求解约束优化问题的一种新的遗传算法

A Novel Genetic Algorithm for Solving Constrained Optimization Problems

提出一种新的求解约束优化问题的遗传算法,算法通过重新定义可行解与不可行解的适应度函数分别对它们进行选择,有效避免了惩罚函数法引入参数所带来的困难,重新设计的交叉算子使得算法对解空间的寻优范围扩大了.数值实验结果表明算法具有较好的鲁棒性,且对最优解位于约束边界上的一类问题具有很大优势.

A novel genetic algorithm for solving constrained optimization is proposed. Feasible and infeasible solution are chosen, separately, via their re-defined fitness function, which effectively avoids the difficulties caused by introducing parameters in penalty function method. Re-designed crossover operators expand the search scope of the proposed algorithm in the optimal solution space. The numerical experience results show that the proposed algorithm has good robust,and has the great advantage for a class of problems whose optimal solution is located on the constraint boundary.