蚁群算法滚动优化的LS-SVM预测控制研究  

Study of LS-SVM predictive control using ant colony algorithm rolling optimization

作　　者：王娟[1] 刘明治[1] 

机构地区：[1]西安电子科技大学机电工程学院,西安710071

出　　处：《控制与决策》2009年第7期1087-1091,共5页Control and Decision

基　　金：国家自然科学基金项目(59775052)

摘　　要：针对非线性过程预测控制的模型预测和滚动优化问题,提出一种蚁群算法滚动优化的最小二乘支持向量机(LS-SVM)新型预测控制器.该控制器以建模简单、精度高的LS-SVM作为预测模型,蚁群算法作为滚动优化策略,避免了滚动优化中复杂的梯度计算.仿真研究表明,该控制器具有良好的非线性控制效果.Aiming at the problems of modeling and optimization in predictive control systems of nonlinear system, a new-style of nonlinear predictive controller is present. This controller utilizes least squares support vector machines as predictive model which has the characteristics of modeling easy and high precision. Ant colony algorithm is used as rolling optimization, so that the complex calculating of the gradient is avoid. The simulation studies show that this controller has favorable effect of nonlinear control.

关 键 词：预测控制 最小二乘支持向量机 蚁群算法 非线性系统 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]