二倍体遺伝的アルゴリズムと境界要素法による射出成型金型冷却孔の最適化

概要

論文の詳細を見る
In this paper, a genetic algorithm(GA) with diploid gene encoding structure is applied to obtain the optimum layout, size, and coolant condition parameters of injection mold.Although GA is effective in searching globally the parameter space of desing variables of optimzation problems, it is costly when such a heavy computation as that by FEM or BEM is required for evaluating the fitness of each individual.In the present work, a diploid gene structure is employed as the encoding scheme for the design variables of an optimization problem for designing cooling channels running through injection molds.In order to evaluate the fitness, which is a function of temperature over the cavity surface of the mold, the boundary element method(BEM)is employed due to its easiness in remeshing the boundary corresponding to design parameters of each individual.Several numerical demonstrations show that the proposed diploid GA can give optimum values of the design variables more repidly with a smaller number of populations than the standard GA with haploidy.
一般社団法人日本機械学会の論文
2000-01-25