Partially observed stochastic shortest path problem - application to sequential paging in cellular networks

Abstract

Polling a roaming mobile user in a cellular network to determine its location is called paging and it requires the use of limited wireless resources. We formulate the paging problem as an optimal sequential search problem for a Markovian target and show that the resulting problem is an instance of a Partially Observed Stochastic Shortest Path (POSSP) problem. Using the theory of POSSP problems, we propose optimal and suboptimal paging algorithms with performance bounds. We then propose a scalable sequential paging architecture for paging multiple mobile stations simultaneously using a finite number of paging resources.