Optimum ConvergeCast Scheduling in Wireless Sensor Networks

Abstract

Target monitoring is an important ConvergeCast application of wireless sensor networks in which sensors monitor a set of targets, and forward the collected data using multi-hop routing to the same location, called the sink. Nearly all previously proposed models only output a set of link transmission configurations, i.e., sets of links that can simultaneously transmit, without providing an ordering of the transmission configurations, nor guaranteeing that such an ordering exists using only the prescribed number of slots. As such, they do not provide a valid schedule to achieve ConvergeCast, and only give a lower bound on the number of slots required for a schedule. In this paper, we propose a first one phase decomposition model and algorithm that outputs a complete and optimum scheduling, i.e., with the output consisting in an ordered sequence of transmission configurations that achieves ConvergeCast. In addition, we show that the resulting transmission graph is not necessarily a tree. The resulting algorithm provides much better schedules, up to 15% less time slots required, than those of the previous best available mathematical programming or heuristic approaches in the literature.