Asymptotic stability analysis of continuous bidirectional associative memory networks

The authors discuss the local asymptotical stability problem of continuous bidirectional associative memory (BAM) and present some sufficient conditions which appear as a set of piecewise-linear inequality relations. Some sufficient conditions for a desired equilibrium to be asymptotically stable are examined. For M library vector pairs to be memorized, the conditions appear as M(N+P) linear equations and M(N+P) piecewise-linear inequalities in N+P+2NP unknowns of entries of a, b, W and V. These results have practical significance in the hardware implementation of the BAM.<
>