Mathematical modelling of metric-driven routing and resource allocation in wireless mesh networks

Abstract

The traditional routing protocols used in wireless mesh networks like AODV are not very efficient since the number of hops to a destination is mainly considered as the routing metric. This may lead to shorter paths; however, the throughput can decrease when traffic is routed through those paths. Various contributions have proposed different metrics or completely different routing algorithms. Nevertheless, for many of these metrics, real-time network performance monitoring is required, increasing their overhead, while the implementation of many of the proposed routing protocols is not feasible in practice because of their complexity. In this paper, we aim to optimize link metrics together with routing. We preserve the shortest path routing principle but with optimizing the link metrics; thus, combining the shortest path algorithm's low overhead with an optimized link metrics scheme. We present a two-phase scheme for the considered problem. During the first phase, we seek for the optimal link scheduling for minimizing the required time slots and then, in the second phase, we present a mixed integer programming model for metric-driven routing design. For comparison, we also present the optimization models for global optimized routing and shortest-hop routing, providing numerical results.