Group selection by using Lotka-Volterra recurrent neural networks

This paper studies the problem of group selection by using Lotka-Volterra recurrent neural networks. The networks are required to be with self-inhibition and lateral inhibition. The group selection is based on the concepts of permitted and forbidden sets. By restricting the strength of the lateral inhibition and self-inhibition to be in some interval, conditions for group selection are obtained. Under these conditions, groups are related to permitted and forbidden sets, i.e., each group is a permitted set. Thus, groups can be selected by the networks. Simulation results further confirm the theory.