An enhanced genetic algorithm approach to the channel assignment problem in mobile cellular networks

Abstract

In mobile cellular networks, frequency channels must be assigned to call requests so that the interference constraints are respected and bandwidth minimized. The problem is known as the Channel Assignment Problem (CAP) and is NP-complete[1]. In this paper we present an Enhanced Genetic Algorithm (EGA) for solving the CAP. The EGA discussed here differs from other genetic algorithm approaches [2,3] in many ways including the use of a partial elitist strategy. Perhaps, the most unique feature of the EGA is its ability to run unsupervised. Users do not have to specify a priori appropriate operational parameters like mutation and crossover rates. These parameters are determined and adjusted automatically during the course of the genetic search. This allows the same EGA to be used for all problem instances. Although the EGA is not guaranteed to find optimal results, we show that it is able to find high-quality solutions in reasonable amounts of time. In addition, we examine how components of the EGA contribute to its effectiveness.