This research presents a Hybrid Particle Swarm Optimization with Bat Algorithm (HPSOBA) based approach to solve Optimal Reactive Power Dispatch (ORPD) problem. The primary objective of this project is minimization of the active power transmission losses by optimally setting the control variables within their limits and at the same time making sure that the equality and inequality constraints are not violated. Particle Swarm Optimization (PSO) and Bat Algorithm (BA) algorithms which are nature-inspired algorithms have become potential options to solving very difficult optimization problems like ORPD. Although PSO requires high computational time, it converges quickly; while BA requires less computational time and has the ability of switching automatically from exploration to exploitation when the optimality is imminent. This research integrated the respective advantages of PSO and BA algorithms to form a hybrid tool denoted as HPSOBA algorithm. HPSOBA combines the fast convergence ability of PSO with the less computation time ability of BA algorithm to get a better optimal solution by incorporating the BA’s frequency into the PSO velocity equation in order to control the pace. The HPSOBA, PSO and BA algorithms were implemented using MATLAB programming language and tested on three (3) benchmark test functions (Griewank, Rastrigin and Schwefel) and on IEEE 30- and 118-bus test systems to solve for ORPD without DG unit. A modified IEEE 30-bus test system was further used to validate the proposed hybrid algorithm to solve for optimal placement of DG unit for active power transmission line loss minimization. By comparison, HPSOBA algorithm results proved to be superior to those of the PSO and BA methods. In order to check if there will be a further improvement on the performance of the HPSOBA, the HPSOBA was further modified by embedding three new modifications to form a modified Hybrid approach denoted as MHPSOBA. This MHPSOBA was validated using IEEE 30-bus test system to solve ORPD problem and the results show that the HPSOBA algorithm outperforms the modified version (MHPSOBA). / Electrical and Mining Engineering / M. Tech. (Electrical Engineering)

Doctor of Philosophy / Department of Electrical and Computer Engineering / Sanjoy Das / Stochastic algorithms are widely used in various modeling and optimization problems. Evolutionary algorithms are one class of population-based stochastic approaches that are inspired from Darwinian evolutionary theory. A population of candidate solutions is initialized at the first generation of the algorithm. Two variation operators, crossover and mutation, that mimic the real world evolutionary process, are applied on the population to produce new solutions from old ones. Selection based on the concept of survival of the fittest is used to preserve parent solutions for next generation. Examples of such algorithms include genetic algorithm (GA) and genetic programming (GP). Nevertheless, other stochastic algorithms may be inspired from animals’ behavior such as particle swarm optimization (PSO), which imitates the cooperation of a flock of birds. In addition, stochastic algorithms are able to address multi-objective optimization problems by using the concept of dominance. Accordingly, a set of solutions that do not dominate each other will be obtained, instead of just one best solution. This thesis proposes a multi-objective GP-PSO hybrid algorithm to recover gene regulatory network models that take environmental data as stimulus input. The algorithm infers a model based on both phenotypic and gene expression data. The proposed approach is able to simultaneously infer network structures and estimate their associated parameters, instead of doing one or the other iteratively as other algorithms need to. In addition, a non-dominated sorting approach and an adaptive histogram method based on the hypergrid strategy are adopted to address ‘convergence’ and ‘diversity’ issues in multi-objective optimization. Gene network models obtained from the proposed algorithm are compared to a synthetic network, which mimics key features of Arabidopsis flowering control system, visually and numerically. Data predicted by the model are compared to synthetic data, to verify that they are able to closely approximate the available phenotypic and gene expression data. At the end of this thesis, a novel breeding strategy, termed network assisted selection, is proposed as an extension of our hybrid approach and application of obtained models for plant breeding. Breeding simulations based on network assisted selection are compared to one common breeding strategy, marker assisted selection. The results show that NAS is better both in terms of breeding speed and final phenotypic level.

