PDF全文下载:2016050562
李欣欣, 张宏立*
(新疆大学 电气工程学院,新疆 乌鲁木齐 830047)
摘要: 针对传统辨识方法辨识非线性系统时存在的辨识精度低、收敛速度慢等问题,引入了一种基于混合引力搜索算法的非线性系统辨识方法。该混合优化算法是将粒子群算法中群体历史最优位置及自身历史最优位置的概念引入到引力搜索算法中,通过帮助粒子接近最优位置,改进了搜索算法中粒子的全局搜索能力,使得该混合算法的开采能力和探索能力得到更好的增强和平衡。对Wiener模型进行辨识,比较分析仿真结果,发现混合优化算法能够提高辨识精度并获得良好的辨识效果,验证了该算法的有效性和可行性。
关键词: 引力搜索算法; 混合优化算法; 全局搜索能力; 非线性系统
中图分类号: TP 202+.7文献标志码: A
Nonlinear System Identification Based on Gravitational Search Algorithm
LI Xinxin, ZHANG Hongli
(College of Electrical Engineering,Xinjiang University,Urumqi 830047,China)
Abstract: In terms of the low identification accuracy and slow convergence speed of the traditional identification methods when the system is nonlinear, this paper introduces the global best and historical best of the particle swarm optimization into the gravitational search algorithm, which helps to approach the optimal position and improves the global searching ability of particles in GSA. The exploration and exploitation abilities of the hybrid algorithm can be enhanced and wellbalanced. According to parameters identification of the Wiener model, the simulation results show that the GSAPSO algorithm can improve the identification accuracy and get good recognition results. The effectiveness and feasibility of the GSAPSO algorithm are verified.
Key words: gravitational search algorithm; hybrid algorithm; global searching ability; nonlinear system
收稿日期: 20150814
作者简介: 李欣欣(1992—),女,硕士研究生.*通信联系人.
文章编号: 16726987(2016)05056205; DOI: 10.16351/j.16726987.2016.05.017