A study of asymptotic stability for delayed recurrent neural networks

This paper addresses the problem of asymptotic stability for discrete-time recurrent neural networks with time-varying delay. The analysis starts with a general assumption that the time-varying delay may be expressed as the lower bound plus the length of an interval over which the delay varies. Then the delay partitioning technique is used to establish a new delay-dependent sufficient condition under which the asymptotic stability of recurrent neural networks with time-varying delay can be guaranteed. The new stability criterion takes the form of linear matrix inequalities, thus lending itself to being readily checkable by the available software package. The obtained theoretical result is further illustrated by numerical results, including their superiority over the existing results on asymptotic stability of delayed recurrent neural networks.