遺伝子の特徴抽出による新しいGA法とその応用

概要

論文の詳細を見る
The Genetic Algorithm (GA), which is optimization method based on the natural evolution theorem, is applicable to the wide scale problems. By using the GA, various evolutionary optimization methods are studied. But, in these methods, the meanings of genetic-code are not considered at all. In this study, the Special Feature Extracting GA (SFE-GA) is proposed. The SFE-GA is based on the idea that the genes possessing the suitable environment extend to all over the chromosome, and is the GA with an operation of stepping up the function. A superior group with the high fitness value includes the schemata of building blocks of good solutions. The schemata is extracted using the t-test, and the group with the low fitness value is overwritten to the extracted schemata. This method is applied to the problem that searches the maximum value in a simple function. The behavior in the search is clarified and it is confirmed to extract the effective schemata sequentially. Moreover, it's applied to the layout design problem of truss structure as the large scale problem.
一般社団法人日本機械学会の論文
2004-10-25