Agrupamento de dados é o processo que consiste em dividir um conjunto de dados em grupos de forma que dados semelhantes entre si permaneçam no mesmo grupo enquanto que dados dissimilares sejam alocados em grupos diferentes. Técnicas tradicionais de agrupamento de dados têm sido usualmente desenvolvidas de maneira centralizada dependendo assim de estruturas que devem ser acessadas e modificadas a cada passo do processo de agrupamento. Além disso, os resultados gerados por tais métodos são dependentes de informações que devem ser fornecidas a priori como por exemplo número de grupos, tamanho do grupo ou densidade mínima/máxima permitida para o grupo. O presente trabalho visa propor o bee clustering, um algoritmo distribuído inspirado principalmente em técnicas de inteligência de enxames como organização de colônias de abelhas e alocação de tarefas em insetos sociais, desenvolvido com o objetivo de resolver o problema de agrupamento de dados sem a necessidade de pistas sobre o resultado desejado ou inicialização de parâmetros complexos. O bee clustering é capaz de formar grupos de agentes de maneira distribuída, uma necessidade típica em cenários de sistemas multiagente que exijam capacidade de auto-organização sem controle centralizado. Os resultados obtidos mostram que é possível atingir resultados comparáveis as abordagens centralizadas. / Clustering can be defined as a set of techniques that separate a data set into groups of similar objects. Data items within the same group are more similar than objects of different groups. Traditional clustering methods have been usually developed in a centralized fashion. One reason for this is that this form of clustering relies on data structures that must be accessed and modified at each step of the clustering process. Another issue with classical clustering methods is that they need some hints about the target clustering. These hints include for example the number of clusters, the expected cluster size, or the minimum density of clusters. In this work we propose a clustering algorithm that is inspired by swarm intelligence techniques such as the organization of bee colonies and task allocation among social insects. Our proposed algorithm is developed in a decentralized fashion without any initial information about number of classes, number of partitions, and size of partition, and without the need of complex parameters. The bee clustering algorithm is able to form groups of agents in a distributed way, a typical necessity in multiagent scenarios that require self-organization without central control. The performance of our algorithm shows that it is possible to achieve results that are comparable to those from centralized approaches.

Orientador: Leonardo de Souza Mendes / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-10T16:36:33Z (GMT). No. of bitstreams: 1 Mattos_EklerPaulinode_M.pdf: 2049372 bytes, checksum: d292b2d52c2a0c0bf5326e8f7e533665 (MD5) Previous issue date: 2007 / Resumo: O Sistema Tutor Inteligente (STI) pertence a uma categoria de sistemas de natureza educacional, utilizado como ferramenta de suporte ao ensino-aprendizagem. Possui uma estrutura modular que tem por finalidade auxiliar o aprendiz na realização de atividades educacionais, bem como a capacidade de adaptar-se de acordo com as necessidades de um aprendiz, o que faz do STI uma arquitetura interessante na construção de softwares educacionais. O trabalho proposto tem por objetivo utilizar a arquitetura de STI, aplicada à área de sistemas de e-Gov como proposta de solução de problemas de natureza distribuída. Como estudo de caso, foi escolhida a área Gestão de Materiais e Medicamentos, justamente por apresentar problema pertinente à distribuição de materiais e medicamentos, nas unidades básicas de saúde (UBS). Cada STI funciona como representante de uma UBS, que tem por função realizar o papel de um agente gestor de estoque (Agente Gestor Tutor - AGT), cargo pouco comum na rede municipal de saúde, auxiliando o administrador de cada setor (visto como o aprendiz) a realizar tarefas complexas de gestão de materiais e medicamentos. Foi realizada uma série de simulações usando o protótipo desenvolvido para testar a sua viabilidade de aplicação com relação ao tratamento do estoque distribuído de uma arquitetura de rede municipal de saúde / Abstract: The Intelligent Tutor Systems (ITS) belong to a category of educational nature systems, used as a tool to support the teaching and learning. It has a modular structure, which aim to help the apprentice in the execution of educational activities, as well as in adapting itself according to the apprentice¿s necessities, what makes the ITS, an interesting architecture in the construction of educational softwares. The proposed work aim to use the ITS architecture in the management of materials, as a solution for the problem of medicine distribution in the health basic units (HBU). Each ITS works as a HBU representative, whose function is to play the role of a managing agent of supply (Tutorial Managing Agent - TMA), a post job not so common in the municipal health¿s network. The TMA assists the administrator of each sector (seen as the apprentice) in executing complex tasks of management of materials and medicines.In this way, many simulations were carried out, using the developed prototype to test its feasibility of application, in relation to the management of the distributed materials of architecture of a municipal health¿s network / Mestrado / Telecomunicações e Telemática / Mestre em Engenharia Elétrica

