Differential back‐pressure routing for single‐queue time‐varying wireless networks

Abstract

Dynamic resource control and routing are important for realising the intelligent control of data transmission in wireless multi-hop networks. It is well known that back-pressure routing based on a max-weight policy maximises network throughput and optimises resource allocation in multi-hop wireless networks with time-varying channels. Due to the slow routing convergence and complex control of data queues, however, back-pressure routing also results in large end-to-end delays and a waste of network resources, particularly when the network loads are light or moderate. In this study, a differential back-pressure routing scheme with single-queue management is proposed to improve the packet delay performance and simplify the management of data queues. Unlike in traditional back-pressure routing, the authors use the differences in the rates of change in data queue length to calculate data pressure. Compared with the method of data backlog calculation based on queue length differences, this method achieves faster routing convergence. They also consider the ceiling problem for a single queue and enhance the effect of the queue cap on the routing metric by means of dynamic weighting. Simulation results show that the proposed routing algorithm achieves a 20% decrease in end-to-end delay in grid and random networks.