Les métaheuristiques sont une famille d'algorithmes stochastiques destinés à résoudre des problèmes d 'optimisation difficile . Utilisées dans de nombreux domaines, ces méthodes présentent l'avantage d'être généralement efficaces, sans pour autant que l'utilisateur ait à modifier la structure de base de l'algorithme qu'il utilise. Parmi celles-ci, l'Optimisation par Essaim Particulaire (OEP) est une nouvelle classe d'algorithmes proposée pour résoudre les problèmes à variables continues. Les algorithmes d'OEP s'inspirent du comportement social des animaux évoluant en essaim, tels que les oiseaux migrateurs ou les poissons. Les particules d'un même essaim communiquent de manière directe entre elles tout au long de la recherche pour construire une solution au problème posé, en s'appuyant sur leur expérience collective. Reconnues depuis de nombreuses années pour leur efficacité, les métaheuristiques présentent des défauts qui rebutent encore certains utilisateurs. Le réglage des paramètres des algorithmes est un de ceux-ci. Il est important, pour chaque probléme posé, de trouver le jeu de paramètres qui conduise à des performances optimales de l'algorithme. Cependant, cette tâche est fastidieuse et coûteuse en temps, surtout pour les utilisateurs novices. Pour s'affranchir de ce type de réglage, des recherches ont été menées pour proposer des algorithmes dits adaptatifs . Avec ces algorithmes, les valeurs des paramètres ne sont plus figées, mais sont modifiées, en fonction des résultats collectés durant le processus de recherche. Dans cette optique-là, Maurice Clerc a proposé TRIBES, qui est un algorithme d'OEP mono-objectif sans aucun paramètre de contrôle. Cet algorithme fonctionne comme une boite noire , pour laquelle l'utilisateur n'a qu'à définir le problème à traiter et le critàre d'arrêt de l'algorithme. Nous proposons dans cette thèse une étude comportementale de TRIBES, qui permet d'en dégager les principales qualités et les principaux défauts. Afin de corriger certains de ces défauts, deux modules ont été ajoutés à TRIBES. Une phase d'initialisation régulière est insérée, afin d'assurer, dès le départ de l'algorithme, une bonne couverture de l'espace de recherche par les particules. Une nouvelle stratégie de déplacement, basée sur une hybridation avec un algorithme à estimation de distribution, est aussi définie, afin de maintenir la diversité au sein de l'essaim, tout au long du traitement. Le besoin croissant de méthodes de résolution de problèmes multiobjectifs a conduit les concepteurs à adapter leurs méthodes pour résoudre ce type de problème. La complexité de cette opération provient du fait que les objectifs à optimiser sont souvent contradictoires. Nous avons élaboré une version multiobjectif de TRIBES, dénommée MO-TRIBES. Nos algorithmes ont été enfin appliqués à la résolution de problèmes de seuillage d'images médicales et au problème de dimensionnement de composants de circuits analogiques / Metaheuristics are a new family of stochastic algorithms which aim at solving difficult optimization problems. Used to solve various applicative problems, these methods have the advantage to be generally efficient on a large amount of problems. Among the metaheuristics, Particle Swarm Optimization (PSO) is a new class of algorithms proposed to solve continuous optimization problems. PSO algorithms are inspired from the social behavior of animals living in swarm, such as bird flocks or fish schools. The particles of the swarm use a direct way of communication in order to build a solution to the considered problem, based on their collective experience. Known for their e ciency, metaheuristics show the drawback of comprising too many parameters to be tuned. Such a drawback may rebu some users. Indeed, according to the values given to the parameters of the algorithm, its performance uctuates. So, it is important, for each problem, to nd the parameter set which gives the best performance of the algorithm. However, such a problem is complex and time consuming, especially for novice users. To avoid the user to tune the parameters, numerous researches have been done to propose adaptive algorithms. For such algorithms, the values of the parameters are changed according to the results previously found during the optimization process. TRIBES is an adaptive mono-objective parameter-free PSO algorithm, which was proposed by Maurice Clerc. TRIBES acts as a black box , for which the user has only the problem and the stopping criterion to de ne. The