Ο σκοπός αυτής της διπλωματικής εργασίας είναι η θεωρητική μελέτη της τεχνολογίας των πρακτόρων λογισμικού και ο τρόπος εφαρμογής τους στο ηλεκτρονικό εμπόριο. Για να γίνει εφικτή η κατανόηση του όρου «τεχνολογία πρακτόρων» ξεκινάμε με τον όρο «πράκτορας». Στην έρευνα η οποία πραγματοποιήθηκε για τον προσδιορισμό των πρακτόρων λογισμικού, παρατηρήθηκε η ύπαρξη μιας πληθώρας διαφορετικών ορισμών, γεγονός το οποίο οφείλεται στην διαφορετική οπτική γωνία με την οποία ο καθένας τους αντιλαμβάνεται και τους μελετά. Ένας απλός ορισμός που θα μας εισάγει στην έννοια των πρακτόρων λογισμικού είναι ο ακόλουθος : «Ο πράκτορας είναι μια λογισμική οντότητα που ξέρει πώς να κάνει ενέργειες που θα μπορούσε να πραγματοποιήσει ένα λογικό ον (άνθρωπος) αν είχε το χρόνο.» Επιπλέον αναφέρουμε ποια χαρακτηριστικά πρέπει να συγκεντρώνουν για να λογίζονται ως ευφυείς πράκτορες λογισμικού. Η αντιδραστικότητα, ο προσανατολισμός στόχου, η δυνατότητα μάθησης, η αυτονομία, η δυνατότητα μεταφοράς, η επικοινωνία/συνεργασία και ο χαρακτήρας είναι τα χαρακτηριστικά εκείνα που είναι απαραίτητα για να θεωρηθεί μια λογισμική οντότητα ως ευφυής πράκτορας λογισμικού. Επίσης, μελετάμε και τις κατηγορίες των ευφυών πρακτόρων. Η κατηγοριοποίηση μπορεί να γίνει είτε με βάση τα χαρακτηριστικά τους, είτε με βάση τους ρόλους που καλούνται να διαδραματίσουν. Στη συνέχεια, προκειμένου να κατανοήσουμε την «τεχνολογία» τους, παρουσιάζουμε τον τρόπο με τον οποίο κατασκευάζονται, μέσα από τις γλώσσες και τα εργαλεία κατασκευής. Ακολούθως αναφέρουμε τα περιβάλλοντα των πρακτόρων, τα οποία μπορούν να ταξινομηθούν ανάλογα με τις ιδιότητες και τα χαρακτηριστικά τους καθώς αυτά έχουν άμεση επίδραση στο σχεδιασμό ενός πράκτορα και στη διαδικασία λήψης των αποφάσεών του. Στη συνέχεια μελετούμε την αρχιτεκτονική ενός πράκτορα, η οποία καθορίζει τα συστατικά του στοιχεία και τις αλληλεπιδράσεις μεταξύ τους. Με τη σειρά τους, τα συστατικά στοιχεία του πράκτορα και οι αλληλεπιδράσεις τους καθορίζουν τις ενέργειες του πράκτορα και τη μελλοντική του κατάσταση, σε σχέση με το ποια είναι τα δεδομένα που προσλαμβάνονται στα αισθητήρια όργανα του πράκτορα και το ποια είναι η τωρινή του εσωτερική του κατάσταση. Η παρούσα διπλωματική εργασία συνεχίζει με την μελέτη των πολυπρακτορικών συστημάτων, του τρόπου οργάνωσης τους, αλληλεπίδρασης και επικοινωνίας τους. Επιπλέον παρουσιάζουμε τις σημαντικότερες εφαρμογές ευφυών πρακτόρων, όπως η Διαχείριση Ηλεκτρονικών Μηνυμάτων (e–mail), η Αναζήτηση Πληροφοριών, τα Προσαρμοστικά Συστήματα Διεπαφής, το Ηλεκτρονικό Εμπόριο, η Διαχείριση Συστήματος και Δικτύου, η Απομακρυσμένη Πρόσβαση, η Συνεργασία και οι Ψυχαγωγικές εφαρμογές. Το ενδιαφέρον μας επικεντρώνεται περισσότερο, λόγω του θέματος της εργασίας μας, στο Ηλεκτρονικό Εμπόριο. Αναλύουμε τα πλεονεκτήματα, τα μειονεκτήματα καθώς και τις μορφές του. Επίσης γίνεται λόγος για τις ηλεκτρονικές αγορές, τη σημασία και τον ρόλο τους. Επιπλέον, καταθέτουμε κάποια στοιχεία για τα νομικά πλαίσια που διέπουν τη δημιουργία πρακτόρων. Γίνεται, επίσης, μια σύντομη αναφορά για τα θέματα ασφαλείας που αφορούν τη λειτουργία των πρακτόρων. Τέλος, παρουσιάζουμε ένα σύστημα ηλεκτρονικού εμπορίου βασισμένο στην τεχνολογία των ευφυών πρακτόρων λογισμικού. Για την κατασκευή του συστήματος χρησιμοποιήθηκε το εργαλείο κατασκευής πρακτόρων Zeus. Η εφαρμογή περιγράφει την υλοποίηση ενός απλού σεναρίου αγοράς, που ονομάζεται Ηλεκτρονική Αγορά. Σκοπός της εφαρμογής είναι να επιτρέπει στους συμμετέχοντες της αγοράς να ανταλλάσσουν ηλεκτρονικά αγαθά μέσω του διαδικτύου με χρήση ευφυών πρακτόρων λογισμικού. Η διπλωματική εργασία ολοκληρώνεται με τους προβληματισμούς που υπάρχουν γύρω από