A Location Management Scheme with Reinforcement Learning for PCS Networks

Mobility management, consisting of location management and handoff management, is a challenge for personal communication services (PCS) networks. Location management mainly involves two operations: location update and paging. A two-stage paging strategy with reinforcement learning is proposed. Under this scheme, each cell in a location area is given a preference. Besides, a small real number is used to decide whether to choose and page cells with higher preferences or with lower ones during the first paging stage so that the scheme can quickly adapt to the moving pattern of mobiles. Simulation results show that the proposed paging strategy can reduce the cost as much as 25%, compared with the paging strategy of the basic location management scheme used in the existing PCS networks.