具有混沌搜索策略的蝙蝠优化算法及性能仿真  

Bat Algorithm with Chaotic Search Strategy and Analysis of Its Property

作　　者：刘长平[1,2] 叶春明[1] 

机构地区：[1]上海理工大学管理学院,上海200093 [2]淮阴工学院经济管理学院,淮安223001

出　　处：《系统仿真学报》2013年第6期1183-1188,1195,共7页Journal of System Simulation

基　　金：国家自然科学基金(71271138);教育部人文社会科学规划基金项目(10YJA630187);教育部高校博士学科点专项科研基金项目(20093120110008);上海市教委科研创新重点项目(12ZS133);上海市重点学科建设资助项目(S30504)

摘　　要：蝙蝠算法是受自然界中的蝙蝠通过回声定位进行搜寻、捕食行为的启发演变而来的一种新颖的仿生群智能优化算法。在分析基本算法仿生原理和局限性的基础上,提出一种改进蝙蝠局部搜索能力的优化算法,通过逻辑自映射函数产生混沌序列,引入到蝙蝠算法中对精英个体进行混沌优化,同时动态收缩搜索空间以加快收敛速度。改进算法有效结合了基本蝙蝠算法的全局优化能力和混沌算法的局部搜索能力,对经典函数的仿真测试表明,改进算法显著提高了优化性能,在寻优精度和全局收敛能力方面优于基本蝙蝠算法,是解决工程应用中复杂函数优化问题的一种有效方法。Inspired by the echolocation behavior of bats,bat algorithm（BA） is developed as a novel bionic swarm intelligence optimization method.An improved bat algorithm was proposed to enhance the local searching ability on the basis of analyzing bionic principle and limitations of basic BA.A series of chaotic variables according to the self-logical mapping function were introduced into the course of BA to optimize the elite of artificial bats,and shrink the search field dynamically.The improved algorithm makes use of the chaotic search to improve the capability of precise search and also keeps the ability of global search of basic BA.Simulation results for benchmark functions show that the proposed algorithm has improved the global optimizing ability remarkably,and has great advantage of accuracy and convergence property compared to BA and PSO.

关 键 词：蝙蝠算法 仿生原理 逻辑自映射 混沌搜索 仿真测试 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]