Optimal aggregated ConvergeCast scheduling with an SINR interference model

Abstract

We consider the scheduling problem for Aggregated ConvergeCast in wireless sensor networks with a physical interference model. Previous work consists of either heuristics without performance guarantees, or approximation algorithms which do not perform well in practice. We propose here a first scalable mathematical SINR (Signal to Interference plus Noise Ratio) model that outputs an optimal Aggregated ConvergeCast schedule. We use large scale optimization techniques, namely a Dantzig-Wolfe decomposition algorithm, to solve it. We perform extensive simulations on networks with upto 70 sensors, and compare our results with the best heuristic in the literature using a SINR model. Results show that schedules output by our new model are significantly better than those output by the best available heuristic, i.e., with TDMA frames that are about 50% shorter.