【摘要】 对约束条件的处理是应用遗传算法求解约束问题所涉及的一个主要内容。本文首先描述了普遍采用的“惩罚函数遗传算法”,针对该算法存在的问题进行了分析,提出了基于梯度修正的模拟退火遗传算法,给出对不可行解的修正方法。算法的实施可分两步,第1步采用修正的遗传算法搜索目标函数的可行解或全局可行最优解;第2步利用模拟退火算法对可行解局部优化。最后以弧齿锥齿轮优化为例,对算法的可行性进行了验证。更多还原

【Abstract】 The constraint handling is one of the major concerns when applying genetic algorithms to solve constrained optimization problems.This paper proposed the integrating repair genetic algorithms and simulated annealing for the constrained problems based on describing the popular penalty for genetic algorithms and analyzing the problems of the algorithms.The repair method is given for the infeasible solution in the proposed procedure.There are two steps for implementing of the algorithm Firstly,the feasible solutions or the feasible optimization solutions of the objective function are gotten by the repair genetic algorithms.Secondly,applying simulated annealing to optimize the feasible solutions locally.Finally the optimum design of bevel gear as an example is strctied to verify the feasibility of algorithms.更多还原