a description of a real situation, there are characters (agents) that compete and collaborate with each other to achieve their goals. Literary narratives, as well as real situations from specific domains, often present sets of typical events and usual patterns in the occurrence of such events. Morever, it is possible to identify relations between the states of each character during the narratives and the progressive appearence of goals. This thesis intends to study the use of Artificial Intelligence and Database techniques for generating plots of narratives. Such use is relevant for the study of literary genres, the creation of new stories and the simulation - both in the real and fictitious worlds - of the behaviour of agents who interact with each other in trying to achieve their goals. We describe an environment created for using behavioural patterns, extrated either from past actions (in case of simulations for helping decision making) or from typical narratives of a literary genre. We use two main techniques for obtaining plots: pattern recognition and simulation. In the first case, we use plan recognition for recognizing typical plots. In the second case, we use a temporal logic both for goal inference and for planning the interaction between agents. Our approach is interactive not only because we take into account the interaction between agents, but also because of the interaction between the user and the context defined by the behavioural patterns. A prototype, implemented in Prolog and using Constraint Programming, is operational. 99_MSc_hester Anna Magdalena HESTER. A ferramenta CGILua em multiplas APIs com o servidor HTTP. M.Sc. Diss. Port. Presentation: 16/07/99 76 p. Advisor: Roberto Ierusalimschy Abstract: This work presents the functioning and design of the tool CGLua is a server-side web tool that dinamically creates web pages from scripts written in the programming language Lua. The tool accommodates different forms of interactions with the WWW server - the multiples APIs - without creating incompabilities with the scripts. The objective of this text is to describe how the software's internal architecture was designed to fit this purpose. The main benefits are a better performance and support to the storage of data based on sessions, through the use of a generic mechanism of data sharing. 99_MSc_albuquerque Antonia Lucinelma Pessoa ALBUQUERQUE. Cenarios virtuais com um estudo de sincronismo de camera. M.Sc. Diss. Port. Presentation: 06/04/99 89 p. Advisors: Jonas de Miranda Gomes (IMPA) and Marcelo Gattass Abstract: Techniques of filming using special effects have existed since the 1920's, long before the advent of computers. Two of them are known as Back Projection - when an actor acts in front of a screen that reproduces other footage (very common in train scenes), and Blue Screen - when an actor acts in front of a blue wall for later composition with another scene. However, it was computer graphics and the computer's advance that made possible great evolution in this area. This work approaches Virtual Sets, describing its conceptualization, and showing its correlation with other areas in computer graphics. The virtual sets' pertinent technologies are identified in computer graphics and have their given solutions and unsolved problems argued. Amongst the presented problems, the work studies one technique based on differential otimization aiming at the synchronism of camera that would allow interaction, in real time, of the real and synthetic images. 99_MSc_gomes Antonio Tadeu Azevedo GOMES. Um framework para provisao de QoS em ambientes genericos de processamento e comunicacao. M.Sc. Diss. Port. Presentation: 31/05/99 162 p. Advisor: Luiz Fernando Gomes Soares Abstract: Over the past few years, an emerging demand for multimedia communication and processing systems has been observed. On these systems, the notion of quality of service (QoS) has evolved either as a way for user requirements specification, or as a reference for the configuration, operation and maintenance of the resource sharing and orchestrating mechanisms. This thesis is motivated by the belief that architectures for the provision of specific services are not flexible enough to support new media treatment technics. As an alternative, the thesis presents a framework that organizes QoS provision in a domain which is sufficiently generic to include communication and processing environment, encompassing from specific subsystems to distributed multimedia systems. With these characteristics, the framework allows the construction of configurable systems, with regard to the introduction of new services, and eases the implementation of resources sharing and orchestrating mechanisms. 99_MSc_pessoa Artur Alves PESSOA. Construcao eficiente de codigos livres de prefixo. M.Sc. Diss. Port. Presentation: 27/08/99 159 p. Advisor: Ruy Luiz Milidiu Abstract: Some of the most efficient methods for representing digital information are based on prefix-free coding. Given a sorted list of weights W=[w1,...,wn], an optimal code is defined to be a prefix-free code that minimizes the weighted sum of its codeword lengths. In this case, the well-known Huffman's algorithm constructs an optimal code for W in O(n) time and using O(n) space. In addition, this code calculation can be done in-place, that is, using only O(1) extra and overwriting the input weights with the output codeword lengths. Let H be the length of the longest codeword in a given prefix-free code. An optimal length-restricted prefix-free code also minimizes the weight sum of its codeword lengths with the additional restriction that we must have H< L for a given value of L. Two approximative algorithms for constructing length-restricted prefix-free codes are the WARM-UP algorithm and the BRCI algorithm. In this work, we present the following new results: an improved in-place implementation for the Huffman's algorithm; the Lazy Traversal algorithm, introducing a new space-economical approach for constructing optimal codes without destroying the input weights; four algorithms based on this new approach: the F-LazyHuff, E-LazyHuff, Runlength F-LazyHuff and B-LazyHuff algorithms, where last one runs in O(n) time, in the worst cas, and requires only O (H) extra space; the homogenization technique: a new technique that can be used to improve the running time of algorithms for constructing prefix-free codes; two efficient implementations for the WARM-UP algorithm, that are based on the inplace Huffman's algorithm and the Runlength F-LazyHuff algorithm, respectively; an in-place implementation for the BRCI algorithm. We also report experimental results comparing the performance of some poposed methods with that of some implementations found in the literature. 99_MSc_carvalho Carla Andrade de Almeida CARVALHO. Um ambiente para apoio ao trabalho em grupo na WWW. M.Sc. Diss. Port. Presentation: 15/09/99 140 p. Advisor: Daniel Schwabe Abstract: This work presents a conceptual environment architecture for supporting workgroups in WWW. This architecture includes groups, workflow, agenda, documents repository services and their interactions, and uses a group-based authorization model to control the access to the resources, reflecting formal and informal organization structures. 99_MSc_mendes Carlos Augusto Teixeira MENDES. DLua: uma ferramenta para desenvolvimento de aplicacoes distribuidas. M.Sc. Diss. Port. Presentation: 12/11/99 132 p. Advisor: Noemi de La Rocque Rodriguez Abstract: DLua is an extension to the Lua language which provides support for developing distributed applications. It's main haracteristic is supporting multiple communications paradigms. The supported paradigms are: RPC, process creation, rendezvous and message passing. This work presents a description of DLua, along with results obtained by a comparison between DLua and C++ implementations of a set of distributed algorithms, covering their performance and expressiveness. 99_MSc_pinto Daniel de Ariosto PINTO. ND: uma linguagem orientada a objetos para processamento paralelo em ambientes de redes de estacoes de trabalho. M.Sc. Diss. Port. Presentation: 24/02/99 145 p. Advisor: Jose Lucas Mourao Rangel Netto Abstract: ND is a totally object oriented parallel language with a virtual shared memory, designed to run in loosely coupled environments and in the SPMD programming model. From the semantical point of view, groups constitute the basic language concept and it is expected they play in a parallel programming environment a role similar to objects in sequential programming. In ND , groups form the basic encapsulation scheme of parallel operations, requiring careful design of classes, given that object groups abstract of the whole processing instead of any single object (class). ND programs are viewed as applications, each one made of one or more units,like in Object Pascal. In the run time, application groups can be created, each one having at each node of the underlying logical network architecture one or more independent copies. Those groups can be used to create object and process groups as desired. Inheritance is considered an essentially sequential feature and conflicts arisen by the inheritance anomaly must be explicitly handled by the programmer. Semaphones and guards are the mechanisms proposed to solve inheritance anomaly cases, guards can be coded at the level of classes and/or methods. Although ND programming style suggests a shared memory environment, the proposed language includes some machanisms for message switching. One of these features is the queue, a feature similar to the homonymous Hybrid facility. 99_PhD_oliveira Denise Aboim Sande e OLIVEIRA. Uma abordagem semiotica para a representacao de conhecimento. Ph.D. Thesis. Port. Presentation: 20/05/99 p. Port. Advisor: Clarisse Sieckenius de Souza Note: Not available. 99_MSc_silva Diva de Souza e SILVA. Uma arquitetura de sistemas de data warehouse usando Heros: um sistema de gerencia de banco de dados heterogeneos. M.Sc.Diss. Port. Presentation: 29/09/99 104 p. Advisor: Rubens Nascimento Melo Abstract: In the development process of Data Warehouse Systems, capturing data is perhaps the most important stage in the process. Data stored in the operational basis of a company and in external sources must be integrated into the Data Warehouse structure, creating a database for applications which will support the company's decision making.The main problem is that data are typically stored in hardware and/or software heterogeneous environments. This work proposes an Architecture for the creation and maintenance of Data Warehouse systems, using a Heterogenous Database Manager System - the HEROS, as a middleware to exploit data from operational systems and external sources in a company. The utilization of a HDBMS as the integrator of the operational bases, translating local data schemes and integrating them into global scheme contributes, in a very positive way, towards the solution of a great problem of Data Warehouse. 99_PhD_laber Eduardo Sany LABER. Codigos de prefixos: algoritmos e cotas. Ph.D.Thesis. Port. Presentation: 08/07/99 117 p. Advisor: Ruy Luiz Milidiu Abstract: The prefix codes play an important role in data compression and data communication. These codes also present relation with search problems. In this thesis, we present new structural and algorithmic results concerning the prefix code class. We theoretically explain results related to the high compression rates of some methods that have been used for practical purposes. We also propose efficient algorithms for constructing optimal prefix codes and some variants. The major results are listed below: a new parallel algorithm for constructing optimal prefix codes; a sharp upper bound for the compression loss introduced due the usage of length restricted prefix codes; an upper bound for the compression loss introduced due the usage of length restricted alphabetic prefix codes; an O (n log n) time approximative algorithm for constructing length restricted prefix code; a O (n log n) time approximative algorithm for constructing length restricted alphabetic prefix code; a strongly polinomial version for the WARM-UP algorithm; a linear time algorithm for recognizing optimal length restricted prefix codes; a proof for Vitter's conjecture about the performance of the Dynamic Huffman Codes constructed by FGK (Faller, Gallenger and Knuth) algorithm. 99_PhD_uchoa Elvira Maria Antunes UCHOA. Framework para integracao de sistemas de bancos de dados heterogeneos. Ph.D.Thesis. Port. Presentation: 11/02/99 241 p. Advisor: Rubens Nascimento Melo Abstract: Heterogeneous database systems have been considered one of the most feasible solutions for the integration of existing systems, both autonomous and different, without the need of alterations. A heterogeneous database system must be able to integrate new types of component database systems, new hardware and software platforms, and new forms of communication, without having to restructure its existing implementation code and design. The framework technique provides the adequate infrastructure to meet this requirement, because a framework is a software (sub)system of a particular domain that can be tailored for individual applications. A framework consists of a large structure that can be reused as a whole for the construction of a new system, which represents, in the heterogeneous database system context, each new federation. This work proposes a framework metamodel, which formally represents in Z specification language, the main concepts relative to the framework technique. Based on this knowledge and the experience acquired in the developing of HEROS - Heterogeneous Database Management System, currently under research at Departamento de Informatica - PUC-Rio, this work presents a framework for integrating heterogeneous database systems and shows how it may be used through a case study. 99_MSc_coelho Enylton Carlos Faria Machado COELHO. Visualizacao interativa de terrenos em uma arquitetura cliente-servidor. M.Sc.Diss. Port. Presentation: 07/07/99 52 p. Advisor: Marcelo Gattass Abstract: Interactive terrain visualization is useful for various applications such as Geographical Information Systems, military training, city planning and tourist information systems. In most of these cases, it is natural to use a Client-Server architecture, in which the data, usually very large, is stored in one or more servers, that respond to requests made by client programs running on the user's machines. This work addresses the various techniques - data storage schemes, network communication, data visualization and interaction control - employed in such systems. Based on this discussion, an application to allow visualization of large terrain data over a local network was developed. This system adopts a multi-resolution representation for the terrain data. This allows fast interaction between user and application, regardless of the size of the data, due to the use of algorithms which select the appropriate level of detail for rendering the scene. System implementation issues are discussed in detail, and performance is assessed. 99_MSc_gonzalezclua Esteban Walter GONZALEZ CLUA. Modelagem procedimental para visualizacao de elementos da natureza. M.Sc.Diss. Port. Presentation: 14/12/99 96 p. Advisor: Marcelo Gattass Abstract: Modeling natural phenomena is an area of active research and growing importance in Computer Graphics. The purposes of this work are to present a thorough bibliography about this subject, to cover fundamental concepts and to propose some new techniques. It focuses mainly on the procedural modeling techniques but it also present specific visualization and ilumination methods. 99_MSc_ silva Fernando Balthar Pereira da SILVA. Acesso as redes ATM baseado em quadros: um estudo do servico da dados da Petrobras. M.Sc.Diss. Port. Presentation: 21/05/99 162 p. Advisor: Luiz Fernando Gomes Soares Abstract: There has been a growing interest in frame based access to ATM WAN networks, both at standardization departments and at manufacturer's forums, as an option for cell based access. This work makes a comparative analysis of proposed technologies for frame access. It stresses aspects and interpretations not covered in the specialized literature and clarifies the advantages and disadvantages of each technology. Based on this study, a proposal for the Petrobras new ATM WAN network is presented. As an additional contribution, this work also presents an implementation of the frame relay service specific convergence sublayer for application servers with ATM network interface card. This implementation will allow frame relay users to access these servers. 99_PhD_varejao Flavio Miguel VAREJAO. DORPA: uma ontologia de design que integra requisitos, processos e artefatos. Ph.D.Thesis. Port. Presentation: 26/04/99 191 p. Advisor: Ana Cristina Bicharra Garcia (UFF) and Clarisse Sieckenius de Souza Abstract: Design has been one of the most frequent themes in Artificial Intelligence research because of its complex and constructive nature. However, different perspectives on design knowledge representation have adopted by the different approaches. As a consequence, it is hard to reuse knowledge from different approaches in the development of knowledge-based design systems. This type of problem has motivated a growing interest on ontologies in knowledge engineering. A design ontology identifies and defines the concepts and relations frequently present in the design problems. The ontology concepts and definitions may be reused by the different approaches, reducing the costs of knowledge-based systems development. This thesis presents a design ontology that integrates requirements, artifacts and process design knowledge. Requirement knowledge describes the desired features of the design product and process. Artifact knowledge describes the components, the design product and their parts. Process knowledge describes the different strategies used by the designers during the design process. The broad view provided by the ontology reveals important aspects to be considered for those interested in acquiring, representing and sharing design knowledge. 99_MSc_silva Geiza Hamazaki da SILVA. Um estudo em sintese construtiva de programas utilizando logica intuicionista. M.Sc.Diss. Port. Presentation: 05/11/99 160 p. Advisor: Edward Hermann Haeusler Abstract: A major issue in computer science is to assure that a program implementation satisfies a formal specification. In order to deal with such problem several methods have been proposed. In this work we show a method, named Constructive Programs Synthesis, which enables us to obtain a program from its specification. Thus, such a program would be correct by construction. The process of synthesis tries to construct a program in an imperative language from its specification, which is used to produce a proof by an Intuitionistic First-Order Theorem Prover using Natural Deduction as deductive system. Once we get the proof, we can extract the program by the association of each inference rule with commands in the imperative program language. Afterwards, we show that the generated program obeys the specification. 99_MSc_gomes Georgia Regina Rodrigues GOMES. Um ambiente para integracao de dados bibliograficos baseado em mediadores. M.Sc.Diss. Port. Presentation: 29/09/99 100 p. Advisor: Rubens Nascimento Melo Abstract: Library systems are an important instrument on all levels of research, whether it be scientific or not. It is becoming more and more necessary to integrate these systems and to make it easier for users to search bibliographical references and in many cases the work itself. Integrating, making available and visualizing data is an area of quite intense research in data banks. This dissertation proposes a software architecture capable of integrating library systems that are different, heterogeneous and widely distributed in their software and hardware environments by offering users a transparent, uniform view of all the libraries that make up the integration model. Our architecture is based on the technology of Mediators and Wrappaers, the former being in charge of managing and integrating the bibliographical data stored in each system and the latter responsible for retrieving the data requested at each data source. 99_MSc_moura Isabela Cordeiro Ribeiro MOURA. Um ambiente para o suporte ao projeto e implementacao de sistemas de informacao baseados na WWW. M.Sc.Diss. Port. Presentation: 30/06/99 272 p. Advisor: Daniel Schwabe Abstract: This paper presents an environment that supports the development of WWW applications designed with OOHDM. This environment allows direct mapping of navigation and interface constructs of OOHDM into a library of functions in the CGI scripting environment CGLua, extended with the DBLua package. Using this environment, hypermedia applications are implemented, as CGI scripts that produce dynamically generated pages, whose contents are fed from a database and integrated with pre-defined templates. The environment also allows the "compilation" of an application, creating an off-line version of the application, suing static pages instead of dynamically generated pages. 99_PhD_duran Juan Eduardo DURAN. Desenvolvimento transformacional de algoritmos em redes baseados em algebras relacionais. Ph.D.Thesis Port. Presentation: 20/10/99 137 p. Advisor: Armando Martin Haeberer Abstract: In this thesis an approach for the transformational development of imperative network algorithms is presented which mainly bases on relations and their algebraic properties. It decomposes a development into three phases, viz. an optimization specification derivation, an abstract recursive algorithm development, and, finally, a concrete algorithm calculation using refinement and datatype representation. Optimization specifications have an explicite form and base mainly on the modeling of datatypes within relational algebra. For the derivation of abstract algorithm existing methods are adapted and extended by a derivation of feasible object set algorithms from an abstract object set specifications. Finally, for the derivation of a concrete algorithm that computes only one optimal solution the refinement of optimization problem specifications, of relational terms, and of abstract algorithms is applied. These refinement steps are combined with transition from relational to functional notation and a suitable representation of datatypes. 99_MSc_fortunato Luidi Xavier FORTUNATO. 2BUYNET - Um framework para instanciacao e administracao de lojas para a internet. M.Sc.Diss. Port. Presentation: 28/07/99 104 p. Advisor : Carlos Jose Pereira de Lucena Abstract: This dissertation presents a software development approach for the construction of electronic commerce websites. At first, problems and feature related to the domain application are discussed. The dissertation also presents some existing software tools for webstore creation, their proposed solutions and the approach they use to impement their features. Based on these comparisons we have developed a framework called 2BuyNet. The framework constitutes an origina software engineering approach for the development of electronic commerce sites. This dissertation describes the design rationale for the framework structure, as well its pros and cons. As an exercise of the use of the framework an application also called 2BuyNet has been instantiated, using the Web as the interface. This application has been published on the Internet and works both as a web store generator service and a web store administrative tool. 99_MSc_esmeraldo Luiselena Luna ESMERALDO. Frameworks para projetos de aplicacao hipermidia. M.Sc Diss. Port. Presentation: 19/05/99 213 p. Advisor: Daniel Schwabe Abstract: This paper describes the notion of framework to hypermidia application designs, suggesting a notation to specify these frameworks and show how to use them by three examples of real hypermidia application projects: forums, virtual stores and on line publications. OOHDM is the method used to design the frameworks. With these frameworks, we intend to show that is possible to reuse the hypermidia applications to minimize the complexity of design hypermidia applications. Another purpose is to legitimate some hypermidia patterns described to navigation projects using them in the proposal frameworks. 99_MSc_suarez Marcela Ozorio SUAREZ. O modelo de I/O baseado em streams aplicado ao desenvolvimento de aplicacoes graficas orientadas a eventos. M.Sc.Diss. Port. Presentation: 16/04/99 63 p. Advisor: Roberto Ierusalimschy Abstract: The objetive of this work is to evaluate the fitness of the stream-based I/O model for the development of event oriented applications. Initially some aspects of the functional programming and its advantages over the procedural programming are presented. The difficulties found in the accomplishment of I/O in functional languages are presented and the well-known alternatives to overcome this problem are described. One of the alternatives described is the stream-based I/O model used in lazy functional languages. An analogy between this I/O model and the event oriented model is presented. This analogy, based on the assynchronous mechanism found in both models, leads to the hypothesis that the stream-based I/O model can be suitable to the development of event oriented applications. In order to prove this hypothesis, a small library was developed for the implementation of graphic event oriented applications. For the accomplishment of this experiment, an interpreter for a lazy functional language with syntax similar to the Scheme language was implemented. The library was implemented in this language called Scheme Lazy. The interpreter construction was based on the lambda calculus and combinators theories. Finally, the difficulties found in the accomplishment of the experiment and the conclusions of this work are presented. 99_PhD_sant'anna Marcelo Henrique B. SANT'ANNA. Circuitos transformacionais. Ph.D.Thesis. Port. Presentation: 14/04/99 157 p. Advisor: Julio Cesar Sampaio do Prado Leite Abstract: Software transformation systems allow the structural and semantic change of software. This category of systems has been successfully applied to several distinct software engineering tasks. Following a pragmatic approach, this research work builds upon Draco paradigm seminal ideas by developing prototypes for transformation systems and applying them to some experiments. Such experiments explore distinct avenues for the evolution of the area, with a particular emphasis on software architecture issues. The results gathered from the experiments are then analyzed and compared to what have been presented by other researchers. Departing from this analysis , an architecture description language for software transformers is proposed. Such language works around the concept of Transformational Circuits (TCs). The idea of a TC is presented on this thesis and it represents an attempt to put together results from the transformation systems and software architecture communities. 99_MSc_rodrigues Marcus Antonio Almeida RODRIGUES. Um framework para a provisao de servico de multicast em ambientes genericos de comunicacao de dados. M.Sc.Diss. Port. Presentation: 21/05/99 136 p. Advisor: Luiz Fernando Gomes Soares Abstract: Since most of the solutions available as candidates to the implementation of a multicast service have been designed to fit specific infrastructure conditions or service characteristics to which they are aimed, or, still, to fit management politics for user groups, the present work introduces a framework for designing a multicast service whose main features are its generality and independence toward possible communication systems and applications. This framework allows the implementation of specific multicast services in a fast and organized way, by reusing the whole generic structure presented, in addition to the configuration of some service components. The framework consists of two basic services: a group management service, and a support service to the construction of the multicast distribution infrastructure. Its use is illustrated by the implementation of broker agent for centralized group communication, applied to a video distribuition service. 99_PhD_fontoura Marcus Felipe Montenegro Carvalho da FONTOURA. Uma abordagem sistematica para desenvolvimento de frameworks. Ph.D.Thesis. Port. Presentation: 23/07/99 167 p. Advisor: Carlos Jose Pereira de Lucena Abstract: No English abstract provided. 