Contributions in Global Derivative-free Optimization to the development of an integrated toolbox of solvers
Autor(a) principal: | |
---|---|
Data de Publicação: | 2021 |
Tipo de documento: | Dissertação |
Idioma: | eng |
Título da fonte: | Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
Texto Completo: | http://hdl.handle.net/10362/141076 |
Resumo: | This dissertation is framed in the research project "BoostDFO: Improving the performance and moving to newer directions in Derivative-Free Optimization", funded by Fundação para a Ciência e Tecnologia, whose objective is to develop efficient and robust algorithms for Global and Multiobjective Derivative-Free Optimization. The demand for algorithms belonging to this class appears in different application areas, such as robotics, electrical engineering, aeronautics or oceanography, where, for several reasons, it is not possible to use derivatives. Global and Local Optimization using Direct Search (GLODS) is an algorithm belonging to this class of optimization methods, which aims at identifying the global minimum of a given problem by computing all the local minima. However, identifying points as global minimizers is always a hard task, being of increased complexity when the function to optimize is computational expensive and time-consuming. Typically, the higher the dimension and complexity of the problem, the more computational effort and time will be required to run the algorithm. The current version of GLODS is developed sequentially, and it is not fully optimized. The present work will analyze in detail the execution and behavior of GLODS, and will propose and evaluate the numerical performance of different parallelization strategies, implemented using the MATLAB Parallel Computing Toolbox (PCT). The parallelization structure will have a primary purpose of allowing to distribute objective function evaluations among different processors of the hardware platforms, both host locally or in public clouds. Three strategies for GLODS parallelization were designed and implemented centered on the poll phase of the algorithm. The first two - parallelization of the poll step with one and two poll centers - were successful and gave good results in terms of solution quality and execution time reduction. The third level of parallelization introduced the possibility of dynamically selecting the number of poll centers, according with the number of processors available, a functionality that can be particularly useful in future work in the area. |
id |
RCAP_c1cc989554f1152a1de51e6041fa928a |
---|---|
oai_identifier_str |
oai:run.unl.pt:10362/141076 |
network_acronym_str |
RCAP |
network_name_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository_id_str |
7160 |
spelling |
Contributions in Global Derivative-free Optimization to the development of an integrated toolbox of solversDerivative-free OptimizationDirectional Direct SearchGLODS algorithmParallel computingCloud computingMATLABDomínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e InformáticaThis dissertation is framed in the research project "BoostDFO: Improving the performance and moving to newer directions in Derivative-Free Optimization", funded by Fundação para a Ciência e Tecnologia, whose objective is to develop efficient and robust algorithms for Global and Multiobjective Derivative-Free Optimization. The demand for algorithms belonging to this class appears in different application areas, such as robotics, electrical engineering, aeronautics or oceanography, where, for several reasons, it is not possible to use derivatives. Global and Local Optimization using Direct Search (GLODS) is an algorithm belonging to this class of optimization methods, which aims at identifying the global minimum of a given problem by computing all the local minima. However, identifying points as global minimizers is always a hard task, being of increased complexity when the function to optimize is computational expensive and time-consuming. Typically, the higher the dimension and complexity of the problem, the more computational effort and time will be required to run the algorithm. The current version of GLODS is developed sequentially, and it is not fully optimized. The present work will analyze in detail the execution and behavior of GLODS, and will propose and evaluate the numerical performance of different parallelization strategies, implemented using the MATLAB Parallel Computing Toolbox (PCT). The parallelization structure will have a primary purpose of allowing to distribute objective function evaluations among different processors of the hardware platforms, both host locally or in public clouds. Three strategies for GLODS parallelization were designed and implemented centered on the poll phase of the algorithm. The first two - parallelization of the poll step with one and two poll centers - were successful and gave good results in terms of solution quality and execution time reduction. The third level of parallelization introduced the possibility of dynamically selecting the number of poll centers, according with the number of processors available, a functionality that can be particularly useful in future work in the area.Esta dissertação enquadra-se no projeto de investigação "BoostDFO: Aumentando o desempenho e abordando novas dimensões em optimização sem recurso a derivadas", financiado pela Fundação para a Ciência e a Tecnologia, cujo objetivo é desenvolver algoritmos eficientes e robustos para Otimização Global e/ou Multiobjectivo sem recurso a derivadas. A procura e utilização de algoritmos pertencentes a esta classe ocorre em diferentes áreas, como a robótica, a engenharia eletrotécnica, a aeronáutica ou a oceanografia, onde, por diversas razões, não é possível a utilização de derivadas. Global and Local Optimization using Direct Search (GLODS) é um algoritmo pertencente a esta classe de métodos de otimização, que visa identificar o mínimo global de um certo problema através do cálculo dos diferentes mínimos locais. No entanto, identificar pontos como minimizantes globais é sempre uma tarefa difícil, sendo de maior complexidade quando a função a otimizar apresenta elevados custos computacionais e tempo de execução. Normalmente, quanto maior a dimensão e a complexidade de um problema, mais recursos e tempo computacional serão necessários para a execução do algoritmo. A versão atual do algoritmo GLODS encontra-se implementada