Master of Science / Department of Civil Engineering / David R. Steward / Accurate modeling of the High Plains Aquifer depends on the availability of good data that represents and quantities properties and processes occurring within the aquifer. Thanks to many previous studies there is a wealth of good data available for the High Plains Aquifer but one key component, groundwater-surface water interaction locations and rates, is generally missing. Without these values accurate modeling of the High Plains Aquifer is very difficult to achieve. This thesis presents methods for simplifying the modeling of the High Plains Aquifer using a sloping base method and then applying mathematical optimization techniques to locate and quantify points of groundwater-surface water interaction. The High Plains Aquifer has a base that slopes gently from west to east and is approximated using a one-dimensional stepping base model. The model was run under steady-state predevelopment conditions using readily available GIS data representing aquifer properties such as hydraulic conductivity, bedrock elevation, recharge, and the predevelopment water level. The Levenberg-Marquardt and particle swarm algorithms were implemented to minimize error in the model. The algorithms reduced model error by finding locations in the aquifer of potential groundwater-surface water interaction and then determining the rate of groundwater to surface water exchange at those points that allowed for the best match between the measured predevelopment water level and the simulated water level. Results from the model indicate that groundwater-surface water interaction plays an important role in the overall water balance in the High Plains Aquifer. Findings from the model show strong groundwater-surface water interaction occurring in the northern basin of the aquifer where the water table is relatively shallow and there are many surface water features. In the central and southern basins the interaction is primarily limited to river valleys. Most rivers have baseflow that is a net sink from groundwater.

[pt] O algoritmo de otimização por enxame de partículas (PSO, do inglês Particle Swarm Optimization) é uma meta-heurística baseada em populações de indivíduos na qual os candidatos à solução evoluem através da simulação de um modelo simplificado de adaptação social. Juntando robustez, eficiência e simplicidade, o PSO tem adquirido grande popularidade. São reportadas muitas aplicações bem-sucedidas do PSO nas quais este algoritmo demonstrou ter vantagens sobre outras meta-heurísticas bem estabelecidas baseadas em populações de indivíduos. Algoritmos modificados de PSO já foram propostos para resolver problemas de otimização com restrições de domínio, lineares e não lineares. A grande maioria desses algoritmos utilizam métodos de penalização, que possuem, em geral, inúmeras limitações, como por exemplo: (i) cuidado adicional ao se determinar a penalidade apropriada para cada problema, pois deve-se manter o equilíbrio entre a obtenção de soluções válidas e a busca pelo ótimo; (ii) supõem que todas as soluções devem ser avaliadas. Outros algoritmos que utilizam otimização multi-objetivo para tratar problemas restritos enfrentam o problema de não haver garantia de se encontrar soluções válidas. Os algoritmos PSO propostos até hoje que lidam com restrições, de forma a garantir soluções válidas utilizando operadores de viabilidade de soluções e de forma a não necessitar de avaliação de soluções inválidas, ou somente tratam restrições de domínio controlando a velocidade de deslocamento de partículas no enxame, ou o fazem de forma ineficiente, reinicializando aleatoriamente cada partícula inválida do enxame, o que pode tornar inviável a otimização de determinados problemas. Este trabalho apresenta um novo algoritmo de otimização por enxame de partículas, denominado PSO+, capaz de resolver problemas com restrições lineares e não lineares de forma a solucionar essas deficiências. A modelagem do algoritmo agrega seis diferentes capacidades para resolver problemas de otimização com restrições: (i) redirecionamento aritmético de validade de partículas; (ii) dois enxames de partículas, onde cada enxame tem um papel específico na otimização do problema; (iii) um novo método de atualização de partículas para inserir diversidade no enxame e melhorar a cobertura do espaço de busca, permitindo que a borda do espaço de busca válido seja devidamente explorada – o que é especialmente conveniente quando o problema a ser