99_PhD_brito Maria Alice Silveira de BRITO. Uma arquitetura orientada a objetos para integracao das facilidades de concorrencia, transacao e persistencia. M.Sc.Diss. Port. Presentation: 17/03/99 222 p. Advisor: Sergio Eduardo Rodrigues de Carvalho Abstract: This work describes an object-oriented architecture for integratingconcurrency, transaction and persistence. In the context of long-lived activities consisting of components controlled by transactions that may commit individually (workflow), showing cooperative characteristics rather than traditional competitive characteristics, we identified two ma or necessities: i) the flexibility to extend a new transaction and to make possible the co-existence oftransactions with different policies and protocols in the same object and ii) the interaction capabilities between application and transactions and among transactions. Existing tools are difficulty to use and are sometimes unsuitable for the combination of paradigms we are considering. We explore abstract data types and inheritance to achieve the expected flexibility, encapsulating all objects employed in mechanisms to integrate transaction at application. We adopt the local atomicity and active object concepts, and design an architecture based in three major classes: object manager, transaction manager and application object, the latter associated to a monitor (a metaobject). To achieve the co-existence among transactions with different policies and protocols we isolated synchronization aspects in the monitor's concurrency control tasks, enabling the representation of the negociation among policies. To simplify interactions we designed architecture elements with flexibility in mind, thus allowing extensions in functionalities and interfaces, ando also avoiding the creation of new programming languages primitives. 99_PhD_mediano Mauricio Riguette MEDIANO. Um metodo para indexacao de estruturas topologicas em banco de dados geograficos. Ph.D.Thesis. Port. Presentation: 23/04/99 92 p. Advisor: Marco Antonio Casanova Abstract: This thesis introduces a new method, called TDA (Topological Data Structure), for the topological indexing of spatial attributes in geographic databases. Basically, the TDA method maintains a topological data structure for each set of spatial attributes for which one wishes to optimize the calculation of topological operators. The topological data structure, which is internally manipulated by the geographic database manager system, is responsible for generating the faces, vertices, and edges that compose the topological representation of each spatial attribute. The TDA method is responsible for updating the topological representation of each attribute whenever the database is updated. The dimension of each of the nine intersection between the inside, the border, and the ouside of two spatial attributes can be efficiently calculated by means of the topological representation of the spatial attributes, and by sweeping the topological structure. The TDA method replaces the calculation of topological operators based on the geometry of spatial attributes. 99_MSc_macedo Nestor Adolfo Mamani MACEDO. Integrando requisitos nao funcionais aos requisitos baseados em acoes concretas do SERBAC. M.Sc.Diss. Port. Presentation: 26/04/99 153 p. Advisor: Julio Cesar Sampaio do Prado Leite Abstract: This essay proposes the integration of NFR (Non-Functional Requirements) to the requirements based on Serbac's Concrete Action. This approach takes into consideration the business as a whole, to which is applied the Business Strategic Planning for Environment Analysis. For requirement elicitation, we make use of the Serbac Model as a baseline. Over this model a specialization is made to the "attribute requirements", the information item of several SERBAC components, into "quality NFRs" and "comprise and operation NFRs". The strategy proposed has two alternatives: (i) to search for the places where it may be possible to find NFRs, using the Business Model by structured analysis, and (ii) to establish a repository of "quality NFRs", organized in accordance to each stakeholder's viewpoint. Each viewpoint has primary quality NFRs (PQN), and each PQN has a set of specific quality NFRs (SQN). On the other hand, each SQN holds a list of satisfaction strategies (LSS). 99_MSc_sanches Newton Cunha SANCHES. Estruturas de dados temporarias para processamento de consultas espaciais. M.Sc.Diss. Port. Presentation: 30/03/99 76 p. Advisor: Marcelo Gattass Abstract: This paper presents a study on efficient query processing in spatial database systems. Spatial queries often demand the chained execution of many selections and joins. This situation poses the need for mechanisms capable of holding results of an operation and sending them to the next one. In this work we briefly present some indexes and algorithms used for executing spatial selections and joins, and also some ways of chaining those operations. Next, we propose a hierarchical data structure with low construction cost which is used to store datasets originated in spatial selections. We compare its performance to some indexes' and to other data structures designed for holding temporary results. This comparison is done using geographic and randomly-generated datasets. 99_MSc_frederick Paula FREDERICK. Visualizacao eficiente de objetos graficos. M.Sc.Diss. Presentation: 01/07/99 75 p. Port. Advisor: Marcelo Gattass Abstract: This work presents a study on the efficient visualization of large figures, such as maps and CAD drawings. These large figures are usually formed by a great number of graphical objects, generally most of them being polygonal lines with many points. In Computer Graphics, conventional techniques for drawing graphical objects are unable to offer interaction times when applied to a huge data volume. Thus, it is necessary to use appropriate mechanisms for speeding up this process. A brief description of graphical objects and of spatial access methods is made. Then some methods are proposed for storing and efficiently retrieving sets of two-dimensional graphical objects in a secondary storage media. These methods consist in using persistent data structures composed by an R-tree, V-trees, and Quadtrees. In order to demonstrate the efficiency of the proposed solutions, some experiments done with real geographic data are shown. 99_PhD_gomes Paulo Cesar Rodacki GOMES. Prototipacao virtual em modelagem de solidos distribuida. Ph.D.Thesis. Port. Presentation: 03/05/99 94 p. Advisor: Bruno Feijo Abstract: This dissertation presents the concept of Virtual Prototyping in Distributed Solid Modeling as a way to develop more effective CAD systems. Questions about geometric modeling, distributed computing, contraint solving and online algorithms are discussed in this work. Also presented is a distributed architecture for the virtual prototype based on the concept of dynamic agents. The dynamic agents are supported by a Geomatry Bus, a CORBA model, and LuaOrb, a binding between the CORBA's Dynamic Invocation Interface and the extension language Lua. According to a special graph defined by the author called the Extended Constraint Graph, reactivity amongst agents can be defined by complex and nonlinear relations which can be changed at run time. This architecture is implemented by an online algorithm that is recursive and distributed. 99_MSc_caroli Paulo Roberto Celidonio CAROLI. Uma metodologia de projeto de software orientado a objetos. M.Sc.Diss. Port. Presentation: 22/12/99 116 p. Advisor: Carlos Jose Pereira de Lucena Abstract: In spite of the increasing utilization of the Object Oriented paradigm, there still exists reluctance on its application in the development of systems as, for example, Web Information Systems. In these type of applications, User Interfaces are in general event oriented and the majority of the Databases are Relational. This work proposes a Three-Tier Architecture for the development of systems that can be naturally divided in User Interface, Business and Persistence layers. It also proposes a Methodology for the development of the Business layer. The objectives of this work are: to help management to keep up with the project development; facilitate the maintenance of large systems subject to modification and evolution; help communication among the persons in the development team; provide Object Oriented development support coexisting with other paradigms; serve as an effective documentation. Applying the concepts here put forth in real projects validated the work. 99_MSc_oliveirajunior Pedro Paulo de M. OLIVEIRA JUNIOR. Exames virtuais utilizando um algoritmo de Ray Casting acelerado. M.Sc. Port. Presentation: 07/12/99 62 p. Advisor: Marcelo Gattass Abstract: This work presents a study on virtual exams, which are techniques for visualizing the inner part of the human body using volume visualization techniques for clinical evaluation purposes. The virtual exams aim at being an alternative for conventional digestive endoscopy, colonscopy, and bronchoscopy exams. We present a comparative study between virtual exams and their analogue conventional exams, analyzing advantages and disadvantages of each of the techinques. We also propose an algorithm for accelerating volume visualization with ray casting, and compare the results obtained with other optimization techniques. 99_MSc_ripper Pedro Santos RIPPER. VMarket: um framework para sistemas de comercio eletronico voltados para mercados virtuais mediados por agentes de software. M.Sc.Diss. Port. Presentation: 15/06/99 77 p. Advisor: Carlos Jose Pereira de Lucena Abstract: Software agent technology is still an emerging technology, and as such, agent based software design is still in its infancy. Software agents have just started to be used in the e-commerce domain, and they are already starting to create a series of new possibilities to this arena. Agents can be used to automate as well as to enhance many stages of the traditional consumer-buying behavior process. Through the minimization of transaction costs, elimination of geographic barriers and time issues, many new markets, not viable before, are now being created, traditional markets are being made more efficient, and the role of the middleman is drastically changing. This dissertation proposes a software approach to the design of agent mediated e-commerce systems, through the definition and implementation of a object oriented framework. This framework focus mainly in e-commerce applications based on virtual marketplaces. A virtual marketplace is nothing more than an internet based system, where software agents interact and negociate, on behalf of their respective users, to buy, sell or find specific goods. In this type of system, all users can be potencial buyers, sellers or both, depending on their specific interests. The main goal of Vmarket is to facilitate the creation of this type of applications, as well as to make them more robust and flexible. It is expected that this approach will greatly enhance the process of experimentation and research on the new possibilities brought by software agents to the e-commerce application domain. Vmarket should give developers the ability to customize virtual marketplaces and define transaction categories on demand incorporating many possible products and services that can be traded online. Its user can create new transaction types and items on individual needs and in the process create customized software agents that can recognize and adapt to the new products and services being offered. Software agents in VMarket will pro-actively broker and negociate with interested and sellers, represented by their respective agents. They can be created with any set of desired behaviors, thereby enabling the consumer to have a virtual presence in the marketplace to further his or her interest, while freeing the consumer from constant monitoring of market progress. 99_PhD_braganca Regina Celia Moreth BRAGANCA. Calculo de sequentes e deducao natural: relacionamentos e consequencias. Ph.D.Thesis. Port. Presentation: 21/05/99 126 p. Advisor: Edward Hermann Haeusler Abstract: We construct a new system of Segment Calculus, the Natural Sequent Calculus - NSC, which allows us to convey a fundamental concept in Natural Deduction to the Sequent Calculus. With this strategy, we may think in the relationship between the Sequent Calculus and Natural Deduction under a new perspective. Using this technique we establish a system of Cut-free Natural Sequent Calculus, to Constant Domain Logic - DC. 99_MSc_siqueira Sean Wolfgand Matsui SIQUEIRA. Uma arquitetura database marketing utilizando Heros - SGBDH. M.Sc.Diss. Port. Presentation: 30/09/99 85 p. Advisor: Rubens Nascimento Melo Abstract: Database Marketing refers to the use of database for supporting marketing activities. Data from several sources, possibly heterogenous, need to be integrated and a set of tools/processes make necessary analysis and marketing strategy implementation (and manage) possible. Integration of heterogenous data sources, in the database community, is treated through the heterogenous database management systems study. In this work, it is proposed a database maketing architecture that uses the heterogenous database management system (HDBMS) technology for the integration of several heterogenous sources. The HDBMS used in the proposed architecture is called HEROS and it is in development in the Computer Science Department at PUC-Rio. The database marketing and HDBMS fundamental theoretical concepts are analyzed and correlated. A systematic of use of the architecture proposed as well as a case study that uses the proposed systematic are presented. 99_PhD_colcher Sergio COLCHER. Um meta modelo para aplicacoes e servicos de comunicacao adaptaveis e com qualidade de servico. Ph.D.Thesis. Port. Presentation: 22/11/99 114 p. Advisor: Luiz Fernando Gomes Soares Abstract: One of the main challenges posed to the telecommunications services operators is to develop a single infrastructure to support the integration of a wide range of services and to allow for the rapid creation, alteration and continuous adaptation of these services to new evolving conditions or demands. The variety os requirements and the rapid evolving technology impose many difficulties, specially when the goal is to develop services based on a limited number of pre-defined options. This work presents a model in which communication services platforms will be constructed to support some kind of adaptability in order to meet the wide range of different demands, the presence of run-time changes in applications and the rapid evolving technology. The model provides some basic abstractions for the representation of multicasting and Quality of Service (QoS) within multimedia communication services and is based on the pipes-and-filters architectural style. Pipes correspond to the explicit representation of connections between components (the filters) that, in turn, represent the active processing elements of a service implementation. QoS specifications can be associated to pipes, which are then called MediaPipes. MediaPipes correspond to an abstract programming entity for the QoS treatment and also allow for the exposure of internal implementation details whenever adaptations are necessary. 99_MSc_cortes Sergio da Costa CORTES. Regras ativas em sistemas de banco de dados: sintaxe, semantica e processamento distribuido. M.Sc.Diss. Port. Presentation: 29/04/99 144 p. Advisor: Sergio Lifschitz Abstract: Active rules enable the development of non-conventional database applications, providing support to a semantics that reflects behaviour based on events. On one hand, many researches were carried out, and active database systems were turned available, but several important proposals were only partially implemented. On the other hand, making use of active rules on a distributed environment enables greater expressiveness, but there is a considerable increase on the system's complexity. In this dissertation, we will discuss a semantic and syntactic extension for the construction of active rules on database systems which comprises several situations involving applications not yet completed by the existing systems. Furthermore, we propose a control scheme for the execution of this extended model of active rules on a distributed database, which satisfies the requirements of transparency and correctness on this environment. 99_PhD_martins Simone de Lima MARTINS. Estrategia de paralelizacao de metaheuristicas em ambientes de memoria distribuida. Ph.D.Thesis. Port. Presentation: 27/07/99 148 p. Advisor: Celso da Cruz Carneiro Ribeiro Abstract: We present an analysis of metaheuristic parallelization strategies in distributed memory environments. Some related issues are studied by implementing some parallel algorithms based on metaheuristics to solve specific problems. We concentrate our work in environments composed of workstations connected by communication networks, due to their low cost and high availability. We compare some software tools useful to develop parallel programs in there in order to show their characteristics in terms of ease of use and performance. Wa analyze three metaheuristics: VNS, GRASP, and tabu search. We compare two software tools using different paradigms: PVM and Linda, used to implement a parallel tabu search for the circuit partitioning problem. We show two parallel GRASP algorithms developed to solve the Steiner problem in graphs using two different local search strategies. We present numerical results showing the quality of solutions obtained, execution times, and the speed up achieved by the parallel GRASP algorithms. 99_PhD_barbosa Simone Diniz Junqueira BARBOSA. Programacao via interface. Ph.D.Thesis. Port. Presentation: 05/05/99 109 p. Advisor: Clarisse Sieckenius de Souza. Abstract: In the past few years, we have witnessed an increase in software functonality as an attempt to meet most users' needs. This approach brings serious usability challenges, due to an increase in application complexity as well. In order to try and meet users' needs, without overloading the application with functionality that is rarely used, there is a tendency to allow end users to configure or program applications, by means of mechanisms that support the so-called end user programming. However, many existing techniques fail to attain acceptable thresholds of usefulness and usability. This work addresses some of the challenges posed by extensible applications. We follow an approach that drops some walls between interface and extension. This approach brings some extension mechanisms to the interface, and readily accessible to end-users, namely extensions based on the semantic-pragmatic resources of metaphors and metonymies. These mechanisms were chosen due to the acknowledgment of the Cognitive Sciences of their critical role in our reasoning processes, especially when we try to describe or understand complex or abstract concepts. (Lakoff e Johnson, 1980; Lakoff, 1987; Lakoff, 1993; Ortony, 1993). We describe an extensible application model that makes use of a knowledge base in which we represent the domain and application elements that may be extended, as well as the necessary classifications for calculating the possible extensions. Our model takes into account the communicative aspects of computer applications, and follows Semiotic Engineering (de Souza, 1993) principles to guarantee the consistency between the original application and the extended one. For that purpose, our model entails the representation of rules that constrain interface amendments, so that extended functionality is adequately reflecte at the resulting interface. 99_PhD_ferreira Simone Maria Bacellar Leal FERREIRA. Uma arquitetura para associacao de ususarios a interfaces. Ph.D.Thesis. Port. Presentation: 31/04/99 142 p. Advisor: Sergio Eduardo Rodrigues de Carvalho Abstract: In order to take care of the necessities of the users, the process of development of user's interface must be centered in the user. The user must be analysed in order to get its features; features these that constitute the "model or profile of users". This model allows the construction of more powerful interfaces, capable to take care many expectations of the users, that is, to offer to different levels of functionality according to the necessities of each person. The present work goals the development of an abstract architecture that makes possible the instantiation of user's interfaces for different types of users and different systems. In the present architecture, profiles of users are part of software. The architecture contributes the design of interfaces with the objective to satisfy the necessities of the users, supplying adaptation mechanisms in time of execution to users that improve its knowledge on the application, becoming more ambitious. To these, the achitecture considers a new interface, thus presenting an agency feature. _______________________________________________________________________ 2000 00_PhD_garcia Alex Vasconcellos GARCIA. Um modelo categorico para traducoes entre linguagens de programacao. Ph.D.Thesis. Port. Presentation: 26/05/00 135 p. Advisor: Edward Hermann Haeusler Abstract: This thesis proposes a categorical model for programming languages. This model is used to establish a formal condition that a translation should satisfy in order to preserve program maintainability throughout the translation process. For this purpose, we start from a categorical model proposed by R.F.C. Walters for context-free languages. In order to establish the criterion for preserving maintainability, it was necessary to introduce semantics for the categorical models of context-free languages. The examples shown illustrate that the arrows semantics (arrows are models of derivation pieces) is related to preservation of program maintainability. We believe that the technique we developed, not only applies to the problem studied in the thesis, but also brings an important insight regarding the semantics of programming languages and the way in which the programmer's intention is related to the syntactical structure of a program. 00_PhD_carvalho Alexandre Plastino de CARVALHO. Balanceamento de carga de aplicacoes paralelas SPMD. Ph.D.Thesis. Port. Presentation: 31/03/00 138 p. Advisor: Celso de Cruz Carneiro Ribeiro Abstract: Parallel algorithms are designed based on two programming models: functional parallelism and data parallelism. The second model, when used in parallel architectures with no instruction level synchronism, is known as SPMD (Single Program, Multiple Data). In SPMD applications, the same code is executed on different processors, over distinct data sets. This model has attracted the attention of programmers and researchers due to its simpler and clearer parallel programming structure. On the other hand, it requires considerable effort in order to identify an ideal load-balancing algorithm for the application under development. The difficulty in this identification is partly related to the large variety of existing strategies. A suitable strategy choice permits controlling efficiently the effect of different dynamic imbalancing factors, present in many parallel applications and environments. This thesis presents instruments to support the development of SPMD applications, with special emphasis on the identification of an appropriate load-balancing strategy. A taxonomy of load-balancing algorithms designed for SPMD applications is proposed. Its standard terminology allows the classification and well-organized comparison of a large number of existing strategies. The central contribution of this thesis is an environment for SPMD applications development, called SAMBA (Single Application, Multiple Load Balancing). The proposed tool generates SPMD applications with load balancing from, basically, the routines that create and execute the SPMD tasks coded by the programmer. The provision of the common structure of SPMD applications and of a load-balancing library allows the programmer to concentrate on his/her specific problem and reduces programming effort. Different versions of the application can be generated, one for each strategy available in the load-balancing library. In this way, the designer can test and compare different strategies for the same application, at no extra programming cost. In order to validate the proposed development environment, the thesis presents the analysis of different load-balancing algorithms in the context of three SPMD applications, developed using SAMBA: Matrix multiplication, function integration and genetic algorithm. The work also includes the definition of a framework for the development of SPMD applications with load balancing, originated from the experience obtained with the development of the proposed environment. 00_PhD_schmidt Ana Elisa Ferreira SCHIMDT. Visualizacao tridimensional combinada de dados volumetricos e modelos poligonais usando o algoritmo Shear-Warp. Ph.D.Thesis. Port. Presentation: 28/04/00 109 p. Advisor: Marcelo Gattass Abstract: This thesis presents a study on the integration between Shear-Warp and Z-Buffer algorithms, and extends the Shear-Warp algorithm to handle scenes composed of both volume and polygonal data. Shear-Warp techniques provide an efficient way to perform the projection and blending stages of the volume-rendering process. They treat volume-data slices in a coherent order and can be naturally integrated with the Z-Buffer algorithm. We present methods to handle opaque/translucent volumes combined with opaque/translucent polygonal models. As volume data usually has a different resolution from that of the final image, in which Z-Buffer renders the polygonal data, several variants for this integration that try to reduce aliasing problems are analyzed. Results are shown to support some conclusions on the trade-off quality-versus-time that can be expected. 00_MSc_moura Ana Lucia de MOURA. Um ambiente de suporte a adaptacao dinamica de aplicacoes distribuidas. M.Sc.Diss. Port. Presentation: 28/09/00 88 p. Advisor: Noemi de La Rocque Rodriguez Abstract: Component-based programming is presently an important trend for the development of distributed applications, allowing complex systems to be constructed by combining several independent an reusable components. These components implement specific services accessed through interfaces that traditionally define only functional requirements - such as operations signature - and abstract the non-functional properties of the offered services, such as performance, security or availability. However, the increasing demand for distributed applications that are able to meet non-functional requirements like the ones listed above, and the dynamic behavior presented by the execution environment of these applications, impose that not only middleware platforms, but also applications, dynamically adapt to the state of their execution environment. Fot that to be possible, it is essential that the applications be able to handle the non-functional properties of the services provided by their components. The purpose of this work is to offer an environment that supports the dynamic adaptation of distributed applications to variations in the non-functional properties of the services provided by their components. For that we developed facilities and mechanisms that allow applications to dynamicallyselect the components that best suit their requirements, verify if these requirements are being met, and react, when appropriate, to variations in the non-functional properties of the services in use. 00_MSc_maianeto Ayrton MAIA NETO. COMMERCEPIPE - Um framework para criacao de canais comerciais consumer to business na internet. M.Sc.Diss. Port. Presentation: 16/06/00 103 p. Advisor: Carlos Jose Pereira de Lucena Abstract: Based on the study of different electronic commerce software applications and principally inspired by the research work and software evolution carried out in the Vmarket framework, this dissertation presents the Commercepipe Object Oriented Framework as an alternative for the instantiation of Consumer to Business Virtual Markets on the Internet. Using Software Agents Technology combined with the WAP (Wireless Application Protocol) as a wireless interface to access Internet services, the framework proposes a new approach for C2B markets, bringing a new vision of how sellers and buyers can interact commercial transactions through the Internet. 00_MSc_muniz Bruno Cavalcanti MUNIZ. Notificacao e controle de versoes para o suporte a autoria cooperativa no sistema HyperProp. M.Sc. Diss. Port. Presentation: 28/08/00 126 p. Advisor: Luiz Fernando Gomes Soares Abstract: it is important to a hypermedia authoring system to support simultaneous document editing, done by several authors, in an adequate and efficient way. To accomplish that, hypermedia systems incorporate version control mechanisms integrated to notification schemes, which inform authors about significant events related to creation and editing of versions. This work specifies notification control mechanisms for the Nested Context Model (NCM), which were implemented within the hypermedia authoring system HyperProp. It is important to observe that these mechanisms can be adapted to other hypermedia models based on compositions and offering version control functionality. In addition, an implementation of the NCM version control is also presented, enabling, among other facilities, the graphical visualization of hyperdocuments history in the HyperProp system. 