2-112 GA の解探索過程における多様性の維持に関する研究

概要

論文の詳細を見る
When the GA is applied to optimization problems, it is important to keep the diversity in designing generation alternation model. Generally, when the diversity is not fully maintained, it is difficult to find good solution, and it is easy to stagnate the premature convergenece. In this paper, we propose the Elite Correlation Selection operator (ECS) as a new selection operator for survival. This selection operator aims to keep the diversity of populations and contributes the high searching ability. This selection operator is an extension of selection operator for survival in the Minimal Generation Gap (MGG). In the selection for survival, this selection operator selects one elite individual and one low simirality between elite individual. As a measure of similarity, we use the haming distance among individuals. The perfomance of the ECS operator is examined with several standard benchmark functions. We compare the performance of the ECS operater with the MGG and SGA. The simulation results show good performance.
一般社団法人日本機械学会の論文
2002-11-14