otimizado envolve restrições ativas no ótimo ou próximas do ótimo; (iv) duas heurísticas de criação da população inicial do enxame com o objetivo de acelerar a inicialização das partículas, facilitar a geração da população inicial válida e garantir diversidade no ponto de partida do processo de otimização; (v) topologia de vizinhança, denominada vizinhança de agrupamento aleatório coordenado para minimizar o problema de convergência prematura da otimização; (vi) módulo de transformação de restrições de igualdade em restrições de desigualdade. O algoritmo foi testado em vinte e quatro funções benchmarks – criadas e propostas em uma competição de algoritmos de otimização –, assim como em um problema real de otimização de alocação de poços em um reservatório de petróleo. Os resultados experimentais mostram que o novo algoritmo é competitivo, uma vez que aumenta a eficiência do PSO e a velocidade de convergência. / [en] The Particle Swarm Optimization (PSO) algorithm is a metaheuristic based on populations of individuals in which solution candidates evolve through simulation of a simplified model of social adaptation. By aggregating robustness, efficiency and simplicity, PSO has gained great popularity. Many successful applications of PSO are reported in which this algorithm has demonstrated advantages over other well-established metaheuristics based on populations of individuals. Modified PSO algorithms have been proposed to solve optimization problems with domain, linear and nonlinear constraints; The great majority of these algorithms make use of penalty methods, which have, in general, numerous limitations, such as: (i) additional care in defining the appropriate penalty for each problem, since a balance must be maintained between obtaining valid solutions and the searching for an optimal solution; (ii) they assume all solutions must be evaluated. Other algorithms that use multi-objective optimization to deal with constrained problems face the problem of not being able to guarantee finding feasible solutions. The proposed PSO algorithms up to this date that deal with constraints, in order to guarantee valid solutions using feasibility operators and not requiring the evaluation of infeasible solutions, only treat domain constraints by controlling the velocity of particle displacement in the swarm, or do so inefficiently by randomly resetting each infeasible particle, which may make it infeasible to optimize certain problems. This work presents a new particle swarm optimization algorithm, called PSO+, capable of solving problems with linear and nonlinear constraints in order to solve these deficiencies. The modeling of the algorithm has added six different capabilities to solve constrained optimization problems: (i) arithmetic redirection to ensure particle feasibility; (ii) two particle swarms, where each swarm has a specific role in the optimization the problem; (iii) a new particle updating method to insert diversity into the swarm and improve the coverage of the search space, allowing its edges to be properly exploited – which is especially convenient when the problem to be optimized involves active constraints at the optimum solution; (iv) two heuristics to initialize the swarm in order to accelerate and facilitate the initialization of the feasible initial population and guarantee diversity at the starting point of the optimization process; (v) neighborhood topology, called coordinated random clusters neighborhood to minimize optimization premature convergence problem; (vi) transformation of equality constraints into inequality constraints. The algorithm was tested for twenty-four benchmark functions – created and proposed for an optimization competition – as well as in a real optimization problem of well allocation in an oil reservoir. The experimental results show that the new algorithm is competitive, since it increases the efficiency of the PSO and the speed of convergence.

