多样性保持的和声搜索算法及其TSP求解  

Diversity maintaining harmony search algorithm and its TSP solution

作　　者：黄鉴[1] 彭其渊[1] 

机构地区：[1]西南交通大学交通运输与物流学院,成都610031

出　　处：《计算机应用研究》2013年第12期3583-3585,共3页Application Research of Computers

摘　　要：为了改善和声记忆库群体多样性,提高算法的全局寻优能力,在度量群体多样性指标的基础上,从参数动态调整方法、和声记忆库更新策略两个方面对基本和声搜索算法进行了改进,提出了多样性保持的和声搜索算法,并将该算法应用于TSP的求解。结合TSP问题特点,设计了基于交换和插入算子的和声微调方法。实例优化结果表明,改进后的算法不容易陷入局部最优,优化性能显著提高。In order to improve population diversity of harmony memory and global optimization ability of harmony search algorithm, this paper proposed a diversity maintaining harmony search algorithm, which improved from two aspects based on the index of population diversity. The one was the method of parameters dynamical adjustment, and the other one was updating strategy of harmony memory. Then TSP was solved by the algorithm. Considering the characteristic of the problem, it designed the method of harmony adjusting based on exchange and insertion operator. The example optimization results show that the proposed algorithm can＇ t easily get into local optimal solution and has obvious better performance.

关 键 词：和声搜索 遗传算法 群体多样性 旅行商问题 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]