A Study of Searching Bias based on a-b Diversity in Genetic Programming

2015 Fiscal Year Final Research Report

• Project/Area Number: 26730133
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Soft computing
• Research Institution: Ryukoku University
• Principal Investigator: Ono Keiko 龍谷大学, 理工学部, 講師 (80550235)
• Research Collaborator: HANADA Yoshiko 関西大学, システム理工学部, 准教授 (30511711)
• Project Period (FY): 2014-04-01 – 2016-03-31
• Keywords: 遺伝的プログラミング / 並列モデル / 遺伝的多様性 / ネットワーククラスタリング

Outline of Final Research Achievements

A parallel model encourages genetic diversity and frequently shows a better search performance than do single population models. To enhance the parallel model, it is important to consider a balance between local and genetic search. In GP, individuals have various features, and, so it is difficult to determine which feature is the most effective. Therefore, we proposed a novel adaptive subpopulation model (cuSASGP). The proposed method automatically generates a correlation network on the basis of the difference between individuals in terms of not only a fitness value but also node size and generates subpopulations by network clustering.
Using three benchmark problems, we demonstrate that performance improvement can be achieved, and that the proposed method significantly outperforms a typical method. Moreover, we verify that genetic diversity can be achieved by adopting subpopulation models such as the island method and cuSASGP in the lighting control problem.

Free Research Field

進化計算