rst objective of this PhD is to make a global study of the behavior of TRIBES under several conditions, in order to determine the strengths and drawbacks of this adaptive algorithm. In order to improve TRIBES, two new strategies are added. First, a regular initialization process is defined in order to insure an exploration as wide as possible of the search space, since the beginning of the optimization process. A new strategy of displacement, based on an hybridation with an estimation of distribution algorithm, is also introduced to maintain the diversity in the swarm all along the process. The increasing need for multiobjective methods leads the researchers to adapt their methods to the multiobjective case. The di culty of such an operation is that, in most cases, the objectives are con icting. We designed MO-TRIBES, which is a multiobjective version of TRIBES. Finally, our algorithms are applied to thresholding segmentation of medical images and to the design of electronic components

Na área de inteligência artificial existem algoritmos de aprendizado, notavelmente aqueles pertencentes à área de aprendizado de máquina AM , capazes de automatizar a extração do conhecimento implícito de um conjunto de dados. Dentre estes, os algoritmos de AM simbólico são aqueles que extraem um modelo de conhecimento inteligível, isto é, que pode ser facilmente interpretado pelo usuário. A utilização de AM simbólico é comum no contexto de classificação, no qual o modelo de conhecimento extraído é tal que descreve uma correlação entre um conjunto de atributos denominados premissas e um atributo particular denominado classe. Uma característica dos algoritmos de classificação é que, em geral, estes são utilizados visando principalmente a maximização das medidas de cobertura e precisão, focando a construção de um classificador genérico e preciso. Embora essa seja uma boa abordagem para automatizar processos de tomada de decisão, pode deixar a desejar quando o usuário tem o desejo de extrair um modelo de conhecimento que possa ser estudado e que possa ser útil para uma melhor compreensão do domínio. Tendo-se em vista esse cenário, o principal objetivo deste trabalho é pesquisar métodos de computação evolutiva multiobjetivo para a construção de regras de conhecimento individuais com base em critérios definidos pelo usuário. Para isso utiliza-se a biblioteca de classes e ambiente de construção de regras de conhecimento ECLE, cujo desenvolvimento remete a projetos anteriores. Outro objetivo deste trabalho consiste comparar os métodos de computação evolutiva pesquisados com métodos baseado em composição de rankings previamente existentes na ECLE. É mostrado que os métodos de computação evolutiva multiobjetivo apresentam melhores resultados que os métodos baseados em composição de rankings, tanto em termos de dominância e proximidade das soluções construídas com aquelas da fronteira Pareto-ótima quanto em termos de diversidade na fronteira de Pareto. Em otimização multiobjetivo, ambos os critérios são importantes, uma vez que o propósito da otimização multiobjetivo é fornecer não apenas uma, mas uma gama de soluções eficientes para o problema, das quais o usuário pode escolher uma ou mais soluções que apresentem os melhores compromissos entre os objetivos / Machine Learning algorithms are notable examples of Artificial Intelligence algorithms capable of automating the extraction of implicit knowledge from datasets. In particular, Symbolic Learning algorithms are those which yield an intelligible knowledge model, i.e., one which a user may easily read. The usage of Symbolic Learning is particularly common within the context of classification, which involves the extraction of knowledge such that the associated model describes correelation among a set of attributes named the premises and one specific attribute named the class. Classification algorithms usually target into creating knowledge models which maximize the measures of coverage and precision, leading to classifiers that tend to be generic and precise. Althought this constitutes a good approach to creating models that automate the decision making process, it may not yield equally good results when the user wishes to extract a knowledge model which could assist them into getting a better understanding of the domain. Having that in mind, it has been established as the main goal of this Masters thesis the research of multi-objective evolutionary computing methods to create individual knowledge rules maximizing