Approximation Algorithm for Base Station Placement in Wireless Sensor Networks

Base station location has significant impact on network lifetime performance for a sensor network. For a multi- hop sensor network, this problem is particular challenging as we need to jointly consider base station placement and data routing strategy to maximize network lifetime performance. This paper presents an approximation algorithm that can guarantee (1 - epsiv) optimal network lifetime performance for base station placement problem with any desired error bound epsiv > 0. The proposed (1 - epsiv) optimal approximation algorithm is based on several novel techniques that enable to reduce an infinite search space to a finite-element search space for base station location. The first technique used in this reduction is to discretize cost parameters (with performance guarantee) associated with energy consumption. Subsequently, the continuous search space can be broken up into a finite number of subareas. The second technique is to exploit the cost property of each subarea and represent it by a novel notion called "fictitious cost point," each with guaranteed cost bounds. This approximation algorithm offers a simpler and in most cases practically faster algorithm than a state-of-the-art algorithm and represents the best known result to this important problem.