sequencialmente e não está totalmente otimizada. O presente trabalho analisará em detalhe a execução e o comportamento deste algortimo, propondo e avaliando o desempenho númerico de diferentes estruturas de paralelização para o mesmo, implementadas recorrendo às ferramentas de paralelismo do MATLAB. A estrutura de paralelização terá como objetivo principal a distribuição das avaliações da função objetivo entre diferentes processadores, num ambiente local ou na cloud pública. Delinearam-se e implementaram-se três estratégias para a paralelização do GLODS, centradas na fase de sondagem do algoritmo. As primeiras duas - paralelização do passo de sondagem com um e com dois centros de sondagem - foram bem sucedidas, permitindo bons resultados em termos de qualidade da solução e também uma redução no tempo de execução do algoritmo. O terceiro nível de paralelização introduziu a possibilidade de selecionar o número de centros de sondagem de foma dinâmica, considerando o número de processadores disponíveis, uma funcionalidade que será particularmente útil no trabalho futuro a desenvolver na área.Medeiros, PedroCustódio, AnaRUNSantos, Nelson Alexandre Charreu2022-06-30T12:25:05Z2021-112021-11-01T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfhttp://hdl.handle.net/10362/141076enginfo:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2024-03-11T05:18:23Zoai:run.unl.pt:10362/141076Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-20T03:49:54.788870Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse |
dc.title.none.fl_str_mv |
Contributions in Global Derivative-free Optimization to the development of an integrated toolbox of solvers |
title |
Contributions in Global Derivative-free Optimization to the development of an integrated toolbox of solvers |
spellingShingle |
Contributions in Global Derivative-free Optimization to the development of an integrated toolbox of solvers Santos, Nelson Alexandre Charreu Derivative-free Optimization Directional Direct Search GLODS algorithm Parallel computing Cloud computing MATLAB Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
title_short |
Contributions in Global Derivative-free Optimization to the development of an integrated toolbox of solvers |
title_full |
Contributions in Global Derivative-free Optimization to the development of an integrated toolbox of solvers |
title_fullStr |
Contributions in Global Derivative-free Optimization to the development of an integrated toolbox of solvers |
title_full_unstemmed |
Contributions in Global Derivative-free Optimization to the development of an integrated toolbox of solvers |
title_sort |
Contributions in Global Derivative-free Optimization to the development of an integrated toolbox of solvers |
author |
Santos, Nelson Alexandre Charreu |
author_facet |
Santos, Nelson Alexandre Charreu |
author_role |
author |
dc.contributor.none.fl_str_mv |
Medeiros, Pedro Custódio, Ana RUN |
dc.contributor.author.fl_str_mv |
Santos, Nelson Alexandre Charreu |
dc.subject.por.fl_str_mv |
Derivative-free Optimization Directional Direct Search GLODS algorithm Parallel computing Cloud computing MATLAB Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
topic |
Derivative-free Optimization Directional Direct Search GLODS algorithm Parallel computing Cloud computing MATLAB Domínio/Área Científica::Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
description |
This dissertation is framed in the research project "BoostDFO: Improving the performance and moving to newer directions in Derivative-Free Optimization", funded by Fundação para a Ciência e Tecnologia, whose objective is to develop efficient and robust algorithms for Global and Multiobjective Derivative-Free Optimization. The demand for algorithms belonging to this class appears in different application areas, such as robotics, electrical engineering, aeronautics or oceanography, where, for several reasons, it is not possible to use derivatives. Global and Local Optimization using Direct Search (GLODS) is an algorithm belonging to this class of optimization methods, which aims at identifying the global minimum of a given problem by computing all the local minima. However, identifying points as global minimizers is always a hard task, being of increased complexity when the function to optimize is computational expensive and time-consuming. Typically, the higher the dimension and complexity of the problem, the more computational effort and time will be required to run the algorithm. The current version of GLODS is developed sequentially, and it is not fully optimized. The present work will analyze in detail the execution and behavior of GLODS, and will propose and evaluate the numerical performance of different parallelization strategies, implemented using the MATLAB Parallel Computing Toolbox (PCT). The parallelization structure will have a primary purpose of allowing to distribute objective function evaluations among different processors of the hardware platforms, both host locally or in public clouds. Three strategies for GLODS parallelization were designed and implemented centered on the poll phase of the algorithm. The first two - parallelization of the poll step with one and two poll centers - were successful and gave good results in terms of solution quality and execution time reduction. The third level of parallelization introduced the possibility of dynamically selecting the number of poll centers, according with the number of processors available, a functionality that can be particularly useful in future work in the area. |
publishDate |
2021 |
dc.date.none.fl_str_mv |
2021-11 2021-11-01T00:00:00Z 2022-06-30T12:25:05Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/masterThesis |
format |
masterThesis |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/10362/141076 |
url |
http://hdl.handle.net/10362/141076 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.format.none.fl_str_mv |
application/pdf |
dc.source.none.fl_str_mv |
reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação instacron:RCAAP |
instname_str |
Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
instacron_str |
RCAAP |
institution |
RCAAP |
reponame_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
collection |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository.name.fl_str_mv |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
repository.mail.fl_str_mv |
|
_version_ |
1799138096575414272 |