求解分数阶混沌系统参数估计问题的和声引力搜索算法  

Harmony Gravitational Search Algorithm for Solving the Problem of Parameter Estimation for Fractional-order Chaotic Systems

作　　者：黄宇[1] 王佳荣[1] 梁伟平[1] 

机构地区：[1]华北电力大学河北省发电过程仿真与优化控制工程技术研究中心,保定071003

出　　处：《系统仿真学报》2016年第5期1045-1053,共9页Journal of System Simulation

基　　金：中央高校基本科研业务费专项资金(2015MS66)

摘　　要：针对分数阶混沌系统的参数估计问题,提出了一种和声引力搜索算法。通过和声搜索算法中的记忆选择与概率搜索对引力搜索算法更新后的粒子进行再创作与调整。为解决粒子接近最优解时出现速度变化过快的问题,提出将和声搜索算法中的局部调整概率值设置为随算法迭代次数增加而非线性减小。采用标准测试函数对该算法进行了测试,测试结果表明本文算法具有良好的稳定性和全局搜索能力。将该算法应用于分数阶R ssler混沌系统与分数阶Lorenz混沌系统的参数估计,计算结果表明了本算法的有效性和鲁棒性。Aiming at parameter estimation for fractional-order chaotic systems, a harmony gravitational search algorithm was proposed. The particles, which were updated by gravitational search algorithm, were recreated and searched through memory selections and probability search in harmony search algorithm. At the same times, in order to solve the problem that the speed changed too fast when the particles closed to the optimal solution, the partial adjustment probability was set to non-linearly decrease with the adsorption capacity increasing. Standard test functions were used to test the algorithm, and the test results show that the algorithm has good stability and global search capability. At last, the harmony gravitational search algorithm is applied in the parameter estimation for fractional-order R ssler system and fractional-order Lorenz system, and the estimation results show that the algorithm is effective and robust.

关 键 词：分数阶混沌系统 参数估计 引力搜索算法 和声搜索 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]