基于惩罚和修复策略的约束优化遗传算法

约束优化问题中最难以解决的就是约束处理问题，将惩罚函数法与修复策略相结合应用于非线性约束优化遗传算法之中，使得约束优化问题在惩罚函数和修复算子的协同作用下收敛于全局最优，有效避免了迭代过程中大量非可行解的产生，解决了在遗传算法约束优化问题中单独使用惩罚和修复方法时一些难以解决的问题。基于随机方向法构造的修复算子作用效果显著，采用多个测试函数对算法进行检验，均能较好地收敛于可行域中的最优解，验证了算法的可靠性。

Constrained optimal genetic algorithm based on strategy of penalty and renovation

The most difficult problem to be solved in constraint optimization is the problem of constraint treatment. Let the combination of penalty function method and renovation strategy be applied to the optimal genetic algorithm with nonlinear constraint, thus let the constrained optimal problem be converged in an overall optimization under the coordinative effect of penalty function and renovation operator. It effectively avoided the generation of large numbered non-feasible solutions in the course of iteration, and solved a number of problems, which are hard to solve while independently using penalty and renovation method in the constrained optimization problems of genetic algorithm. On the basis of remarkable result of the effect of renovation operator constructed by random direction method, tests were carried out on the algorithm using many testing functions, which shows that the problems of constrained optimization could all be converged into the optimal solution of feasible region, thus verified the correctness of the algorithm.