Department   Hiroshima shudo University  The Faculty of Commercial Sciences
   Position   Professor
Language Japanese
Publication Date 2018/11
Type Articles
Peer Review With peer review
Title Grouping of Genes According to Correlation Coefficients and Grouping-Based Crossover for Adaptive Differential Evolution
Contribution Type Co-Authored Publication
Journal Extended Abstract of The 50th ISCIE Inter. Sympo. on Stochastic Systems Theory and Its Applications
Journal TypeAnother Country
Publisher The Institute of Systems, Control and Information Engineers
Volume, Issue, Pages 115-116頁
Number of pages 2
Author and coauthor Takahama Tetsuyuki and Setsuko Sakai
Details In this study, we propose to use correlation coefficients of search points in order to detect such distribution. The correlation coefficients between decision variables can be obtained from the search points. The highly correlated variables (genes) are grouped and the grouped genes are crossed (or not crossed) simultaneously. However, if only GBX is used as a crossover operation, the diversity of the search points tends to be lost rapidly. In order to keep the diversity, we propose to adjust the probability of applying GBX adaptively. In this study, GBX and adaptive adjustment of the probability of applying GBX is introduced to JADE where JADE is extended by group learning.