Made available in DSpace on 2016-08-17T14:53:04Z (GMT). No. of bitstreams: 1 Paulo Jose Melo Gomes Correa.pdf: 5684127 bytes, checksum: c4aae2b2365a502a84da14364cdf5d4d (MD5) Previous issue date: 2009-09-14 / This research presents studies involving Electricity Engineering and Oil and Biofuel Analytical Chemistry areas, whose objective is the perfectioning of chemical analysis steps for the Fuel Quality Monitoring Program, instituted by the National Agency of Petroleum and Biofuel - ANP, executed in the Maranhão State by the Laboratory of Analyses and Research in Analytical Chemistry of Petroleum and Biofuel - LAPQAP. For this, improvements were proposed for the Fuel Quality Control Multiagente System, and help for the decisions taking of the laboratory. The inclusion of new agents to the multiagent society is still considered, where the objective is to add a new technique automatized for chemical analyses, beyond additional functionalities. A fuel ontology was considered for the communication mechanism that will be shaped using 101 methodology, whose objective is the representation of domain knowledge of chemical analyses beyond supply of a communication language among the society agents. In order to reach the objectives we used Artificial Intelligence techniques, inference motor JESS (Java Expert System Shell), an ontology technology to represent the domain knowledge serving as vocabulary of the communication process, the middleware JADE (Java Agent DEvelopment framework) for environment execution with their improvements and the development methodology of multiagent systems named PASSI for the system modeling. / Esta pesquisa apresenta estudos envolvendo as áreas de Engenharia de Eletricidade e Química Analítica de Petróleo e Bicombustíveis, tendo como objetivo o aperfeiçoamento das etapas de análise químicas do Programa de Monitoramento da Qualidade de Combustíveis (PMQC), instituído pela Agência Nacional de Petróleo e Biocombustíveis ANP, executado no Estado do Maranhão pelo Laboratório de Análises e Pesquisa em Química Analítica de Petróleo e Biocombustíveis - LAPQAP. Para isto, são propostas melhorias no Sistema Inteligente de Monitoramento e Controle da Qualidade de Combustíveis SIMCQC utilizado pelo laboratório no auxilio a tomada de decisões. Propõe-se a inclusão de novos agentes à sociedade multiagente, tendo como objetivo aumentar a quantidade de técnicas análises químicas no SIMCQC. Para o mecanismo de comunicação é mostrada a criação de uma ontologia de combustíveis que foi modelada utilizando-se da metodologia 101, cujo objetivo é a representação do conhecimento do domínio de análises químicas e o fornecimento de uma linguagem de conteúdo para o mecanismo de comunicação dos agentes da sociedade. Para o alcance dos objetivos foram utilizadas técnicas de Inteligência Artificial, motor de inferência JESS (Java Expert System Shell), a tecnologia de Ontologia para representar o conhecimento do domínio e servir como vocabulário do processo de comunicação, o middleware JADE (Java Agent DEvelopment Framework) para execução do ambiente com suas melhorias e a metodologia de desenvolvimento de sistemas multiagente PASSI para a modelagem do sistema.

