Robust semidefinite relaxations for a quadratic OFDMA resource allocation scheme

This paper proposes two robust binary quadratically constrained quadratic programs (BQCQP) for wireless Orthogonal Frequency Division Multiple Access (OFDMA) networks. The first one is based on a scenario uncertainty approach from Kouvelis and Yu [1] and the second is based on an interval uncertainty approach from Bertsimas and Sim [2]. Both robust models allow to decide what modulations and what sub-carriers are going to be used by a particular user in the system depending on its bit rate requirements. Thus, we derive two robust semidefinite relaxations to compute lower bounds. Our numerical results show, in average near optimal integrality gaps of 4.12% and 1.15% under the scenario and interval approach when compared to the optimal solution of the problem derived by linearizing the two quadratic models with Fortet linearization method. Some comparison between the two robust approaches is also provided.