基于概率约束的DMC预测控制算法研究

Study on Predictive Control Algorithm of Dynamic Matrix Control Based on Probabilistic Constrains

动态矩阵控制(DMC)是模型预测控制(Model Predictive Control)的典型算法,能有效处理系统输入输出约束和抑制随机干扰。对于存在约束的线性稳定系统,本文将条件均值和状态估计误差协方差引入系统的约束表达式中,通过高斯近似把概率约束指标转换为确定性线性约束,建立科学的有约束 DMC 预测控制系统,使被控对象能够更好的满足约束。这种方法简单,易于实现,Matlab 进行仿真结果证明其可行性与优越性。

Dynamic matrix control （DMC）, an important control algorithm of model predictive control, is of interest because it is an effective method to restrain stochastic disturbances and handle constrains. In this paper, using a Gaussian approximate for a linear system with probabilistic constraints, the original problem is approximated by deterministic-constrained DMC problem. The efficiency of this approach is demonstrated by simulation results using Matlab.