sets of arbitrary user-defined criteria. This is achieved by employing the class library and knowledge rule construction environment ECLE, which had been developed during previous research work. A second goal of this Masters thesis is the comparison of the researched evolutionary computing methods against previously existing ranking composition methods in ECLE. It is shown in this Masters thesis that the employment of multi-objective evolutionary computing methods produces better results than those produced by the employment of ranking composition-based methods. This improvement is verified both in terms of solution dominance and proximity of the solution set to the Pareto-optimal front and in terms of Pareto-front diversity. Both criteria are important for evaluating the efficiency of multi-objective optimization algorithms, for the goal of multi-objective optimization is to provide a broad range of efficient solutions, so the user may pick one or more solutions which present the best trade-off among all objectives

Im Mittelpunkt unserer Untersuchungen steht das mehrkriterielle Optimierungsproblem, in einer beliebigen nichtleeren Menge eines halbgeordneten endlich dimensionalen Raumes. Zu dessen Lösung betrachten wir ein Dialogverfahren, in dem der Entscheidungsträger in jedem Schritt seine Wünsche äußert. Bei der Bestimmung einer Lösung, die den Entscheidungsträger zufriedenstellt, müssen wir ein im Allgemeinen nichtkonvexes und nicht triviales skalares Optimierungsproblem lösen. Zur Lösung dieses Problems haben wir zwei Klassen einparametrischer Optimierungsprobleme (Einbettungen) konstruiert. Mit Hilfe der Projektion auf den konvexen Ordungskegel haben wir gezeigt, dass diese Einbettungen wohldefiniert sind. Im Gegensatz zu der in der Literatur untersuchten Standardeinbettung, sind die in dieser Arbeit betrachteten Einbettungen durch die Skalarisierungen der Vektoroptimierungsprobleme mittels streng monotoner skalarisierender Funktionen motiviert. Diese Untersuchung wird unter dem Gesichtspunkt der Theorie der einparametrischen Optimierungsprobleme für den Fall eines beliebigen spitzen polyedrischen Ordnungskegels durchgeführt. Sie umfasst z.B. Fragestellungen nach der Art der Singularitäten, die für die verschiedenen Einbettungen auftreten können, nach den Bedingungen, unter denen eine Zusammenhangskomponente in der Menge stationärer oder verallgemeinerter kritischer Punkte mit Hilfe von Kurvenverfolgungsmethoden numerisch beschrieben werden kann und nach den hinreichenden Bedingungen für die Existenz einer Lösungskurve. Anschließend haben wir das von Guddat und Jongen eingeführte Konzept der strukturellen Stabilität eines skalaren Optimierungsproblems in der Vektoroptimierung verallgemeinert und einen Zusammenhang zur strukturellen Stabilität eines Minimaxproblems erstellt. Dieses Minimaxproblem steht in starker Beziehung zur Skalarisierungsmethode der Vektoroptimierungsprobleme. / In this work we consider the multiobjective optimization in a subset of a partially orded finite dimensional space. In order to solve this problem we use a dialogue procedure in which the decision maker has to determine in each step the aspiration and reservation level expressing his wishes (goals). This leads to an optimization problem which is not easy to solve in the nonconvex case. We solve it proposing two classes of one-parametric optimization problems (embeddings). Using the projection in the ordering cone, we show that these embeddings are well defined, i.e. the corresponding constraint sets depending on real-valued parameters are not empty. Contrary to the very known standard embedding the proposed embeddings are motivated by the use of strongly monotonically increasing functions, which play an important role by the scalarization of multiobjective optimization problems. The two classes of embeddings are investigated from the point of view of parametric optimization considering a pointed polyhedral cone. This investigation includes the determination of the kind of singularities which can appear, the conditions under which a connected component in the set of stationary or generalized critical point can be numerically described using pathfollowing methods and a solution curve may exist. Finally, we extend the concept of structural stability by Guddat and Jongen to the multiobjective optimization problems and establish a connection to the problem of Minimax type, which is related to the scalarization of multiobjective optimization problems.

