【摘要】 介绍了一种基于递归模糊神经控制器的控制系统,它是通过在模糊神经网络控制器的第2层引入动态递归环节,从而使其具有动态映射能力。文中还提出了递归模糊神经控制器的混合学习算法,即先采用免疫遗传算法的“粗”学习,再采用BP梯度算法的“细”学习。通过对锅炉主汽温控制的仿真表明了该网络结构和训练方法的可行性和有效性。更多还原

【Abstract】 A novel control system based on the recursive fuzzy-neural controller is presented. The system achieves the ability of dynamic mapping by adding dynamic recurrence to the second layer of the fuzzy-neural network. The mixed learning algorithm of the recursive fuzzy-neural controller is also put forward, including the “thick” learning of the immune genetic algorithm first, and then the “thin” learning of the BP algorithm. The network structure and the training method are proved feasible and effective through simulation of main steam temperature control for boilers.更多还原