Multiple Sink Placement Strategy for Underwater Wireless Sensor Networks

We propose a mathematical model for multiple sink placement in Underwater Wireless Sensor Networks (UWSNs). The model is constructed for a 3D mesh topology. Our objective is to minimize the overall number of hops of every source-closest sink pair of cells. We analyze the constructed model, and prove its hardness by reducing the k-median problem, a well-known NP-hard problem to our problem. We then use the Partitioning Around Medoid (PAM) approximation algorithm to find a near optimal solution to our sink placement problem. We test the solutions obtained by this approximation algorithm by comparing them with the optimal solutions obtained using a brute force method. We also compare the approximate solutions with solutions obtained using a random-based heuristic. Results show that the proposed PAM-based approximation achieves remarkably near-optimal solutions.