Minimum Latency Aggregation Scheduling for Arbitrary Tree Topologies under the SINR Model

摘要

Almost all the existing wireless data aggregation approaches need a topology construction step before scheduling. These solutions assume the availability of flexible topology controls. However, in real scenarios, lots of factors (impenetrable obstacles, barriers, etc.) limit the topology construction for wireless networks. In this paper we study a new problem called Minimum-Latency Aggregation Scheduling for Arbitrary Tree Topologies (MLAT). We first provide an NP-hardness proof for MLAT. Second, we draw an important conclusion that two frequently used greedy scheduling algorithms result in a large overhead compared with the optimal solution: the scheduling latency generated by these two greedy solutions are 1/2 n times the optimal result, where n is the total number of links. We finally present an approximation algorithm for MLAT which works well for the tree with a small depth. All the above results are based on the SINR (Signal-to-Interference-plus-Noise Ratio) model.