• Entrar
    Ver item 
    •   Página inicial
    • BIBLIOTECA DIGITAL: Teses & Dissertações
    • 40001016043P4 Programa de Pós-Graduação em Engenharia Elétrica
    • Teses
    • Ver item
    •   Página inicial
    • BIBLIOTECA DIGITAL: Teses & Dissertações
    • 40001016043P4 Programa de Pós-Graduação em Engenharia Elétrica
    • Teses
    • Ver item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Otimização multiobjetivo utilizando metaheurísticas para modelos em engenharia

    Thumbnail
    Visualizar/Abrir
    R - T - MARIANE GAVIOLI BERGAMINI FERNANDES.pdf (13.17Mb)
    Data
    2024
    Autor
    Bergamini, Mariane Gavioli
    Metadata
    Mostrar registro completo
    Resumo
    Resumo: Na área da engenharia, ao se trabalhar com um problema de otimização, pode haver necessidade de se otimizar três ou mais objetivos como, por exemplo, no projeto de geradores em engenharia elétrica, de motores em engenharia mecânica e de estruturas na engenharia civil. As metaheurísticas multiobjetivo são técnicas bastantes atrativas quando se lida com problemas de múltiplos objetivos conflitantes, pois conseguem explorar os tradeoffs entre diferentes objetivos e alcançar as soluções em um tempo computacional razoável para problemas não lineares/grande escala (dimensão). A busca por novas técnicas para a otimização de sistemas resulta no desenvolvimento de novas metaheurísticas para suprir as lacunas que as técnicas anteriores não foram tão eficientes (como exemplo, técnicas com tempo computacional razoável), neste sentido, este trabalho apresenta duas novas metaheurísticas multiobjetivo. A primeira metaheurística desenvolvida utiliza como base o Algoritmo do Leão e os conceitos de dominância e decomposição, o qual será denominado MOLA (Multi-objective Optimization Lion Algorithm). A segunda metaheurística desenvolvida utiliza o Algoritmo do Leão e a Busca Local Iterativa trabalhando de forma cooperativa e o conceito de vetores suporte, o qual será denominado MOLAILS (Multi-objective Optimization Lion Algorithm Iterated Search Local). As novas metaheurísticas propostas foram submetidas a um conjunto de funções de teste para otimização multiobjetivo e seus desempenhos foram comparados com o de outros algoritmos de otimização por meio da distância geracional invertida (IGD), pelo hipervolume (HV) e por inferência não paramétricas. Os resultados dos testes em benchmarks mostraram que os algoritmos MOLA e o MOLSLA superam ou pelo menos tem desempenho semelhante ao Multi-objective Grey Wolf Optimizer (MOGWO), Multiple Objective Particle Swarm Optimization (MOPSO), Multiobjective Salp Swarm Algorithm (MSSA), Multi-objective Evolutionary Algorithm Based on Decomposition (MOEA/D) e Algoritmo Genético de Sorting Não Dominado III (NSGA-III). Os algoritmos propostos foram então aplicados em problemas de modelagem da área da engenharia a estimação de parâmetros da malha hidráulica de uma usina hidroelétrica e a a estimação de parâmetros do sistema massa-mola-amortecimento. Os resultados médios dos índices de desempenho mostraram que as metaheurísticas propostas MOLA e MOLAILS obtevem resultados semelhantes ou superiores do algoritmo MOGWO.
     
    Abstract: In the area of engineering, when working with an optimization problem, there may be a need to optimize three or more objectives, for example, in the design of generators in electrical engineering, engines in mechanical engineering and structures in civil engineering. Multi-objective metaheuristics are very attractive techniques when dealing with problems with multiple conflicting objectives, as they can explore trade-offs between different objectives and reach solutions in a reasonable computational time for non-linear/large-scale problems. The search for new techniques for system optimization results in the development of new metaheuristics to fill the gaps that previous techniques were not as efficient (for example, techniques with reasonable computational time). In this sense, this work presents two new multi-objective metaheuristics. The first metaheuristic developed uses the Lion Algorithm and the concepts of dominance and decomposition as a basis, which will be called MOLA (Multiobjective Optimization Lion Algorithm). The second metaheuristic developed uses the Lion Algorithm and Iterative Local Search working cooperatively and the concept of support vectors, which will be called MOLAILS (Multi-objective Optimization Lion Algorithm Iterated Search Local). The proposed new metaheuristics were subjected to a set of test functions for multi-objective optimization and their performance was compared with that of other optimization algorithms using inverted generational distance (IGD), hypervolume (HV) and non-parametric inference. Benchmark test results showed that MOLA and MOLSLA algorithms outperform or at least have similar performance to Multi-objective Gray Wolf Optimizer (MOGWO), Multiple Objective Particle Swarm Optimization (MOPSO), Multiobjective Salp Swarm Algorithm (MSSA), Multi-objective Evolutionary Algorithm Based on Decomposition (MOEA/D) and Non-Dominated Sorting Genetic Algorithm III (NSGA-III). The proposed algorithms were then applied to engineering modeling problems, the estimation of parameters of the hydraulic network of a hydroelectric plant and the estimation of parameters of the mass-spring-damping system. The average results of the performance indexes showed that the proposed metaheuristics MOLA and MOLAILS obtained similar or higher results to the MOGWO algorithm.
     
    URI
    https://hdl.handle.net/1884/90010
    Collections
    • Teses [41]

    DSpace software copyright © 2002-2022  LYRASIS
    Entre em contato | Deixe sua opinião
    Theme by 
    Atmire NV
     

     

    Navegar

    Todo o repositórioComunidades e ColeçõesPor data do documentoAutoresTítulosAssuntosTipoEsta coleçãoPor data do documentoAutoresTítulosAssuntosTipo

    Minha conta

    EntrarCadastro

    Estatística

    Ver as estatísticas de uso

    DSpace software copyright © 2002-2022  LYRASIS
    Entre em contato | Deixe sua opinião
    Theme by 
    Atmire NV