τη δημιουργία πραγματικών εφαρμογών πρακτόρων στο ηλεκτρονικό εμπόριο. / The aim of this diploma thesis is the theoretical study of the software agents technology and their way of application in e-commerce. In order to be feasible the comprehension of term “technology of agents” we begin with the term “agent”. In the research which was realised for the determination of software agents, existence of abundance of different definitions was observed, something which is owed in the different optical corner with which each one conceives them and studies them. A simple definition that will import us in the significance of agents of software is the attaché: “The agent is a computational entity that knows how to make energies that a reasonable being (person) could realise a if it had the time.” Moreover we report which characteristics they should assemble in order to characterize them as intelligent software agents. Reactiveness, the orientation of the goal, the possibility of learning, the autonomy, the possibility of transport, the communication/collaboration and the character are the characteristics that are essential in order to consider a computational entity as intelligent software agents. Also, we study the categories of intelligent software agents. The categorisation can be done either based to their characteristics, either based to the roles that they are called to play. Afterwards, in order to understand their “technology”, we present the way they are manufactured, through the languages and the tools of manufacture. Followingly we report the environments of agents, which can be categorized depending on the attributes and their characteristics, also they have direct effect in the planning of agent and in the process of his decision-making. Afterwards we study the architecture of agent, which determines his constitutive elements and the interactions from each other. At their time, the constitutive elements of an agent and their interactions determine the energies of agent and his future situation, in combination which are the data that are engaged in the sensory bodies of agent and which is his actual internal situation. The present diploma thesis continues with the study of multiagent systems, their way of organisation, interaction and their communication. Moreover we present the most important applications of intelligent software agents, as the Management of Electronic Messages (e-mail), the Search of Information, the Adoptive Systems of Contact, the Electronic Commerce, the Management of System and Network, the Removed Access, the Collaboration and the Recreational applications. Our interest is most focused, because of the subject of our diploma thesis, in the Electronic Commerce. We analyze the advantages, the disadvantages as well as his forms. Also, we report the electronic markets, their importance and their role. Moreover, we deposit certain elements for the legal frames that condition the creation of agents. Also, a short report, on the subjects of safety that concern the operation of agents, is done. Finally, we present a system of electronic commerce based on the technology of intelligent software agents. For the manufacture of system was used the tool of manufacture of agents Zeus. The application describes the concretisation of simple script of market, that is named Electronic Market. Aim of application is to allow in participating the market to exchange electronic goods via the internet with use of intelligent software agents. The diploma thesis is completed with the reflections that exist round the creation of real applications of agents in the electronic commerce.