Submitted by Jadson Francisco de Jesus SILVA (jadson@uefs.br) on 2018-07-18T21:55:12Z No. of bitstreams: 1 Disserta??o.pdf: 2639155 bytes, checksum: af49bdcdf83d4a063546324a223124a4 (MD5) / Made available in DSpace on 2018-07-18T21:55:12Z (GMT). No. of bitstreams: 1 Disserta??o.pdf: 2639155 bytes, checksum: af49bdcdf83d4a063546324a223124a4 (MD5) Previous issue date: 2018-03-27 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior - CAPES / Wireless visual sensor networks can provide valuable information for a lot of moni- toring and control applications, which has driven much attention from the academic community in last years. For some applications, a set of targets have to be covered by visual sensors and sensing redundancy may be desired in many cases, especially when applications have availability requirements or demands for multiple coverage perspectives for viewed targets. For rotatable visual sensors, the sensing orientations can be adjusted for optimized coverage and redundancy, with different optimization approaches available to address this problem. Particularly, as different optimization parameters may be considered, the redundant coverage maximization issue may be treated as a multi-objective problem, with some potential solutions to be conside- red. In this context, two different evolutionary algorithms are proposed to compute redundant coverage maximization for target viewing, intending to be more efficient alternatives to greedy-based algorithms. Simulation results reinforce the benefits of employing evolutionary algorithms for adjustments of sensors? orientations, poten- tially benefiting deployment and management of wireless visual sensor networks for different applications. / As redes de sensores visuais sem fio podem obter, atrav?s de c?meras, informa??es importantes para aplica??es de controle e monitoramento, e tem ganhado aten??o da comunidade acad?mica nos ?ltimos anos. Para algumas aplica??es, um conjunto de alvos deve ser coberto por sensores visuais, e por vezes com demanda de redund?ncia de cobertura, especialmente quando h? requisitos de disponibilidade ou demandas de m?ltiplas perspectivas de cobertura para os alvos visados. Para sensores visuais rotacion?veis, as orienta??es de detec??o podem ser ajustadas para otimizar cobertura e redund?ncia, existindo diferentes abordagens de otimiza??o dispon?veis para solucionar esse problema. Particularmente, como diferentes par?metros de otimizac?o podem ser considerados, o problema de maximiza??o de cobertura redundante pode ser tratado como um problema multiobjetivo, com algumas solu??es potenciais a serem consideradas. Neste contexto, dois algoritmos evolutivos diferentes s?o propostos para calcular a maximiza??o de cobertura redundante para visualiza??o de alvos, pretendendo ser alternativas mais eficientes para algoritmos gulosos. Os resultados da simula??o refor?am os benef?cios de empregar algoritmos evolutivos para ajustes das orienta??es dos sensores, potencialmente beneficiando a implanta??o e o gerenciamento de redes de sensores visuais sem fio para diferentes aplica??es.

