Genetic Algorithmic Approach to Mitigate Starvation in Wireless Mesh Networks

Wireless mesh networks (WMNs) have a set of self-organized and dynamically self-configurable nodes and each node may act as a router and a host. The performance of the applications running on the nodes that are more than one-hop away from the gateway suffers by starvation. Starvation occurs in a situation that nodes close to the gateway capture the resources (channel) rather than giving the chance to the nodes situated in longer distance to the gateway. In this paper, to avoid starvation, a cross-layer technique is employed to identify optimal contention window (CW) for individual nodes. The size of CW is based on the network layer quality of service (QoS) parameters and the available channel. In this work, the scheduling method uses a genetic algorithm (GA) to mitigate the starvation by selecting optimal CW. Simulations are conducted using multimedia traffic and results are compared with the priority-based scheduling method. Performance of the proposed algorithm is evaluated in terms the parameters such as throughput, packet delivery ratio, end-to-end delay and number of cache replies used. The comparison shows that GA optimization to mitigate the starvation works better than the priority-based starvation avoidance method.