求解约束优化问题的一种新的进化算法

分析了现有的约束优化进化算法的一些不足之处，提出了一种处理约束优化问题的新算法。新算法将多目标优化思想与全局搜索和局部搜索机制有机地结合起来；在全局搜索过程中，作为一种小生态遗传算法，排挤操作利用Pareto优劣关系比较个体并接受具有相似性的父代个体和予代个体中的优胜者；在局部搜索过程中，首先对局部群体中的个体赋予Pareto强度，然后根据Pareto强度选择个体。通过一个复杂高维多峰测试函数验证了新算法的有效性。

A new evolutionary algorithm for solving constrained optimization problems

After analyzing the drawbacks of existing constrained optimization evolutionary algorithms(COEAs),a new algorithm,which effectively combines multi-objective optimization with global and local search,was proposed to handle constrained optimization problems(COPs).In the process of the global search,as a niching genetic algorithm(NGA),crowding operator accepts the winner of the similar parent and offspring as a member of the next population based on Pareto dominance.During the local search,firstly,Pareto strength value is computed for each individual,which depends on the number of individuals in the subpopulation dominated by the current individual,and subsequently the selection of individual is based on the Pareto strength value.The experimental results show that the algorithm proposed is very suitable for functions with high dimension and multimodality.In addition,it is clear that the new algorithm is steadier than other COEAs from the literature in dealing with COPs.