一种基于二次变异策略的改进型遗传算法  

Improved genetic algorithm based on double mutation operators

作　　者：马福祥[1] 马秀娟[1] 

机构地区：[1]青海师范大学计算机学院,西宁810008

出　　处：《计算机工程与应用》2014年第13期62-65,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.61164005);教育部春晖计划项目(No.Z2012101);青海师范大学青年创新项目(No.12948)

摘　　要：通过对基本遗传算法采用单点位变异和倒置变异两次变异操作进行改进,并把该算法应用到TSP问题的求解中。仿真结果表明,改进后的算法提高了种群的多样性,增强了算法的局部搜索能力,从而使最终找到的解比基本遗传算法更优。另外,二次变异的改进遗传算法对种群规模的敏感性比非二次变异的基本遗传算法更强,相同条件下当增大种群规模时,二次变异的改进算法能得到更优的解。Simple genetic algorithm is improved by using single point mutation and inversion mutation operators. The algorithm is applied to Traveling Salesman Problem（TSP）. Simulation results show the diversity of population can be also improved by uising the modified algorithm. The local search capacity of the algorithm is effectively improved. The algo-rithm can find better solution than the simple genetic algorithm. In addition, the improved genetic algorithm has higher sensitivity for population size. Under the same conditions, when the population size increases, the improved algorithm can get a better solution.

关 键 词：遗传算法 二次变异 旅行商问题(TSP) 种群多样性 搜索能力 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]