基于粒子群的多BP网络并行非线性预测控制

针对传统预测控制算法在解决非线性系统控制问题时，存在难以建立精确的数学模型、控制精度不高等缺点，提出一种新的非线性系统预测控制方案。以多BP神经网络作为并行预测模型，克服误差积累以及网络规模庞大的缺点；运用粒子群优化(PSO)算法完成非线性预测控制的滚动优化。仿真表明，该方案的控制效果比常规动态矩阵控制效果有所提高，该方案是可行和有效的。

Multi-BP Neural Network Parallel Nonlinear Predictive Control Based on Particle Swarm Optimization

It is difficult to build an accurate model using the traditional predictive control algorithm for the nonlinear system and the control accuracy is lower. The paper proposes a new nonlinear system predictive control scheme. A multi-BP neural network predictive model is obtained, and it can cover the shortages of error cumulation and large network scale. The PSO is used for the rolling optimization procedure of the system. The simulation results illustrate that the scheme obtains better control effects than normal dynamic matrix control algorithm and it is feasible and effective.