[pt] Estudos empíricos na área de séries temporais indicam que combinar modelos preditivos, originados a partir de diferentes técnicas de modelagem, levam a previsões consensuais superiores, em termos de acurácia, às previsões individuais dos modelos envolvidos na combinação. No presente trabalho é apresentada uma metodologia de combinação convexa de modelos estatísticos de previsão, cujo sucesso depende da forma como os pesos de combinação de cada modelo são estimados. Uma Rede Neural Artificial Perceptron Multi-camada (Multilayer Perceptron - MLP) é utilizada para gerar dinamicamente vetores de pesos ao longo do horizonte de previsão, sendo estes dependentes da contribuição individual de cada previsor observada nos dados históricos da série. O ajuste dos parâmetros da rede MLP é efetuado através de um algoritmo de treinamento híbrido, que integra técnicas de busca global, baseadas em computação evolucionária, junto com o algoritmo de busca local backpropagation, de modo a otimizar de forma simultânea tanto os pesos quanto a arquitetura da rede, visando, assim, a gerar de forma automática um modelo de ponderação dinâmica de previsores de alto desempenho. O modelo proposto, batizado de Neural Expert Weighting - Genetic Algorithm (NEW-GA), foi avaliado em diversos experimentos comparativos com outros modelos de ponderação de previsores, assim como também com os modelos individuais envolvidos na combinação, contemplando 15 séries temporais divididas em dois estudos de casos: séries de derivados de petróleo e séries da versão reduzida da competição NN3, uma competição entre metodologias de previsão, com maior ênfase nos modelos baseados em Redes Neurais. Os resultados demonstraram o potencial do NEWGA em fornecer modelos acurados de previsão de séries temporais. / [en] Empirical studies on time series indicate that the combination of forecasting models, generated from different modeling techniques, leads to higher consen+sus forecasts, in terms of accuracy, than the forecasts of individual models involved in the combination scheme. In this work, we present a methodology for convex combination of statistical forecasting models, whose success depends on how the combination weights of each model are estimated. An Artificial Neural Network Multilayer Perceptron (MLP) is used to generate dynamically weighting vectors over the forecast horizon, being dependent on the individual contribution of each forecaster observed over historical data series. The MLP network parameters are adjusted via a hybrid training algorithm that integrates global search techniques, based on evolutionary computation, along with the local search algorithm backpropagation, in order to optimize simultaneously both weights and network architecture. This approach aims to automatically generate a dynamic weighted forecast aggregation model with high performance. The proposed model, called Neural Expert Weighting - Genetic Algorithm (NEW-GA), was com- pared with other forecaster combination models, as well as with the individual models involved in the combination scheme, comprising 15 time series divided into two case studies: Petroleum Products and the reduced set of NN3 forecasting competition, a competition between forecasting methodologies, with greater emphasis on models based on neural networks. The results obtained demonstrated the potential of NEW-GA in providing accurate models for time series forecasting.

Problemas ocasionados por perturbações na qualidade da energia elétrica (QEE) podem provocar sérios prejuízos, tanto de cunho social, quanto financeiros, aos clientes conectados ao sistema elétrico de potência como um todo. Neste contexto, os clientes que mais sofrem são os clientes industriais, pois estes possuem cargas sensíveis a vários distúrbios associados à falta da QEE. Sendo assim, para adoções de medidas preventivas, ou corretivas, que melhorem os índices de QEE, faz-se necessário um monitoramento dos sistemas elétricos que permita um melhor acompanhamento da ocorrência dos distúrbios. Nesta pesquisa é proposta a modelagem do problema de alocação ótima de monitores de QEE em sistemas de distribuição com múltiplos objetivos, os quais são: minimização do custo do monitoramento, minimização da ambiguidade topológica, maximização do monitoramento das cargas, maximização da quantidade de ramais monitorados, minimização da quantidade de afundamentos não monitorados, e maximização da redundância do monitoramento dos afundamentos. Na resolução do problema foi utilizado o Algoritmo Evolutivo Multiobjetivo com Tabelas (AEMT), adotado por ter boa capacidade de resolução com muitos objetivos. Os resultados obtidos permitiram observar que o AEMT forneceu as fronteiras de Pareto com soluções diversificadas e bem distribuídas ao longo da mesma, mostrando-se de grande relevância para o planejamento de sistemas de monitoramento da QEE em sistemas de distribuição de energia. A principal contribuição desta tese é o fornecimento de um modelo que permite às empresas de energia avaliar os investimentos que farão nos seus sistemas de monitoramento considerando seis critérios distintos, permitindo uma maior flexibilidade no estabelecimento do plano de monitoramento e uma melhor análise do custo/benefício considerando os seis aspectos abordados. / Problems arising from disturbances in power quality (PQ) can cause serious damage, both social, and financial, to customers connected to the electrical power distribution systems as a whole. In this context, the customers who suer most are industrial customers, as they have loads sensitive to various disturbances associated with the lack of PQ. Thus, in order to adopt preventive or corrective measures to improve PQ rates, it is necessary to monitor electrical systems to allow better oversight of the occurrence of disturbances. In this research, the proposal is to model the problem of optimal allocation of power quality monitors in distribution systems with multiple objectives. The multiple objectives are: minimizing the monitoring cost, minimizing ambiguities in topology, maximizing the load monitoring, maximizing the area monitoring, minimizing the voltage sag unmonitored, and maximizing the redundancy in the sag monitoring. In solving the problem, a Multiobjective Evolutionary Algorithm with Tables (MEAT) was adopted due to ability to deal with many objectives. The results show that the AMET finds a set of ecient solutions that are diversified and well-distributed along the Pareto Front, and that they are highly relevant for planning of PQ monitoring systems in electrical power distribution systems. The main contribution of this thesis is to provide a model that allows utilities better evaluate investments that they will make in their monitoring systems comprising six dierent criteria, allowing greater flexibility in establishing the monitoring plan and a better analysis of cost/benefit considering the six aspects.

