The Stability Analysis Of Neural Networks With Impulses And Time-delays

Pulse and Delay is not an uncommon phenomenon that exists in various fields of mod-ern science and technology. Studies on neural networks with pulse and delay is one of animportant research field of neural networks. It is also the hot topic discussed in this field.In particular, pulse and delay are two important aspects for considering the states of neuralnetworks system. They can indicate the changeability of system more accuracy and system-atically.In this paper, firstly, we mainly research on two types of network systems: the Bidi-rectional Associative Memories (BAM) neural networks with time delay and random pulsesand Hopfield neural networks with time delay and variable pulse. Secondly, pulse behavioroccurs regularly in the neural network system of Central Africa, the existence and stabilityconditions of equilibrium of the system. Lastly, we discuss the stability of system by con-structing a suitable Lyapunov function, using the inequality technique, reductio ad absurdum,mathematical induction, and obtain the new results.In practical system, Random pulse arrivals usually follow the Poisson process. Wediscuss the equilibrium point and the stability of BAM neural networks system as the arrivalinterval is exponential distribution by numerical simulation, and also as the strength of pulsearrival is variable.