一种人工蜂群算法改进方案  

Improved artificial bee colony algorithm

作　　者：梁静[1] 葛宇[2] 冉晓娟[3] 李琦[1] 

机构地区：[1]成都工业学院网络中心,成都610031 [2]四川师范大学基础教学学院,成都610068 [3]四川旅游学院信息技术系,成都610000

出　　处：《计算机应用研究》2015年第11期3295-3299,共5页Application Research of Computers

基　　金：四川省教育厅资助项目(12ZB112)

摘　　要：为提高人工蜂群算法的寻优效率,对算法中跟随蜂和侦察蜂的搜索策略进行了改进,提出了一种反映个体拥挤程度的crowd指标,并基于该指标为跟随蜂设计了针对优秀个体的自适应邻域搜索策略;针对侦察蜂的搜索行为,设计了一种较差个体重置机制,以保持种群多样性,防止算法可能出现的早熟收敛。通过八个典型测试函数的仿真结果表明,相比原算法以及实验中列出的同类算法,改进算法在收敛速度和寻优精度上取得了较好的效果,说明了改进方案的有效性。In order to improve the efficiency of artificial bee colony algorithm, this paper improved search strategy of onlooker bees and scout bees. Specifically, the improved algorithm presented crowd indicators and designed an adaptive neighborhood search strategy for onlooker bees based on crowd indicators. Meanwhile, in order to keep the diversity of population and avoid premature convergence, this paper designed poor individual reset mechanism for scout bees. The simulation results of eight typical functions show that the improved algorithm has achieved good results in convergence speed and accuracy, compared with the basic artificial bee colony algorithm and other similar algorithms.

关 键 词：人工蜂群算法 跟随蜂 侦察蜂 邻域搜索 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]