基于粒子群优化的混合模型预测控制研究  

Hybrid model predictive control of nonlinear system based on PSO

作　　者：高鹏[1] 刘浩然[1] 郝晓辰[1] 郭峰[1] 史鑫[1] 

机构地区：[1]燕山大学电气工程学院,河北秦皇岛066004

出　　处：《机电工程》2011年第2期231-234,共4页Journal of Mechanical & Electrical Engineering

摘　　要：为解决复杂非线性系统的控制精度不高,稳定性难于保证等问题,将预测控制技术应用于非线性控制中,提出了一种基于粒子群优化和混合模型的预测控制算法。混合模型预测控制算法使用模糊聚类和最小二乘法建立了系统的复合模型,由带压缩因子的粒子群算法优化获得了非线性控制系统的控制量,在非线性模型上对模糊预测控制算法和混合模型预测控制算法进行了分析,对2种方法进行了Matlab仿真,比较了两种算法的控制精度和稳定性。试验结果表明混合模型预测控制算法不仅具有很高的控制精度,并且具有很好的鲁棒性。In order to solve the problems of control accuracy is not high and stability is difficult to guarantee in complex nonlinear system,a hybrid model predictive control algorithm based on swarm optimization was introduced to nonlinear system.Fuzzy clustering and least squares method were used to set up the hybrid model,and the control values of nonlinear control systems were computed with the particle swarm optimization algorithm with compress factor.Fuzzy predictive control algorithm and hybrid model predictive control algorithm were separately evaluated on the nonlinear control systems.The Matlab simulations to the two algorithms were carried on to compare control accuracy and stability.The experimental results show that the hybrid model predictive control algorithm not only has high control accuracy,but also has robust.

关 键 词：非线性系统 混合模型预测控制 模糊神经网络 最小二乘法 粒子群 

分 类 号：TP271.72[自动化与计算机技术—检测技术与自动化装置]