An estimate of impulse bounds in delayed BAM neural networks

This paper further studies the exponential stability of delayed bidirectional associative memory neural networks and focuses on the impulse effect on the exponential stability property. It is shown that if the corresponding impulse-free DBAM is globally exponentially stable the impulsive analog will remain its stability property even if the measurements of the states are magnified to some extent at the impulse instants. Furthermore, the admissible upper bound of impulse is estimated in terms of exponential convergence degree of the corresponding impulse-free DBAM and the length of impulse interval.