Dynamic characteristic analysis of BSB neural models with time-delay

In this paper, dynamic characteristic analysis of BSB neural models is carried out by introducing a time-delay term in this model. The dynamic performances of this kind of BSB with symmetric weight matrices are analyzed step by step. It has been proved in the paper that under certain assumptions, its generic trajectory, starting in an open hypercube, hit an edge and stays on that edge until it reaches an extreme point. Further more, an example is used to show that the BSB models with and without time delay have different attraction domain of stored patterns.