带拥塞控制的多种群二元蚁群算法  

Multi-population binary ant-colony algorithm with congestion control strategy

作　　者：严彬[1] 熊伟清[1] 程美英[1] 叶青[1] 

机构地区：[1]宁波大学计算机科学与技术研究所,浙江宁波315211

出　　处：《控制理论与应用》2009年第4期387-394,共8页Control Theory & Applications

基　　金：国家自然基金资助项目(60773072);浙江省自然基金资助项目(Y106080);宁波市自然科学基金资助项目(2007A610051)

摘　　要：二元蚁群算法在函数优化中有着良好的表现,但仍存在易陷入局部最优和在多峰函数求解中无法同时得到多个解的缺陷.使用拥塞控制策略改善算法的全局寻优能力,同时引入多种群的思想,提出了带拥塞控制多种群二元蚁群算法.通过对几个不同函数(包括单峰与多峰)的测试,实验结果表明该改进算法在保证较好的全局搜索能力的基础上,拥有很好的多目标求解能力。Binary ant-colony algorithm has good performance in the function optimization problem. However, this algorithm is easy to be trapped into the local optimization region and can not get all the solutions of the multi-modal function. An improved binary ant-colony algorithm based on congestion control strategy and multi-population is proposed to overcome these disadvantages. The congestion control strategy and the multi-populations method are respectively introduced to improve the global optimization ability and its multi-modal optimization ability. Tests with various functions （including single-modal and multi-modal functions） demonstrate that the improved algorithm has good global search ability and ensures a better effect to the multi-objective problems.

关 键 词：二元蚁群 拥塞控制策略 多种群 多峰函数 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]