Abstract
Backpressure based scheduling has revealed remarkable performance in wireless multihop networks as reported in a lot of previous work. However, its lack of consideration on energy use efficiency is still an obstacle for backpressure based algorithms to be deployed in resource-constrained wireless sensor networks (WSNs). In this paper, we focus on studying the design of energy efficient backpressure based algorithm. For this purpose, we propose a gradient-assisted energy-efficient backpressure scheduling algorithm (GRAPE) for WSNs. GRAPE introduces a new link-weight calculation method, based on which gradient information and nodal residual energy are taken into account when making decisions on backpressure based transmission scheduling. According to the decisions made by this new method, packets are encouraged to be forwarded to nodes with more residual energy. We theoretically prove the throughput-optimality of GRAPE. Simulation results demonstrate that GRAPE can achieve significant performance improvements in terms of energy use efficiency, network throughput, and packet delivery ratio as compared with existing work.
Highlights
Backpressure algorithm was proposed by Tassiulas and Ephremides in [1] and it has been proven to be throughput optimal
Backpressure algorithm is purely queue length based and it works in a way such that packet transmission scheduling decisions are made based on queue backlog differentials between neighboring nodes
It is seen that gradient-assisted energy-efficient backpressure scheduling algorithm (GRAPE) outperforms other backpressure based algorithms (i.e., BP and enhanced dynamic backpressure routing algorithm (EDR)) in terms of energy use efficiency, which validates the high energy use efficiency of our new link-weight calculation method used in GRAPE
Summary
Backpressure algorithm was proposed by Tassiulas and Ephremides in [1] and it has been proven to be throughput optimal. Backpressure algorithm is purely queue length based and it works in a way such that packet transmission scheduling decisions are made based on queue backlog differentials between neighboring nodes. International Journal of Distributed Sensor Networks two endpoints In such a way of transmission scheduling, packets are always pushed away from network hot-spots (by the so-called back pressure), no matter whether such transmissions lead to routing detours or even loops. Extensive simulation results demonstrate that GRAPE can yield significant performance improvements in terms of energy use efficiency, network throughput, and packet delivery ratio performance as compared with existing work such as the classical backpressure algorithm [1], enhanced dynamic back-pressure routing algorithm (EDR) [3], and min-hop routing.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have