An Evolutionary Model for the Study of Wireless Networks

In this paper, we explore the applicability of a well known biological modeling technique to the problem of studying dynamic resource allocation in wireless networks. We utilize a variant of Kauffman´s NK model for epistatic evolution, where epistasis is taken to be a measure of the degree to which network links share a common resource, for example time or bandwidth. In our model, alleles correspond to bandwidth assignments, and a given genotype can be seen as are source allocation solution applied across the entire network. A genetic algorithm is employed to evaluate the fitness of possible resource allocation solutions. We believe that this work represents an important step in the direction of understanding network dynamics and wireless resource allocation strategies for autonomic networks.