M.Sc. (Information Technology) / The purpose of the research presented in this dissertation is to explore the uses of an adaptive multi-agent system for critical information infrastructure protection (CIIP). As the name suggests, CIIP is the process of protecting the information system which are connected to the infrastructure essential to the continued running of a country or organisation. CIIP is challenging due largely to the diversity of these infrastructures. The dissertation examines a number of artificial intelligence techniques that can be applied to CIIP; these techniques range from multi-agent systems to swarm optimisation. The task of protection is broken into three distinct areas: preventing unauthorised communication from outside the system; identifying anomalous actions on computers within the system; and ensuring that communication within the system is not modified externally. A multi-agent learning model, MALAMANTEAU, is proposed as a way to address the problem of CIIP. Due to various problems facing CIIP, multi-agent systems present good opportunities for solving these many problems in a single model. Agents within the MALAMANTEAU model will use diverse artificial and computational intelligence techniques in order to provide an adaptable approach to protecting critical networks. The research presented in the dissertation shows how computational intelligence can be employed alongside multi-agent systems in order to provide powerful protection for critical networks without exposing further security risks.

[pt] Sistemas multiagentes foram introduzidos como um novo paradigma para a conceituação, concepção e implementação de sistemas de software que estão se tornando cada vez mais complexos, abertos, distribuídos, dinâmicos, autônomos e altamente interativos. No entanto, a engenharia de software orientada a agentes não tem sido amplamente adotada, principalmente devido à falta de linguagens de modelagem que não conseguem ser expressivas e abrangentes o suficiente para representar abstrações relacionadas aos agentes de software e apoiar o refinamento dos modelos de projeto em código. A maioria das linguagens de modelagem não define como essas abstrações devem interagir em tempo de execução, mas muitas aplicações de software precisam adaptar o seu comportamento, reagir à mudanças em seus ambientes de forma dinâmica, e alinhar-se com algum tipo de comportamento individual ou coletivo de aplicações normativas (por exemplo, obrigações, proibições e permissões). Neste trabalho, foi proposta uma abordagem de metamodelo e uma arquitetura para o desenvolvimento de agentes adaptativos normativos. Acredita-se que a abordagem proposta vai avançar o estado da arte em sistemas de agentes de modo que tecnologias de software para aplicações dinâmicas, adaptáveis e baseadas em normas possam ser projetadas e implementadas. / [en] Multi-agent systems have been introduced as a new paradigm for conceptualizing, designing and implementing software systems that are becoming increasingly complex, open, distributed, dynamic, autonomous and highly interactive. However, agent-oriented software engineering has not been widely adopted, mainly due to lack of modeling languages that are expressive and comprehensive enough to represent relevant agent-related abstractions and support the refinement of design models into code. Most modeling languages do not define how these abstractions interact at runtime, but many software applications need to adapt their behavior, react to changes in their environments dynamically, and align with some form of individual or collective normative application behavior (e.g., obligations, prohibitions and permissions). In this paper, we propose a metamodel and an architecture approach to developing adaptive normative agents. We believe the proposed approach will advance the state of the art in agent systems so that software technologies for dynamic, adaptive, norm-based applications can be designed and implemented.

The thesis is dealing with traffic management using telemetry networks. The problematic of telemetry networks and multiagent systems. A simulation model is proposed in Java which enables configuration simulation and assessment.

<p dir="ltr">The growth of electric vehicles (EVs) offers several benefits for air quality improvement and emissions reduction. Nonetheless, EVs also pose several challenges in the area of highway transportation. These barriers are related to the limitations of EV technology, particularly the charge duration and speed of battery recharging, which translate to vehicle range anxiety for EV users. A promising solution to these concerns is V2V DWC technology (Vehicle to Vehicle Dynamic Wireless Charging), particularly mobile energy disseminators (MEDs). The MED is mounted on a large vehicle or truck that charges all participating EVs within a specified locus from the MED. However, current research on MEDs offers solutions that are widely considered impractical for deployment, particularly in urban environments where range anxiety is common. Acknowledging such gap in the literature, this thesis proposes a comprehensive methodological framework for optimal MED deployment decisions. In the first component of the framework, a practical system, termed “ChargingEnv” is developed using reinforcement learning (RL). ChargingEnv simulates the highway environment, which consists of streams of EVs and an MED. The simulation accounts for a possible misalignment of the charging panel and incorporates a realistic EV battery model. The second component of the framework uses multiple deep RL benchmark models that are trained in “ChargingEnv” to maximize EV service quality within limited charging resource constraints. In this study, numerical experiments were conducted to demonstrate the MED deployment decision framework’s efficacy. The findings indicate that the framework’s trained model can substantially improve EV travel range and alleviate battery depletion concerns. This could serve as a vital tool that allows public-sector road agencies or private-sector commercial entities to efficiently orchestrate MED deployments to maximize service cost-effectiveness.</p>

