Multiple source transfer learning for dynamic multiobjective optimization
Recently, dynamic multiobjective evolutionary algorithms (DMOEAs) with transfer learning have become popular for solving dynamic multiobjective optimization problems (DMOPs), as the used transfer learning methods in DMOEAs can effectively generate a good initial population for the new environment. H...
| Autores: | ; ; ; ; ; |
|---|---|
| Tipo de documento: | Artículo |
| Estado: | Published version |
| Fecha de publicación: | 2022 |
| País: | España |
| Institución: | Basque Center for Applied Mathematics (BCAM) |
| Repositorio: | BIRD. BCAM's Institutional Repository Data |
| OAI Identifier: | oai:bird.bcamath.org:20.500.11824/1562 |
| Acceso en línea: | http://hdl.handle.net/20.500.11824/1562 |
| Nivel de acceso: | Acceso abierto |
| Palabra clave: | Domain adaptation Dynamic optimization Evolutionary algorithm Multiobjective optimization Transfer learning |
| _version_ | 1878730373328797696 |
|---|---|
| author | Ye, Y. |
| author2 | Lin, Q. Ma, L. Wong, K. C. Gong, M. Coello, C.A. |
| author2_role | author author author author author |
| author_browse | Coello, C.A. Gong, M. Lin, Q. Ma, L. Wong, K. C. Ye, Y. |
| author_facet | Ye, Y. Lin, Q. Ma, L. Wong, K. C. Gong, M. Coello, C.A. |
| author_role | author |
| collection | BIRD. BCAM's Institutional Repository Data |
| description | Recently, dynamic multiobjective evolutionary algorithms (DMOEAs) with transfer learning have become popular for solving dynamic multiobjective optimization problems (DMOPs), as the used transfer learning methods in DMOEAs can effectively generate a good initial population for the new environment. However, most of them only transfer non-dominated solutions from the previous one or two environments, which cannot fully exploit all historical information and may easily induce negative transfer as only limited knowledge is available. To address this problem, this paper presents a multiple source transfer learning method for DMOEA, called MSTL-DMOEA, which runs two transfer learning procedures to fully exploit the historical information from all previous environments. First, to select some representative solutions for knowledge transfer, one clustering-based manifold transfer learning is run to cluster non-dominated solutions of the last environment to obtain their centroids, which are then fed into the manifold transfer learning model to predict the corresponding centroids for the new environment. After that, multiple source transfer learning is further run by using multisource TrAdaboost, which can fully exploit information from the above centroids in new environment and old centroids from all previous environments, aiming to construct a more accurate prediction model. This way, MSTL-DMOEA can predict an initial population with better quality for the new environment. The experimental results also validate the superiority of MSTL-DMOEA over several competitive state-of-the-art DMOEAs in solving various kinds of DMOPs. |
| eu_rights_str_mv | openAccess |
| format | article |
| id | ES_1afc8ea2e087c84ecadbf179d9cbc0b1 |
| institution | Basque Center for Applied Mathematics (BCAM) |
| instname_str | Basque Center for Applied Mathematics (BCAM) |
| network_acronym_str | ES |
| network_name_str | España |
| oai_identifier_str | oai:bird.bcamath.org:20.500.11824/1562 |
| publishDate | 2022 |
| publishDateSort | 2022 |
| reponame_str | BIRD. BCAM's Institutional Repository Data |
| spelling | Multiple source transfer learning for dynamic multiobjective optimization Ye, Y. Lin, Q. Ma, L. Wong, K. C. Gong, M. Coello, C.A. Domain adaptation Dynamic optimization Evolutionary algorithm Multiobjective optimization Transfer learning Recently, dynamic multiobjective evolutionary algorithms (DMOEAs) with transfer learning have become popular for solving dynamic multiobjective optimization problems (DMOPs), as the used transfer learning methods in DMOEAs can effectively generate a good initial population for the new environment. However, most of them only transfer non-dominated solutions from the previous one or two environments, which cannot fully exploit all historical information and may easily induce negative transfer as only limited knowledge is available. To address this problem, this paper presents a multiple source transfer learning method for DMOEA, called MSTL-DMOEA, which runs two transfer learning procedures to fully exploit the historical information from all previous environments. First, to select some representative solutions for knowledge transfer, one clustering-based manifold transfer learning is run to cluster non-dominated solutions of the last environment to obtain their centroids, which are then fed into the manifold transfer learning model to predict the corresponding centroids for the new environment. After that, multiple source transfer learning is further run by using multisource TrAdaboost, which can fully exploit information from the above centroids in new environment and old centroids from all previous environments, aiming to construct a more accurate prediction model. This way, MSTL-DMOEA can predict an initial population with better quality for the new environment. The experimental results also validate the superiority of MSTL-DMOEA over several competitive state-of-the-art DMOEAs in solving various kinds of DMOPs. http://hdl.handle.net/20.500.11824/1562 |
| spellingShingle | Multiple source transfer learning for dynamic multiobjective optimization Ye, Y. Domain adaptation Dynamic optimization Evolutionary algorithm Multiobjective optimization Transfer learning |
| status_str | publishedVersion |
| title | Multiple source transfer learning for dynamic multiobjective optimization |
| title_full | Multiple source transfer learning for dynamic multiobjective optimization |
| title_fullStr | Multiple source transfer learning for dynamic multiobjective optimization |
| title_full_unstemmed | Multiple source transfer learning for dynamic multiobjective optimization |
| title_short | Multiple source transfer learning for dynamic multiobjective optimization |
| title_sort | Multiple source transfer learning for dynamic multiobjective optimization |
| topic | Domain adaptation Dynamic optimization Evolutionary algorithm Multiobjective optimization Transfer learning |
| topic_facet | Domain adaptation Dynamic optimization Evolutionary algorithm Multiobjective optimization Transfer learning |
| url | http://hdl.handle.net/20.500.11824/1562 |