In order to succeed, it is imperative that all e-commerce systems include an effective and reliable trust and reputation modeling system. This is particularly true of decentralized e-commerce systems in which autonomous software engage in commercial transactions. Many researchers have sought to overcome the complexities of modeling a subjective, human concept like trust, resulting in several trust and reputation models. While these models each present a unique offering and solution to the problem, several issues persist. Most of the models require direct experience in the e-commerce system in order to make effective trust decisions. This leaves new agents and agents who only casually use the e-commerce system vulnerable. Additionally, the reputation ratings of agents who are relatively new to the system are often indistinguishable from scores for poorly performing agents. Finally, more tactics to defend against agents who exploit the characteristics of the open, distributed system for their own malicious needs are required. To address these issues, a new metric is devised and presented: seniority. Based on agent age and activity level within the e-commerce system, seniority provides a means of judging the credibility of other agents with little or no prior experience in the system. As the results of experimental analysis reveals, employing a reputation model that uses seniority provides considerable value to agents who are new agents, casual buyer agents and all other purchasing agents in the e-commerce system. This new metric therefore offers a significant contribution toward the development of enhanced and new trust and reputation models for deployment in real-world distributed e-commerce environments.

Software agents have been heralded as the most important emerging technology of the decade. As software development firms eagerly attempt to integrate these autonomous programs into their products, researchers attempt to define the concept of agency and to develop architectures that will improve agent capabilities. Decision Support System (DSS) researchers have been eager to integrate agents into their applications, and exploratory works in which agents have been used within a DSS have been documented. This dissertation attempts to further this exploration by studying the agent features and underlying architectures that can lead to the successful integration of agents in DSS. This exploration is carried out in three parts. In the first part, a review of the relevant research streams is provided. The history and current status of software agents is first discussed. Similarly, a historical and current view of DSS research is provided. Lastly, a historical and tutorial-type of discussion is provided on the topic of Artificial Intelligence (AI) planning. This review of the relevant literature provides a general background for the conceptual analyses and implementations that are carried out in the next two sections. In the second part, the literature on software agents is synthesized to develop a definition of agency applicable to DSS. Using this definition, an agent-integrated DSS that supports variance-analysis is designed and developed. Following this implementation, a general framework for agent-enabling DSS is suggested. The use of this framework promises to raise some DSS to a new level of capability whereby "what-if" systems are transformed into real-time, proactive systems. The third part utilizes this general framework to agent-enable a corporate-planning system DSS and extends the framework in the second section through the introduction of an automated-planning agent. The agent uses AI planning to generate decision-making alternatives, providing a means to integrate and sequence the models in the DSS. The architecture used to support this planning agent is described. This new kind of DSS enables not only the monitoring of goals, but also the maintenance of these goals through agent-generated plans. The conclusion summarizes the contributions of this work and outlines in considerable detail potential research opportunities in the realm of software agents, DSS, and planning. / Ph. D.

This research is an effort to test our hypothesis that a software agent based architecture will provide a better response time and will be more maintainable and reusable than the present J2EE based architecture of DLNET (Digital Library Network for Engineering and Technology). We have taken a portion of the complete DLNET application for our study, namely the Content Review Process, as our test bed. In this work, we have explored the use of software agents in the current setup of DLNET for the first time, specifically for the Content Review part of the application and tried to evaluate the performance of the resulting application. Our work is a novel approach of doing content review using software agent architecture. The proposed system is an automated process that will asynchronously look for suitable reviewers based on content (the input) and create logs for the administrator to view and analyze. In the first part of the thesis we develop a new system that is parallel to the existing DLNET Content Review Process. In the second part, we compare the newly developed Content Review Process with the baseline (old Content review Process) by designing comparison tests and measuring instruments. This part of the thesis includes the selection of dependent variables, design of various measurement instruments, execution of the quasi-experiments and analysis of the empirical results of comparisons tests. The quasi-experiments are done to measure the response time, maintainability, scalability, correctness, reliability and reusability of the two systems. The results show that the proposed software agent based system gives better response time (an improvement ranging from 57% to 82%) and is more maintainable (an improvement ranging from 16% to 67%) and more reusable (an improvement ranging from 1% to 26%). The improvement in the response time may be attributed to the fact that the agent based systems are inherently multithreaded while the existing content review system is a serial application. Both the systems, however, give comparable results for other dependent variables. / Master of Science