Problemas ocasionados por perturbações na qualidade da energia elétrica (QEE) podem provocar sérios prejuízos, tanto de cunho social, quanto financeiros, aos clientes conectados ao sistema elétrico de potência como um todo. Neste contexto, os clientes que mais sofrem são os clientes industriais, pois estes possuem cargas sensíveis a vários distúrbios associados à falta da QEE. Sendo assim, para adoções de medidas preventivas, ou corretivas, que melhorem os índices de QEE, faz-se necessário um monitoramento dos sistemas elétricos que permita um melhor acompanhamento da ocorrência dos distúrbios. Nesta pesquisa é proposta a modelagem do problema de alocação ótima de monitores de QEE em sistemas de distribuição com múltiplos objetivos, os quais são: minimização do custo do monitoramento, minimização da ambiguidade topológica, maximização do monitoramento das cargas, maximização da quantidade de ramais monitorados, minimização da quantidade de afundamentos não monitorados, e maximização da redundância do monitoramento dos afundamentos. Na resolução do problema foi utilizado o Algoritmo Evolutivo Multiobjetivo com Tabelas (AEMT), adotado por ter boa capacidade de resolução com muitos objetivos. Os resultados obtidos permitiram observar que o AEMT forneceu as fronteiras de Pareto com soluções diversificadas e bem distribuídas ao longo da mesma, mostrando-se de grande relevância para o planejamento de sistemas de monitoramento da QEE em sistemas de distribuição de energia. A principal contribuição desta tese é o fornecimento de um modelo que permite às empresas de energia avaliar os investimentos que farão nos seus sistemas de monitoramento considerando seis critérios distintos, permitindo uma maior flexibilidade no estabelecimento do plano de monitoramento e uma melhor análise do custo/benefício considerando os seis aspectos abordados. / Problems arising from disturbances in power quality (PQ) can cause serious damage, both social, and financial, to customers connected to the electrical power distribution systems as a whole. In this context, the customers who suer most are industrial customers, as they have loads sensitive to various disturbances associated with the lack of PQ. Thus, in order to adopt preventive or corrective measures to improve PQ rates, it is necessary to monitor electrical systems to allow better oversight of the occurrence of disturbances. In this research, the proposal is to model the problem of optimal allocation of power quality monitors in distribution systems with multiple objectives. The multiple objectives are: minimizing the monitoring cost, minimizing ambiguities in topology, maximizing the load monitoring, maximizing the area monitoring, minimizing the voltage sag unmonitored, and maximizing the redundancy in the sag monitoring. In solving the problem, a Multiobjective Evolutionary Algorithm with Tables (MEAT) was adopted due to ability to deal with many objectives. The results show that the AMET finds a set of ecient solutions that are diversified and well-distributed along the Pareto Front, and that they are highly relevant for planning of PQ monitoring systems in electrical power distribution systems. The main contribution of this thesis is to provide a model that allows utilities better evaluate investments that they will make in their monitoring systems comprising six dierent criteria, allowing greater flexibility in establishing the monitoring plan and a better analysis of cost/benefit considering the six aspects.