00_MSc_martin Carlos Bazilio MARTINS. Deteccao de paralelismo a partir da semantica denotacional e de grafos de dependencias. M.Sc.Diss. Port. Presentation: 12/05/00 91 p. Advisor: Edward Hermann Haeusler Abstract: This work propose a generic architecture that generates parallel automatic code from sequential programs. In literature this subject is called implicit parallelism. The architecture has a target language's denotational semantic specification and an example code in such a language as its input. The output is a parallel code in the same abstraction level as the original sequential code. Intermediate code generation with parallel constructs (cobegin/coend, forall,...) is one of the intermediate steps. This code is processed to be instantiated to a particular machine. Our implemented prototype processes the C language and generates multithreaded programs from sequential ones. 00_PhD_cassino Carlos Roberto Serra Pinto CASSINO. Distribuicao de carga em sistemas WEB controlada por ferramentas de construcao de paginas dinamicas. Ph.D.Thesis Port. Presentation: 19/12/00 81 p. Advisor: Roberto Ieruslamischy Abstract: Load distribution among several servers is a scalability typically adopted in Web sites. E-commerce applications, among others, use dynamic pages to offer sophisticated services, demanding even more processing power from those servers. This sophistication imposes some constraints to the load distribution mechanism, such as fixing some clients on some servers, diminishing the performance of the distribution. Our thesis is that integrating the load control into the dynamic page construction tool allows the use of internal information of the tool, and therefore a more effective control of the distribution and more efficiency. In this work, we propose a load distribution model in which the page construction tools controls, by manipulating the embedded links of the generated pages, the distribution of the clients among the servers. Since this load distribution model differs from other existing models, a relevant part of our work is the implementation and analysis of a prototype. With this prototype, we could analyze the performance of some load sharing algorithms, the behavior under overloads, and others. This analysis shows that this distribution model acheives a good performance when compared to other models commonly adopted. Moreover, this integrating model is able to handle the constraints imposed by the dynamic applications, such as client fixing and distribution guided by classes of requests. 00_MSc_renteria Christian Jacques RENTERIA. CN: deducao natural para CLT. M.Sc.Diss. Port. Presentation: 09/05/00 74 p. Advisor: Noemi de La Rocque Rodriguez Abstract: CLT (Computation Tree Logic) is a used both for the synthesis and for the verification of concurrent programs. Usually presented in the Hilbert style (axiomatic), it has been adapted for sequent calculus, but without normalization. The aim of the present work is to present a Natural Deduction system for CLT. For that reason the original language has been modified, and labeled formulae have been used. It has been proved that the system is correct and complete. Then the possibility of normalizing the proofs is studied and it is seen, as expected, that strong normalization is achieved in the system without the induction rule. 00_MSc_guaranys Clecio Radler dos GUARANYS. e-BusinessCard: um framework para Smart Cards. M.Sc.Diss. Port. Presentation: 28/09/00 p. Advisor: Carlos Jose Pereira de Lucena. Note: Not available yet. 00_MSc_mesquita Fernanda Victal MESQUITA. Suporte ao desenvolvimento de aplicacoes SPMD. M.Sc.Diss. Port. Presentation: 07/01/00 60 p. Advisor: Noemi de La Rocque Rodriguez Abstract: This work describes the implementation of a prototype of SAMBA (Single Program Multiple Balancing). SAMBA is a tool that supports the development of SPMD parallel programs. It is basically intended for applications where the execution of each task is totally independent of the others. SAMBA also supplies a library of load balancing algorithms to manage the distribuition of the workload between the parallel processes. SAMBA also supplies a library of load balancing algorithms to manage the distribuition of the workload between the parallel processes. 00_MSc_amaral Fernando Naufel do AMARAL. RETOOL - uma logica de acoes para sistemas de transicao temporizados. M.Sc.Diss. Port. Presentation: 09/05/00 117 p. Advisor: Edward Hermann Haeusler Abstract: This dissertation defines RETOOL, an action logic featuring an operator to denote necessary conditions and postconditions of actions in a timed transmission system (an extension of the formalism of transition systems meant to model real-time reactive/concurrent computational systems). A semantics for RETOOL is presented and compared to previous proposals. An adequate axiomatization is given, along with detailed correctness and weak completeness proofs. 00_MSc_freitas Flavio Correa de FREITAS. Uma proposta de arquitetura para comparacao de juncoes paralelas. M.Sc.Diss. Port. Presentation: 27/09/00 70 p. Advisor: Sergio Lifschitz. Abstract: The performance comparison of parallel strategies for the join operation is difficult because they are implemented in many distinct environments. Benchmarks and test environments mentioned in the literature make use, in general, of both not opened hardware and software or even simulations of DBMS' operations. This work discusses the existing environments that evaluate and compare parallel join strategies, including their implementation resources, software tools and available comparison, named ARCOJP (Architecture for Comparison of Join in Parallel), is poposed here. It may instantiate many functionalities of the known approaches and uses a local DBMS (Minibase) at each processing node. An implementation of ARCOJP is detailed, which runs in a PC cluster running Linux and MPI. Some preliminary comparison results are also given for a few selected strategies. 00_MSc_gomes Flavio Spolidoro Ferreira GOMES. Uma arquitetura distribuida para a construcao de aplicacoes de gerencia de rede. M.Sc.Diss. Port. Presentation: 18/12/00 108 p. Advisor: Noemi de La Rocque Rodriguez Abstract: The growth of corporative networks and the different kinds of network platforms require more and more the use of scalable network management applications and make possible the integrated management of different platforms. By exploring the resources of an interpreted language, this dissertation presents a distributed architecture for building network management applications with monitoring functionalities and asynchronous events reception functionalities. It also explores the advantages of a Web-based network management application, which we built aiming at testing this architecture. 00_PhD_cafezeiro Isabel Leite CAFEZEIRO. Compartilhamento e semantica denotacional. Ph.D.Thesis. Port. Presentation: 12/05/00 113 p. Advisor: Edward Hermann Haeusler Abstract: This thesis presents the fundamentals of a metalanguage for Denotational Semantics, with the purpose of expressing the concept of sharing. The approach consists in a generalization of some of the usual operators of Denotational Semantics to a shared version. Category Theory is used to the definition of the operators as continuous bifunctors. The categories used have, as objects, Scott-Ershov Domains. The fixed point for equations written with the operators is ensured. Two cases studies are presented, showing that the operators are compatible with the ususal presentation of Denotational Semantics. The first case study considers sharing at the syntactical level, whereas the second concerns the semantic level. 00_MSc_sabatneto Jaime de Melo SABAT NETO. Integrando requisitos nao funcionais a modelagem orientada a objetos. M.Sc.Diss. Port. Presentation: 29/03/00 206 p. Advisor: Julio Cesar Sampaio do Prado Leite Abstract: Non-functional requirements (NFRs) are restrictions or quality attributes of systems or system development processes. NFRs were presented a long time ago and are now part of several requirements engineering syllabus. However, the majority of software engineering methods in general and requirements engineering methods in particular don't take into account NFRs. Hence, we propose, in this work, the OONFR strategy. This strategy uses, as input, a Language Extended Lexicon of the Universe of Discourse (LEL of UofD) and produces a class diagram which has indications of classes, attributes, operations and relationships that specializes NFR satisfaction strategies and satisfies NFRs. The OONFR strategy includes the following activities: build a Language Extended Lexicon of the Universe of Discourse-NFR (LEL of UofD-NFR), build scenarios and build a class diagram. In this work, we also present a tool, that gives support to the usage of the OONFR strategy, and three case studies, which carried out in order to achieve an initial validation of the strategy. 00_MSc_macedo Jose Antonio F. de MACEDO. Um estudo de SGBDs baseado em agentes. M.Sc.Diss. Port. Presentation: 27/09/00 72 p. Advisor: Sergio Lifschitz Abstract: There are applications that ask for extensions and configurations of DBMSs functionalities in order to supply new demands. The construction of DBMSs that satisfy these multiple needs is difficult due to the diversity of application domains. The software agents research are deals with definition of systems that execute in heterogeneous, distributed and dynamic environments, mostly used in application domains where the reactivity factor is important. These agents properties can be applied to the DBMSs context so to facilitate their extension and configuration. The present dissertation proposes a new paradigm for DBMSs that consider the use of agents in their architectures: Agent-based Databases. This work studies and proposes three DBMSs architectures based on agents called integration architectures. One of these architectures is implemented and discussed so to evaluate in practice our approach in the context of the execution of parallel joins and load balancing. 00_PhD_breitman Karin Koogan BREITMAN. Evolucao de cenarios. Ph.D.Thesis. Port. Presentation: 12/05/00 153 [+53] p. Advisor: Julio Cesar Sampaio do Prado Leite Abstract: In recent years, a growing number of researchers have turned to scenarios to improve communication among the stakeholders of a computer-based system under construction. Scenarios are informal descriptions of situations in the system environment and help the elicitation and validation of information during all stages of system development. On the other hand, the informality of scenarios allow them to be used without any specific methodological support. This aspect, that has been pointed out specially by the Requirements community has been given little attention and makes the systematic use of scenarios in software practice more difficult. To understand what is required for effective use of scenarios in system development, we have conducted a series of empirical case studies that shed light on the complexities of scenario evolution. As a result of these case studies, we have produced a model of scenario evolution that explains the dynamics of the evolutionary process while providing a multi-tier approach to scenario evolution, based on operations and relationships of the model. This thesis proposes an organization that allows an engineering approach to scenario evolution. The organization is based on domain knowledge embodied in the scenario evolution model and takes into consideration other software management issues, such as configuration management and traceability. The organization implies a framework for scenario evolution that can be used in any open-ended, general software development environment. This framework can be configured, through the use of hot spots, to tailor specific user needs. We also offer a prototype tool, called SET (Scenario Evolution Tool), that provides automated support to the framework. 00_MSc_rodrigues Leandro Marques RODRIGUES. Integracao de documentos SMIL ao sistema HyperProp e desenvolvimento de ferramentas para exibicao de objetos com relacionamentos de sincronizacao. M.Sc.Diss. Port. Presentation: 21/11/00 118 p. Advisor: Luiz Fernando Gomes Soares Abstract: The use of the World-Wide Web to distribute multimedia documents with temporal and spatial synchronization relationships is a research topic that has recently gained a lot of attention. The SMIL language had its recommendation published by the W3C for declarative authoring of these documents, but it has some limitations. In order to overcome these limitations, this dissertation presents the integration of SMIL documents with HyperProp, a hypermedia system based on a conceptual model called NCM (Nested Context Model). This integration allows SMIL documents to be refined with NCM facilities and manipulated in a hypermedia system with version control and high-level authoring of NCM documents with temporal synchronization relations. The dissertation also proposes some extensions to NCM, adding some concepts found in SMIL, in order to simplify the authoring of NCM documents with temporal synchronization relations. Finally, the dissertation presents the development of presentation tools, integrated with the HyperProp system, to control the presentation of continuous media (audio and video) and static images, supporting all temporal and spatial relations defined in the NCM model, besides user interaction relations. 00_PhD_moura Leonardo Mendonca de MOURA. Um framework para analise e verificacao de programas. M.Sc.Diss. Port. Presentation: 05/05/00 199 p. Advisor: Carlos Jose Pereira de Lucena Abstract: Presently, program analysis and verification is at best achieved by a hand-crafted tool specific to a programming/specification language. Since the manual construction of tools is expensive, they are hard to obtain and not often used, limiting the quality of the implemented software. We have proposed a framework for the automated generation of simulation, analysis and verification tools for programs/specifications based on machine-processable formal definitions of the language's operational semantics. We also use a modular approach for language semantics that allows us to reuse semantic modules and fragments of already defined tools. 00_PhD_riveraescriba Luis Antonio RIVERA ESCRIBA. Animacao baseada em fisica com modelos geometricos em multi-resolucao. Ph.D.Thesis. Port. Presentation: 29/09/00 150 p. Advisors: Paulo Cezar Pinto de Carvalho (IMPA) and Marcelo Gattass Abstract: The use of physical laws in the generation of animations permits the production of (or aims at producing) real effects of movements and automatic treatment of collisions. However, a number of difficulties are encountered for the efficient execution of the needed numeric computations and for the production of interesting effects, due to the geometric complexity of the objects in the animation. In this thesis, we formulate a model to generate animations based on physical models of objects with arbitrary geometry. The irregular details of the object surfaces are incorporated in the calculation of the object contacts which are found to be in collision. We use multiresolution theory based on the wavelet transformation to establish a relationship between the details of the object surfaces and the surfaces rugosity degree, using statistical estimators. This approach leads to the hierarchical construction of oriented-retangular adapted envelopes (MOBBtree - Multiresolution Oriented Bounding Boxe tree). They preserve the surfaces details, allowing for a fast detection of possible surface segments in collision, that are later used to determine the tolerated contact points. The contact dynamics uses the impulse-based paradigm that, simultaneously with the rugosity degree of the local collision point, generate interesting and visually realistic animations. 00_MSc_pinto Luis Arthur Ferreira PINTO. Autoria grafica de estruturas de documentos Hipermidia no sistema HyperProp. M.Sc.Diss. Port. Presentation: 30/08/00 119 p. Advisor: Luiz Fernando Gomes Soares Abstract: One of the most important challenges in current hipermidia systems is to provide a GUI that helps the visualization of document structure, in order to avoid the common user desorientation. This thesis proposes a graphical tool for structural navigation and authoring of hypermedia documents based on the Nested Context Model (NCM). The key features of this tool, a component of the Hyperprop, are the information filtering mechanisms and the graph layout techniques, both used to improve the structural view legibility. 00_PhD_franca Luiz Paulo Alves FRANCA. Um processo para construcao de geradores de artefatos. Ph.D.Thesis. Port. Presentation: 17/08/00 169 p. Advisor: Arndt von Staa. Abstract: An artifact generator is a software tool which produces an artifact from its high-level specification. An artifact is any item created as part of definition, maintenance or utilization of a software process. Artifact generators are a way to reduce production costs and time to market, and to increase reliability of artifacts within a given application domain. However, in order to achieve a more widespread usage of artifact generators, the difficulty and the cost of their development and maintenance must be reduced. This thesis presents a low-cost artifact generator development and maintenance process. The process departs from an example of a simple artifact within the target application domain. Given this example and considering all possible artifacts of the domain, all commonalties and variabilities are identified, as well as the properties that the specification of each specific target artifact must satisfy. The example is then modified in order to contain specific generator tags at all variable points. These tags establish the transformation rules that must be applied to the specification in order to generate the application. We have used a CASE tool, which allows the programming of the specification editors and of the generators. By means of another tool, the tagged example artifact is transformed into a component, which is combined with the transformation library, yielding the generator code to be internalized by the CASE tool. The process has been successfully used to transform specifications into applications, components and documentation. 00_MSc_silva Marcelo Cabral da SILVA. Indentificacao de estilos de arquiteturas: um processo dirigido por conhecimento. M.Sc. Diss. Port. Presentation: 18/09/00 206 p. Advisor: Julio Cesar Sampaio do Prado Leite Abstract: One of the most recent subjects in the software engineering area, the software architecture, tries to systematize the process of design software, supplying a solid base for the reuse. As more complex systems appear, the importance of system specification and design increases. Therefore software architecture becomes an important discipline [Shaw+94]. Software architecture is related to the system macro-structure. The architectural vision is an abstract vision of the relationship among the elements comprised by a system, as well as the topology they make up. Algorithm detailment or data representation are not the goals of this thesis. Instead, the composition of the elements that form the architectures and their restrictions will be analyzed [Bass+98]. According to our understanding, software architecture is a collection of components, connectors and their behavior restrictions. In other words, software architecture is the collection of the first decisions to be taken during the design phase of a system. In general, these decisions are the most difficult to be taken. After system development, these decisions are the most difficult ones to be modified. Our objective is to provide a method to select architectures based on a requirement list. For such we will count on a knowledge base, as well as on a method. In order to generate the knowledge base, a taxonomy will be defined and based on it relationships among elements will be identified. The collection of relationship among elements will generate architecture styles. In parallel to the architecture style identification study, we will also use concepts of the Requirements Engineering. Therefore, requirements will be used as elements of architecture classification and subsequent selection. A correlation between requirements and elements will be made, starting from each of the identified architectural style. 00_MSc_bastos Marcelo Pereira BASTOS. Busca tabu reativa paralela com religamento para o problema de Steiner em grafos. M.Sc.Diss. Port. Presentation: 24/04/00 52 p. Advisor: Celso da Cruz Carneiro Ribeiro Abstract: Let G=(V,E) be an undirected graph, where V is the set of nodes and E the set of edges. Let c : E - be a function which associates a weight [symbol] to each edge [symbol] . Given a subset X [symbol] V of terminal nodes, the Steiner problem in graph defined by G, X and c consists in finding a connected subgraph of G spanning all nodes in X such that the sum of its edge weights is minimum. In this work, we present a tabu search algorithm that uses advanced techniques such as reactive tabu search and path-relinking. The reactive tabu search handles dynamically the tabu tenure in order to avoid cycling. The reaction and escape mechanisms used by the reactive tabu search are described. Path-relinking is an intensification strategy that seeks to incorporate attributes of high quality solutions (elite solutions) visited along the search trajectory. We define randomized algorithm based on the shortest path heuristic is proposed for diversification. Computational results are reported for benchmark instances of the Steiner problem in graphs found in the literature. The results are compared with those obtained by other heuristics. We also present a parallelization strategy for the proposed algorithm and computational results obtained by this strategy are reported. 00_PhD_prais Marcelo PRAIS. Estrategias de variacao de parametros em procedimentos Grasp e aplicacoes. Ph.D.Thesis. Presentation: 07/24/00 101 p. Advisor: Celso da Cruz Carneiro Ribeiro Abstract: GRASP is a greedy randomized adaptive search meta-heuristic for combinatorial optimization problems. It is an iterative process, that consists of two phases: in the first phase, a feasible solution is constructed; in the second phase, a local search is performed in order to improve the solution found at the first phase. One of the objectives of this thesis is to apply GRASP to the solution of the matrix decomposition problem that arises in the context of TDMA traffic assignment in satellite communications. Two original neighborhoods are convinced and explored in the local search phase of GRASP. Additionally, a new procedure named Reactive GRASP is proposed, where the main GRASP parameter is self-adjusted according to quality of the solutions obtained in the previous iterations. Based on the results obtained with the Reactive GRASP for the matrix decomposition problem, the behavior of GRASP is extensively studied for some of the classical problems. The performance of the algorithm is evaluated for different strategies of parameter variation, in contrast with single and fixed value of the parameter along the iterations. The performance of GRASP is evaluated taking into account the matrix decomposition problem, set covering, W-MAXSAT, and graph planarization. 00_MSc_machado Marcos de Carvalho MACHADO. Segmentacao de dados sismicos via Hyperstack para visualizacao. M.Sc.Diss. Port. Presentation: 27/04/00 115 p. Advisor: Marcelo Gattass Abstract: Applying volume rendering techniques to seismic data makes it possible to visualize the whole data volume at once, which may represent again for seismic interpretation when compared to traditional techniques based on two-dimensional cross-sections. However, it has been verified that the usual opacity function definition scheme is not suitable to this task due to its inability to isolate an individual seismic event. The approach followed in this work in order to generate a good seismic volume rendering image consists in investigating automatic or semi-automatic segmentation methods to allow isolating seismic events. This work studies the utilization of the Hyperstack and the Probabilistic Hyperstack methods, both based on Scale Space Theory concepts for multidimensional image segmentation. The work also proposes a variation for both methods which incorporates the Region Growing strategy. An analysis of the methods is done using a series of two-dimensional synthetic images. This makes possible to understand the basic behavior of the algorithms in situations where the image complexity is controlled and the inspection of the results is simple. Finally, the methods are tested with two-dimensional real seismic data. 00_MSc_antonacci Meire Juliana ANTONACCI. NCL: Uma linguagem declarativa para especificacao de documentos hipermidia com sincronizacao temporal e espacial. M.Sc.Diss. Port. Presentation: 19/04/00 100 p. Advisor: Luiz Fernando Gomes Soares Abstract: The main goal of this work is the creation of a declarative language for the specification of hypermedia documents. At first, an analysis of several existing languages is made. Some fundamental characteristics of hypermedia languages are then recognized. Indeed these characteristics derive from hypermedia models the languages are based on. The more expression power models have the more authoring resources they provide allowing the development of complex documents. After analyzing these languages and describing the in models characteristics an analysis and refinement of the NCM hypermedia conceptual model is made with the purpose of providing NCM with all desirable characteristics found in the language based on NCM will then provide the same characteristics. The language developed, called NCL, was specified using the XML standard, which defines a markup language recommended by W3C. Through NCL it is possible to specify generic hypermedia documents containing the previously mentioned characteristics, including time and space synchronization relations among their components - this is the main contribution of the present work. Another contribution to be pointed out is the proposal for an extended SMIL, a hypermedia language standardized by W3C, aiming at repairing some SMIL shortcomings and offering the same advantages of NCL. 00_MSc_lemos Melissa LEMOS. Gerenciamento de memoria para comparacao de biossequencias. M.Sc.Diss. Presentation: 09/14/00 127 p. Advisor: Sergio Lifschitz Abstract: Many genome projects are currently being developed. The number of nucleotide and aminoacid sequences and related annotations are increasing very fast due to recent advances in the underlying technology. There is a very large amount of information obtained from many organisms, including the human genome. Thus, it becomes very important to use a database management system for efficient storage and access to all this information. Besides this, the development and efficient use of data analysis' algorithms are also important research problems. Among these, one of the most important is sequence comparison, as this operation is the basis for other more complex manipulations. There are two main algorithm families for sequence comparison with the sequences stored in databases: FAST and BLAST. The latter is the family most used in practice. The goal of this work is to present sequence storage structures for memory and buffer management, which use a page replacement ad-hoc policy that we call here Ring Management. As so, we claim that the BLAST family comparison algorithm performance may be improved. Indeed, we have implemented a simulation of the Ring Management and the experimental results obtained confirm it. 00_MSc_sa Michelle Santos SA. Uma abordagem online para comparacao de juncoes paralelas. M.Sc.Diss. Presentation: 15/12/00 71 p. Advisor: Sergio Lifschitz Abstract: Different approaches for the parallel execution of the relational join operator were proposed in the literature. Implementations were developed in order to compare and to evaluate these strategies in a parallel environment. However, there is no agreement concerning the strategies, with or without load balancing, with respect to several real conditions where these strategies are executed. This work presents an approach to the parallel join problem from the online algorithms theory point of view. We use competitive analysis to evaluate and compare the behavior of strategies compared to the offline (optimal) one. Instead of an analytical study, common to the online context, this dissertation presents an implementation and a simulation to evaluate an online parallel join strategy with load balancing with respect to workload distribution. As adversaries we consider a variation in the number of parallel processors available and the skew on join attribute data values. 