摘要
为了有效地解决人工蜂群算法容易陷入局部最优的缺陷,提出了一种改进蜂群算法。利用反向学习方法构建初始种群,以提高初始化解的质量。同时,利用分布估计算法构造优秀个体解空间的概率模型来进行邻域搜索,以改善算法的搜索性能并防止陷入局部最优。对连续空间优化问题进行了仿真实验,结果表明改进算法具有较快的收敛速度,全局寻优能力显著提高。
This paper proposed a modified artificial bee colony algorithm to tackle the dilemma of easily trapping in local optimum in original artificial bee colony.Firstly,it applied an opposition-based learning method for the initial population generation,which aimed to improve the quality of initial solutions.Meanwhile,it employed the estimation of distribution metaheuristic to established the probability model of solution domain about good individuals for neighbor search.This operation was capable of improving the searching performance and avoiding local optimum.Finally,it conducted the simulations to tackle the continuous space optimization problems.The experimental results demonstrate that the modified algorithm has fast constringency speed and the global optimization capability is enhanced.
作者
王永琦
吴飞
孙建华
Wang Yongqi;Wu Fei;Sun Jianhua(School of Electronic&Electrical Engineering Science,Shanghai University of Engineering,Shanghai 201620,China;College of Information Science&Engineering,Hunan University,Changsha 410082,China)
出处
《计算机应用研究》
CSCD
北大核心
2018年第3期658-660,704,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(F020207)
上海市科委资助项目(13510501400)
上海市工程技术大学<信号与系统>平台课程建设项目(k201602004)
关键词
人工蜂群算法
连续空间优化
反向学习
分布估计算法
artificial bee colony(ABC)algorithm
continuous space optimization
opposition-based learning
estimation of distribution algorithm(EDA)
