A comparison of randomized and evolutionary approaches for optimizing base station site selection

Abstract

It is increasingly important to optimally select base stations in the design of cellular networks, as customers demand cheaper and better wireless services. From a set of potential site locations, a subset needs to be selected which optimizes two critical objectives: service coverage and financial cost. As this is an NP-hard optimization problem, heuristic approaches are required for problems of practical size. Our approach consists of two phases which act upon a set of candidate site permutations at each generation. Firstly, a sequential greedy algorithm is designed to commission sites from an ordering of candidate sites, subject to satisfying an alterable constraint. Secondly, an evolutionary optimization technique, which is tested against a randomized approach, is used to search for orderings of candidate sites which optimize multiple objectives. The two-phase strategy is vigorously tested and the results delineated.