Game Theoretic Analysis of Resource Allocation in Multi-tiered Networks

We study resource allocation in wireless multi-tiered OFDMA cellular systems, like the current 5G networks. We compare coordinated versus distributed methods of assignment of resource blocks to maximize signal strength and throughput to ensure quality of service. The distributed assignment, where every wireless client optimizes its utility, is modeled by a strategic game with the cellular users as players. We study the existence of Nash equilibrium in this game and provide bounds for the inefficiency of this strategic game via analysis of the Price of Anarchy (PoA). Our social objective is chosen to provide fair throughput by measuring the minimum data rate. We show that approximating the social optimum, i.e. maximizing the minimum data rate, to within any given factor is NP-Hard and provide an algorithm that is illustrated to perform well in practice.