Near optimal multicoloring for inter-cluster interference avoidance

In this paper, we consider conflict-free resource assignment in clustered wireless mesh networks. Transmissions within a cluster of neighboring nodes can be scheduled by the cluster-head, a leader in the cluster. Avoiding interferences between neighboring clusters is usually addressed as a coloring problem by the literature. We first propose a new formulation of this problem, and express it as a multicoloring problem, each color representing a slice of conflict-free resources (e.g. a set of TDMA frames). We further propose CCSA (Clustering Configuration for slots assignment), a novel distributed algorithm that implements a heuristics for the multicoloring problem. We provide simulation results on networks of 100 nodes that compare CCSA to super-frame, the closest state-of-the art solution and to an upper bound provided by a linear program solving the multicoloring problem. Simulations results show that CCSA provides double capacity over super-frame and reaches 80% of the upper bound.