Os Sistemas de Armazenamento a Baterias (SAEB) têm demonstrado uma grande flexibilidade de aplicações em melhorias e resoluções de problemas em Sistemas de Distribuição de Energia Elétrica (SDEEs). Grandes variações no valor de tensão dentro de um período, seja diário ou semanal, são observados devido à predominante topologia radial dos SDEEs e o constante aumento da utilização de Geradores Distribuídos (GDs). Pelas características de operar como carga ou geração, os SAEBs podem ser utilizados para melhorar o perfil de tensão. No entanto, as restrições de operação desses dispositivos tornam dificultoso identificar bons momentos de atuação e barramentos de alocação para este propósito. Geralmente, essa atuação é tratada de uma forma dependente dos GDs, porém essa abordagem não permite que os SAEBs possam operar em momentos independentes a fim de melhorar o perfil de tensão do SDEE. Desta forma, neste trabalho é desenvolvida uma abordagem para o planejamento da alocação e atuação do SAEB, utilizando uma modificação no algoritmo Particle Swarm Optimization (PSO) de forma que o SAEB possa atuar independente dos GDs e ser alocado em outras barras, com o objetivo de melhorar o perfil de tensão. As soluções são analisadas através de simulações de Monte Carlo para investigar o comportamento em diversas situações de curva de carga. Os resultados demonstram que a abordagem proposta busca encontrar boas alocações e atuações e que os parâmetros técnicos dos SAEBs, como capacidade de energia armazenada e potência nominal do inversor, influenciam diretamente nos resultados. / Battery Energy Storage Systems (BESS) have demonstrated great flexibility of applications in improvements and problem-solving in Electrical Distribution Systems (DSs). Significant variations in the nominal voltage value within a period, either daily or weekly, are observed due to the predominant radial topology of the DSs and the constant increase of the use of Distributed Generators (DGs). By the characteristics of operating as load or generation, SAEBs can be used to improve the voltage profile. However, the restrictions of these devices make it difficult to identify good operating moments and allocation buses for this purpose. Generally, this operation is treated in a way dependent on the DGs, but this strategy does not allow the BESS to operate at independent moments to improve the voltage profile of the SDEE. Thus, in this work an approach is developed for SAEBs allocation and operation planning, using a modification in the Particle Swarm Optimization (PSO) algorithm so that the SAEB can operate independently of the GDs and be allocated in other bars, with the objective to improve the voltage profile. The solutions are analyzed through Monte Carlo simulations to investigate the behavior in various load curve situations. The results demonstrate that the proposed approach seeks to find proper allocations and actions and that the technical parameters of SAEBs directly influence the results.

Submitted by Marlucy Farias Medeiros (marlucy.farias@unipampa.edu.br) on 2017-10-02T18:12:15Z No. of bitstreams: 1 Robson André Domanski - 2016.pdf: 5137261 bytes, checksum: 1e4aac0a601a8fb57b3a32e21268568a (MD5) / Approved for entry into archive by Marlucy Farias Medeiros (marlucy.farias@unipampa.edu.br) on 2017-10-04T17:34:54Z (GMT) No. of bitstreams: 1 Robson André Domanski - 2016.pdf: 5137261 bytes, checksum: 1e4aac0a601a8fb57b3a32e21268568a (MD5) / Made available in DSpace on 2017-10-04T17:34:54Z (GMT). No. of bitstreams: 1 Robson André Domanski - 2016.pdf: 5137261 bytes, checksum: 1e4aac0a601a8fb57b3a32e21268568a (MD5) Previous issue date: 2016-12-13 / O projeto de circuitos integrados analógicos, dentro da indústria da microeletrônica tem a sua evolução ditada pela grande necessidade da integração de circuitos mistos. Esta evolução faz com que os dispositivos semicondutores sejam cada vez mais miniaturizados, o que implica na complexidade cada vez maior no processo de fabricação, resultando em uma grande variabilidade de parâmetros. Esta complexidade no projeto está diretamente ligada ao dimensionamento dos dispositivos que compõem o circuito, já que o espaço de projeto é altamente não-linear. O dimensionamento de circuito analógico pode ser modelado como um problema de otimização e resolvido por heurísticas de otimização. A solução resultante é dependente da estratégia de modelagem e na estimativa de desempenho, o que é feito, em geral, por simulação elétrica. Neste contexto, foi desenvolvida a ferramenta UCAF. No entanto, a solução otimizada cai na fronteira do espaço de projeto, onde uma pequena variação nos parâmetros do dispositivo afeta o desempenho do circuito. Isso conduz à inclusão de simulação Monte Carlo no circuito de otimização, aumentando o esforço computacional. O objetivo principal deste trabalho é analisar dois métodos diferentes de amostragem, a fim de reduzir o número de rodadas Monte Carlo, e a inserção da heurística de otimização Particle Swarm Optimization, visando a minimização do tempo necessário para o dimensionamento do circuito. A