求解组合优化问题的双群体结构遗传算法

Bi-group Genetic Algorithm For Solving Combinatorial Optimization Problem

提出基于督导群体和进化群体的双群体遗传算法。区别于一般的遗传算法,双群体遗传算法充分利用了督导群体的监督导向作用和问题的先验知识;同时,算法设计考虑加入了邻域函数产生一定数量相异性较大的新个体,从而大大提高了算法的全局搜索性能。以(MR)TSP为例,大量数值实验表明,该算法能迅速收敛到问题的最优解。

A bi-group genetic algorithm based on supervising group and evolving group is designed. Different from simple genetic algorithm,hi-group genetic algorithm makes good use of the supervision and guiding of supervising group and foregone knowledge of problem,and in the process of designing,neighborhood function is fused in to generate a number of dissimilar individuals,so that the global search performance of this algorithm is greatly enhanced.Taking （MR） TSP as examples,the bi-group genetic algorithm could,showed by a great many numerical experiments,convergent rapidly to the optimum of problem.