適応型遺伝的アルゴリズムによるトラス構造物の最適化

概要

論文の詳細を見る
In the present paper, an adaptive genetic algorithm (GA) that population size and mutation rate are controlled dynamically in the search process is proposed. The search performance of the adaptive GA realized by personal computer (PC) cluster is discussed by applying it to the optimal design of truss structures. One of the problems in GA is that many experiences are required to decide the parameters such as the population size, the crossover rate, the mutation rate. The only parameter of the adaptive GA is an initial mutation rate, and the search performance is robust for the initial mutation rate. Therefore, the number of the numerical experiments to adjust the parameter decreases, and the adaptive GA is effective in the optimal design of structures that the amounts of calculation of the objective function increase. The search performance of GA which adjusted parameters to the best condition, and the adaptive GA was almost equal.
一般社団法人日本建築学会の論文
2003-03-20