结合解耦思想研究多变量系统预测函数控制。通过引入神经网络补偿环节 ,对多变量系统解耦 ,在此基础上 ,对解耦后各子系统进行单变量预测函数控制 ,以确定各个控制量。同时基于系统脉冲传递函数 ,得到单变量PFC控制器的显式表达式。仿真表明 ,该算法有较好的跟踪特性 ,对解决多变量系统的优化和控制具有一定的适用性 Study of Multivariable Predictive Functional Control Based on Decoupling. By introducing the compensation of neural network, the multivariable system is decoupled. On this basis, the univariate predictive function control of each subsystem is decoupled to determine the control variables. At the same time, the explicit expression of single variable PFC controller is obtained based on the system pulse transfer function. The simulation results show that the proposed algorithm has better tracking characteristics and has some applicability to the optimization and control of multivariable systems