ラディアルベーシス関数ネットワークと領域適応型遺伝的アルゴリズムを用いた最適設計 (第2報, 制約条件のある場合における検討)

抄録

In this paper, we use the radial basis function network in order to approximate the fitness function of the genetic algorithms and try to obtain the approximate optimum results within the relatively small number of function call. In the first report, we have shown the effectiveness of the proposed method in unconstrained optimization. In the constrained optimization, we have to take care of constraints. In genetic algorithms, we usually treat them by including into fitness function with penalty function. However, when we estimate fitness function by using penalty function, giving good penalty becomes very difficult. Moreover, when we give a large penalty to them, approximation will be influenced by those penalities and it becomes almost impossible to obtain real optimum solutions. In order to use the benefits of the method in unconstrained problems, we ignore data those which are not included in feasible region and assume to have unconstrainted optimum problem. From famous 25 bar truss problem, we will show the effectiveness of the method.