A new adaptive genetic algorithm for fixed channel assignment

This paper presents a new genetic algorithm (GA) with good convergence properties and a remarkable low computational load. Such features are achieved by on-line tuning up the probabilities of mutation and crossover on the basis of the analysis of the individuals' fitness entropy. This way, a brand new method to control and adjust the population diversity is obtained. The resulting GA attains quality solutions, thus offering an interesting alternative to other global search techniques, such as simulated annealing, Tabu search and neural networks, as well as to standard GAs. The new algorithm is applied to solve the problem of frequency reuse in mobile cellular communication systems, where the main aim is to obtain a conflict-free channel assignment among the cells such that the resulting bandwidth is close to the minimum channel span required for the whole network. The algorithm performance has been tested and compared by making use of a selection of the most well-known benchmark instances; optimal bandwidth solutions have been achieved within a reasonable computation time. (C) 2007 Elsevier Inc. All rights reserved.