00_MSc_rabelo Patricia Garces RABELO. Um provador interativo de teoremas. M.Sc.Diss. Presentation: 25/10/00 103 p. Advisor: Edward Hermann Haeusler Abstract: Developing formal specification of a program has become a common practice in recent years due to the software technological advance. Thus, it increases the necessity of making theorem proofs to assure the system specification consistency. However, it is not only formal specification developing complex, but also the theorem proving. Nevertheless, the literature is weak on offering those tools with real good user interfaces. Thinking about this problem, we have constructed an interactive intuitionist first order theorem prover, attending major interactive human-computer principles. The proposed theorem prover, shows a graphical representation of the proof process. The inreaction between the user and the prover is characterized by the choice of the strategy among the ones suggested in order to finalize the proof. Concluding this work, we present several case studies showing that our proposed goals were met. 00_PhD_cerqueira Renato Fontoura de Gusmao CERQUEIRA. Um modelo de composicao dinamica entre sistemas de componentes de software. Ph.D.Thesis. Port. Presentation: 21/08/00 104 p. Advisor: Roberto Ierusalimschy Abstract: Different component systems, such as CORBA, COM, and Java, have different object models and type systems. Such differences make the interoperability between components of distinct systems more difficult, and thus are an obstacle for component reuse. In this dissertation, we argue that an interpreted language with a specific set of reflexive mechanisms, together with a type system with structural compatibility, offers a composition mechanism suitable for dynamic component connection and for interoperability between different component systems. This composition mechanism performs at runtime the tasks of verifying types, connecting adapting and implementing components, and handles components of different systems in a uniform way, allowing them to be connected transparently. The proposed composition mechanism is based on a model that favors flexibility at runtime. This composition model is composed of two major elements. The first one is an object model, defined in order to represent components of the different systems addressed in this dissertation. Thus, this object model performs the role of a unifying model, that is, a model in which objects from different systems can interact and be represented transparently. The second element of our composition model is a design pattern to implement bindings between interpreted languages and component systems. This design pattern, named Dynamic Language Binding, does not use the traditional stubs technique. Instead of this, it uses reflection and dynamic typing to implement generic proxies, which can represent any component of a specific system, and generic adopters, which allow component implementations using the composition language itself. In order to validate our proposal, we describe the LuaOrb system, which is an implementation of our composition model. LuaOrb uses the interpreted language Lua as its dynamic composition language, and integrates the systems CORBA, COM and Java. 00_MSc_assis Rodrigo Lemos de ASSIS. Facilitando a percepcao em ambientes virtuais de aprendizado atraves da abordagem groupware. M.Sc.Diss. Port. Presentation: 18/04/00 144 p. Advisor: Hugo Fuks Abstract: Recent research on groupware and learningware presents some approaches for managing awareness information in order to enhance cooperation on distributed systems. This work proposes ways to support awareness on learningware environments. The study of awareness in groupware is the first step of this work, which leads to relevant aspects regarding awareness in learningware. Some of these concepts are implemented in the learningware environment AulaNet. 00_MSc_toledo Rodrigo Penteado R. de TOLEDO. QUADLOD: uma estrutura para a visualizacao interativa de terrenos. M.Sc.Diss. Port. Presentation: 11/04/00 122 p. Advisor: Marcelo Gattass Abstract: This work focuses on the use of hierarchical structures to interactively visualize terrain. Our purpose is to achieve real time three-dimensional visualization of terrain models obtained from a geographic information system. These models can represent a broad area of the Earth's surface with detailed information, thus yielding very large data sets. For this reason, it is necessary to use a compatible structure to store and query these data. In this work, we present a new structure which combines advantages of different structures provided in the literature. We also discuss an implementation of these ideas and series of tests, drawing several conclusions. 00_PhD_pinto Sergio Crespo da Silva PINTO. Composicao em WebFrameworks. Ph.D.Thesis. Port. Presentation: 28/08/00 150 p. Advisor: Carlos Jose Pereira de Lucena Abstract: This thesis presents all identified problems related to the composition of WebFramework in the Web Platform. The problem was approached through a domain specific language (DSL) called WebCompose developed for this purpose. The theoretical background used for the design of WebCompose comes from the Abstract Design Views and Abstract Design Object which emphasizes separation of concerns and define a set of properties for the realization of flexible design. The case study illustrates the use of all the commands in the DSL and all composition situations identified. 00_MSc_canuto Suzana de Abreu Accioly CANUTO. Heuristicas de busca local para o problema da arvore de Steiner com premios. M.Sc.Diss. Port. Presentation: 17/03/00 58 p. Advisor: Celso da Cruz Carneiro Ribeiro Abstract: Given an indirected graph with prizes associated with its nodes and costs associated with its edges, the prize-collecting Steiner tree problem consists in finding a subtree of this graph which minimizes the sum of the costs of its edges plus the sum of the prizes of the nodes not spanned. This problem has great importance in telecommunication local access network design. One of its applications in this field consist in balancing the potential revenue that can be obtained by providing services to customers and the cost of building the network to do so. Despite its importance, few algorithms have been proposed to solve the prize-collecting Steiner tree problem. This dissertation presents a multi-start local search algorithm using perturbed prizes. Path-relinking is used to improve solutions found by the local search and variable neighborhood search is used as a post-optimization procedure. Computational experiments reported show that the algorithm finds optimal solution on nearly all of the instances tested. A parallel algorithm that uses the MPI standard to provide message passing is also proposed. This parallel algorithm is an extension of the sequential version and finds even better solutions with reasonable processing times. 00_PhD_batista Thais Vasconcelos BATISTA. Luaspace: um ambiente para desenvolvimento de aplicacoes baseadas em componentes. Ph.D.Thesis. Port. Presentation: 20/22/00 166 p. Advisor: Noemi de La Rocque Rodriguez Abstract: Object integration models, also known as component models, present important abstractions to facilitate the development of distributed applications formed by reusable components. These models offer support to reduce the complexities associated with distributed programming and with the use of heterogeneous components. In spite of this support, the component models do not provide facilities to describe the global organization of an application. In order to address this issue, we have adopted the configuration-based approach to define the structure of component-based applications. In the configuration-based approach configuration languages are used to define the structure of an application. Classically, configuration-based systems follow the Configuration Paradigm, which uses compoenets whose interface descibe the services required by clients and those provided by servers. Therefore, component interaction are described in the configuration program, merely binding required and provided services. This configuration model is not compatible with components whose interface do not provide information about required services and presents limitations regarding dynamic evolution of application and capacity of adapting components to bridge compositional mismatches. In this work, we argue that a different approach, using an interpreted and procedural language as a tool for application configuration, can address this aspect and provide support for an important requirement of nowadays applications: dynamic reconfiguration. We present an environment, Luaspace, for configuring and reconfiguring component-based application. Luaspace combines the use of heterogeneous and distributed components provided by component models with an interpreted and procedural language, Lua and a set of tools based on this language. The environment offers mechanisms for composition and evolution of aplications. In this environment, components, scripts and glue code are the elements that form an application. Luaspace allows adaptation of components that need to bridge compositional mismatches and offers support for programmed, ad-hoc, structural and functional reconfiguration and also for dynamic installation of new components. 00_MSc_kraiser Tula KRAISER. Servico de tolerancia a falhas para o ambiente CORBA. M.Sc.Diss. Port. Presentation: 17/11/00 112 p. Advisor: Noemi de La Rocque Rodriguez Abstract: The component programming model has been shown as a good paradigm to program distributed systems. The CORBA middleware, which is specified by OMG allows the programming of components in several programming languages and for diverse operational environments. The components programmed in CORBA benefits from the use of services like the transaction service, name service, and event service. The OMG has specified a new fault tolerant service to serve applications that require a high degree of reliability. The objective of this work is to evaluate the OMG proposal for the fault tolerant service. To achieve this goal, we have developed a prototype of a fault toleranc infra-structure for active replication. Using communication group technology and CORBA services, we have implemented the necessary components to the infra-structure using the interfaces defined by OMG. 00_MSc_silva Viviane Torres da SILVA. ContentNet: um framework para interoperabilidade de conteudos educacionais utilizando o padrao IMS. M.Sc.Diss. Port. Presentation: 24/03/00 92 p. Advisor: Carlos Jose Pereira de Lucena Abstract: The World Wide Web original organization assumed that its stared information was to be read by machines without concern to the possibility of machines interpreting this information. The growth of the amount of information available on the Web made it more complex to search, have access to and to obtain information on the Web. The cope with this problem the platform IMS has been proposed. A consortium formed by academic institutions, companies and government agencies is developing the platform. The goal of the platform is to promote the viability and growth of online activities in the area of education. One of the most important activities of the consortium in Internet based learning with the possibility of interoperability among different environments for Web-based education. Based on the proposed platform and the need for content exchange among different AulaNet servers we have developed on object oriented framework to enable the description, localization and use of educational content available in servers which are compatible with the IMS platform. _______________________________________________________________________ 2001 01_MSc_medeiros Adriana Pereira de MEDEIROS. Especificacao declarativa e implementacao de aplicacoes hipermidia na web. M.Sc.Diss. Port. Presentation: 08/06/01 222 p. Advisor: Daniel Schwabe Abstratc: Hypermidia application development is composed of two main steps: the application specification by a design method (such as OOHDM) and its implementation in a programming language (using a support environment). The design specification of a hypermidia application is generally composed of models whose information must be mapped onto the implementation environment primitives. In many environments this mapping is done through the manual translation of the models into application code. This paper presents the declarative language OOHDM-ML, allowing the specification of hypermidia applications designed with OOHDM, and the OOHDM-XWEB environment, created to support the implementation of these applications. The main objective of this environment is to automatically generate the OOHDM design description of a hypermidia application into the format required by the OOHDM-Web 2.0 environment, using its OOHDM-ML specification. This description is generated by the OOHDM-Translation stylesheet implemented in XSLT, a transformation language for XML documents. 01_PhD_barbosa Alvaro Cesar Pereira BARBOSA. Middleware para integracao de dados heterogeneos baseado em composicao de frameworks. Ph.D.Thesis. Port. Presentation: 03/05/01 154 p. Advisor: Rubens Nascimento Melo Abstract: Developing heterogeneous data integration systems in the Internet is a current and complex problem. There are many users with different skills and interests, needing to access different data types and data models, and requiring different functionality from the integration systems. In literature, there are a large number of integration systems reported. However, such systems are either specifically developed according to one application, or extremely generic to cope with different situations. Both cases bring problems when it is necessary to customise them to a new application. In order to conciliate a great variety of existing data, it is necessary to research new models for data integration systems, which could be tailored to a specific context. This is the goal of the present work: the development of the CoDIMS (Configurable Data Integration Middleware System), a flexible and configurable environment for generating configured systems for the integration of distributed and heterogeneous data. The essence of the CoDIMS environment is the Control component, which implements mechanisms for physical and logical configuration. The physical configuration is obtained through the selection and integration of an adequate set of components that implement Data Integration Middleware Services - DIMS. DIMS Components are based on typical data management services, like the existing ones in the DBMSs. In order to provide more flexibility and facility in the adaptation to a specific application, DIMS Components are modelled using the framework technique. The logical configuration is obtained using an original approach, where DIMS services execution schedule is abstraced in a representaion based on the workflow concept. 01_MSc_leite Angela Cristina Juliao LEITE. Weblearning: um ambiente para autoria de aplicacoes de apoio ao aprendizado na Web. M.Sc.Diss. Port. Presentation: 26/10/01 131 P. Advisor: Daniel Schwabe Abstract: The objective of this work is the devepopment of an environment for the authorship of applications to support learning on the Web, named WebLearning. The features of this environment are based on requirements culled from many Instructional Design theories, used by educators in traditional teaching. Consequently, WebLearning is able to support several pedagogical approaches, enabling authors to create content on WebLearning using different instructional styles. The navigational structure of the applications developed with WebLearning is easy to use. The environment also provides tools to make this navigation more efficient, such as annotations - to enable the student to add personalized annotations to the content -, and navigational sessions, structures that allow the student to keep up some navigation he/she made previously. These structures enable the authors to follow the progress of each student while navigating the content. WebLearning's model was buil using the OOHDM methodology. 01_PhD_paiva Anselmo Cardoso de PAIVA. Um metodo para a compressao de dados volumetricos baseado na transformada do cosseno local. Ph.D.Thesis. Port. Presentation: 25/04/01 77 p. Advisor: Marcelo Gattass Abstract: The volume visualization techniques present two classical computer problems: long execution time and large memory requirements. The memory requirements becomes more investigated with the popularization of the distributed applications. This problem is important also, to make possible the access to volume visualization techniques from a personal computer or low end workstation with limited memory. In this thesis we propose a new lossy compression scheme for volumetric data based on the local cosine transform. This method is appropriated for further volume visualization application because it provides good compression rates, minimizes reconstruction errors, and allows local decompression of the volume. We analyze the compression results and estimate some scheme parameter variations, investigating the adaptivity properties and different space decomposition arrangements. 01_MSc_sampaiojunior Antonio Benedito Coimbra SAMPAIO JUNIOR. Especificacao de um protocolo de comunicacao hipermidia com qualidade de servico e modularidade funcional. M.Sc.Diss. Port. Presentation: 12/06/01 176 p. Advisor: Luiz Fernando Gomes Soares Abstract: In recent years an increasing demand for the transmission of multimedia information with guaranteed Quality of Service (QoS) over the Internet has arisen. This has been driven mainly by the fact that since its conception, the world wide network offers only the best-effort service without guarantees of delivery of the data packages transmitted through internet applications. On situations of congestion, the referred packages can be lost or suffer undetermined delays in their transmission what is not desirable in multimedia applications. Driven by the fact that the hypermedia protocol currently used on the WWW, the HTTP, does not offer adequate support for multimedia transmission with guaranteed QoS, this research was developed with the objective of presenting the complete specification of a new hypermedia communication protocol which offers Quality of Service (QoS) treatment, functional modularity and independence between the transport and resource reservation protocols. To develop the protocol here proposed, an analysis of the various hypermedia and multimedia protocols developed by IETF, ITU and WAPForum, was carried out in order to single out their core characteristics and limitations. Based on the findings of this analysis, a new hypermedia communication protocol which possesses the desirable characteristics of the other protocols but does not inherit their limitations was proposed. Various scenarios of use for the proposed protocol were presented, within these, particular emphasis was given to the scenario of document request with Quality of Service; this being the area where the main contribution of this research lies. 01_MSc_alvarez Beatriz Silva Villa ALVAREZ. Edicao tridimensional de fotografias arquitetonicas. M.Sc.Diss. Port. Presentation: 28/08/01 68 p. Advisors: Paulo Cesar Pinto Carvalho (INPA) and Marcelo Gattass Abstract: This work proposes an interactive system that allows modifying a 3D photograph depicting an architectutral scene so that the user may evaluate the aesthetic effects and the impact the changes may cause. The proposed method is based on the existence of three main mutually orthogonal directions of interest in architectural photos. The identification of vanishing points for those directions allows the calibration of the camera used to take the pictures and the insertion of new elements to the scene. 01_PhD_barbosa Carlos Alberto Moreno BARBOSA. Modelagem dinamica de produto baseada na historia de design. Ph.D.Thesis. Port. Presentation: 02/05/01 165 p. Advisors: Marcelo Dreux (DEM/PUC) and Bruno Feijo Abstract: The distributed object approach is applied to many researches in different areas. However, few references can be found regarding its application to design studies. Even fewer are found related the collaborative aspect of design in strongly dynamic situations. The objective of this study is to propose new models and mechanisms to deal with design under a collaborative approach, based on the same stage of the artifact evolution and dynamic features of the processes. Firstly, a model for the artifact properties classification, properties of which are organized in generic and spatial groups, is described. All the model elements are objects. A primary element of information has a value a measure unit to establish its order of magnitude, when necessary. The associations between objects are realized with flexible mechanisms, which keep information about its hierarchy within a tree and allow their organization according to the focus of interest. The artifact evolution is studied and represented through the design history, in which the reassessment of properties where more than one alternative can be followed. Besides object models, studies about paradigms relevant to this research are also developed: interpretive languages, geometric bus, database, and middleware for distributed objects. A tier model, elaborated as a basis for the design integrated environment, is proposed and the principles for such an environment are formulated. A model aiming at integrating design professionals in an organization structure, the activities plans to be developed and design specification are also described. The participation of each of these elements in a job can be observed in the object virtual tree. More than one object tree can be instantiated to the same client and a bi-directional mapping between this representation and the database in a remote server guarantees the consistency of the operations. In the proposed collaborative environment there is no central or local schema showing where the objects are stored in some of the machines. The generated messages can automtically find the addressed objetcs, which allows flexibility to the model in order to support the artifact specification dynamically proposed by the designer. A framework instance is implemented where the elements involved in the integration can be dynamically created to a specific client and stored in any remote machines connected to the network. 01_PhD_baptista Carlos Eduardo dos Santos BAPTISTA. Dificuldades na quebra de codigo de prefixo. Ph.D.Thesis. Port. Presentation: 20/12/01 169 p. Advisor: Ruy Luiz Milidiu Abstract: Huffman's method is widely used for construction of prefix codes for data compression. Huffman's algorithm gives a binary tree of minimum cost among all binary trees whose leaves are the elements of a dictionary. In general, large files that need to be safe against the action of intruders, have to be submitted to two different and successive processes: data compression and encryption. However, data compression methods together with some strategy can also be used as encryption. Huffman's codes have minimum redundancy compared with other prefix free codes, wich suggests the use of Huffaman's algorithms also for encryption. In this report, we examine and revise some of the main results available in the literature related to the difficulty in breaking encryptions based on prefix free codes of variable length, which includes Huffman's codes. During our analyses, we've found some mistakes in the proofs of the NP-Completeness of the problems Subsequence Prefix Code (SPC), Restricted Prefix Encoding (RPE) and Prefix Encoding (PE). In the case of the SPC, an argument used to prove the indutive step of the Claim that affirms that the partition of a plaintext contains only maximal meta_characters is wrong. We've decided to substitute the SPC for the Maximal Subsquence Prefix Code (MSPC) problem, in which we impose that the meta_characters of the elements of the partition have to be maximal. In the case of the RPE, the application of the technic of the proof by induction, used in the main Claim, is also wrong. But in the RPE case, the mistake was solved, because we could correct the proof of the mentioned Claim. The results confirming the NP_Completeness of the Maximal Subsequence Prefix Code (MSPC), Restricted Prefix Encoding (RPE) and Prefix Encoding (PE) show evidences that certain variants of Huffman's algorithm added to certain stategies can also be used for encryption of text files. 01_PhD_cunha Cecilia Kremer Vieira da CUNHA. Um modelo semiotico dos processos de comunicacao relacionados a atividade de extensao a aplicacao por usuarios finais. Ph.D.Thesis. Port. Presentation: 17/09/01 155 p. Advisor: Clarisse Sieckenius de Souza Abstract: Applications that are extensible by end-users represent a proposal to cope with the issue that it is improbable that a software will attend every specific need of each different user within a domain. One of the research areas approaching extensible applications is End-User Programming (EUP). EUP empirical investigations evidence the existence of collaborative practices related to the process of application extension. In these practices, people communicate with each other with various purposes: to help themselves deal with their problems regarding computing, to share ready-made extensions or to build extensions together. Our work complements these investigations by offering a deep and theoretically motivated treatment of the communication phenomena related to the observed collaborative practices. Based on Semiotics theory and Semiotic Engineering, we propose a model of these phenomena, describing its structure and behavior thus supporting a better understanding of the issues related to them. According to the model, we designed a computable language for the representation of extensions in association with corresponding communicative discourse. 01_PhD_braga Christiano de Oliveira BRAGA. Logica de reescrita como um framework semantico para semantica operacional estrutural modular. Ph.D.Thesis. Port. Presentation: 28/09/01 116 p. Advisor: Edward Hermann Haeusler Abstract: The abstraction level provided by programming languages constructs in the past decade has increased substantially. We moved from low level machine programming in card processing systems to executable mathematical specifications using functions, logic specifications with inference rules, or algebraic specifications with abstract datatypes, all running on laptops. The motto is becoming more and more "what" instead of "how". A byproduct of this technology evolution is the transformation technology which moved from assembly to machine level code complication towards-level specification processing. This is accomplished by transformations from that high level language to another language following a refinement process, that is, a process that preserves semantic properties from the source language into the language, in a potential network of transformations. Among the high level frameworks that relate to the above mentioned process, are actions semantics (AS), modular strutural operational semantics (MSOS), and rewriting logic (RWL). AS is a framework for specifying the semantics of programming languages in a very readable way without loosing its mathematical precision. MSOS is an extension to structural operational semantics framework, that manages the pragmatic issue of modularity in a simple and elegant way. RWL is a generic framework to which many logics and specification languages have been mapped to. This thesis applies state-of-the-art transformation technology to the scope of semantic-driven generation of programming languages interpreters. The main contribution of this work is a proven correct mapping from MSOS to RWL and the MSOS-SL Interpreter, an implementation of an interpreter for MSOS specifications. As a case study for the MSOS-SL Interpreter, we have implemented the Maude Action Tool (MAT), an executable environment for AS using the MSOS specification of action notation. We have coded our tools in Maude, an efficient implementation for rewriting logic, with a fairly intensive use of its meta-programming capabilities. 01_PhD_szwarcman Dilza de Mattos SZWARCMAN. Dead reckoning orientado a metas para personagens reativos. Ph.D.Thesis. Port. Presentation: 14/03/01 113 p. Advisor: Bruno Feijo Abstract: This work proposes a new type of dead reckoning for share state management of networked virtual environments. By exploring the idea of clone autonomy, the goal-oriented dead reckoning is particularly advantageous for the state management of reactive characters. In this new version, state predictions, typical of the dead reckoning technique, become a matter of goal fulfillment. The goal-oriented state management maintains physical and mental state consistency through mechanisms that favor scalability and visual soundness - a concept proposed in this work to better evaluate the quality of a network virtual environment. As a generalization, the new dead reckoning formulation includes previous versions as special cases. 