Max-min fuzzy Hopfield neural networks and an efficient learning algorithm

摘要

We set up a dynamical fuzzy neural network system, i.e. the so-called max-min fuzzy Hopfield network in the paper, and prove the Lyapunov stability of the equilibrium points (attractor) of the system. Also, we discuss the uniform stability of the system and show some sufficient conditions, with which the given fuzzy pattern is the attractor of the system. Moreover, we obtain a nontrivially attractive basin of the attractor. Therefore, our models have good fault-tolerance. With an analytic method, we design an efficient learning algorithm for connected weights of the networks. Finally, simulation examples demonstrate our conclusions.