Optimizing the Location Areas Planning in the SUMATRA Network with an Adaptation of the SPEA2 Algorithm

摘要

This paper presents our adaptation of the Strength Pareto Evolutionary Algorithm 2 (SPEA2, a Multi-Objective Evolutionary Algorithm) to optimize the Location Areas Planning Problem. Location Areas is a strategy widely used to manage one of the most important issues of the Public Land Mobile Networks: the mobile location management. In contrast to previous works, we propose a multi-objective approach with the goal of avoiding the drawbacks associated with the linear aggregation of the objective functions. The main advantage of a multi-objective approach is that this kind of algorithm provides a wide range of solutions among which the network operator could select the solution that best adjusts to the network real state at each moment. Furthermore, in order to obtain realistic results, we apply our proposal to the SUMATRA network, a test network that stores real-time information of the users' mobile activity in the San Francisco Bay (USA). Experimental results show that our proposal outperforms the results obtained in other works and, at the same time, it achieves a great spread of solutions.