01_PhD_barboza Eduardo Uchoa BARBOZA. Algoritmos para problemas de Steiner com aplicacoes em projeto de circuito VLSI. Ph.D.Thesis. Port. Presentation: /05/01 100 p. Advisor: Marcus Vinicius Soledad Poggi de Aragao Abstract: The first part of this thesis addresses the Packing of Two Steiner Trees Problem, wich is NP-hard and has applications on VLSI layout. We present a new formulation and a theoretical investigation on the corresponding polyhedra. New families of valid inequalities, including proven facets, are identified. We also discuss branch-and-cut implementation issues. The resulting code has a good practical performance and can solve some instances with up to 10.000 vertices in minutes. The second part of the thesis addresses the Steiner Problem in Graphs, wich is also NP-hard. VLSI layout applications yield Steiner instances considered hard to be solved by current methods. In particular, preprocessing techniques developed for Steiner problems over general graphs are not likely to reduce significantly such instances. We propose a new preprocessing procedure effective for VLSI problems. Testing this procedure on real-world instances available on the web, significant reductions were obtained within reasonable computational times. Finally, we study the exact solution of the Problem in Graphs. We analyze a well-known dual heuristic for this problem and propose three new dual heuristies to improve its performance. These heuristics lead to two exact algorithms, a branch-and-bound without linear programming and a branch-and-cut. The algorithms were tested over instances from the literature and obtained good results, solving several instances for the first time. 01_PhD_porto Fabio Andre Machado PORTO. Estrategias para execucao paralela de consultas em bancos de dados cientificos distribuidos. Ph.D.Thesis. Port. Presentation: 16/04/01 152 p. Advisors: Rubens Nascimento Melo and Patrick Valduriez (INRIA) Abstract: This work was conceived based on the data integration architecture proposed by the Le Select system which is an integration middleware developed to support scientific applications. The goal of this work is to propose an execution strategy for distributed scientific queries that minimize their elapsed-time and maximize resource utilization. As a result of this research, we propose an original strategy for the generation of parallel execution plans. Considering the assumption that the execution of scientific programs present a per tuple cost extremely higher than of traditional relational operations, we adopt a strategy that aims into reducing the number of programs invocation during the execution. To that end we conceive some special execution techniques. The result of applying such techniques are reflected in the query execution plan where the placement of the expensive program invocation is flexibilized, making it possible to adopt heuristics for parallel query plans generation, as well as the hybrid strategy for expensive predicates. This strategy combines pipelined and independent inter-operator parallelism providing an execution profile that is adpated to variations of estimated parameters during query execution. As an alternative to the hybrid strategy, we propose the Picking Cherry approach for the execution of expensive predicates based on the analysis of input data dependency. 01_PhD_szenberg Flavio SZENBERG. Acompanhamento de cenas com calibracao automatica de cameras. Ph.D.Thesis. Port. Presentation: 19/12/01 159 p. Advisors: Marcelo Gattass and Paulo Cesar Pinto Carvalho (INPA) Abstract: In the casting of sports events, it has become very common to insert synthetic elements to the images in real time, such as adds, marks on the field, etc. Usually, this insertion is made using special cameras, previously calibrated and provided with features that record their movements and parameter changes. With such information, inserting new objects to the scene with the adequate projection is a simple task. In the present work, we will introduce an algorithm to retrieve, in real time and using no additional information, the position and parameters of the camera in a sequence of images containing the visualization of previously-known models. For such, the method explores the exixtence in these images of straight-line segments that compose the visualization of the model whose positions are known in the three-dimensional world. In the case of a soccer match, for example, the respective model is composed by the set of field lines determined by the rules that define their geometry and dimensions. Firstly, methods are developed to extract long straight-line segments from the first image. Then an image of the model is located in the set formed by such segments based on an interpretation tree. With such information, the segments that compose the visualization of the model are readjusted, resulting in the obtainment of interest points which are then passed to a proceeding able to locate the camera responsible for the model's visualization. For the second image on, only a part of the algorithm is used, taking into account the coherence between the frames, with the propose of improving performance to allow real-time processing. Among several applications that can be employed to evaluate the performance and quality of the proposed method, there is one that captures images with a camera to show the on-line functioning of the algorithm. By using image capture, we can test algorithm in a great variety of instances, including different models and environments. 01_MSc_dahis Gilda DAHIS. Um modelo para a especificacao de cenarios de interacao. M.Sc.Diss. Port. Presentation: 14/10/01 177 p. Advisors: Clarisse Sieckenius de Souza, Simone Diniz Junqueira Barbosa Abstract: Since there are many complementary approaches that aim to contribute to support human-computer interaction design, it is worth use the best of some of them to treat questions less explored. The lack of research about modeling in abstraction levels intermediary between task and interface levels in addition with the fact that despite of the constant development of new technologies the existing models privilege specifc interface styles I-O devices make interesting the investigation of modeling in an abstraction level intermediary between taks and interface, that focus on representations less dependent of implementation technologies. An abstraction level that represents the possible conversations between user system, wich we call interaction level, seems to satisfy those requirements. We believe that the conversation can represent an interactive system execution, independently of the interface implementation tecnology, because, in fact, this execution always consists in the information exchange between this system and its user or group of users. By means of the recognized value of scenarios as design tools, in this work we present a scenario specification model in this abstraction level little explored. As the majority of the scenario specification models are more concerned with the use of scenarios on requirements elicitation activities, the design process part that focus on user/client to designer communication, we concentrate on exploring the use of scenarios on the other side of this communication. Following the Semiotic Engineering approach, wich considers that software applications are messages, sent from designers to users, we propose a model for the specification of scenarios as part of the content of these messages. 01_MSc_sardinha Jose Alberto Rodrigues Pereira SARDINHA. VGROUPS: um framework para grupos virtuais de consumo. M.Sc.Diss. Port. Presentation: 20/02/01 107 p. Advisor: Ruy Luiz Milidiu Abstract: The main goal of this framework is to create instantiated applications that explore the power of the virtual consuming groups. Many benefits can be obtained with the union of these consumers, such as: reduced prices; gifts; auto-financing; time-sharing; and much more. An online retailer is able to create marketing campaings more efficiently with the framework VGroups. Nevertheless, we had to combine some theories in order to achieve these results. The theories we used were: network marketing; specialized offerings; segmentation; and product aggregation and desegregation. With this combination, we were able to create two group applications,LotOnline and Bundless.com. LotOnline is an online Brazilian lottery that allows the creation of two different consuming organizations called bolao. The results of some tests were able to show us how important the consuming organizations can influence the adhesions to tha groups. The application Bundless.com used some data mining techniques and evolutionary software agents to build a tool that automatically generates offerings. This tool proved to be very useful, because it is able to select the best offerings to the future consuming groups. 01_MSc_fernandes Jose Rodrigues FERNANDES. O algoritmo homofonico canonico para cifragem e compressao simultaneas. M.Sc.Diss. Port. Presentation: 30/03/01 131 p. Advisor: Ruy Luiz Milidiu Abstract:In this work, we propose the HC algorithm that simultaneously compresses and encrypt. It aims at the transmission and storage of large textual files. It is based on variable length Homophonic Substitution. We use canonical Huffman coding to represent the homophonic symbol set. Furthermore, we adopt a secret key at the final permutation stage of the encryption scheme. We compare HC against a conventional file compressor. Our experiments indicate a compression loss smaller than 6%. The time performance of HC is close to the similar measure for the conventional compressor. These results indicate that HC provides a good balance of quality and efficiency. 01_MSc_nogueira Leticia NOGUEIRA. Um ambiente de gerencia de aplicacoes CORBA. M.Sc.Diss. Port. Presentation: 29/03/01 100 p. Advisor: Noemi de La Rocque Rodriguez Abstract: Component-based programming is nowadays an important trend for distributed systems development. Using component-based programming, we can subtitute or add new components in applications and also reuse tested components. Among other component systems, CORBA has become the most popular. It's widely accepted because it's an open specification, supported by a large number of programming languages, available from a large number of vendors (academics and commercials). However, the success of CORBA distributed systems depends upon its reliability and so depends upon management support. Despite the large number of ORBs available, few of them are concerned with the growing necessity of administration tools to help systems development and maintenance. This work presents Cometa, a LuaCorba tool for managing CORBA applications that supports the administration and reconfiguration of dynamic systems. With Cometa, we can use, in a simple and integrated manner, CORBA services that are important to administrate systems: the Interface Repository, the Trading Service and the Implementation Repository. 01_MSc_miceli Lucia Aparecida MICELI. GROUPLORE: um ambiente de apoio ao compartilhamento de conhecimentos. M.Sc.Diss. Port. Presentation: 21/11/01 137 p. Advisor: Clarisse Sieckenius de Souza Abstract: Grouplore is a tool for interdisciplinary research groups. Grouplore supports the organization, storage, retrieval and maintenance of an annotation base spontaneously composed by group members, directly or indirectly, about the content of documents. One of the distinctive features of Grouplore is the annotation classification mechanism. Grouplore has two important characteristics aiming at specific needs of the kind of user it is aimed at. The first, the classification of the annotations is extracted from a group model. The second, since it deals with research groups, its features are neither fixed, nor definitive. Both the annotator's or the group profile can change. These characteristics point out at an important facet of Grouplore, the retrieval mechanism. As the purpose of the environment is to support the dynamic nature of a research group, Grouplore works with the notion of Eras. An Era is characterized by a period of time in which certain features are kept invariable. 01_MSc_callado Luiz Fernando Martins CALLADO. WebBBD-XML - Uma arquitetura de um sistema de banco de dados baseada em XML para extracao de conteudo da web. M.Sc.Diss. Port. Presentation: 21/12/01 127 p. Advisor: Rubens Nascimento Melo Abstract: The extraction of Web contents has been the target of investments made by the scientific community and by many software development companies, not only in Brazil but also abroad. The aim of Web pages content extraction is to identify relevant information in a subject domain. The extraction technique allows for building Web-oriented database systems targeting a specific subject domain. These databases are known as webbases. Building them is a hard job as long as those information retrieved from the Web must be stored in the database. As a consequence its maintenance is extremely difficult because the Web pages are changed frequently. As the volume of those information is extremely large and there is a lack of criteria to organize them, accessing those information in the webbase is not an easy task as well. This master thesis proposes the use of EebBD-XML, a XML-based database system architecture that aims to allow Web contents extration and storage in webbases. The architecture is totally XML-based and has two major components: the first one is reponsible for building the webbase and is called Crawler Module, once it implements a semantic crawler to select the information needed; the seccond one analyses and selects the information stored in the webbase, in order to answer users search requirements. The information analysis in the webbase is made through the use of an Ontology Model wich reflects the concepts and their relationships pertaining to the domain. An Ontology Model, wich is proposed for each domain, supports the extration process by detecting interesting information. The architecture is flexibilized towards its support to differnt subject domains through the specification of a specific Ontology Model. 01_PhD_cysneiros Luiz Marcio CYSNEIROS. Requisitos nao funcionais: da elicitacao ao modelo conceitual. Ph.D.Thesis. Port. Presentation: 08/02/01 224 p. Advisor: Julio Cesar Sampaio do Prado Leite Abstract: The development of complex information systems calls for conceptual models that model aspect beyond entities and activities. In particular, recent research has pointed out that conceptual models need to model goals, in order to capture the intentions which underline complex situations within an organizational context. One particular class of goals is named non-functional (NFR) which need to be captured and analyzed from the very early phases of the software development process. Errors caused by not conveniently dealing, not dealing at all with NFR are among the most difficult and expensive to correct. In spite of that, a few proposal are met on the literature to systematically deal with NFRs. This thesis broaches two aspects on dealing with NFR that were not covered in the literature: how to elicit NFR and how to integrate them to the conceptual models. To do that, we propose a strategy that deals with NFR during the early phases of software development and integrate these NFR to conceptual models. The thesis was validated by three case studies. The results of these case studies suggest, that the use of the proposed strategy can lead to a final conceptual model with better quality as well as to a more productive software development process. 01_MSc_moura Marcel Stanley Albuquerque de MOURA. Relacoes espaciais em documento hipermidia. M.Sc.Diss. Port. Presentation: 31/08/01 152 p. Advisor: Luiz Fernando Gomes Soares Abstract: Hypermidia Authoring is a very complex process. Along with the definition of the document's logic struture, the specification of its temporal and spatial synchronization plays a major role. Spatial synchronization deals with the document's objects visual placement within their presentation devices and the definition of the spatial dynamics that should take place on those objects throughout a presentation. This work presents a spatial synchronization model for hypermidia documents along with a definition of a set of spatial relationships. An instantiation of this model into NCM (Nested Context Model), and its extension through a definition of spatio-temporal relationships and animations are also presented. Finally, some features of a synchronization tool are showed. 01_MSc_amaral Marcelo Jaccoud AMARAL. Mantendo codigo-fonte em XML - rumo ao hiperfonte. M.Sc.Diss. Port. Presentation: /05/01 240 p. Advisor: Edward Hermann Haeusler Abstract: The file format used today to maintain source code hampers both manual and automatic edition. Using the Extensible Markaup Language (XML) and an appropriate document model designed for the programming language, the source code can be turned into a hypersource which exhibits several avantages over the plain text version. We present methods for designing the XML document models, creating translation programs to import traditional source code (using TXL), and building other tools to work with the hypersource. Examples of document models for TXL, Fortran and C++ are given, together with the import programs, style sheets and several transformations written in XSLT, a language designed to transform XML files. 01_MSc_bezerra Marcia Martyres BEZERRA. Um modelo hierarquico para deteccao de regioes regulatorias em sequencias de DNA. M.Sc.Diss. Port. Presentation: 28/11/01 98 p. Advisor: Marcus Vinicius Poggi Abstratc: The main purpose of this dissertation is to study methods for the detection of regulatory regions in DNA sequences. It presents a hierarchical method for this purpose, where mechanisms for introducing correlations between different components of these regions are available. This dissertation also presents a language for a precise description of the signs that characterize regulatory regions. The method proposed was implemented and tested on primate lentivirus DNA sequences. The results reported lead to the conclusion that the methodology proposed is adequate, moreover they indicate that it supports throughout analysis. 01_PhD_lima Maria Julia Dias de LIMA. ORFEO: programacao distribuida orientada a eventos com funcoes e continuacoes como valores de primeira classe. Ph.D.Thesis. Port. Presentation: 20/07/01 108 p. Advisors: Noemi Rodriguez, Roberto Ierusalimschy Abstract: In this work, we support the thesis that functions and continuations as first-class values constitute a good basis for abstractions that benefit from the asynchronous behavior found in event-oriented distributed programming. We have proposed and implemented ORFEO, a system based on the Lua programming languages, wich assigns to remote functions the same first-class status of local functions. ORFEO also has the property of manipulating continuations as first-class values, allowing the capture of what remains to be executed in the processing of an event. The union of these two properties - remote functions and continuations as first-class values - in an event-oriented communication context allows a developer to build new abstractions, such as distributed objects and synchronization mechanisms, without the need of any special primitive in the system. The result combines the expressive power of a procedural language that offers remote functions and continuations as first-class values with the flexibility of functional techniques and extension mechanisms. 01_MSc_jacyntho Mark Douglas de Azevedo JACYNTHO. Um framework e uma arquitetura para implementacao de aplicacoes hipermidia na Web. M.Sc.Diss. Port. Presentation: 20/09/01 341 p. Advisor: Daniel Schwabe Abstrat: In the competitive environment of the Information Economy, programming productivity is indispensable for success. In this context, the creation of tools that automate software development is crucial. This dissertation proposes an object (component) oriented framework and an architecture for the implementation of hypermidia applications in the World Wide Web, using the J2EE (Java 2 Enterprise Edition) plataform. This framework is based on the OOHDM method and on the traditional Model-View-Controller architecture. In addition, this work brings important Software Engineering principles to web application development, allowing the production of organized, maintainable, and reusable code. 01_PhD_martinez Marta Emilia Barria MARTINEZ. Algoritmos para QoS em redes de computadores. Ph.D.Thesis. Port. Presentation: 19/03/01 173 p. Advisor: Luiz Fernando Gomes Soares Abstract: One of the problems related to the evaluation of networks that provide guarantees of QoS is the lack of simple models that would generate a first approximation of QoS parameters for the different types of traffic transported by the network. The overall objective of this thesis is to suggest some improvements in the techniques used in the transmission of different types of traffic across a computer network in order to achieve a better QoS. To this end, various algorithms, models and performance analysis methodologies are proposed, and the latter are then applied to the performance analysis of the proposed algorithms, calculating the most elevant performance measures in each particular case. The algorithms analyzed include those of the type used in packet-switching networks, such as TCP/IP, as well as those used in label switching networks, exemplified by the ATM network. The QoS parameters of greatest interest were the probability of loss of cells and system delay. Other performance measures were also derived, including various probability distribution function. 01_MSc_mota Oscar Thyago Jose Duarte Dantas Lisboa MOTA. Uma arquitetura adaptavel para provisao de QoS na Internet. M.Sc.Diss. Port. Presentation: 29/05/01 112 p. Advisor: Luiz Fernando Gomes Soares Abstract: The variety of intserv and diffserv possible configurations as well as the various QoS provision techniques used in subnetworks makes it difficult to identify which model and subnetwork technology should be used to accomplish a specific service. Additionally, the fast technological evolution suggests the development of flexible architectures that accommodate, in execution time, adaptations nowadays only possible by less dynamic procedures such as hardware or firmware updates. This paper proposes an adaptable architecture for Internet QoS provision that is independent of service models and provision mechanisms, besides the used provider subnetwork technology. Through a generic framework definition the paper shows how hot-spots can be specialized to implement the two mentioned models used for Internet QoS provision. Following this, the paper proposes a general architecture that allows the QoS provision representation, in host and routers, independent from service models and exisitng subnetworks. 01_MSc_ribeiro Paula Clark RIBEIRO. Modelagem e implementacao OO de sistemas multi-agentes. M.Sc.Diss. Port. Presentation: 26/10/01 106 p. Advisor: Carlos Jose Pereira de Lucena Abstract: This work proposes a mapping between agent-based software engineering elements and object oriented software engineering elements. The mappings are illustrated through the modelling of an Internet marketplace. During the analysis and design phases, the Gaia methodology for agent-oriented analysis and design is used. Gaia methodology views the system as an organization of roles that interact. Two mappings are proposed between the Gaia methodology elements and object oriented programming models. The first one is done to the MAS Framework, a framework to build multi-agent systems, and the other is done to a design pattern to construct multi-agent systems. MAS Framework is applicable to reactive agents and the design pattern is concerned with the development of cognitive agents. Both of them were develped at Teccomm labs. At the end, an analysis of the advantages and disadvantages of mapping Gaia methodology artifacts is presented based on two approaches of developing OO systems. The case studies involve the development of the Babilonia marketplace. 01_MSc_gorini Renata Abrantes de Campos GORINI. Um ambiente desuporte a autoria cooperativa de documentos hipermidia. M.Sc.Diss. Port. Presentation: 11/09/01 132 p. [xiv p.] Advisor: Luis Fernando Gomes Soares Abstract: This thesis describes how a cooperative work can be organized using a task structure. Based on this structure, some mechanisms will be defined in order to support a coordinated collaboration between participants of a work. In this context, a task will represent a cooperative work, while a set of tasks will represent a complex cooperative work. Based on tasks, an environment was proposed to support the cooperativa authoring on hypermidia documents. This environment will allow definitions of a task structure, task relationships (to define rules for the execution of a work, represented by tasks) analogous to workflow systems, and the definition of collaboration, comunication and coordination mechanisms, wich are essentials during cooperative processes. The environment was specialized and implemented for a particular hypermidia system, called HyperProp, using some hypermidia concepts (nodes and links), more precisely some concepts defined by the NCM model (a hypermidia model). 01_MSc_werneck Renato Fonseca Furquim WERNECK. Problema de Steiner em grafos: algoritmos primais, duais e exatos. M.Sc.Diss. Port. Presentation: /08/01 142 p. Advisor: Marcus Vinicius Soledad Poggi de Aragao Abstract: The Steiner problem in graphs is one of the most important NP-hard problems in combinatorial optimization, with applications in VLSI design, networks, logistics and computational biology, among others. We present a computational assessment of several strategies to deal with this problem in practice. The methods are divided into a number of classes: constructive heuristics, local search procedures, metaheuristics, dual heuristics and exact algorithms. Each class has its own way of dealing with the problem: running times, solution qualities and guarantees provided differ among them. We discuss the most relevant algorithms described in the literature and propose new ones. The result is a collection of tools that represent the state-of-the-art in practical algorithms for the Steiner problem in graphs. Several open instances in the literature were solved or had their bounds improved by these methods. 01_MSc_kostin Sergio KOSTIN. Modelo para gerenciamento de dados em um ambiente de comando e controle. M.Sc.Diss. Port. Presentation: 27/04/01 148 p. Advisor: Marco Antonio Casanova Abstract: The development of Command and Control Information Systems is a trend in worldwide Armed Forced. Military Operations are very mobile, subject to interferences of the environment and of the enemy. In short, the communications are not trustworthy. The organization of Military Operations is hierarchical and can change in time. The goal of this dissertation is the definition of a model of data management of mobile computing that meets the requirements of a command and control system for military application. Specifically, the main contributions of the dissertation are: a data replication model; the definition of techniques for treatment of the search and diffusion of the system's information, considering the mobility characteristics, the customer's profile and the location of the stations; and a framework for the implementation of proposed techniques. 01_PhD-silva Sergio Roberto Pereira da Silva. Um modelo semiotico para programacao por usuarios finais. Ph.D.Thesis. Port. Presentation: 30/05/01 208 p. Advisor: Clarisse Sieckenius de Souza Abstract: One of many noticeable trends in the software industry has been to create resources to configure or extend applications available to end-users, in an attemp to broaden the applications' usability, thus satisfying the real need of these users. Unfortunately, simply making such resources available does not enable end-users to extend their applications, for they normally lack the knowledge of how to use them. Adler and Winogrand [ADLER'92] propose that the criterion for the usability of software should be the extent to which it supports the understanding and learning of its working, adaptation and extension. The area of End-User Programming could be seen as an answer to such demand. However, to the best of our knowledge, there is no agreement regarding a theoretical model that reasonably supports the development of extensible applications. In this work we take the view of software as a meta-communication artifact, proposed in the Semiotic Engineering [DE SOUZA'93], and cast its use as a communication process between the software designer and the end-user. Taking up this view, we make use of Jakobson's Verbal Communication Model [JAKOBSON'60] to understand the effect of the functions of language on this process. Supported by this knowledge and by the Semiotic Continuum and Interpretive Abstraction Principles, proposed by Semiotic Engineering [DE SOUZA'01A], we present a theoretical model for the end-user programming task that pursues Adler and Winograd's proposal. This model is composed by a process for the creation of software extensions; a type-language for end-user programming that uses communicative resources to maintain the textual cohesion and coherence, such as the use of anaphors, metaphors and metonymy; and a software architecture with specific mechanism to support the interpretation of instances of that type-language and to maintain the Semiotic Engineering principles previously cited. 01_PhD_fiorini Soeli Teresinha FIORINI. Arquitetura para reutilizacao de processos de software. Ph.D.Thesis. Port. Presentation: /04/01 200 p. [42 p.] Advisor: Julio Cesar Sampaio do Prado Leite Abstract: This thesis present an architecture for the organization and description of software processes. Its main goal is to enable software process reuse. The architecture is organized by three different process types (standard, pattern and usual) and by the concept of process framework. Processes are described according to their process type and by special languages, wich are based on XML (Extensible Markup Language). In order to help the retrieval of process information, the architecture provides a facet classification schema, a 3C-based reuse guide, and a process pattern classification (community, family and individual). A prototype tool implements the architecture. A requirements engineering process framework was built in order to explore the architecture capabilities. A case study on the reuse of requirements engineering process was conducted. 