自适应神经元非模型多变量优化补偿控制

Adaptive Neural Non-Model Optimal Compensating Control for MIMO System

根据解耦补偿和优化控制的思想，本文提出了一种完全不依赖于对象模型的自适应神经元多变量优化补偿器模型，给出了神经无权系数的在线学习方法，分析了其工作机理．进一步给出了在某多侧线精馏塔和连续搅拌釜式反应器（CSTR）上的仿真结果．

Based on the principle of the decoupling compensating and optimal control,an adaptive neural optimal compensator independent on object model and the on-line learning method of neuron weights are presented. The principle of the compensator is analysed. Further,the simulation results in a distillation column with multi-outputs and in a CSTR are shown.