基于SAPSO算法的非线性预测控制方法研究

摘要

Neural network model is widely used in nonlinear systems control predictive, but there is a problem that predictive control law is difficult to strike. In this paper SAPSO algorithm is proposed to optimization solution. Based on analysis of PSO algorithm and SAPSO algorithm, neural network predictive control strategy is optimized by using SAPSO optimization algorithms. Through simulation experiments, prediction performance of PSO algorithm and SAPSO algorithm have been compared. The simulation results show that SAPSO optimization algorithm can effectively reduce the number of iterations and improve the accuracy of convergence.%神经网络模型在非线性系统预测控制中得到广泛地应用,但是存在预测控制律难以求取的问题,文章提出模拟退火粒子群优化(SAPSO)算法来进行优化求解.在对PSO算法与SAPSO算法进行分析的基础上,采用SAPSO优化算法对神经网络预测控制策略进行了优化,再通过仿真实验对PSO算法与SAPSO算法的预测性能进行了比较.仿真结果表明SAPSO优化算法能有效减少迭代次数、提高收敛精度.