一种约束输入输出的隐式广义预测控制新算法

基于CARIMA模型提出了一种约束输入输出的隐式广义预测控制算法.针对广义预测控制问题,在整个预测时域和控制时域,对输入幅值,输入增量和输出幅值施加了约束,引入了输入输出柔化系数,从而简化了目标函数,减小了计算量,该算法不必求解逆矩阵;并采用了隐式广义预测自校正控制算法,利用并列预测控制器间的特点,直接辨识输出预测器中的参数,从而避免了在线求解Diophantine方程.该算法占用内存小,计算速度快,仿真结果表明该算法具有良好的控制性能.

A New Algorithm of Implicit Generalized Predictive Control with Input/Output Constraints

In this paper, a new implicit generalized predictive control algorithm with input/output constraints based on CARIMA model is presented. Input, input increment and output are constrained by focusing on the generalized predictive control in whole prediction horizon and control horizon. The input/output flexible coefficients are used, so the cost function is simplified, calculation is reduced and the inverse matrix does not require calculating. It avoids solving Diophantine equation on - line by adopting implicit generalized predictive self - tuning control algorithm. The parameters of the output predictive equation are directly identified by the characteristic of the parallel predictors. The method needs less computer memory and the computational speed is higher. Simulation results show superior performance of this algorithm.