2-408 協調搬送問題における共進化型ニューラルコントローラの適応度景観の推定に関する一考察

概要

論文の詳細を見る
The fitness landscapes describe the search space as a multidimensional landscape defined by the genotype-to-fitness mapping through which evolution moves. Smith introduced fitness evolvability portraits to represent the features of landscape (modality, ruggedness and neutrality) in easy discrete problem. In this paper, we deal with the more complex problem, cooperative carrying problem by means of three coevolutionary approaches. We apply the fitness evolvability portraits for representation of the each fitness landscapes and to compare with the real problem and the theoretical test functions. Like result here, we show the various features of fitness landscapes and fitness values relevant to different coevolutionary approaches.
一般社団法人日本機械学会の論文
2002-11-14