论文部分内容阅读
针对训练模糊神经网络时收敛时间慢，难以实时实现的缺点，将Ｈｏｐｆｉｅｌｄ网络引入模糊神经网络权值的优化问题中，从而融合了Ｈｏｐｆｉｅｌｄ网络可进行实时优化和模糊神经网络可引入专家知识化权值的优点。理论分析和模拟实验表明，这种网络可以在电路时间常数数量级内给出优化后的模糊神经网络权值，并且具有Ｌｙａｐｕｎｏｖ意义下的稳定性，为模糊神经网络权值的实时优化提供了一条新途径。
To overcome the shortcoming of slow convergence time and hard real-time training in the training of fuzzy neural network, the Hopfield neural network is introduced into the optimization problem of fuzzy neural network weights, so that Hopfield neural network can be real-time optimized and fuzzy neural network can be used to introduce expert knowledge weight The advantages. Theoretical analysis and simulation experiments show that this kind of network can give the optimal weight of the fuzzy neural network within the order of the time constant of the circuit, and it has the stability in the sense of Lyapunov, providing a new real-time optimization for the weight of the fuzzy neural network way.