Abstract

Assignment problems in radio networks, like the channel assignment, may be solved by graph coloring algorithms. In this paper we compare the performance of several recent techniques on a simple coloring problem: the search for a bipartite subgraph with the maximum number of edges of a given graph. The algorithms considered are a neural network, a genetic algorithm, simulated annealing and two heuristics. The results show that one heuristic, a new proposed multiagent system, and the standard simulated annealing technique are faster outperforming the algorithms in the literature. We also show how the algorithms may be easily adapted to deal with the k- coloring of a graph and can therefore be used for solving assignment problems in telecommunications. c #1995 by Lawrence Erlbaum Assoc. Inc. Pub., Hillsdale, NJ 07642 Applications of Neural Networks to Telecommunications 2 pp. 49--56 (1995). ISBN 0-8058-2084-1 1 Introduction Several issues related to the design of radio networks, as...

Citations