Maximum asset performance is one of the major goals for electric power distribution system operators (DSOs). To reach this goal minimal life cycle cost and maintenance optimization become crucial while meeting demands from customers and regulators. One of the fundamental objectives is therefore to relate maintenance and reliability in an efficient and effective way. Furthermore, this necessitates the determination of the optimal balance between pre¬ventive and corrective maintenance, which is the main problem addressed in the thesis. The balance between preventive and corrective maintenance is approached as a multiobjective optimization problem, with the customer interruption costs on one hand and the maintenance budget of the DSO on the other. Solutions are obtained with meta-heuristics, developed for the specific problem, as well as with an Evolutionary Particle Swarm Optimization algorithm. The methods deliver a Pareto border, a set of several solutions, which the operator can choose between, depending on preferences. The optimization is built on component reliability importance indices, developed specifically for power systems. One vital aspect of the indices is that they work with several supply and load points simultaneously, addressing the multistate-reliability of power systems. For the computation of the indices both analytical and simulation based techniques are used. The indices constitute the connection between component reliability performance and system performance and so enable the maintenance optimization. The developed methods have been tested and improved in two case studies, based on real systems and data, proving the methods’ usefulness and showing that they are ready to be applied to power distribution systems. It is in addition noted that the methods could, with some modifications, be applied to other types of infrastructures. However, in order to perform the optimization, a reliability model of the studied power system is required, as well as estimates on effects of maintenance actions (changes in failure rate) and their related costs. Given this, a generally decreased level of total maintenance cost and a better system reliability performance can be given to the DSO and customers respectively. This is achieved by focusing the preventive maintenance to components with a high potential for improvement from system perspective. / QC 20100810

No description available.

Made available in DSpace on 2014-12-17T15:47:03Z (GMT). No. of bitstreams: 1 SilviaMDMM_TESE.pdf: 3010194 bytes, checksum: 43610ec3f0a30c2e5ef7fb5c0b2dc5b0 (MD5) Previous issue date: 2013-12-16 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / The Quadratic Minimum Spanning Tree Problem (QMST) is a version of the Minimum Spanning Tree Problem in which, besides the traditional linear costs, there is a quadratic structure of costs. This quadratic structure models interaction effects between pairs of edges. Linear and quadratic costs are added up to constitute the total cost of the spanning tree, which must be minimized. When these interactions are restricted to adjacent edges, the problem is named Adjacent Only Quadratic Minimum Spanning Tree (AQMST). AQMST and QMST are NP-hard problems that model several problems of transport and distribution networks design. In general, AQMST arises as a more suitable model for real problems. Although, in literature, linear and quadratic costs are added, in real applications, they may be conflicting. In this case, it may be interesting to consider these costs separately. In this sense, Multiobjective Optimization provides a more realistic model for QMST and AQMST. A review of the state-of-the-art, so far, was not able to find papers regarding these problems under a biobjective point of view. Thus, the objective of this Thesis is the development of exact and heuristic algorithms for the Biobjective Adjacent Only Quadratic Spanning Tree Problem (bi-AQST). In order to do so, as theoretical foundation, other NP-hard problems directly related to bi-AQST are discussed: the QMST and AQMST problems. Bracktracking and branch-and-bound exact algorithms are proposed to the target problem of this investigation. The