Minimum-time aggregation scheduling in multi-sink sensor networks

Data aggregation is an essential operation in wireless sensor networks and aggregation time is critical in many sensor network applications. This paper focuses on the minimum-time aggregation scheduling problem in multi-sink sensor networks (MS-MTAS problem). To the best of our knowledge, this is the first work on the MS-MTAS problem, which is NP-hard. In this paper, we first present a lower bound for MS-MTAS. Then, we design two approximation algorithms for MS-MTAS based on the maximal independent sets. The upper bounds of the aggregation time are, respectively, (min{26k-2, 2Δ})(max^k_i=1{R_i}-l)+2Δ and 23 max^k_i=1 {R_i}+2Δ-23, where k is the number of sinks, A is the maximum vertex degree of the network graph and R_i is the radius of the induced subgraph G[V(T_i)], with respect to sink S_i. Here T_i refers to the tree generated by the corresponding algorithm and V(T_i) is the vertex set of T_i. We prove that the second algorithm is a nearly constant approximation. This demonstrates that nearly constant approximation can be achieved for MS-MTAS, which is the same as MTAS. Both algorithms are implemented by simulation and the simulation results verify the effectiveness of the proposed algorithms.