An Application of Statistical Search Theory toLocation Management in Wireless Networks

This paper describes an application of statistical search theory to the important problem of location management in wireless networks. Statistical search theory is concerned with the optimal allocation of search eeort to identify the location of an object hidden in a subset of a given domain. In wireless networks , to establish a connection or route a packet to a mobile terminal, the exact location of the target terminal must rst be determined. Accordingly, a search procedure is required to discover the location of mobile terminals whose whereabouts are unknown. We focus on the search component of location management and show how the methodology of statistical search theory can adapted to the wireless environment. A major advantage of the approach described here is that the peculiar attributes of wireless networks such as limited bandwidth and intermittent disconnections are taken into consideration in formulating and solving the search problem. An interesting result of the general theory is that under realistic hypothesis regarding the network connguration, there exists a greedy, Bayesian strategy that is optimal for a variety of performance measures.