01_PhD_oliveira Toacy Cavalcante de OLIVEIRA. Uma abordagem sistematica para a instanciacao de frameworks orientados a objetos. Ph.D.Thesis. Port. Presentation: 05/10/01 166 p. Advisor: Carlos Jose Pereira de Lucena Abstract: The use of techniques that can improve quality and productivy in the software development process is a key point to Software Engineering. Among those techniques the reuse of well documented past experiences is one of the most efficient to reduce the resources, such as time and cost, used. In this scenario, object oriented frameworks represent a step ahead to reuse in the way that they represent knowledge to a certain application domain where reuse can be broadly obtained. The framework reuse process, called instantiation, is driven by a set of reuse activities that due to their complexity, should be systematically executed in order to produce a significant advantage to the software development process. The main goal of this work is to report a set of techniques such as, framework design documentation and reuse process description language that are integrated in a reuse environment that guide the reuser through the instantation process. _______________________________________________________________________ 2002 02_PhD_silva Adelailson Peixoto da SILVA. Extracao de malhas adaptativas em multi-resolucao a partir de volumes, usando simplificacao e refinamento. Ph.D.Thesis. Port. Presentation: 04/10/02 104 p. Advisors: Marcelo Gattass and Luiz Carlos Pacheco R. Velho Abstract: This work presents a method for extracting multiresolution adaptive polygonal meshes, from volumetric objects. Main aplications of this work are related to medical area, fluid dynamics, geosciences, metheorology and others. In these areas data may be represented as volumetric objects. Volumetric datasets are implicit representations of objects, so it's very difficult to apply directly any process to these objects. Mesh extraction obtains an explicit representation of the object, such that it's easier to process directly the objects. The presented method extracts the mesh from two main processes: Simplification and Refinement. The simplification step extracts a supersampled representation of the object (in the same volume resolution), and simplifies it in such a way to obtain a base mesh (or coarse mesh), in a low resolution, but containing the correct topology of the object. Refinement step uses the distance transform to obtain a multiresolution representation of the mesh, it means that at each instant it's obtained an adaptive higher resolution mesh. The final mesh presents a set of important properties, like good triangle aspect ratio, convergency to the object surface, may be applied as to objects with boundary and as to objects with multiple connected components, among others properties. 02_MSc_pfeifer Alesio Leoni PFEIFER. ALua: um sistema multi-canal baseado em eventos e mobilidade de codigo. M.Sc.Diss. Port. Presentation: 11/07/02 58 p. Advisor: Noemi de La Rocque Rodriguez Abstract: Current interest in wide-area distributed computing has highlighted the need for an adequate programming model for this environment. Because of its asynchronous nature, event-driven programming provides a suitable model for dealing with the failures and delays that are frequent in this contex. Besides that, new services and uses of computer networks often demand a high degree of customizability and increasingly demand more flexibility and extensibility. Recent research has explored the mobile code paradigm, investigating how dynamic code exchange between processess can enhance the flexibility of services and applications. In this work we propose an architecture for distributed multimedia applications based on an event-driven programming model. To avoid the synchronization problems that are inherent to multi-threaded programming, the propose architecture is based on a single-threaded structure. Instead of multithreading, we opted for the event-oriented approach allied to multiple communication channels with user-defined handling procedures to allow the application to deal concurrently with control and data streams. We discuss this programming model, present the system we have implemented based on this model, and describe the experience we have had with this system. 02_MSc_mele Alexandre MELE. Um framework para simulacao de redes moveis Ad Hoc. M.Sc.Diss. Port. Presentation: 30/12/02 109 p. Advisor: Markus Endler Abstract: A mobile ad hoc network consists of a set of portable computational devices, equipped with a wireless communication interface, that are randomly and dynamically distributed in space, and where each host serves as a router for the other hosts by discovering and maintaining multi-hop routes among the hosts. In general, protocols for mobile ad hoc networks are more complex than equivalent protocols for static networks, due to several properties of such networks, such as its dynamic topology, the mutual interference, concurrent access and smaller communication bandwidth of the wireless links, as well as, operation with restricted amount of energy, and scarce resources of the mobile devices. A main focus of research in mobile ad hoc networks has been the development, analysis and comparison of routing protocols for such networks. Therefore, there is some demand for environments that facilitate the rapid prototyping, the simulation and the debugging of protocols at the network and other layers for such networks. These environments should preferably be flexible, easy to use, and allow for the definition of different levels of abstractions for modeling the main characteristics of the mobile network, such as the pattern of mobility, the wireless links, the energy consumption, etc. This thesis describes the design and implementation of a framework for the simulation of mobile ad hoc networks, which aims at supporting the development of concrete simulation environments for prototyping, testing and doing the complexity and performance analysis of protocols for such networks. 02_MSc_azambuja Antonio Carlos Pereira de AZAMBUJA. Multi-resolucao para a visualizacao de reservatorios naturais de petroleo. M.Sc.Diss. Port. Presentation: 06/05/02 138 p. Advisors: Marcelo Gattass and Waldemar Celes Filho Abstract: Current flow-simulation meshes of natural oil reservoirs (NOR) are composed of hundreds of thousands of hexahedral cells. The visualization of the geometry of these cells superimposed with color attributes to represent properties and flow results requires the rendering of an unstructured mesh of millions of triangles. Current graphics hardware does not allow for an interactive visualization of such meshes. As computers and graphics boards increase their processing capacity, simulation meshes also grow and the solution to the rendering problem usually includes acceleration techniques, one of which is multi-resolution (MR). However, currently known MR models are not applicable to NOR simulation meshes due to this field's specific requirements, such as the preservation of the hexahedral-cell model and discontinuities among cells. In fact, MR techniques tend to focus on Realistic Visualization, while the NOR problem is one of Scientific Visualization, for which generic solutions still do not exist. The present work proposes a specific MR model for the visualization problem concerning NOR simulation meshes, in which discontinuous space partition, hexahedral-cell-based semantics and the problem's visualization characteristics are taken into account. The proposed model allows an efficient construction of a MR structure, from which, in real time, adaptive meshes can be extracted that depend on: (a) the geometric error approximation, (b) the view, and (c) the polygon budget. This model can also be used combined with another acceleration technique, frustum culling, which allows for the hierarchical elimination of regions in the mesh that are out of the view volume. The proposed model was implemented in a system on which extensive testing was performed, providing results that allowed us to draw some conclusions and recommendations. 02_MSc_barbosa Clarissa Maria de Almeida BARBOSA. MetaCom-G*: especificacao da comunicacao entre membros de um grupo. M.Sc.Diss. Port. Presentation: 27/03/02 111 p. Advisors: Clarisse Sieckenius de Souza and Raquel Oliveira Prates Abstract: The complexity inherent in the design activity and the challenges proper to the design of multi-user interfaces motivate the development of models, methods, techniques and tools for the support of the design process of multi-user interfaces. MetaCom-G [Prates,1998] is one of such models. Within the Semiotic Engineering framework [de Souza,1993]. Prates proposes that the designer be offered an environment which enables him to describe how members of a group organize themselves in order to perform the group activity. In addition, this environment should provide the designer with qualitative indicators of his description, which will assist him in making his design decisions. Nevertheless, MetaCom-G only allows the designer to describe communication among group members at a high level of abstraction. Consequently, the environment proposed in this model holds little information about the group's communication process to effectively help the designer make his future design decisions. In order to expand MetaCom-G's support to the designer, we extend this model and propose MetaCom-G*. With the extended model, we enable the designer to describe more accurately the communication process of a group. As the extension is based on the Speech Act Theory [Austin,1962; Searle,1969,1979] and Discourse Analysis [Brown&Yule, 1983], the information provided by the environment for the support of the design process of multi-user interfaces proposed in MetaCom-G* leads the designer to reflect on the effect of communication on group members. By doing so, it helps him make his future design decisions regarding this communication. 02_MSc_nehab Diego Fernandes NEHAB. A implementacao da linguagem de programacao Sloth. M.Sc.Diss. Port. Presentation: 12/07/02 58 p. Advisor: Roberto Ierusalimschy Abstract: This work describes a new functional programming language and its implementation. Sloth is a non-strict purely functional interpreted programming language, wich supports high order functions as first class values, structured data types and pattern-matching based function definitions. Sloth was created for academical use in activities related to the use and to the implementation of functional programming languages. Therefore, there has been a great concern with the simplicity of both the language specification and its implementation. As a result, Sloth's syntax is modest, but expressive enough for practical applications. Furthermore, the great simplicity and portability of its implementation has no serious impact on its performance. The main contributions of this work are the design and implementation of Sloth, an optimization to combinator graph reduction as an alternative to the use of Turner combinators, a new architecture for the implementation of interpreted functional programming languages, based on the cooperation between two programming languages, and the performance comparison between Sloth and similar languages. 02_MSc_andrade Diogo Vieira ANDRADE. Um modelo para determinacao de precos para leiloes. M.Sc.Diss. Port. Presentation: 14/08/02 69 p. Advisor: Marcus V. S. Poggi de Aragao Abstract: This work presents a linear programming model for virtual markets. The main objective is to encourage sellers to offer lower prices and buyers to pay more for products. This is done through a system where the agents are awarded for acting this way. The basic version of the model works as an auction where buyers and sellers can make their bids. Some extensions were developed for this basic model, to allow situations like: transportation cost between buyers and sellers; scale economy for the buyer's price; multiples products in the same auction. For the basic model and each of its extensions, the system of awards has been developed based on Game Theory principles. In each case, the model was reinterpreted as a game and its core solutions were found. It was also verified how these solutions determine market's behavior. 02_MSc_poyart Eduardo Ribeiro POYART. IEngine - uma interface abstrata para motores de jogos 3D. M.Sc.Diss. Port. Presentation: 26/04/02 64 p. Advisor: Bruno Feijo Abstract: 3D game engines are APIs that help the development of games with real time 3D graphics. It is convenient to use a 3D game engine when developing games, because all the low-level functionality is already built. However, each 3D engine API is constructed in a different way, causing engines to be incompatible. The learning process is also difficult, because the APIs of some engines are not well designed. This work analyzes this issue by doing a critic survey of some existing 3D engines and proposes an abstraction layer over them, called IEngine. The IEngine layer gives engines a standard API. An application built over IEngine can be compiled with any of the 3D engines adapted to IEngine, and that application could switch engines without any code change. The adaptation of a new engine to IEngine is also facilitated by IEngine's structure of abstract classes (interfaces) and is done by inheritance. In this work, an implementation of IEngine over the open-source and multiplatform engine Crystal Space is presented. Also, it is presented a way to implement IEngine over a second engine, Fly3D. Finally, the possibility of IEngine receiving an interface layer with other programming languages such as Java and Lua is studied, in such a way to allow the quick development of good visual quality 3D games. 02_MSc_rodrigues Elisette Maria Figueiredo de Almeida RODRIGUES. Aplicando o modelo MOF e o XMI para a integracao de bancos de dados heterogeneos. M.Sc.Diss. Port. Presentation: 25/03/02 77 p. Advisor: Rubens Nascimento Melo Abstract: The Information Technology Boom during the last decades, consequence of technology maturation, resulted in a large diversity of implementations technologies. The current software vendors' strategies pass now, due to that variety of implementation technologies used during the "IT rush", through an integration path. The heterogeneity of systems to be integrated, allied to the massive use of the Internet, specially for publishing businesses information, providing the user with a simple and complete assessment of the subjects, and also reducing the information storm that has to be handled, while improving their quality, asks for solutions that allow an easy and sound legacy data integration to new applications implemented using modern technologies. Heterogeneous Database Systems are viewed as a good legacy systems' integration solution, since they can be coupled "as is". However, the majority of newly developed applications use object technology, usually not used in DBMS's. From this, arises the need for integrating OO applications with DBMS's using distinct technologies. This work aims on presenting a model for systems integration, considering the characteristics above, developing a framework to facilitate the construction of specific solutions for integrating OO applications to legacy systems. The framework is based on the MOF Model and XMI, both OMG standards. 02_MSc_santos Fabio Marcos de Abreu SANTOS. Plataformas para jogos de guerra com computacao movel. M.Sc.Diss. Port. Presentation: 10/05/02 117 p. Advisors: Marco Antonio Casanova and Roberto de Beauclair Seixas (IMPA) Abstract: Mobile applications attracted great interest in the last years. Several applications, based on totally new concepts have been created by taking advantage of the obiquitous computation. War games traditionally have helped military training in many countries, bringing to the classroom simulations of the operation theater. Military field exercises are equally necessary to prepare the troop. The use of war games with mobile equipment represents a new field trainning facility for the Brazilian Marines. We belive that war games can not be limited to classrooms, but be extended to the field, where the realism of the operation theater helps the Brazilian Marines to receive a fuller and better training, compatible with the importance of its mission. The main goal of this work is to propose some ways to use mobile devices on Brazilian Marines field exercices integrating it with the war game, wich runs in classrooms. Nowadays, the field exercices use only wired workstations and the server needs to be transported to the exercice's place. The data entry is manual and data are received by radio communications. By developing a WAP application and using tecniques of data summarization and active databases, we will receive the response of each requested message by SMS. 02_MSc_silva Flavio Heleno Bevilacqua e SILVA. Coordenacao de sistemas multiagentes via equilibrio de mercado economico. M.Sc.Diss. Port. Presentation: 19/04/02 134 p. Advisor: Ruy Luiz Milidiu Abstract: Multiagent systems need efficient strategies of coordination to achieve their objectives. When there is a task distributed among agents, their interaction involves an exchange of messages. This problem can be harmful depending on the volume. The strategies of coordination must reduce the volume of messages exchanged and the number of centralized elements without prejudice the system efficacy. The objective here uses the balance of the economic market to develop coordination techniques for multiagent systems with the market-oriented modelling. These techniques reduce the interaction of agents and don't need a central coordinator. We explain this model by using two applications: a resource allocation and a component realocation in consumer groups. Preliminary results compared with another technique of coordination present advantages in some aspects, essentialy in the search for the best possible solution and the measurement of the agents performance. We conclude that this modelling is an efficient strategy of coordination, with easy implementation that, depending on the kind of application, generates better solutions than other existing coordination techniques. 02_MSc_guimaraes Francisco Jose Zamith GUIMARAES. Utilizacao de ontologias no dominio B2C. M.Sc.Diss. Port. Presentation: 19/12/02 195 p. Advisor: Carlos Jose Pereira de Lucena Abstract: The main difficulty associated with the B2C domain is increasing the usefulness of WWW for the electronic trade through the improvement of the services provided to the consumer. Even through the WWW allows the buyer to have access to a great amount of information, to obtain the information from the right supplier that sells the desired product by a reasonable price can be a very expensive task. One of the ways of improving the web functionality is through the use of intelligent agents for search of information, that is, the introduction of purchase agents that aid the buyers to find products of their interest. For that to happen we need to overcome an inherent difficulty of the WWW: the mixture of natural language, images and layout information in HTML is one of the greatest barries for the automation of the electronic trade, because the semantics of the information is only comprehensible for human beings. To solve this problem we hope to produce more efficient purchase agents by associating them to the use of ontologies, and virtual stores that have special annotations that follow ontologies. In the present dissertation we make a study of the main technologies related to ontologies development in computer science. We also develop a case study about the ontologies application to the B2C domain, seeking in this way to evaluate potential and existing difficulties for the development of this type of application. 02_MSc_pierre Gustavo Moreira PIERRE. Ambiente integrado para posicionamento em operacoes militares. M.Sc.Diss. Port. Presentation: 01/07/02 65 p. Advisors: Marcelo Gattass and Roberto de Beauclair Seixas (IMPA) Abstact: With technological development, we have been witnessing a growing popularization of satellite-location equipment (GPS receivers) and pocket computers (PDA) in several areas: air, sea and earth transports, mapping and geoprocessing, sports and entertainment. We can see hikers, for example, acquiring their own GPS receivers. In the military environment, this popularization is also occurring: the Armed Forces in several countries are obtaining increasing numbers of GPS receivers and are testing equipment to interact with them. The purpose of this work is to integrate GPS receivers and PDAs, with the intention of employing them in military operations in order to obtain quick responses for a faster and more efficient use of fractions of involved troops. This work involves the generation of files readable by PDAs, the communication of GPS receivers with PDAs, and an easy and fast visualization of the information provided by the receivers. It also includes a comparison between the use of map in two different representations, raster and vector, both allowing for zooming and movement of the image, as well as the possibility of measuring distances between points of the map in the PDA screen, and the inclusion, exclusion and alteration of several observations. The great advantage of using raster images instead of vector images in PDAs for this purpose was verified, as well as some disadvantages in the use of PDAs, common in military operations, with general-purpose GPS receivers. 02_PhD_mathiasfilho Ivan Mathias Filho. A documentacao e a instanciacao de frameworks orientados a objetos. Ph.D.Thesis. Port. Presentation: 17/04/02 184 p. Advisor: Carlos Jose Pereira de Lucena Abstract: The extreme competitiveness of contemporary economy generates a huge demand for cheaper, efficient and reliable software products, wich often are developed under great pressures of time and budget. These premises suggests that software development must take place in an environment where proved solutions can be modified, combined and adapted to be used in the development of new products. Therefore, the use o object oriented frameworks, or frameworks for short, seems to be one of the most promising techniques for code and design reuse. Nevertheless, one of the obstacles that have to be removed before the widespread use of frameworks is the great amount of time for study necessary to be proficient in the use of a specific framework. This situation occurs due to the inherent complexity of the design of the frameworks, wich are conceived to fulfill the requirements of an entire application family. In this contex, this work presents a detailed description of a high-level framework instantiation process based on the Features Model. This model is responsible for presenting to the application developers a simplified view of the main functional and technological characteristics of a framework. In a following step, the instantiation process modifies the original design of the framework to produce a new design for a specific member of an application family. All of the modification steps will be based on choices available in the Features Model. 02_MSc_magalhaes Joao Alfredo Pinto de MAGALHAES. Um framework multi-agentes para busca e flexibilizacao de algoritmos de classificacao de documentos. M.Sc.Diss. Port. Presentation: 27/09/02 103 p. Advisor: Carlos J. P. de Lucena Abstract: We are living in the information age, where knowledge is constantly being created in a rate that was never seen before. This is mainly due to Internet, that changed all the information exchange paradigms between people. Through the net, it is possible to publish or exchange whole works, reaching an audience impossible to be reached through other means. However, excess of information can be harmful: having too much information can be equal to having no information at all. Our work was to build a multi-agent framework for search and flexibilization of textual document classification algorithms of a specific domain. We have built an infra-structure that separates the concerns of document search and selection (plataform) from the concerns of document classification (an application of the separation of concerns concept). It is possible not only to use existing algorithms, but also to generate new ones that consider domain-specific characteristics of documents. We generated four instances of the framework, a webclipping application, a knowledge management component, a search engine for websites and an application for the semantic web. 02_MSc_cunha Leonardo Magela CUNHA. Formacao de grupos de trabalho utilizando agentes de software. M.Sc.Diss. Port. Presentation: 17/04/02 104 p. Advisor: Hugo Fuks Abstract: Nowadays, the complexity of work and the dissemination of the information and communication technologies value and make group work a potential allied for the organizations. The computer support offered under the umbrella of groupware systems is based on the research from Computer Supported Cooperative Work (CSCW). The support for the workers should be given since the group formation until its dissolution, going through the group work itself. In Software Engineering, multi-agent systems provide a properly level of abstraction for the treatment of complex and distributed problems. One of these problems is characterized by environments for group work and/or learning in groups. In this dissertation it is presented the study and how a multi-agent system was implemented in AulaNet to help group formation. 02_MSc_lima Luciana dos Santos LIMA. Um framework para provisao de QoS em redes moveis sem fio. M.Sc.Diss. Port. Presentation: 23/08/02 142 p. Advisor: Luiz Fernando Gomes Soares Abstract: In the last decades, increasing interest has been observed in technologies related to wireless mobile communication environments. Great part of such interest follows the growth in the telecommunications industry, more specifically of cellular mobile telephony systems. Users are quickly absorbing these technologies and generating new necessities, such as the use of services of multimedia data, which demand quality assurance. In this context, one of the challenges is the efficient use of the limited frequency spectrum available, providing users quality services. This thesis proposes an architecture for the provision of end-to-end QoS in wireless mobile networks, taking into account inherent characteristics of these environments. To reach this purpose, modifications were proposed to the frameworks for QoS provisioning in generic processing and communication environments, aiming to fulfill the necessities brought by mobile networks, generating an adaptable architecture to offer QoS in environments that support mobile connections. In order to validate the proposal, an instantiation of frameworks for QoS provisioning in wireless mobile networks is described by means of a mobility scenario in which an infrastructure network of integrated services is simulated working over Mobile IP. 02_MSc_martins Lucimar Candido MARTINS. Personalizacao de visoes sobre documentos hipermidia. M.Sc.Diss. Port. Presentation: 10/06/02 125 p. Advisors: Marco Antonio Casanova and Carlos J. P. de Lucena Abstract: One of the most difficult tasks on dealing with the large amount of available data on the Web is to deliver the right data to a specific user. Many applications related to e-learning, e-commerce and information retrieval have been developed for this purpose. Some examples are search engines, recommendation systems and personalization systems. In spite of the great effort to develop personalized presentation and filtering systems there are challenging problems to be resolved. Some of them are directly related to information filtering, packaging and formatting, wherever this information needs to meet the user goals, interests and preferences. The main purpose of this dissertation is to define a framework, using the separation of concerns concepts, which contemplates personalization issues such as filtering, packaging and formatting processes. The framework facilitates the implementation and fine-tuning of algorithms addressing these issues. 02_PhD_seibel Luiz Fernando Bessa SEIBEL. Bio-AXS: uma arquitetura para integracao de fontes de dados e aplicacoes de biologia molecular. Ph.D.Thesis. Port. Presentation: 20/05/02 169 p. Advisor: Sergio Lifschitz Abstract: There are many molecular biology databases, also known as genome databases, that need to be integrated with all related applications and other data sources. This work proposes the use of an object-oriented framework for genome data access and manipulations. It is an innovative approach for an integration tool, which deals well with these application domain requirements, such as flexibility, reusability and extensibility. This works also presents a conceptual schema for the central dogma of molecular biology, defining an ontology which can be used within the data sources and compared to previous approaches. The framework, called Bio-AXS, is presented in details using class and sequence diagrams that explore both the structural and dynamic parts of the architecture. The Bio-AXS implementation issues are discussed, as well as some functionalities of the implemented prototype. Indeed, many case studies were explored in order to show the adequacy and applicability of the proposed architecture. 