On the performance of distributed power allocation and scheduling in multi-cell OFDMA

In next generation wireless systems such as Long Term Evolution (LTE), Radio Resource Management with Inter-cell Interference Coordination (ICIC), is a key issue to be addressed. Although centralized resource allocation (RA) and collaborative processing can optimally perform ICIC, the overall required complexity suggests the consideration of distributed techniques. In this paper we investigate the behavior of distributed RA strategies aimed at maximizing the weighted average sum-rate of a multi-cell clustered system in presence of power constraints by evaluating throughput gap and fairness with respect to their respective centralized implementations. In the distributed RA strategies the inter-cell interference is partially coordinated through the use of power planning schemes (fractional frequency reuse - FFR - is the case considered here), leading to a reduction in both signaling and feedback requirements. The results obtained show that the distributed RA with aggressive frequency reuse is able to approach t performance of the centralized RA when the number of users is large. However, FFR is useful to improve fairness (or cell-edge user performance) at the expense of sum-rate when the scheduling weights are those of proportional fair scheduling (PFS), whereas FFR and in general power planning techniques are able to improve throughput while preserving the fairness when the scheduling weights are set to guarantee equal average rate to users.