A parallel insular model for location areas planning in mobile networks

The main interest of this paper is the optimization of the location areas planning in cellular radio networks. It is well known that the quality of service in mobile networks depends on many parameters, among them an optimal location area planning. Furthermore, it is more interesting to provide a logical organization for the already deployed networks. In this paper, we propose the use of heuristics strategies and hybrid metaheuristics strategies to solve the location areas planning problem. The latter is formulated as a constrained planar graph partitioning problem by using a mathematical model which is based on a very realistic specification. Heuristics strategies are based on greedy algorithms while hybrid metaheuristics are based on genetic algorithms. New genetic operators have been designed to this specific problem. Moreover, parallel approaches have been proposed to improve the quality of solutions and speedup the search. Results obtained on real-life benchmarks show the effectiveness of the developed optimization algorithms.