Design of personalized location areas for future Pcs networks

In Global Systems for Mobile Communications (GSM), always-update location strategy is used to keep track of mobile terminals within the network. However future Personal Communication Networks (PCS) will require to serve a wide range of services (digital voice, video, data, and email) and also will have to support a large population of users. Under such demands, determining the exact location of a user by traditional strategies would be difficult and would result in increasing the signaling load imposed by location-update and paging procedures. The problem is not only in increasing cost, but also in non-efficient utilization of a precious resource, i.e., radio bandwidth; In this thesis, personalized Location Areas (PLAs) are formed considering the mobility patterns of individual users in the system such that the signaling due to location update and paging is minimized. We prove that the problem in this formulation is of NP complexity. Therefore we study efficient optimization techniques able to avoid combinatorial search. Three known classes of optimization techniques are studied. They are Simulated Annealing, Tabu Search and Genetic Search. Three algorithms are designed for solving the problem. Modeling does not assume any specific cell structure or network topology that makes the proposal widely applicable. The behavior of mobile terminals in the network is modeled as Random Walk with an absorbing state and the Markov chain is used for cost analysis; Numeric simulation carried out for 25 and 100 hexagonal cell networks have shown that Simulated Annealing based algorithm outperforms other two by indicators of the runtime complexity and signaling cost of location management. The ID\u27s of cells populating the calculated area are provided to the mobile terminal and saved in its local memory every time the mobile subscriber moves out its current location area. Otherwise, no location update is performed, but only paging. Thus, at the expense of small local memory, the location management is carried more efficiently