基于菌群优化算法的非线性系统预测控制器设计  

Predictive controller design for nonlinear system based on BSFO algorithm

作　　者：徐志成[1] 陶国正[1] 王树青[2] XU Zhicheng;TAO Guozheng;WANG Shuqing(College of Electrical Engineering,Changzhou Institute of Mechatronic Technology,Changzhou 213164,China;Institute of Advanced Process Control,State Key Laboratory of Industrial Control Technology,Zhejiang University,Hangzhou 310027,China)

机构地区：[1]常州机电职业技术学院电气工程学院,常州213164 [2]浙江大学工业控制技术国家重点实验室先进控制研究所,杭州310027

出　　处：《江苏科技大学学报（自然科学版）》2018年第3期393-398,共6页Journal of Jiangsu University of Science and Technology:Natural Science Edition

基　　金：国家自然科学基金资助项目(70471052);科技部科技型中小企业技术创新基金(13C26243202171);江苏高校品牌专业建设工程资助项目(PPZY2015C238)

摘　　要：作为一种先进的计算机控制算法,预测控制具有良好的自适应性和鲁棒性,一直是自动控制领域的研究热点.针对非线性系统,文中提出了一种基于菌群优化算法和模型学习的单步预测控制算法,算法在不需要系统数学模型前提下,根据非线性系统输入输出数据,通过最小二乘支持向量机建立系统数学模型,获得系统输出估计值,利用反馈校正减少预测误差,最后利用菌群优化算法进行滚动优化得到控制量,得到有效的预测控制器.由单变量和多变量非线性系统的仿真结果证明了算法鲁棒性和自适应性较好.As a kind of advanced computer control algorithm,predictive control has been a research hotspot in the field of automatic control because of its good adaptiveness and robustness. For the nonlinear system,a predictive control algorithm based on BSFO（ bacterial swarm foraging for optimization） algorithm and model learning is proposed in this paper. According to the system input and output data,the algorithm establishes mathematical model by the LS-SVM（ least squares support vector machine）,forecasts the output value,and reduces the error through the feedback and error correction with unknown mathematical models. The control values are obtained by the rolling optimization of BSFO algorithm,and effective controllers are obtained. The simulation results for univariate and multivariate nonlinear systems show that the proposed algorithm has an excellent adaptive ability and robustness.

关 键 词：预测控制 菌群优化算法 支持向量机 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]