A Variable Neighborhood Search Algorithm for Assigning Cells to Switches in Wireless Networks

Abstract

The problem of assigning cells to switches in wireless networks consists of minimizing the total operating cost, that is, the cost of linking cells to switches and the cost of handover from one cell to another, by taking into account factors such as network topology, switch capacity and traffic load in the entire network. Such a problem is well known in the literature as NP-hard, such that exact enumerative approaches are not suitable for solving real-size instances of this problem. Thus, heuristics are recommended and have been used for finding good solutions in reasonable execution times. Tabu Search (TS) is one of the best heuristics used to solve this problem. This research proposes a hybrid heuristic approach for further improving the quality of solutions obtained from TS. This approach applies TS in combination with variable neighborhood search, a recent metaheuristic that is based on the principle of systematic change of neighborhood during the local search.. A key element in the success of this approach is the use of several neighborhood structures that complement each other well and that remain within the feasible region of the search space.