Study of genetic search for the Frequency Assignment Problem

The goal of this paper is twofold. First, we present an evolutionary approach to a real world application: the Frequency Assignment Problem (FAP) in Cellular Radio Networks. Second, we present an empirical study on the effectiveness of crossover for solving this problem. Experiments carried out on a set of real-size FAP instances (up to 300 cells, 30 frequencies and 30,000 interference constraints) show the interest of EAs. At the same time, empirical evidence suggests that the contribution of crossover is marginal for this application.