全局较优解引导的人工蜂群算法  

Artificial Bee Colony Algorithm Guided by Global Better Solutions

作　　者：王冰[1,2] 

机构地区：[1]牡丹江师范学院理学院,黑龙江牡丹江157011 [2]北京理工大学数学与统计学院,北京100081

出　　处：《数学的实践与认识》2015年第13期160-172,共13页Mathematics in Practice and Theory

基　　金：黑龙江省教育厅科学技术研究项目(12541840)

摘　　要：人工蜂群算法(ABC)是一种模拟蜜蜂群体寻找优良蜜源的群体智能优化算法.针对人工蜂群算法收敛速度较慢、探索能力较强而开发能力偏弱等问题,提出一种改进的蜂群算法.算法利用更多的较优蜜源位置的信息来引导采蜜蜂和跟随蜂的搜索行为.为了提高算法的全局收敛速度,通过基于混沌策略的方式生成初始化种群,并且在每一代侦察蜂阶段后对全部新蜜源进行反向再搜索.另外,每次蜜蜂邻域搜索之后,采用比较新旧蜜源的花蜜值(而非适应度值)的方法来更新蜜源位置.通过对14个标准测试函数进行仿真实验,结果表明所提出的算法能有效加快收敛速度,提高开发能力和解的精度.Artificial bee colony （ABC） algorithm is an intelligent optimization algorithm which simulates the foraging behaviour of honey bee colony. For the shortcomings of artificial bee colony algorithm, such as good at exploration but poor at exploitation and poor conver- gence rate, an improved algorithm is proposed. The solution search equations take advantages of the information of some of the best solutions to guide the search behaviour of employed bees and onlookers. In order to accelerate convergence speed, an initialization strategy based on chaos is applied instead of a pure random initialization~ and after the scout phase a novel opposition-based searching method is employed to generate new alternative solutions. In addi- tion, a more robust calculation to compare and determine the quality of alternative solutions is used. Experimental results on 14 benchmark functions show that proposed algorithms can accelerate effectively convergence speed, its local optimization ability and accuracy of the solution are significantly improved.

关 键 词：人工蜂群算法 种群初始化 搜索方程 一般的反向学习 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]