论文部分内容阅读
研究了基于人工神经元网络模型的非线性预测控制。所采用的网络为一种将线性模型与多层前向网络相结合的ＤＬＦ网络。仿真结果表明，该“混合网络”易训练，收敛速度可大大加快。在ＤＬＦ模型的基础上，本文研究了一种非线性预测控制算法，它的显著特点是在线计算量小。对于一非线性过程──球形罐液位的仿真结果表明，基于ＤＬＦ的非线性预测控制效果颇佳。
The nonlinear predictive control based on artificial neural network model is studied. The network used is a DLF network that combines a linear model with a multi-layer forwarding network. Simulation results show that the “hybrid network” is easy to train and convergence speed can be greatly accelerated. On the basis of DLF model, this paper studies a nonlinear predictive control algorithm, which is characterized by a small amount of online computation. For a non-linear process ─ ─ spherical tank level simulation results show that non-linear predictive control based on DLF is quite good.