タカハマ セツコ
TAKAHAMA Setsuko 高濱 節子 所属 広島修道大学 商学部 職種 教授 |
|
言語種別 | 英語 |
発行・発表の年月 | 2017/03 |
形態種別 | 著書 |
標題 | A Comparative Study on Grouping Methods for an Adaptive Differential Evolutuion |
執筆形態 | 分担 |
掲載誌名 | Challenging Researches in Economic Sciences |
掲載区分 | 国内 |
出版社・発行元 | Kyushu University Press |
巻・号・頁 | pp.51-91 |
頁数 | 41 |
著者・共著者 | Setsuko Sakai, Tetsuyuki Takahama |
概要 | In this study, group-based learning of algorithm parameters is proposed, where individuals are divided into plural groups according to some criteria and the parameters are learned for each group. Three criteria including the objective value for each individual, the distance between the best individual and each individual, and the distance between the centroid of individuals and each individual are proposed and tested.
DE with group learning is applied optimization of various 13 functions . It is shown that proposed method is effective compared with JADE. Especially, the criterion using the objective value is most effective among three criteria. Also, it is shown that parameters for good individuals are controlled to intensify convergence and parameters for bad individuals are controlled to keep divergence. |
ISBN | 978-4-7985-0206-9 |