基于遗传算法的最小生成树的参数优化研究

应用遗传算法求最小生成树问题是一种可行而有效的新方法 .为了提高算法的进化效率和收敛性 ,通过数值模拟试验 ,对算法中主要控制参数的优化组合模式进行研究 ,获得了主要参数的优化组合规律 :群体规模取 10～ 30 ;换位率在 0 .5～ 0 .9之间选取 ;选择率可采用自适应方式确定 ,进化初期采用较大值 ,随着进化过程的前进可逐渐减小 ;编码长度不应过长等

Study on the Parameters Optimization of Minimum Spanning Tree Algorithms Based on Genetic Algorithms

It's a new feasible and efficient method to solve the minimum spanning tree problems by genetic algorithms. In order to improve the evolutionary efficiency and convergence of the algorithms,this paper makes a thorough research to the combination model of the main controlling parameters of the algorithms by numerical simuliation tests.According to the results of the research, the rules of optimal parameter combinations are acquired.The population size can be chosen between 10 and 30. the transpose probablity ( P t ) can be chosen between 0.5 and 0.9;the select probability( P s )can be choosed by self_adapting method. P s can be larger at the beginning of the process of evolutionary,and then becomes smaller as the evolutionary process goes ahead,and the coding length shouldn't be too long.Those rules provide a valuable reference for the application of the algorithms.