A dynamic programming approach to maximizing a statistical measure of the lifetime of sensor networks

Abstract

The inherent many-to-one flow of traffic in wireless sensor networks (WSNs) produces a skewed distribution of energy consumption rates, leading to the early demise of those sensors that are critical to the ability of surviving nodes to communicate their measurements to the base station. Numerous previous approaches aimed at balancing the consumption of energy in wireless networks are either too complex or do not address problems unique to the flow of traffic in WSNs. In this article, we propose the use of a dynamic programming algorithm (DPA), an operational, low-complexity algorithm, used in conjunction with four different route discovery algorithms. We perform complexity analysis, statistical evaluation of changes in power consumption rates effected, and verify spatial redistribution of energy consumption of sensors in the network. Our results on multihop networks of 100 randomly placed nodes show that, on average, the two best performing variants of DPA yield a reduction of up to 28% and 36% in power consumption rate variance at the cost of raising average power consumption by 15% and 21%, respectively. Computational complexities of DPA variants range from O ( N 3 ) to O ( N 4 ), which is significantly lower than linear search of the solution space of O ( N ! N i ). Analysis by diffusion plots shows that DPA reduces power consumption of sensors that experience the highest power consumption under the shortest path routes.