EXPONENTIAL STABILITY OF HOPFIELD-TYPE DELAY IMPULSIVE DISCRETE NEURAL NETWORKS AND COMPUTER SIMULATION

Abstract

A discrete Hopfield-type neural network with constant delays, instan-taneous switching topologies at certain times and time variable connection weights isstudied. Some criteria for exponential stability are derived. The obtained results areillustrated on an example with different activation functions such as tanh, Swish, andthe error function. The example is computer realized by coding the correspondingalgorithms for calculating the values of the solution for each step.