A genetic algorithm for the global planning problem of UMTS networks

The ultimate objective of network planning is to achieve the best performance and profitability with the minimum network investment. In this paper, we investigate the use of the genetic algorithm in order to find "good" feasible solution for the global topology planning problem of Universal Mobile Telecommunications System (UMTS) networks. The latter has been shown to be NP-hard as it is composed of three different subproblems (each one being NP-hard): the cell planning problem, the access network planning problem and the core network planning problem. As a result, we concentrate our effort on the development of an approximate approach based on the genetic algorithm. Numerical results show that quasi-optimal solutions (on average, within 2.56 ± 1.01% of the optimal solution) can be found by using the genetic algorithm with a relatively short computation time compared to CPLEX