<p dir="ltr">The concept of reinforcement learning has become increasingly relevant in learning- based applications, especially in the field of autonomous navigation, because of its fundamental nature to operate without the necessity of labeled data. However, the infeasibility of training reinforcement learning based autonomous navigation applications in a real-world setting has increased the popularity of researching and developing on autonomous navigation systems by creating simulated environments in game engine platforms. This thesis investigates the comparative performance of Unity and Unreal Engine 5 within the framework of a reinforcement learning system applied to autonomous race car navigation. A rudimentary simulated setting featuring a model car navigating a racetrack is developed, ensuring uniformity in environmental aspects across both Unity and Unreal Engine 5. The research employs reinforcement learning with genetic algorithms to instruct the model car in race track navigation; while the tools and programming methods for implementing reinforcement learning vary between the platforms, the fundamental concept of reinforcement learning via genetic algorithms remains consistent to facilitate meaningful comparisons. The implementation includes logging of key performance variables during run times on each platform. A comparative analysis of the performance data collected demonstrates Unreal Engine's superior performance across the collected variables. These findings contribute insights to the field of autonomous navigation systems development and reinforce the significance of choosing an optimal underlying simulation platform for reinforcement learning applications.</p>

<p dir="ltr">Autonomous vehicles require motion forecasting of their surrounding multiagents (pedestrians</p><p dir="ltr">and vehicles) to make optimal decisions for navigation. The existing methods focus on</p><p dir="ltr">techniques to utilize the positions and velocities of these agents and fail to capture semantic</p><p dir="ltr">information from the scene. Moreover, to mitigate the increase in computational complexity</p><p dir="ltr">associated with the number of agents in the scene, some works leverage Euclidean distance to</p><p dir="ltr">prune far-away agents. However, distance-based metric alone is insufficient to select relevant</p><p dir="ltr">agents and accurately perform their predictions. To resolve these issues, we propose the</p><p dir="ltr">Semantics-aware Interactive Multiagent Motion Forecasting (SIMMF) method to capture</p><p dir="ltr">semantics along with spatial information and optimally select relevant agents for motion</p><p dir="ltr">prediction. Specifically, we achieve this by implementing a semantic-aware selection of relevant</p><p dir="ltr">agents from the scene and passing them through an attention mechanism to extract</p><p dir="ltr">global encodings. These encodings along with agents’ local information, are passed through</p><p dir="ltr">an encoder to obtain time-dependent latent variables for a motion policy predicting the future</p><p dir="ltr">trajectories. Our results show that the proposed approach outperforms state-of-the-art</p><p dir="ltr">baselines and provides more accurate and scene-consistent predictions. </p>

<p dir="ltr">The original definition of Nash Equilibrium applied to normal form games, but the notion has now been extended to various other forms of games including leader-follower games (Stackelberg games), extensive form games, stochastic games, games of incomplete information, cooperative games, and so on. We focus on general-sum stochastic Stackelberg games in this work. An example where such games would be natural to consider is in security games where a defender wishes to protect some targets through deployment of limited resources and an attacker wishes to strategically attack the targets to benefit themselves. The hierarchical order of play arises naturally since the defender typically acts first and deploys a strategy, while the attacker observes the strategy ofthe defender before attacking. Another example where this framework fits is in testing during epidemics, where the leader (the government) sets testing policies and the follower (the citizens) decide at every time step whether to get tested. The government wishes to minimize the number of infected people in the population while the follower wishes to minimize the cost of getting sick and testing. This thesis presents a learning algorithm for players to converge to their stationary policies in a general sum stochastic sequential Stackelberg game. The algorithm is a two time scale implicit policy gradient algorithm that provably converges to stationary points of the optimization problems of the two players. Our analysis allows us to move beyond the assumptions of zero-sum or static Stackelberg games made in the existing literature for learning algorithms to converge.</p><p dir="ltr"><br></p>