02_PhD_ribeiro Marcelo Blois RIBEIRO. Web life - uma arquitetura para a implementacao de sistemas multi-agentes para a Web. Ph.D.Thesis. Port. Presentation: 19/12/02 203 p. Advisor: Carlos Jose Pereira de Lucena Abstract: The resolution of distributed problems by applying computer systems composed by agents and organized towards a common objective is attracting the scientific community's attention. This is shown by the increasing number of initiatives for the development of methods and platforms to help constructing multi-agent systems. Some platforms concentrate efforts on supporting some multi-agent systems' aspects while others try to offer a complete infrastructure for the agents and organizations development. The Web Life architecture offers an implementation platform for developing Web-based multi-agent systems. The architecture is focused on the provision of a complete multi-agent infrastructure with native support for communication, coordination, decision-taking and task-oriented behavior. The incorporation of current under development standards and tools for solving specific problems in multi-agent systems is one of the main objectives of this work. These standards and tools are integrated by the Web Life framework. The architecture try to help the evolution of the Web into a Web with semantic contents that may be automatically processed - the Semantic Web. 02_PhD_carneiro Marcelo Medeiros CARNEIRO. Interfaces assistidas para deficientes visuais utilizando dispositivos reativos e transformadas de distancia. Ph.D.Thesis. Port. Presentation: 30/10/02 162 p. Advisors: Luiz Velho and Marcelo Gattass Abstract: The natural evolution of user interface models that occurred in the last few decades ended up popularizing a standard model based almost exclusively on visual metaphors. This process has left visually impaired users unable to use computers and to access new technologies. Some actions have been made to revert this scenario. Most of them were based on adapting the existing models instead of creating specific solutions for the visually impaired community. The development of applications for such users requires the use of new technologies, tools and communication media. This thesis proposes the use of force feedback devices in the project and implementation of assistive user interfaces, helping blind users in simple 2D interaction tasks. By exploring the sense of touch, such devices can be used to improve the efficiency of the communication between the user and the interface. Also, this work investigates the use of distance transforms as a powerful mechanism to support many 2D interaction tasks. 02_MSc_gerosa Marco Aurelio GEROSA. Categorizacao e estruturacao de mensagens textuais em ambientes virtuais de colaboracao. M.Sc.Diss. Port. Presentation: 28/02/02 85 p. Advisor: Hugo Fuks Abstract: The telecommunications technologies are improving the way of transmiting and generating information. But these technologies also lead to a non-manageable growth of the quantity, nature and quality of the available information. This is a critical problem in collaboration, where group members have to intensively communicate in order to resolve tasks. A way to minimize the information overload in textual communication is to organize, structure and categorize messages. The obejtive of this work is to analyze whether by structuring and categorizing the messages, there is a reduction in the information overload, and to figure out what other effects take place when the communication and collaboration of a group is realized through textual messages. The categorization of messages was implemented in the asynchronous communication services of the learning environment AulaNet and used during four semesters in a couse. 02_MSc_silva Marcos Aurelio Citeli da SILVA. Simulacao do escoamento de agua e oleo em meios porosos. M.Sc.Diss. Port. Presentation: 11/09/02 67 p. Advisors: Marcus Vinicius Soledade Poggi de Aragao and Marcus Vinicius Sarkis Martins Abstract: Many problems arising from real world can be represented by systems of partial differential equations (PDE's). However, the resulting discrete equations produce large and frequently bad conditioned matrices. This work implements the mixed finite element method to numerically solve a system of PDE's coming from a multiphase flow in porous media model and improve its performance by preconditioners and parallel processing. 02_MSc_silva Mauricio Hofmam da SILVA. Tratamento eficiente de visibilidade atraves de arvores de volumes envolventes. M.Sc.Diss. Port. Presentation: 08/03/02 154 p. Advisors: Marcelo Gattass and Waldemar Celes Abstract: Rendering complex three-dimensional engineering models has been a challenge for Computer Graphics ever since its origin, as detailed models are often composed of millions of polygons while current graphic stations are able to display, at interactive rates, only dozens or hundreds of thousands of polygons. A way to increase the performance of viewers of three-dimensional models is to reduce the number of polygons passed to the rendering pipeline by eliminating large groups of polygons classified as non-visible for being out of the viewing frustum or hidden by other polygons. In this work, we study the use of bounding volumes to determine sets of polygons which are potentially visible, propose ways to automatically structure such polygons in a hierarchy so as to restrict the necessary computations for this purpose, and compile a series of results which allow us to take some conclusions on the use of bounding volumes and model structuring. 02_PhD_silveira Milene Selbach SILVEIRA. Metacomunicacao designer-usuario na interacao humano-computador: design e construcao do sistema de ajuda. Ph.D.Thesis. Port. Presentation: 18/12/02 147 p. Advisor: Clarisse Sieckenius de Souza Abstract: This work is based on the Semiotic Engineering theory of Human-Computer Interaction. This theory views the application interface as a meta-communication act, a message from designer to user, representing the designers' view of the artifact he developed. Since the designer cannot be present in the interface, he is represented by his deputy, who is endowed with a communicative capacity that allows it to carry out a discourse that is complete and exclusively regarding the designers' final conclusions about the needs, preference, capacities and opportunities that he understands the user has. Being impossible to predict all the interpretations that each user can generate for the application, the designer needs, through his deputy, to explain everything that he did (and why he did it), and to give the users a chance to clarify their doubts. The best alternative for meta-communicacion is, certainly, the help system. In this way, the work herein proposed extends the current theoretical model of Semiotic Engineering, making explicit the presence of the help system and its communicative role. The purpose of this extension is to provide software designers with epistemic tools to support them when constructing their application's help systems. These tools allow designers to explore help systems' communicative power, supporting them in the reflection about the materials available for its elaboration (design models, design rationale, among others), thus supporting them both in the construction of the discourse embedded in the help system, and in the elaboration of the possible ways the user will be able to express himself within this discourse. 02_MSc_silva Otavio Rezende da SILVA. Uma arquitetura para sistemas multi-agentes baseada em espacos de tuplas reflexivos. M.Sc.Diss. Port. Presentation: 18/12/02 79 p. Advisor: Carlos Jose Pereira de Lucena Abstract: Software technology is undergoing a transition from monolithic architectures, constructed according to a single overall design, into open architectures composed of conglomerates of collaborative, heterogeneous, and independently designed agents and multi-agent systems (MAS). These architectures are driven by additional system-wide properties, such as communication, coordination, adaptability, mobility and persistence. However, controlling these properties and managing different application agents is not an easy task, requiring a more sophisticated software architectural approach. In this contex, we propose the Reflective Blackboard architectural pattern, which is the result of the composition of tho other well-known architectural patterns: the Blackboard pattern and the Reflection pattern. The pattern allows a better separation of concerns, supporting the separate handling of control strategies, by means of the computational reflection technique. An infrastructure, which implements the proposed pattern and provides support to multiple system-level properties, such as mobility, communication, coordination and persistence, was also developed. The Reflective Blackboard pattern was applied together with this infrastructure to the development of a Marketplace application in order to validate de concepts proposed in this work. 02_PhD_vilain Patricia VILAIN. Modelagem da interacao com o usuario em aplicacoes hipermidia. Ph.D.Thesis. Port. Presentation: 30/01/02 221 p. Advisor: Daniel Schwabe Abstract: Hypermidia applications are inherently interactive, with wich users interact in an intensive way. This work proposes a technique dedicated exclusively to represent the information exchanged in the interaction between users and applications. This technique is called UID (User Interaction Diagram) and it can be used to represent the exchange of information contained in textual descriptions wich express the requirements of given hypermedia application, without considering specific aspects of user interface and design. The proposal presented here is to use UIDs in several phases of the development cycle of applications where an intensive exchange of information with users takes place. In requirements gathering, as UIDs present a simple notation and are comprehensible by users, they are used to aid the specification and validation of requirements. In conceptual design, as UIDs present data items participate of the interaction, they are used for identifying the classes, attributes and associations of a preliminary class diagram. In navigation design, UIDs are used for identifying the indexes and contexts through which users can navigate. A UID extension that supports navigation aspects, called UND (User Navigation Diagram), can also be used in this phase. This work proposes a method for requirements gathering where UIDs are used to represent user tasks. Further, guidelines to define a preliminary class diagram, indexes and contexts, from UIDs, are presented. To test and improve the proposed guidelines, three case studies were developed, on which the examples used during the text are based. Besides being used in these case studies, the proposals presented here have been currently applied in the method OOHDM (Object-Oriented Hypermedia Design Method), thus showing the viability of such proposals. 02_MSc_lucena Paula Salgado LUCENA. Expressive talking heads: um estudo de fala e expressao facial em personagens virtuais. M.Sc.Diss. Port. Presentation: 25/06/02 80 p. Advisors: Marcelo Gattass and Luiz C. Velho Abstract: The human face is interesting and challenging mainly because of its familiarity. Essentially, it is the part of the human body that is used to recognize individuals. As well as the face, the speech is an important instrument for human communication, allowing the exteriorization of thoughts and the definition of emotions. Together, speech and face are the main elements of interactivity among human beings. However, the natural and faithful reproduction of the peculiarities of these elements in the computational universe is not a simple task, constituting topics of research in diverse areas, particularly in facial animation. Among the diverse types of facial animation systems developed, those that involve the facial animation of the virtual character combined with speech synchronization are distinguished as directly related to this work. These kinds of systems are known as talking head or talking face. For the development of a talking head system, it is necessary to identify the possible approaches for the speech and face modeling. The models used will influence not only the way that the animation is performed, but will also affect the system's interactivity. An important contribution of the present master thesis is the study of several possible approaches for the main elements and the proposal of taxonomy for the classification of the talking head systems. From the proposed taxonomy and making use of one approach for each analyzed parameter, an application was developed that receives as input a text composed by the character's speech and genus, language and emotion parameters, and it generates as output, in real time, the animation of a virtual character uttering the input text with speech synchronization and expressiveness. The system developed, called "Expressive Talking Heads", explores the naturalness of facial animation and it seeks to offer the user a real-time interactivity interface. The "Expressive Talking Heads" system can run as a stand-alone application or connected to web browsers. It was designed and developed to provide a platform - and operating system-independent solution. 02_PhD_aiex Renata Machado AIEX. Uma investigacao experimental da distribuicao de probabilidade do tempo de solucao em heuristicas GRASP e sua aplicacao na analise de implementacoes paralelas. Ph.D.Thesis. Port. Presentation: 05/08/02 243 p. Advisors: Celso Carneiro Ribeiro and Mauricio G. C. Resende Abstract: GRASP (Greedy Randomized Adaptive Search Procedure) is a multi-start metaheuristic for combinatorial optimization problems. GRASP has been used to find quality solutions of several combinatorial optimization problems. In this work we describe a methodology for analysis of GRASP. Hybrid strategies of GRASP with path relinking are also proposed. These strategies are studied for the 3-index assignment problem (AP3) and for the job-shop scheduling problem (JSP) and are analyzed according to the methodology proposed. The methodology for analysis of GRASP is used to predict quanlitatively how the quality of the solution varies in a parallel independent GRASP, using the data of the GRASP sequential version as input. The GRASPs for the AP3 and for the JSP are parallelized and the computational results are compared to the results obtained using the methodology proposed. 02_PhD_noya Ricardo Choren NOYA. Uma linguagem de modelagem para sistemas baseados em agentes. Ph.D.Thesis. Port. Presentation: 20/12/02 121 p. Advisor: Carlos Jose Pereira de Lucena Abstract: The agent technology is gaining acceptance, both in academy and industry, with regards to distributed systems development. Modeling languages and development processes were created to formalize the development of agent based systems. Nevertheless, existing modeling languages are either based on previous development methods (object oriented) or they do not have artifacts (models) that show the agency characteristics that exist in a system. This thesis proposes a modeling language, for agent based systems, that generates artifacts that model agency aspects, such as adaptation, learning, interaction and autonomy. The language has artifacts that allow the modeling of the non-agent part of an agent based system, using ontology. 02_MSc_guimaraes Robson Mattos GUIMARAES. Considerando o usuario em aplicacoes hipermidia. M.Sc.Diss. Port. Presentation: 25/04/02 224 p. Advisor: Daniel Schwabe Abstract: This work aims at characterizing the problem in using the user as part of the modeling of a hypermedia project and at extending the OOHDM method to represent such problems. To illustrate the motivation, several cases of Web site personalization and information security applied to information systems based on Web are presented. The OOHDM method is extended using new primitives, such as user and role, that are also used in the definition of the main components of this method: navigational classes, links, access structures and contexts. The hypermedia application framework patterns raised to customization are described. Besides that, it is presented how the OOHDM-WEB 2.0 can be altered in order to contemplate the proposed extensions. 02_MSc_costa Rogerio Luis de Carvalho COSTA. Alocacao de dados e distribuicao de carga para execucao paralela da estrategia blast de comparacao de sequencias. M.Sc.Diss. Port. Presentation: 27/03/02 75 p. Advisor: Sergio Lifschitz Abstract: The BLAST (Basic Local Alignment Search Tool) is the sequence comparison strategy mostly used in computational biology. While molecular biology databases are getting larger, performance has become an important problem to be solved. Indeed, depending on the processing environment, sequences complexity and the database size, a BLAST execution for a given sequences input set may last for hours or even days. Among all distinct solutions that aim at obtaining better performances, we are interested here in the use of multiprocessor machines and parallel tecniques. In this dissertation we study the use of distributed database design methods for executing BLAST in a shared-nothing parallel architecture. We propose specific data allocation approaches similar to conventional fragmentation and replication techniques. We give some practical results obtained from a PC cluster-based implementation with MPI and a WU-BLAST version, which is kept unchanged. The main parameters evaluated include the number and size of input sequences and also the database size. Besides presenting detailed results for performance and speedup, we discuss tasks distribution strategies to deal with load balancing issues. 02_PhD_bergmann Ulf BERGMANN. Evolucao de cenarios atraves de um mecanismo de rastreamanto baseado em transformacoes. Ph.D.Thesis. Port. Presentation: 18/12/02 224 p. Advisor: Julio Cesar Sampaio do Prado Leite Abstract: The importance of traceability in software evolutions was quite established, but despite the effort spent, they are many issues that difficult they effective use. These issues are related to the generating and validating of trace information about the relationships and interdependencies between the real world, the requirements and the software system. This information is often mistrusted since they may have become obsolete due to separate evolution of models and systems. This thesis study the use of the transformational technology in the implementation of a traceability mechanism that automatically obtain trace information by identifying the changes made in consecutive artifact versions and store it as transformations. The main advantages gained with it uses are the elimination of the problems of the lack of trace information updating, the systematization of the trace process and the decrease of the cost to obtain, update and validate trace information. Besides these benefits, the storage of trace information in the form of transformations incorporates a larger knowledge on the accomplished modeling, unilike other mechanisms that only store the artifacts and their static inter-relationships. 02_MSc_braconi Viviane Monteiro BRACONI. Heuristicas multifluxo para roteamento de produtos em redes dutoviarias. M.Sc.Diss. Port. Presentation: 26/04/02 78 p. Advisor: Ruy Luiz Milidiu Abstract: This master's dissertation deals with the Problem of Petroleum Derived Products Transportation through Pipelines(PTDPO). We focus on Petrobras nets located in the State of Sao Paulo. These networks represent the most intense operations of Petrobras. Each network is composed of a set of areas(bases) and segments. The bases are responsible for the production and delivery of the products. The areas are not sef-sufficient in the supply of their consumers' markets so it is necessary to transport products through pipes. Considering that initial configuration and consumption values are already available, the problem consists of determining how to operate the network of pipes to satisfy the demand within a established time interval without violating the technological restrictions.The PTDPO is a problem of NP-Complete cass. As this is a difficult problem and because of the implementation difficulties and acquisition of data, we decided to simplify the problem to obtain better results. In order to find to solution to the problem, we propose a heuristic approach separated in two phases: planning of the products to be pumped and scheduling of the products to be pumped. In the first phase of the proposed heuristic we adopted a relaxation of the problem that makes possible the use of linear programming for definition of routes and products necessary to satisfy the demand of the network . In this phase the proposed model is derived from the network flows model with multiperiods. Due to the difficulty in obtaining an integer result in a reasonable execution time, we restricted the solution to the continous model. In the scheduling phase we defined for each segment the order of pumping of the products. This order could be modified if some restriction of tank's capacity were violated. If it is necessary, we can accomplish interactions between the heuristic phases to find a feasible solution for PTDPO. In our experiments using a Pentium III 850MHz and the XPRESS-MP application, version 1.2, we found feasible solutions for 5 instances pertaining to two different network of Petrobras, for programming intervals varying between 5 and 30 operational days. Most of these instances were built on the real data of the Petrobras. As a whole,the processing time used for obtaining these results did not exceed 20 minutes. In searching for a feasible scheduling, this work promoted a detailed study of the PTDPO and its intrinsic difficulties. Among other aspects, it made possible development of more efficient linear formulation provided a better understanding of the problem as a result of the performed experiments. _______________________________________________________________________ 2003 03_PhD_alvim Adriana Cesario de Faria ALVIM. Uma heuristica hibrida de melhoria para o problema de bin packing e sua aplicacao ao problema de escalonamento de tarefas. Ph.D.Thesis. Port. Presentation: 13/06/03 136 p. Advisor: Celso Carneiro Ribeiro Abstract: We propose in this work a hybrid improvement procedure for the bin packing problem. This heuristic has several components: lower and upper bounds; reductions; construction of initial solutions by reference to the dual problem; heuristics for load redistribution based on dominance, differencing, and unbalancing; and tabu search. We also investigate the application of this hybrid heuristic to the problem of task scheduling on identical parallel processors. Computational results on hundreds of benchmark test problems are represented. 03_MSc_pigatti Alexandre Altoe PIGATTI. Modelos e algoritmos para o problema de alocacao generalizada (PAG) e aplicacoes. M.Sc.Diss Port. Presentation: 25/07/03 74 p. Advisors: Marcus Vinicius Soledade Poggi de Aragao and Eduardo Uchoa Abstract: This dissertation tackles the Generalized Assignment Problem (PAG), models and algorithms are studied and proposed. This work was motivated by a real world application/ the Truck Loading Problem (PCC). Research was done on approximated (metaheuristics) and exact algorithm for solving the PAG. The approximated algorithms proposed were based on a recent idea from Fischetti and Lodi (2003). It uses integer programming to explore wider neighborhoods. The results were compared to the best known, while demanding much less implementation effort and using less cpu time. The exact algorithm proposed is a branch-and-cut-and-price developed from the branch-and-price algorithm of Savelsbergh (1997). We used stabilized column generation techniques similar to the one by Du Merle, Villeneuve, Desrosiers and Hansen (1999), and devised experiments with different implementations of this mechanism. The resulting algorithm proved its efficiency by solving to optimality open instances from the literature. Finally, experiments with the PCC turned possible the evaluation of the codes developed on real problems. 03_PhD_abelem Antonio Jorge Gomes ABELEM. Difusao seletiva em inter-redes IP baseadas em redes opticas. Ph.D.Thesis. Port. Presentation: 02/04/03 130 p. Advisors: Michael Anthony Stanton and Noemi La Rocque Rodriguez Abstract: Multicast communication and recent advances in optical technology, most specifically in Wavelength Division Multiplexing (WDM), allied with the consolidation of IP as the dominant protocol of convergent networks, offer new perspectives for the next generation Internet. This thesis utilises these technologies to propose a set of adaptations, called MIRROR, to multicast communication, specifically IP Multicast, in labelled burst-switched optical networs. MIRROR proposes modifications to traditional IP Multicast in order to improve its scalability as a function of the number of simultaneously active groups, as well as marking it more appropriate for use in optically switched networks. Basically, MIRROR includes new proposals for handling state information about the multicast distribution tree, as well as for the establishment of label-based multicast paths. In order to evaluate this proposal, two approaches are followed, one based on a comparative analysis between MIRROR and a number of other alternatives to IP Multicast proposed in the literature, and the other based implementation of a prototype in the simulation environment provided by NS (Network Simulator). The comparative analysis evaluates such parameters as: state requirement information, control overhead, packet processing efficiency and tree cost. The prototype implementation implements a new node structure and alters existing NS modules (OBS e MPLS), to make possible the simulation of labelled burst-switched optical networks in the multicast context. 03_PhD_pessoa Artur Alves PESSOA. Dois problemas de otimizacao em grafos: transporte em redes de dutos e busca com custos de acesso. Ph.D.Thesis. Port. Presentation: 23/05/03 119 p. Advisors: Ruy Luiz Milidiu and Eduardo Sany Laber Abstract: We consider two combinatorial optimization problems: the pipeline transportation problem (PTD) and the problem of searching with different access costs (PBC). In PTD, we are given a directed graph G = (N,A) where each arc corresponds to a pipeline. We are also given a set of batches, each batch being initially located at an arc or node and having a destination node. A subset of these batches are considered as further batches. Our aim is to find a sequence of pipeline operations leading all non-further batches to their corresponding destination nodes. First, we show that PTD is NP-hard, even for the case where G is acyclic. Next, we present a polynomial algorithm called BPA. This algorithm solves PTDS, a variation of PTD, for general graphs. For acyclic graphs, BPA also minimizes a general cost function. For the case of makespan minimization for PTDS, we prove that there is no (formula) approximate algorithm for any E > 0, unless P = NP, where n is the instance size. The previous result also holds if G is both acyclic and planar. In PBC, we are given an ordered vector with n elements and the corresponding access costs. Our aim is to find a search strategy that minimizes either the average cost with uniform probabilities (PBCM) or the worst case cost (PBPC). In both cases, the best known exact algorithm requires in O(n³) time and O(n²) space. For PBCM, we present the ratio algorithm, that requires O(n²) time and O(n) space. This algorithm always obtains a search strategy with average cost at most 4ln(n+1)/n, assuming the sum of all access cost to be 1. Furthermore, we introduce approximation algorithms for both PBCM and PBPC. Both of them give (2+ E+0(1))-approximate solutions, for any E>0, in O(n) time and space. 03_MSc_soaresneto Carlos de Salles SOARES NETO. Descricao arquitetural da provisao de QoS em ambientes genericos de processamento e comunicacao. M.Sc.Diss. Port. Presentation: 11/08/03 113 p. Advisor: Luiz Fernando Gomes Soares Abstract: The increased demand for platforms with support for multimedia applications raised the importance of mechanisms for Quality of Service provisioning, since each media has its own processing and communication requirements. SCM Model (Service-Composition Model) provides abstractions for the representation and programming of QoS aspects and multicast in communication services. According to its terminology, the QoS provisioning can be seen as a service provider, where QoS negotiation and QoS maintenance meta services act upon it. The QoS negotiation are the mechanisms responsible for the admission of new user flows, while the QoS maintenance meta service is responsible for maintaining the negotiated level of service during the service operation. Such meta services had been previously described as frameworks modelled in UML. The present work focuses on the architectural description of these meta services using Wright architecture description language (ADL), which allows the use of its analysis and formal verification tools to infer properties. To smooth this approach, a domain-specific language (DSL) called LindaQoS is proposed as a notation closest to the abstraction level of the problem domain, specifically designed to define hierarchies of negotiation and maintenance subsystems. Moreover this work presents a compiler allowing the translation of LindaQoS specifications into architectural descriptions (currently, using Wright) and into programming languages (JAVA in the future). 