A Cluster-merge Algorithm for Solving the Minimum Power Broadcast Problem in Large Scale Wireless Networks

In this paper, we address the minimum power broadcast problem in wireless networks. Assuming nodes are equipped with omni-directional antennas, the inherently broadcast nature of wireless networks can be exploited to compute power efficient routing trees. We propose a 2-stage cluster-merge algorithm for computing minimum power broadcast trees. The cluster phase is a look-ahead variant of the Broadcast Incremental Power algorithm [1] and the merge phase is a probabilistic positive reinforcement search procedure, as used in swarm intelligence algorithms. A local treeimprovement procedure is incorporated as an optional step in the merge phase to boost the performance of the algorithm. A key advantage of such a cluster based approach is significant reduction in time complexity. Simulations show that the algorithm is able to generate high quality solutions in relatively little computational time.