[pt] Teste de software é uma atividade crítica no processo de desenvolvimento de sistemas, principalmente devido ao aumento da complexidade das aplicações atuais e pelo impacto que tais aplicações podem gerar. Relacionado a essa complexidade, o paradigma de sistemas multiagentes tem sido utilizado especialmente quando entidades pró-ativas, autônomas, autoadaptáveis e distribuídas precisam ser desenvolvidas. Para acompanhar a criação, manutenção e execução dos testes nesses sistemas, torna-se necessária a adoção de um processo de gerência, pois prevê a realização das atividades de planejamento, projeto, execução e acompanhamento dos testes. Visando ajudar nessa gerência, a tese apresenta o Java self-Adaptive Agent Framework for Self- Test (JAAF mais T), framework que permite a criação de agentes autoadaptativos capazes de realizar autoteste, isto é, coordenar a execução dos testes necessários para validar suas autoadaptações. Como diversas informações são usadas para ajudar na gerência desses testes, documentá-las ajuda a entender como evoluir e executá-los. Baseada nessa preocupação, a tese oferece uma nova linguagem de modelagem chamada de UML Testing Profile for Coordination (UTP-C), perfil (profile) da UML que permite a modelagem dessas informações. Por fim, para automatizar o processo de gerência dos testes executados por agentes autoadaptativos, a tese apresenta ferramentas capazes de gerar de forma automática artefatos usados pelo JAAF mais T baseados em modelos UTP-C. / [en] Testing software systems has become a critical activity of software development over time. Especially when the development of complex systems, which are able to self-adapt theirs behaviors when necessary, is becoming extremely common. In this context, the multi-agent system (MAS) paradigm has been used especially when distributed, autonomous and pro-active entities are represented. Aiming to control the creation, maintenance and execution of tests on these systems, it is necessary to adopt a management process that considers the activities of planning, design, execution and monitoring of the tests. In order to help this management, the thesis presents the Java self-Adaptive Agent Framework for Self-Test (JAAF more T), that aims to allow the creation of selfadaptive agents that perform self-tests. Since several data are used to help the management of the tests, documenting them helps to understand how to evolve and execute them. Based on this concern, the thesis offers a new modeling language called UML Testing Profile for Coordination (UTP-C), profile of the UML that allows the modeling of these data. At last, but not least important, tools, which were created to automate the management of the tests executed for selfadaptive agents, are presented to generate useful artifacts used for instances of the JAAF more T based on UTP-C models.

[pt] Este trabalho apresenta um modelo de avaliação de aprendizagem em debates síncronos e presenciais mediados pela web. O modelo é construído para atender as necessidades da avaliação do ensino levado a efeito segundo três paradigmas metodológicos: a aprendizagem centrada no aprendiz, a aprendizagem colaborativa e a aprendizagem guiada por problemas. Uma máquina de estados é proposta para suporte à avaliação da participação do aprendiz, segundo a Teoria do Vínculo de Enrique Pichon Rivière, permitindo assim um juízo de valor sobre comportamentos observáveis e aferição de aprendizagem de conhecimentos em um processo de construção coletiva e baseada em problemas colocados como desafios para os aprendizes. O ambiente AulaNetTM é utilizado como referencial, dadas as suas características de servidor de conteúdo e por se constituir em um ambiente para ensino à distância cuja concepção permite o uso da metodologia de ensino cooperativo. O acoplamento dos sistemas foi concebido de modo a atender as diretivas de simplicidade e dependência estrutural mínima. Para o módulo de avaliação é suficiente acessar a base do AulaNetTM e transferir os dados para a base do mesmo, a fim de garantir a independência dos sistemas. Um padrão do tipo quadro-negro (blackboard) é projetado para implementar essas características. / [en] This work presents a model for learning evaluation to be used in web-based learning systems. This model is constructed to provide conditions for a correct use of three learning paradigms: Learner centered learning, Cooperative Learning and Problem Based Learning. A finite state machine is proposed to support the evaluation of the apprentice`s participation under the Vinculum Theory by Enrique Pichon Rivière. As designed, the teacher can do a judgement of the observable behavior of each student and make an evaluation of the knowledge learning in a collective construction process based on the PBL (Problem Based Learning) strategy. The AulaNetTM environment is used in this work because it has a well-designed content server structure and is strongly oriented towards collaborative learning. The binding of the proposed model with the AulaNetTM environment follows the guidelines of simplicity and minimal dependence of both systems. A blackboard architecture is implemented to get the characteristics of independence and low complexity described above.