03_MSc_rocha Cristiano Braz ROCHA. Inferencias semanticas na recuperacao de informacoes para aplicacoes hipermidia. M.Sc.Diss. Port. Presentation: 24/07/03 134 p. Advisors: Marcus Vinicius Soledade Poggi de Aragao and Daniel Schwabe Abstract: The information overload problem is one of the most challenging problems being faced today. In order to solve this problem, different areas such as Knowledge Management, Semantic Web and Hypermedia Applications Modeling have used similar solutions that consist basically of semantically structuring the information so it can better accessed. This dissertation porposes an infra-structure based on classic algorithms and techniques of Artificial Inteligence that utilizes the availability of domain specific models to enable the applications where they are defined to make inferences about these particular domains. These inferences enable the creation of new functionalities in these applications. Four new functionalities were proposed and implemented, the most important being a semantic search. The new functionalities presented were successfully tested in two existing applications: the website of the Computer Science Department of PUC-Rio and the Portinari Knownledge Portal that presents all the work of the famous brazilian painter Candido Portinari. 03_MSc_vasconcelos Davi Romero de VASCONCELOS. Analise de estrategias utilizando verificacao formal de modelos. M.Sc.Diss. Port. Presentation: 25/02/03 136 p. Advisor: Edward Hermann Haeuler Abstract: In formal methods, one of the approaches that have been successful lately is Model Checking, which consists in a technique to achieve automatic verification about a system behavior. Nowadays, the model checking is very frequently employed in computer science to formal verification of software and hardware, but it is not used in other knowledge fields, such as mathematics and economics. The purpose of this research is to apply model checking in economics problems, using Game Theory. Some inadequancies have been observed. Therefore, it is necessary to create new definitions: a generic and qualitative definition of game that uses one logic language called Game Analysis Logic (GAL); a new language to describe game called RollGame (Romero + All Game); a translation from RollGame to a language of model specification; a translation from game definition to Kripke structure. It is also been observed that the use of model checking makes it possible to analyze players' strategies. One tool, called StratAn-RollGame (Strategy Analyzed using RollGame), makes the translation from RollGame to model Checking automatic. Thus, the present research has demonstrated that is possible indeed to use model checking in other knowledge fields. 03_PhD_saade Debora Christina Muchaluat SAADE. Relacoes em linguagens de autoria hipermidia: aumentando reuso e expressividade. Ph.D.Thesis. Port. Presentation: 28/03/03 215 p. Advisor: Luiz Fernando Gomes Soares Abstract: This work is related to hypermedia authoring and execution environments, and its main focus is declarative document authoring. Starting from studies about architectural description languages (ADL), which are used for specifying software system architectures, this thesis identified facilities found in ADLs that could be applied to the hypermedia domain, with advantages. Aiming at improving the expressiveness and reuse in the specification of relations in hypermedia authoring languages, this work introduced the concept of hypermedia connector, which has a role similar to ADL connectors, that is, representing relations among components of a document. Besides connectors, this work also introduced the concept of hypermedia composite template, which has a role similar to architectural styles in ADLs, that is, representing generic structures of nodes and links that can be reused in several distinct documents. As a validation of the proposed concepts, the 2.0 version of the NCL - Nested Context Language - hypermedia authoring language, based on the NCM - Nested Context Model - conceptual model, was developed and integrated to the HyperProp hypermedia system, incorporating the new facilities. The NCL 2.0 language was developed using a modular structure, following the principles adopted by the W3C - World-Wide Web Consortium. Thus, its modules for the specification of connectors and templates, respectively called XConnector and XTemplate, can be incorporated to other existent languages, such as XLink, XHTML and SMIL, used for web document authoring. This thesis also proposes extensions to these languages, exemplified by the incorporation of XConnector and XTemplate facilities into the XLink standard. 03_MSc_souza Edinalda Maria de SOUZA. Um estudo sobre um algoritmo para visualizacao de terrenos. M.Sc.Diss. Port. Presentation: 04/07/03 78 p. Advisor: Marcelo Gattass Abstract: Algorithms for the interactive visualization of terrains are very complex and, at the same time, of great importance to many applications, such as games and activity-planning over terrains. Due to such complexity and importance, in the past decade this subject has received great attention by researchers on Computer Graphics. As a consequence, a number of strategies have been developed. Among the most successful strategies, one can highlight recent works by Lindstrom and Pascucci. The algorithm proposed by these authors has various implementations available in the Internet and deserves to be reevaluated. The present work makes such reevaluation by means of an independent implementation developed by the author and tested over a base or real terrains. With the purpose of making this analysis more complete and to support some conclusions, comparative results with other algorithms in the area are also presented. 03_MSc_dias Fabio Meira de Oliveira DIAS. Execucao de workflow em ambientes com desconexao. M.Sc.Diss. Port. Presentation: 02/02/03 60 p. Advisor: Marco Antonio Casanova Abstract: Workflow management systems are frequently used for modeling, monitoring and controlling the coordinated execution of activities performed by workgroups in a variety of contexts. With the widespread use of portable computers and their growing computational power, conventional systems have often proved to be overly restrictive, effectively limiting the level of autonomy of the users involved. The primary goal of this work is to identify and analyze different flexibilization techniques and mechanisms that can be employed in a workflow management system aimed at supporting disconnected operation. The main challenge is to provide a satisfactory degree of independence among individuals in cooperating teams who share a common goal and work in disconnected environments. In order to test the viability of the ideas discussed in this dissertation, a system was built whose design met the requirements presented in the text and which allows the exploration of specific features of different kinds of workflow so as to enhance execution flexibility, without compromising the predefined structure. 03_PhD_ayres Fausto Veras Maranhão AYRES 03_PhD_lima Fernanda LIMA. Modelagem semantica de aplicacoes na WWW. Ph.D.Thesis Port. Presentation: 28/03/03 128 p. Advisor: Daniel Schwabe Abstract: In this thesis we present a method for the design and implementation of web applications for the Semantic Web. Based on the "Object Oriented Hypermedia Design Method" approach, we used ontology concepts to define an application conceptual model, extending the expressive power of the original method. The navigational models definitions use a query language capable of querying both schema and instances, enabling the specification of flexive access structures. Additionally, we propose the use of faceted access structures to improve the selection of navigational objects organized by multiple criteria. Finally, we present an implementation architecture that allows the direct use of the application specifications when deriving a final application implementation. 03_MSc_ferreira Francisco Eduardo dos Reis FERREIRA. Desenvolvimento de aplicacoes baseadas em servicos na Web Semantica. M.Sc.Diss. Port. Presentation: 26/03/03 103 p. Advisors: Carlos Jose Pereira de Lucena and Daniel Schwabe Abstract: This work address issues on the development of service-oriented applications on the semantic web. It presents a platform conceived to support information sharing among different services accessed by different devices and hardware infrastructure. Using open standards and a totally distributed approach, Everyware allows users to publish their services without need of configuration and administration. Services can also be semantically tagged and integrated to the Semantic Web, so they can be easily processed by machines. Everyware platform offers a standard architecture to application development and an API to the development of coodination mechanisms. It integrates most concerns that are present in service-based applications. 03_MSc_carvalho Gustavo Robichez de CARVALHO. M.Sc.Diss. Port. Presentation: 22/08/03 88 p. Advisors: Carlos Jose Pereira de Lucena and Arndt von Staa Abstract: Component Based Software Engineering is an approach for reusing software artifacts when developing applications. In order to develop solutions using this approach, it is necessary to compose software using components that have already been developed. After putting those pieces together, we need to coordinate the interdependencies established among those compositions to fulfill the requirements, needed to resolve a problem. This dissertation proposes a software architecture that separates and structures the concepts of coordination, composition and software components in different architectural layers. Using this approach, we expect that specific modifications in layer constructions have the minimum impact on the others layers. ACCA (Architecture for Coordination and Composition of software Artifacts) must be understood as a conceptual struture that is used to organize component based software development. It also presents a composition framework, the reification process for ACCA and a software development process organized using this approach. 03_PhD_rosseti Isabel Cristina Mello ROSSETI. Estrategias sequenciais e paralelas de GRASP com reconexao por caminhos para o problema de sintese de redes a 2-caminhos. Ph.D.Thesis. Port. Presentation: 24/07/03 120 p. Advisor: Celso Carneiro Ribeiro Abstract: Let G = (V,E) be a connected undirected graph, where V is the set of nodes and E denotes the set of edges. A 2-path between nodes (s,t) is a sequence of at most two edges connecting them. Given a non-negative weight function associated with edges of G and a set D of origin-destination pairs of nodes, the 2-path network design problem (2PNDP) consists in finding a minimum weighted subset of edges containing a 2-path between the extremities of every origin-destination pair in D. Applications can be found in the design of communication networks, in which paths with few edges are sought to enforce high reliability and small delays. The GRASP metaheuristic is a multistart process, in which each iteration consists of two phases: construction and local search. The best solution found after a fixed number of iterations is returned. Path-relinking is applied as an attempt to improve the solutions found at the end of each GRASP iteration. Parallel implementations of metaheuristics are very robust. Typical parallelizations of GRASP correspond to multiple-walk independent-thread strategies, based on the balanced distribution of the iterations over the processors. In the case of multiple-walk cooperative-thread strategies, the processors exchange and share information collected along the trajectories that they investigate. In this thesis, sequential and parallel heuristics are developed for 2PNDP. Variants and combinations of GRASP with path-relinking are analysed by comparing the results of the proposed algorithms with those obtained by others algorithms described in the literature. Parallel GRASP with path-relinking heuristics are compared to investigate the influence of the cooperation among processors in terms of solution quality and processing time. We also explore differents strategies to optimize the parallel implementations, to make better use of the computational resources and to reduce communication and memory conflicts. 03_MSc_imperial Juliana Carpes IMPERIAL 03_MSc_rezende Juliana Lucas de REZENDE. Aplicando tecnicas de conversacao para a facilitacao de debates no ambiente AulaNet. M.Sc.Diss. Port. Presentation: 20/03/03 171 p. Advisor: Hugo Fuks Abstract: It is difficult to coordinate participants simultaneously interacting and generating ideas in a textual chat session. During several editions of ITAE course - Information Technologies Applied to Education - the following difficulties were perceived by the mediators: ensure participation from all users, keep the focus of the conversation and the flow of the chat. This research investigates the use of conversation techniques to facilitate the dynamics of textual chats. A tool (Mediated Chat 2.0) implementing those techniques was developed and made available to help mediators conduct chat sessions in a course using the AulaNet environment. Currently, the standard application used in the AulaNet environment is the Mediated Chat 1.0, an instance of the Communication Channels Framework developed in the Software Engineering Laboratory at PUC-Rio in 2000. The Mediated Chat 2.0 is an instance of the same framework. To evaluate the developed application, chats had been carried out in ITAE 2002.2 where both Mediated Chat 1.0 and 2.0 were used and compared. The chat sessions had been analyzed to evaluate whether the mediators managed to successfully apply the dynamics supported by the Mediated Chat 2.0. The Mediated Chat 2.0 will be incorporated into the AulaNet software. 03_MSc_duarte Julio Cesar DUARTE 03_MSc_kurizky Konstatin Kurizky 03_MSc_moreno Lorenza Leao Oliveira MORENO. Codigos de prefixo de rapida decodificacao. M.Sc.Diss. Port. Presentation: 21/03/03 84 p. Advisor: Ruy Luiz Milidiu Abstract: Even with the evolution of communication and storage devices, the use of complex data structures, like video and hypermedia documents, keeps increasing the demand for efficient data compression mechanisms. Prefix codes are one of the most know compressors, since they are executed by some compression methods that group different algorithms, besides presenting a good performance when used separately. A lot of approaches have been tried to improve the decoding speed of these codes. One major reason in that files are compressed and update just a few times, whereas they have to be decompressed each time they are accessed. This work presents prefix codes and their decoding techniques in order to introduce a new coding scheme. In this scheme length-retricted codes are used to control the space requirements of the Look-up table, an efficient and fast prefix codes decoding method. Since restricted codewords are used, a small loss of compression efficiency is admitted. Empirical experiments indicate that this loss in the coded text is smaller than 11% if a character based model is used, and the observed average decoding speed is five times faster than the one for canonical codes. For a word based model, the average decoding speed is 3,5 times faster than a canonical decoder, but it decrease when a large number of symbols is used. Hence, this method is very suitable for applications where a character based model is used and extremely fast decoding is mandatory. 03_MSc_paula Maira Greco de PAULA. Projeto da interacao humano-computador baseado em modelos fundamentados na engenharia semiotica: construcao de um modelo de interacao. M.Sc.Diss Port. Presentation: 27/03/03 87 p. Advisor: Simone Diniz Junqueira Barbosa Abstract: Due to the propagation of personal computers, it is increasingly important to build highly usable interfaces, taking into account users' characteristics, preferences, and needs. Diverse models have been proposed to cope with the complexity of human-computer interaction (HCI) design. However, most of them deal at once with elements that should be addressed by distinct models. Moreover, many of these models are based on cognitive theories, which focus mainly on the individual interacting with an application, without exploring the fact that an application is the product of a rational decision-making process carried out by a designer. This gap is deal with by Semiotic Engineering, a theory of HCI which views the interface as a designer-to-users message, representing the designer's solution to what he believes are the users' problems, needs, and preferences. In this message, he is telling users, directly or indirectly, what he had in mind when he conceived the applicaton. Within Semiotic Engineering, this work extends scenarios, adapts an existing task model and builds a model which treats interaction as conversation. Our goal is to conceive models and representations that serve, each under a clear perspective, as epistemic tools that support the designer's reflection about the interactive solution being conceived. A small case study was conducted to evaluate the quality of an interactive solution designed using the proposed task and interaction models, in comparison with CTT, a widely used task model. 03_MSc_moreira Marcos Magalhaes MOREIRA 03_MSc_leal Marcus Amorim LEAL. LuaTS: um espaco de tuplas reativo orientado a eventos. M.Sc.Diss. Port. Presentation: 24/02/03 87 p. Advisor: Noemi de La Rocque Rodriguez Abstract: The widespread use of the Internet along with the rapid growth and acceptance of the Web as a general application platform impose new requirements associated with the integration and coordination of autonomous and heterogeneous software compoments. The specific needs of this contex led to the development of new coordination mechanisms, among which the reactive tuple space. In this work we present LuaTS, a reactive, event oriented tuple space that supports only asynchronous calls. The system, developed in Lua, provides functionaties that allow programmers to extend its basic semantics and also support a more flexible tuple search and retrieval process. We describe the implementation of LuaTS and explore its features through different examples that include classic concurrent and distributed programming problems. We show that the uncoupling provided by the tuple space model, together with an event oriented programming dynamics, simplify inter-process synchronization and yield a clear execution stream, improving, in many cases, the development process of distributed applicacoes. 03_PhD_cortes Mariela Ines CORTES. Suporte computacional a evolucao de frameworks. Ph.D.Thesis. Port. Presentation: 17/06/03 127 p. Advisor: Carlos Jose Pereira de Lucena Abstract: Framework development is expensive not only because of the intrinsic difficulty related to the elicitation of domain knowledge but also because of the lack of methods and techniques to suport its evolution and interactive development. The present thesis proposes the use of two complementary techniques to support framework evolution: refactoring and extension rules. The refactoring technique has been developed to enable software re-structuring in a way to produce more readable and reusable code. Extension rules have been proposed to change the structure of the framework variation points by allowing the addition of new design functionalities. Both techniques preserve the observable behavior of programs. This property is formally verified in this work by using CCS approach to model checking. The proposed approach has been tested by means of a tool specially developed to support the application of the defined rules. 03_MSc_leite Matheus Costa LEITE. Um modelo de computacao para circuitos de objetos. M.Sc.Diss. Port. Presentation: 11/04/03 68 p. Advisor: Carlos Jose Preira de Lucena Abstrac: Object Oriented Programming is a mature, well established software modeling technique. Nevertheless, the importance of its role has the same magnitude as the consensus in respect to its weakness and limitations. OO is not a panacea, and, should it fail, alternatives must be found - some hybrid, while others entirely new. In this work, we argue that the parallel between OO and electric circuits is an interesting hybrid solution, for some of the basic features found in such circuits are the same as the ones sought after as the Holy Grail of Software Engineering - concurrency, modularity, robustness, scalability, etc. - and that are not always achieved only with the traditional OO approach. Hence, our proposal is the establishment of a correlation between electric circuits and object oriented programming. From the former, comes the circuit: closed path where information flows and is processed. From the second, comes the object: abstract entity that constitutes the information flowing within the circuit. Finally, from their union, arises a new model of computation - the object circuit - where it is supposed the benefits brought by each part are complementary. We motivate our discussion with a collection of simple - albeit elucidative - examples, followed by a case study in the simulation field. In order to ratify the functioning of these circuits, an object circuit's implementation was built on top of the Java programming language. 03_PhD-mamanimacedo Nestor Adolfo MAMANI MACEDO. Criando uma arquitetura de memoria corporativa baseada em um modelo de negocio. Ph.D.Thesis. Port. Presentation: 03/07/03 172 p. Advisors: Julio Cesar Sampaio do Prado Leite and Teresia Diana Lewe van Macedo-Soares Abstract: This work presents a proposal for building corporate memory architecture. It has three layers: sources, middleware and repositories. First, we create an conceptual enterprise model, named Organizational Baseline (OB), based on theories of business management and organized in accordance with an approach of ontology. Second, we use multi agents system as a means to organize our proposal of architecture. The OB is the main component, where we store whole relevant information for the organization; it can be extracted from legacy systems, databases, Internet sites or human beings. The conceptual enterprise model is based on constructs elicited from positioning and emphasizing efficiency views of strategic analysis, together with function analysis and total quality management. 03_PhD_renteria Raul Pierre RENTERIA. Algoritmos para regressao por minimos quadrados parcias. Ph.D.Thesis. Port. Presentation: 19/03/03 80 p. Advisor: Ruy Luiz Milidiu Abstract: The purpose of many problems in the machine learning field is to model the complex relationship in a system between the input X and output Y variables when no theoretical model is available. The Partial Least Squares (PLS) is one linear method for this kind of problem, for the case of many input variables when compared to the number of samples. In this thesis we present versions of the classical PLS algorithm designed for large data sets while keeping a good predictive power. Among the main results we highlight PPLS (Parallel PLS), a parallel version for the case of only one output variable, and DPLS (Direct PLS), a fast and approximate version, for the case of more than one output variable. On the other hand, we also present some variants of the regression algorithm that can enhance the predictive quality based on a non-linear formulation. We introduce LPLS (Lifted PLS), for the case of only one dependent variable based on the theory of kernel functions, KDPLS, a non-linear formulation for DPLS, and MKPLS, a multi-kernel algorithm that can result in a more compact model and a better prediction quality, thanks to the use of several kernels for the model building. 03_PhD_rodrigues Rogerio Ferreira RODRIGUES. Formacao e controle de apresentacoes hipermidia com mecanismos de adaptacao temporal. Ph.D.Thesis. Port. Presentation: 27/03/03 156 p. Advisor: Luiz Fernando Gomes Soares Abstract: The development of hypermedia/multimedia systems requires the implementation of an element responsible for receiving the specification of a hyperdocument (structure, relations and presentation descriptions) and controlling the presentation. This execution control module is commonly known as hypermedia formatter. The goal of this thesis is to discuss the aspects related to the design and implementation of hypermedia formatters, proposing a hyperdocument execution model and a framework to be reused in the implementation of hypermedia systems. In particular, this work deals with document adaptation issues, both before document presentation and on-the-fly, prefetching mechanisms, integration with presentation tools, and integration with the infrastructure (in special, operating system and network). The proposal presented follows an object-oriented specification approach, defining an architecture for temporal and spatial synchronization in hyperdocument presentations. The proposal can also be used in other application domains that need real-time synchronization. 03_MSc_maffra Sergio Alvares Rodrigues de Souza MAFFRA. Propagacao de som em ambientes acusticos virtuais bidimensionais. M.Sc.Diss. Port. Presentation: 28/04/03 96 p. Advisors: Marcelo Gattass and Luiz Henrique de Figueiredo Abstract: For a long time, computational simulation of acoustic phenomena has been used mainly in the design and study of the acoustic properties of concert and lecture halls. Recently, however, there is a growing interest in the use of such simulations in virtual environments in order to enhance users' immersion experience. Generally, we can say that a virtual acoustic environment must be able to accomplish two task: simulating the propagation of sound in an environment and reproducing audio with spatial content, that is, in a way that it allows the recognition of the direction of sound propagation. These tasks are the topic of the present dissertation. We begin with a revision of the most common algorithms for the simulation of sound propagation and, briefly, of the reproduction of audio with spatial content. We then present the implementation of a virtual acoustic environment, based on beam tracing algorithms, which simulates the propagation of sound waves in two-dimensional environments. As most of the computation is made in a pre-processing stage, the virtual acoustic environment implemented is appropriate only for spatially fixed sound sources. The propagation paths computed are made of specular reflections and of diffractions. 03_PhD_cruz Sylvia de Oliveira e CRUZ. Identificando objetos atraves de pronomes. Ph.D.Thesis. Port. Presentation: 31/03/03 125 p. Advisor: Carlos Jose Pereira de Lucena Abstract: The need to know the object identification to request it's services is a restriction in the construction of object oriented system. To reduce this restriction we have proposed a form of generic reference to an object named pronoum. Pronouns allow the programmer to specify how to send a message without knowing the destination object by name. With pronouns, the visibility of an object increases without breaking the rule of encapsulation. With this feature decoupled server classes may be constructed and used in different contexts. This work define some pronouns, show its utilities through of examples and suggest a framework to new pronouns creation. 03_MSc_melo Taciana Melcop L. de MELO. Estrategia multi-agente para leiloes simultaneos de bens relacionados. M.Sc.Diss. Port. Presentation: 20/03/03 119 p. Advisor: Ruy Luiz Milidiu Abstract: This work presents a multi-agent system to trade in simultaneous auctions of related goods. The dissertation describes the multi-agent architecture, and also the analysis and development of strategies for trading in simultaneous auctions, where the purchase of combined goods is required. Some well known problems in trading were identified in order to design the architecture, such as price prediction, good allocation, decision making, reasoning under uncertainty, and demand segmentation. Each agent that composes the system is concerned with one of those trading subproblems. This makes possible to apply different computational techniques to separately solve the subproblems and then combine the solutions. The Trading Agent Competition (TAC) is used to illustrate our approach. TAC was chosen to test the developed heuristics since it presents a set of characteristics that adequately fits the problem domain. Each heurist developed was tested and had its results compared to TAC previous editions. Finally, the system shows a high performance on very competitive scenarios tested by using the TAC server environment.