Probabilistic routing for multiple flows in wireless multi-hop networks

Maximizing network throughput is one of the main objectives in wireless multi-hop networks. However, small link bandwidth and severe interference become two great obstacles in network throughput improvement. Single route always encounters great congestion, especially for multiple flows with the same source and destination. Multi-path routing can solve the bandwidth shortage issue to some extent, but conventional routing algorithms of this type have some limitations during route selection and using, which makes bandwidth improvement and interference reduction unfeasible at the same time. In this paper we propose a new metric called effective bandwidth to describe the real bandwidth that a flow can get through a specific route. Based on this metric, we present a probabilistic multi-path routing protocol for multiple flows, which can improve the network throughput by selecting route with bigger effective bandwidth using higher probability. The simulation results show that probabilistic multi-path routing has high superiority over single-path routing in improving network throughput.