Research Article
Multi-factorial Evolutionary Algorithm Using Objective Similarity Based Parent Selection
@INPROCEEDINGS{10.1007/978-3-030-92163-7_5, author={Shio Kawakami and Keiki Takadama and Hiroyuki Sato}, title={Multi-factorial Evolutionary Algorithm Using Objective Similarity Based Parent Selection}, proceedings={Bio-Inspired Information and Communications Technologies. 13th EAI International Conference, BICT 2021, Virtual Event, September 1--2, 2021, Proceedings}, proceedings_a={BICT}, year={2022}, month={1}, keywords={Multi-factorial optimization Evolutionary algorithms Objective function similarity}, doi={10.1007/978-3-030-92163-7_5} }- Shio Kawakami
Keiki Takadama
Hiroyuki Sato
Year: 2022
Multi-factorial Evolutionary Algorithm Using Objective Similarity Based Parent Selection
BICT
Springer
DOI: 10.1007/978-3-030-92163-7_5
Abstract
This work proposes a multi-factorial evolutionary algorithm encouraging crossovers among solutions with similar target objective functions and suppressing crossovers among solutions with dissimilar target objective functions. Evolutionary multi-factorial optimization simultaneously optimizes multiple objective functions with a single population, a solution set. Each solution has a target objective function, and sharing solution resources in one population enhances the simultaneous search for multiple objective functions. However, the conventional multi-factorial evolutionary algorithm does not consider similarities among objective functions. As a result, solutions with dissimilar target objectives are crossed, and it deteriorates the search efficiency. The proposed algorithm estimates objective similarities based on search directions of solution subsets with different target objective functions in the design variable space. The proposed algorithm then encourages crossovers among solutions with similar target objectives and suppresses crossovers among solutions with dissimilar objectives. Experimental results using multi-factorial distance minimization problems show the proposed algorithm achieves higher search performance than the conventional evolutionary single-objective optimization and multi-factorial optimization.