heuristic algorithms developed are: Pareto Local Search, Tabu Search with ejection chain, Transgenetic Algorithm, NSGA-II and a hybridization of the two last-mentioned proposals called NSTA. The proposed algorithms are compared to each other through performance analysis regarding computational experiments with instances adapted from the QMST literature. With regard to exact algorithms, the analysis considers, in particular, the execution time. In case of the heuristic algorithms, besides execution time, the quality of the generated approximation sets is evaluated. Quality indicators are used to assess such information. Appropriate statistical tools are used to measure the performance of exact and heuristic algorithms. Considering the set of instances adopted as well as the criteria of execution time and quality of the generated approximation set, the experiments showed that the Tabu Search with ejection chain approach obtained the best results and the transgenetic algorithm ranked second. The PLS algorithm obtained good quality solutions, but at a very high computational time compared to the other (meta)heuristics, getting the third place. NSTA and NSGA-II algorithms got the last positions / O problema da ?rvore Geradora M?nima Quadr?tica (AGMQ) ? uma vers?o do problema da ?rvore Geradora M?nima na qual se considera, al?m dos custos lineares tradicionais, uma estrutura de custos quadr?tica. Tal estrutura quadr?tica modela efeitos de intera??o entre pares de arestas. Os custos lineares e quadr?ticos s?o somados para compor o custo total da ?rvore geradora, que deve ser minimizado. Quando as intera??es s?o restritas ?s arestas adjacentes, o problema ? denominado ?rvore Geradora M?nima Quadr?tica em Adjac?ncia de Arestas (AGMQA). A AGMQA e a AGMQ s?o problemas NP-dif?ceis que modelam diversos problemas de projeto de redes de transporte e distribui??o. Em geral, a AGMQA emerge como um modelo mais apropriado para a modelagem de problemas reais. Embora, na literatura, os custos lineares e quadr?ticos sejam somados, em aplica??es reais, os custos linear e quadr?tico podem ser conflitantes. Neste caso, seria mais interessante considerar os custos separadamente. Neste sentido, a Otimiza??o Multiobjetivo prov? uma modelagem mais realista para os problemas da AGMQ e da AGMQA. Uma revis?o do estado da arte, at? o momento, n?o foi capaz de encontrar qualquer trabalho que investigue esses problemas sob um ponto de vista biobjetivo. O objetivo desta Tese ?, pois, o desenvolvimento de algoritmos exatos e heur?sticos para o Problema Biobjetivo da ?rvore Geradora Quadr?tica em Adjac?ncia de Arestas (AGQA-bi). Para tanto, como fundamenta??o te?rica, discutem-se outros problemas NP-dif?ceis diretamente relacionados ? AGQA-bi, a saber: AGMQ e AGMQA. Algoritmos exatos backtracking e branch-and-bound s?o propostos para o problema-alvo desta investiga??o. Os algoritmos heur?sticos desenvolvidos s?o: busca local Pareto Local Search, Busca Tabu com ejection chain, Algoritmo Transgen?tico, NSGA-II e uma hibridiza??o das duas ?ltimas propostas mencionadas denominada NSTA. Os algoritmos propostos s?o comparados entre si por meio da an?lise de seus desempenhos em experimentos computacionais com casos de teste adaptados da literatura da AGMQ. No que se refere aos algoritmos exatos, a an?lise considera, em especial, o tempo de execu??o. No caso dos algoritmos heur?sticos, al?m do tempo de execu??o, a qualidade do conjunto de aproxima??o gerado ? avaliada. Indicadores de qualidade s?o empregados para aferir tal informa??o. Ferramentas estat?sticas apropriadas s?o usadas na an?lise de desempenho dos algoritmos exatos e heur?sticos. Para o conjunto de inst?ncias utilizado e considerando os crit?rios de qualidade dos conjuntos de aproxima??o gerados e tempo de execu??o dos algoritmos, os experimentos mostraram que o algoritmo de Busca Tabu com ejection chain obteve melhores resultados e que o algoritmo transgen?tico ficou em segundo lugar. A busca local PLS obteve solu??es de qualidade, mas a um tempo computacional muito alto se comparado ?s outras (meta)heur?sticas. Nesse sentido, ocupa a terceira coloca??o. Por fim, ficaram os algoritmos NSTA e NSGAII