amostragem por hipercubo latino, a qual requer um número menor de amostras para um nível de confiança razoável, é utilizado nas primeiras iterações do processo de otimização. Depois de um certo ponto, o método de amostragem é alterado para a amostragem aleatória tradicional. A heurística Particle Swarm Optimization foi implementada na ferramenta UCAF, devido ao seu baixo custo computacional. A metodologia é aplicada para o dimensionamento de um amplificador de transcondutância operacional OTA Miller e um amplificador Telescopic, mostrando vantagens em termos de tempo de processamento e desempenho do circuito. Pode-se demonstrar que a utilização de uma nova heurística, e diferentes métodos de amostragem para a simulação Monte Carlo no processo de otimização produz uma busca mais eficiente no espaço de projeto com um ganho em relação ao esforço computacional. / The analog integrated circuit design within the microelectronics industry has its evolution dictated by the great need for integration of mixed circuits. This trend makes the semiconductor devices are increasingly miniaturized, which implies the increasing complexity in the manufacturing process, resulting in a large variability op parameters. This complexity is directly linked to the design of devices that compose the circuit, since the design space is highly nonlinear. The design of analog circuit can be modeled as an optimization problem and solved by optimization heuristics. The resulting solution is dependent on modeling strategy and performance estimation, which is done generally by electrical simulation. In this context, the UCAF tool was developed. However, the optimized solution falls on the border of the design space where a small variation in device parameters affect circuit performance. This leads to the inclusion of Monte Carlo simulation on the circuit optimization, increasing the computational effort. The main objective of this study is to analyze two different methods of sampling, in order to reduce the number of Monte Carlo runs, and the inclusion of a new heuristic optimization, in order to minimize the time required for the design of the circuit. The Latin hypercube sampling, which requires a smaller number of samples for a reasonable confidence level is used in the first iteration of the optimization process. After a certain point, the sampling method is changed to the traditional random sampling. Heuristic Particle Swarm Optimization was implemented in UCAF tool, due to its low computational cost. The methodology is applied for the design of a Miller and a Telescopic operational transconductance amplifier, showing advantages in terms of processing time and circuit performance. We can demonstrate that the use of a new heuristic, and different methods of sampling for Monte Carlo simulation in the optimization process produces a more efficient search of the design space, and advantages in relation to computational effort.

Esta tese consiste em apresentar um método para localização e identificação de consumidores com alta contribuição para a distorção harmônica de tensão medida em subestações de sistemas de distribuição de energia elétrica. Cabe comentar que a etapa de localização visa obter uma lista das possíveis posições onde possa estar alocado o consumidor que possua cargas não lineares com grande consumo de potências harmônicas. Partindo-se desta lista, realiza-se a etapa de identificação, em que são estimadas as amplitudes de cada harmônica na posição selecionada. Por fim, um algoritmo para ajuste/sintonia do método de localização é empregado com o intuito de se realizar uma possível correção com relação à posição do consumidor. Desta forma, por meio de estudos de caso (simulados), os resultados obtidos procuram validar a metodologia proposta. / This thesis provides a method for location and identification of consumers with larger contribution to harmonic distortion of voltage in power distribution substations. It is worth to mention that the stage of consumers location must furnish a list of possible positions where there may be consumers, who have nonlinear loads with high consumption of harmonic power. From this list, the identification stage is performed in order to estimate the amplitude of each harmonic from the location selected. Finally, a method for improve the location algorithm is employed in order to refine the consumer position. Therefore, by means of simulated case studies, the results obtained for these stages seek to validate the methodology proposed.

