A metaheuristic based fair dynamic spectrum allocation policy

Abstract

This work proposes a novel dynamic and adaptive yet fair spectrum allocation policy to be adopted by a service provider (SP). Here, we consider a scenario where a single SP owns multiple access networks (ANs) of different (not necessarily competing) technologies. We conceive that an SP employs an entity, called local spectrum controller (LSC), which manages the common pool of spectrum and is responsible for distributing the spectrum to individual ANs in a fair manner. Since the available spectrum is insufficient to satisfy the aggregate demand from all ANs simultaneously, LSC has to resort to a dynamic spectrum allocation technique. We have modeled the problem as an n-player cooperative bankruptcy game and have solved the problem with the help of Shapley value. We have proposed a suitable utility function for the AN with respect to its received spectrum and have identified possible objectives of LSC (i.e., SP) namely, maximizing equality of distribution (MED) and maximizing proportional fairness (MPF). We have proposed a novel equal utility distribution (EUD) algorithm that maximizes our MED objective. Finally, we have studied performances of EUD, Shapley value and max-min fairness algorithm with respect to the above objectives and have compared with nucleolus solution concept.