TAKAHAMA Setsuko
Department Hiroshima shudo University The Faculty of Commercial Sciences Position Professor |
|
Language | English |
Publication Date | 2024/02 |
Type | books |
Title | A Study on Crossover Operations Using Hill-Valley Detection for Multimodal Optimization by Particle Swarm Optimization with Graph-Based Speciation |
Contribution Type | Co-Authored Publication |
Journal Type | Japan |
Publisher | Kyushu University Press |
Volume, Issue, Pages | pp.73-93 |
Number of pages | 21 |
Author and coauthor | Setsuko Sakai and Tetsuyuki Takahama |
Details | Speciation using proximity graphs has been studied as a method for solving multimodal optimization problems. In this study, we propose to introduce crossover operations to the PSO with graph-based speciation using βRNG to improve the search accuracy for functions with a large number of optimal solutions and a regular distribution of the solutions.Good solutions, which are obtained by hill-valley detection, are used in the crossover operations to enhance the effectiveness of the operations. We also propose an adaptive method to adjust the crossover rate to avoid the loss of search efficiency for other functions due to the crossover operations. |
ISBN | 978-4-7985-0371-4 |