Submitted by JOSIANE SANTOS DE OLIVEIRA (josianeso) on 2019-03-13T12:19:46Z No. of bitstreams: 1 Lucas Schmidt Goecks_.pdf: 2315393 bytes, checksum: 3985e874dfb67958cda4aa69697f671f (MD5) / Made available in DSpace on 2019-03-13T12:19:46Z (GMT). No. of bitstreams: 1 Lucas Schmidt Goecks_.pdf: 2315393 bytes, checksum: 3985e874dfb67958cda4aa69697f671f (MD5) Previous issue date: 2018-12-30 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Como uma das atividades mais importantes na engenharia de produção, o planejamento de instalações consiste na tomada de decisões relativas ao leiaute dos setores, unidades de produção/ fabricação, locais de armazenamento, e assim por diante. Conceito que é apoiado pela variabilidade dos processos produtivos, que muda de um período de produção para outro e de uma empresa para a outra. Atualmente, a literatura apresenta abordagens de como solucionar o problema de leiaute para empresas de pequeno e médio porte com modelos de planejamento, e de tomada de decisão multicritérios, ou meta-heurísticos. A literatura aborda estes dois métodos de forma separada. Inclusive, não existem relatos de comparações entre eles desde o conhecimento do autor. Como resposta à esta lacuna de pesquisa, definiu-se o seguinte objetivo: "identificar um método para planejamento de leiautes aplicável em empresas de pequeno e médio porte". A meta foi desenvolver uma ferramenta de modelagem genérica e que atenda à diferentes necessidades. Sendo assim, este trabalho abordou o Systematic Layout Planning (SLP) e o Particle Swarm Optimization (PSO) para planejamento de leiautes, avaliando a melhor proposta pelo Analytic Hierarchy Process (AHP). Em decorrência de interesses práticos que visam à aplicação de ferramentas para a solução de problemas específicos, este trabalho classifica-se como pesquisa aplicada de abordagem quantitativa, embasado por processos de tomada de decisão e de modelagem. Os resultados obtidos demonstram que o SLP fornece melhores propostas de leiautes que o PSO, para pequenas e médias empresas. O SLP respeita a alocação adjacente dos setores de acordo com o fluxo de material, enquanto que o PSO distribui aleatoriamente as áreas produtivas, o que proporciona maior variabilidade nas propostas de leiautes. O SLP demandou maior tempo de planejamento e um método auxiliar (AHP) para definição da melhor proposta de leiaute. Já o PSO forneceu o melhor leiaute sem uma ferramenta de suporte e a simulação foi mais rápida após estruturação do modelo do algoritmo. Implicações práticas à esta pesquisa encontram-se na análise da redução de custos com dados reais. Foram identificados na literatura objetivos de otimização e restrições mais usuais. Quanto ao tipo de leiaute, conforme as características da empresa a ser explorada, será considerado o tipo job-shop/funcional. Esta pesquisa contribui ao meio acadêmico no âmbito de sintetizar dois métodos, distintos, para planejamento de leiautes e compará-los com uma ferramenta de tomada de decisão multicriterial. Ao meio empresarial, a mesma fornece métodos que podem ser incorporados ao cotidiano das empresas no que diz respeito ao planejamento de leiautes e tomada de decisões. / As one of the most important activities in production engineering, facility planning consists of making decisions regarding the layout of the sectors, production/manufacturing units, storage locations, and so on. This concept is supported by the variability of production processes, which changes from one period of production to another and from one company to another. Currently, the literature presents approaches of how to solve the problem of layout for small and medium-sized companies with models of planning, and decision-making multi-criteria, or metaheuristics. The literature addresses these two methods separately. In fact, there are no reports of comparisons between them since the knowledge of the author. In response to this research gap, the following objective was defined: "to identify a method for layout planning applicable to small and medium-sized enterprises". The objective was to develop a generic modeling tool that meets different needs. Thus, this work approached Systematic Layout Planning (SLP) and Particle Swarm Optimization (PSO) for the layout planning, evaluating the best proposal by the Analytic Hierarchy Process (AHP). Because of practical interests that aim at the application of tools for the solution of specific problems, this work is classified as applied research of quantitative approach, based on processes of decision-making and modeling. The results obtained demonstrate that SLP provides better layout proposals than the PSO, for small and medium enterprises. The SLP respects the adjacent allocation of the sectors according to the material flow, while the PSO randomly distributes the productive areas, which provides greater variability in the layout proposals. The SLP required greater planning time and an auxiliary method (AHP) to define the best layout proposal. The PSO provided the best layout without a support tool and the simulation was faster after structuring the algorithm model. Practical implications of this research lie in the analysis of cost reduction with real data. Optimization objectives and constraints that are more usual have been identified in the literature. As for the type of layout, according to the characteristics of the company, and because it is a single case study, the job-shop type will be considered. This research contributes to the academic environment in the context of synthesizing two distinct methods for planning layouts and comparing them with a multi-criteria decision-making tool. In the business environment, it provides methods that can be incorporated into companies’ day-to-day planning and decision making.

