Some Neighbourhood Approaches for the Antenna Positioning Problem

The problem of positioning antennas in cellular networks is a known problem in the field of telecommunications. It consists of selecting from a set of candidate sites, the best locations to install the base stations in order to maximize the network coverage while minimizing the number of the used stations. In theory, the problem is NP-hard. To solve it in practice, we propose in this work the adaptation of two metaheuristics based on the local search that are the Iterated Local Search (ILS) and the Breakout Local Search (BLS) and provide a new algorithm inspired from both the ILS and the BLS algorithms. The latter is based on a local search process and a perturbation in the exploration of the search space. It is distinguished by its mechanism of reinitialization of the search and by its way of generating a new starting solution. To validate our approach, we have implemented, tested and compared these algorithms to several other methods on a real instance of the problem. The experimental results obtained show that the proposed approach improves the performance of these methods in most of the cases.