This thesis presents a study of the fragmented nature of activities involved in the specification and procurement of construction products. Existing Internet-based approaches have perpetuated fragmentation in this domain and this has resulted in a problem known as 'digital anarchy.' The thesis describes the design and development of an application based on the metaphor of software agents to address this problem. The thesis has validated the view that failures in the existing approaches have resulted in the 'digital anarchy' problem. Such approaches have failed to eradicate paper-based communication patterns and its related sequential work methods. The thesis has established the proof of concept of using agents for the management of distributed information in the selected domain. The work has also synthesised agent technology, XML and Web Services into a cohesive architecture, demonstrating that the ultimate solution to the 'digital anarchy' problem will emerge from a synergy of the various streams of research in Internet-related technologies. The work has established that there is significant scope for systematic improvement in the specification and procurement of construction products listed on the Internet based on the agent paradigm.

Submitted by Rosivalda Pereira (mrs.pereira@ufma.br) on 2017-08-11T20:35:45Z No. of bitstreams: 1 CleberPereira.pdf: 3089056 bytes, checksum: ab85b5af73443dbf91e821cbac7192ab (MD5) / Made available in DSpace on 2017-08-11T20:35:45Z (GMT). No. of bitstreams: 1 CleberPereira.pdf: 3089056 bytes, checksum: ab85b5af73443dbf91e821cbac7192ab (MD5) Previous issue date: 2010-08-09 / The proposed modeling system for assessing undergraduate degree courses, according to the parameters required by ENADE emerges from the need of the IES adjust its methodology and assessment procedures to the terms set by the exam. The proposed solution does not address a specific undergraduate course and may be applied for any knowledge area. The use of software agents is justified on the grounds of a classical computer application not being able to adapt to the reality of each student, delivering the second behavioral questions identified during the evaluation process.The statistics generated from this reasoning will serve both for students and for teachers in order to reevaluate the teaching and learning process. The modeling was performed according to the UML 2.0 TROPOS methodology was used for the modeling of agents and their implementation was done in JADE. / A proposta de modelagem de sistema de avaliação dos cursos superiores de graduação, segundo os parâmetros exigidos pelo ENADE, emerge da necessidade das IES adequarem sua metodologia e procedimentos de avaliação aos moldes estabelecidos pelo exame. A solução proposta não se dirige a um curso específico de graduação, podendo ser utilizada para qualquer área do conhecimento. O uso de agentes de software em seu desenvolvimento justifica-se em razão de uma aplicação computacional clássica não ser capaz de adaptar-se à realidade de cada aluno, entregando quesitos segundo o comportamento identificado durante o processo de avaliação. As estatísticas geradas a partir deste raciocínio servirão tanto para os alunos como para os docentes no sentido de reavaliar o processo de ensino aprendizagem. A modelagem foi efetuada segundo a UML 2.0, foi utilizada a metodologia TROPOS para a modelagem dos agentes e sua implementação foi realizada no JADE.