Dans le cadre de l'optimisation multiobjectif, les métaheuristiques sont reconnues pour être des méthodes performantes mais elles ne rencontrent qu'un succès modéré dans le monde de l'industrie. Dans un milieu où seule la performance compte, l'aspect stochastique des métaheuristiques semble encore être un obstacle difficile à franchir pour les décisionnaires. Il est donc important que les chercheurs de la communauté portent un effort tout particulier sur la facilité de prise en main des algorithmes. Plus les algorithmes seront faciles d'accès pour les utilisateurs novices, plus l'utilisation de ceux-ci pourra se répandre. Parmi les améliorations possibles, la réduction du nombre de paramètres des algorithmes apparaît comme un enjeu majeur. En effet, les métaheuristiques sont fortement dépendantes de leur jeu de paramètres. Dans ce cadre se situe l'apport majeur de TRIBES, un algorithme mono-objectif d'Optimisation par Essaim Particulaire (OEP) qui fonctionne automatiquement,sans paramètres. Il a été mis au point par Maurice Clerc. En fait, le fonctionnement de l'OEP nécessite la manipulation de plusieurs paramètres. De ce fait, TRIBES évite l'effort de les régler (taille de l'essaim, vitesse maximale, facteur d'inertie, etc.).Nous proposons dans cette thèse une adaptation de TRIBES à l'optimisation multiobjectif. L'objectif est d'obtenir un algorithme d'optimisation par essaim particulaire multiobjectif sans paramètres de contrôle. Nous reprenons les principaux mécanismes de TRIBES auxquels sont ajoutés de nouveaux mécanismes destinés à traiter des problèmes multiobjectif. Après les expérimentations, nous avons constaté, que TRIBES-Multiobjectif est moins compétitif par rapport aux algorithmes de référence dans la littérature. Ceci peut être expliqué par la stagnation prématurée de l'essaim. Pour remédier à ces problèmes, nous avons proposé l'hybridation entre TRIBES-Multiobjectif et un algorithme de recherche locale, à savoir le recuit simulé et la recherche tabou. L'idée était d'améliorer la capacité d'exploitation deTRIBES-Multiobjectif. Nos algorithmes ont été finalement appliqués sur des problèmes de dimensionnement des transistors dans les circuits analogiques / Meta-heuristics are recognized to be successful to deal with multiobjective optimization problems but still with limited success in engineering fields. In an environment where only the performance counts, the stochastic aspect of meta-heuristics again seems to be a difficult obstacle to cross for the decision-makers. It is, thus, important that the researchers of the community concern a quite particular effort to ease the handling of those algorithms. The more the algorithms will be easily accessible for the novices, the more the use of these algorithms can spread. Among the possible improvements, reducing the number of parameters is considered as the most challenging one. In fact, the performance of meta-heuristics is strongly dependent on their parameters values. TRIBES presents an attempt to remedy this problem. In fact, it is a particle swarm optimization (PSO) algorithm that works in an autonomous way. It was proposed by Maurice Clerc. Indeed, like every other meta-heuristic, PSO requires many parameters to be fitted every time a new problem is considered. The major contribution of TRIBES is to avoid the effort of fitting them. We propose, in this thesis, an adaptation of TRIBES to the multiobjective optimization. Our aim is to conceive a competitive PSO algorithm free of parameters. We consider the main mechanisms of TRIBES to which are added new mechanisms intended to handle multiobjective problems. After the experimentations, we noticed that Multiobjective-TRIBESis not competitive compared to other multiobjective algorithms representative of the state of art. It can be explained by the premature stagnation of the swarm. To remedy these problems, we proposed the hybridization between Multiobjective-TRIBES and local search algorithms such as simulated annealing and tabu search. The idea behind the hybridization was to improve the capacity of exploitation of Multiobjective-TRIBES. Our algorithms were finally applied to sizing analogical circuits' problems

Tese de mestrado integrado. Engenharia Electrotécnica e de Computadores (Área de Especialização de Energia). Faculdade de Engenharia. Universidade do Porto. 2012