https://scholars.lib.ntu.edu.tw/handle/123456789/636088
標題: | Adaptive donor selection mixing for multi-objective optimization: An enhanced variant of mo-gomea | 作者: | Liao, Hsu Chen Fang, Wen Zhong TIAN-LI YU |
關鍵字: | donor selection | evolutionary algorithms | model building | multi-objective optimization | scalability | 公開日期: | 15-七月-2023 | 來源出版物: | GECCO 2023 - Proceedings of the 2023 Genetic and Evolutionary Computation Conference | 摘要: | The multi-objective gene-pool optimal mixing evolutionary algorithm with interleaved multi-start scheme (MO-GOMEA) is a powerful, parameterless model-based genetic algorithm that excels at solving multi-objective combinatorial optimization problems. In this paper, we propose a new mixing mechanism, adaptive donor selection mixing (ADSM) and further integrate it into MO-GOMEA to form a new variant, ADSM-MO-GOMEA. The proposed ADSM mechanism adaptively switches between cluster-guided and elitist-guided mixing, with the latter having a customized donor selection for the receiver based on empirical observations and mathematical derivation. The empirical results on multiple benchmark problems indicate that ADSM-MO-GOMEA improves the effectiveness over the original MO-GOMEA and achieves a lower inverted generational diversity and higher front occupation within the given limited number of evaluations. |
URI: | https://scholars.lib.ntu.edu.tw/handle/123456789/636088 | ISBN: | 9798400701191 | DOI: | 10.1145/3583131.3590403 |
顯示於: | 電機工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。