Minimum Data-Latency-Bound $k$-Sink Placement Problem in Wireless Sensor Networks

In this paper, we propose a new multiple-sink positioning problem in wireless sensor networks to best support real-time applications. We formally define this problem as the
k
-Sink Placement Problem (
k
-SPP) and prove that it is APX-complete. We show that an existing approximation algorithm for the well-known
k
-center problem is a constant factor approximation of
k
-SPP. Furthermore, we introduce a new greedy algorithm for
k
-SPP and prove its approximation ratio is very near to the best achievable, 2. Via simulations, we show our algorithm outperforms its competitor on average.