论文部分内容阅读
针对非线性强时滞系统,传统的预测控制算法难以建立精确模型,其控制精度不高。提出一种基于最小二乘支持向量机(LS-SVM)的非线性模型预测控制算法,该算法通过LS-SVM对非线性系统输入输出数据序列的训练学习,构建其离线的预测模型,然后运用量子粒子群优化(QDPSO)算法来完成整个滚动优化的过程。仿真结果表明基于LS-SVM的非线性模型预测控制比动态矩阵控制具有更好的控制品质。
For nonlinear strong time-delay systems, the traditional predictive control algorithm is difficult to establish an accurate model, and its control accuracy is not high. A non-linear model predictive control algorithm based on least square support vector machine (LS-SVM) is proposed. This algorithm uses LS-SVM to train and study the input and output data series of nonlinear system and builds its offline prediction model. Quantum Particle Swarm Optimization (QDPSO) algorithm to complete the entire rolling optimization process. Simulation results show that nonlinear model predictive control based on LS-SVM has better control quality than dynamic matrix control.