A heuristic for fast convergence in interference-free channel assignment using D1EC coloring

Abstract

This work proposes an efficient method for solving the Distance-1 Edge Coloring problem (D1EC) for the assignment of orthogonal channels in wireless networks with changing topology. The coloring algorithm is performed by means of the simulated annealing method, a generalization of Monte Carlo methods for solving combinatorial problems. We show that the simulated annealing-based coloring converges fast to a suboptimal coloring scheme. Furthermore, a stateful implementation of the D1EC scheme is proposed, in which network coloring is executed upon topology changes. The stateful D1EC reduces the algorithm's convergence time by one order of magnitude in comparison to stateless algorithms. © 2011 Springer Science+Business Media B.V.