A gestão de cursos de graduação, no âmbito da coordenação de curso, está relacionada à execução de práticas que visam à melhoria da qualidade de ensino, pesquisa e extensão. No entanto, o que se observa, na maioria dos casos, é que os gestores não apresentam conhecimento aprofundado sobre os processos que estão sob sua responsabilidade, no que se refere ao fluxo de atividades e informações, devido, usualmente, a pouca qualificação ao assumirem cargos administrativos, ocorrendo uma necessidade de profissionalização da gestão. Esta tese propõe a capacitação de coordenadores de cursos de graduação, no que concerne aos seus processos de negócio, através da utilização de Jogos de Empresa. A mesma apresenta o jogo SPOIE – Simulador de Processos Organizacionais baseado em Instituições Eletrônicas, que pretende auxiliar o coordenador de curso a conhecer melhor o fluxo dos processos, os participantes envolvidos em cada processo, bem como as interações que devem ocorrer durante sua realização. A abordagem de Instituições Eletrônicas e Agentes de Software é utilizada de forma a prover mecanismos de representação das diversas interações existentes nos processos institucionais. Por fim, é apresentada uma análise da aplicação do SPOIE à coordenadores de cursos de graduação de diferentes Instituições de Educação Superior. Com a implementação da ferramenta e a aplicação realizada, percebe-se que é possível desenvolver um treinamento autônomo dos coordenadores de curso de graduação, no que concerne aos seus processos administrativos, com a abordagem supracitada. / The undergraduate programs management (course coordination) concerns in the practices implementation to qualify the teaching, research and extension. However, in most cases, the managers do not have the essential knowledge about the processes that are under their responsibility. Usually, the information flow and their activities are not so precise to the managers. A possible answer is the professor poor skills when he needs to assume an administrative position. This thesis proposes an undergraduate coordinators programs trainee. This program was focused in the business process through an Enterprise Games approach. The article presents the SPOIE - Organizational Process Simulator-based on Electronic Institutions - game, which was developed to assist the coordinator to understand the workflow, the participants involved in each case, as well as the interactions that must occur during its implementation. The software agents and the Electronic Institutions approaches are used to provide a formal representation among the interactions in the existing institutional processes. Finally, we present the SPOIE analysis held in different universities. Considering the above assertions, it is possible to conclude that the game can be used in the coordinators autonomous training of their administrative processes.

A tecnologia de agentes surge como uma solução que fornece flexibilidade e robustez para lidar com domínios dinâmicos e complexos. Tal flexibilidade pode ser alcançada através da adoção de abordagens já existentes baseadas em agentes, como a arquitetura BDI, que provê agentes com características mentais de crenças, desejos e intenções. Essa arquitetura é altamente personalizável, deixando lacunas a serem preenchidas de acordo com aplicações específicas. Uma dessas lacunas é o algoritmo de seleção de planos, responsável por selecionar um plano para ser executado pelo agente buscando atingir um objetivo, e tendo grande influência no desempenho geral do agente. Grande parte das abordagens existentes requerem considerável esforço para personalização e ajuste a fim de serem utilizadas em aplicações específicas. Nessa dissertação, propomos uma abordagem para seleção de planos apta a aprender quais planos possivelmente terão os melhores resultados, baseando-se no contexto atual e nas preferências do agente. Nossa abordagem é composta por um meta-modelo, que deve ser instanciado a fim de especificar metadados de planos, e uma técnica que usa tais metadados para aprender e predizer resultados da execução destes planos. Avaliamos nossa abordagem experimentalmente e os resultados indicam que ela é efetiva. Adicionalmente, fornecemos uma ferramenta para apoiar o processo de desenvolvimento de agentes de software baseados em nosso trabalho. Essa ferramenta permite que desenvolvedores modelem e gerem código-fonte para agentes BDI com capacidades de aprendizado. Um estudo com usuários foi realizado para avaliar os benefícios de um método de desenvolvimento baseado em agentes BDI auxiliado por ferramenta. Evidências sugerem que nossa ferramenta pode auxiliar desenvolvedores que não sejam especialistas ou que não estejam familiarizados com a tecnologia de agentes. / Agent technology arises as a solution that provides flexibility and robustness to deal with dynamic and complex domains. Such flexibility can be achieved by the adoption of existing agent-based approaches, such as the BDI architecture, which provides agents with the mental attitudes of beliefs, desires and intentions. This architecture is highly customisable, leaving gaps to be fulfilled in particular applications. One of these gaps is the plan selection algorithm that is responsible for selecting a plan to be executed by an agent to achieve a goal, having an important influence on the overall agent performance. Most existing approaches require considerable effort for customisation and adjustment to be used in particular applications. In this dissertation, we propose a plan selection approach that is able to learn plans that provide possibly best outcomes, based on current context and agent’s preferences. Our approach is composed of a meta-model, which must be instantiated to specify plan metadata, and a technique that uses such metadata to learn and predict plan outcomes. We evaluated our approach experimentally, and results indicate it is effective. Additionally, we provide a tool to support the development process of software agents based on our work. This tool allows developers to model and generate source code for BDI agents with learning capabilities. A user study was performed to assess the improvements of a tool-supported BDI-agent-based development method, and evidences suggest that our tool can help developers that are not experts or are unfamiliar with the agent technology.