Centralized vs Distributed Resource Allocation in Multi-Cell OFDMA Systems

Abstract

Radio Resource Management with Inter-cell Inter-\nference Coordination (ICIC), is a key issue under investigation\nfor next generation wireless systems such as Long Term Evolu-\ntion (LTE). Although centralized resource allocation (RA) and\ncollaborative processing can optimally perform ICIC, the overall\nrequired complexity suggests the consideration of distributed\ntechniques. In this paper we propose and compare a centralized\nRA strategy aimed at maximizing the sum-rate of a multi-cell\nclustered system in presence of power and fairness constraints, and\na distributed RA strategy where inter-cell interference is partially\ncoordinated through power planning schemes with preassigned\npower (an example is fractional frequency reuse). The distributed\nRA strategy reduces both signaling and feedback requirements,\npreserves intra-cell fairness and jointly works with a load bal-\nancing algorithm to support inter-cell fairness. We show in the\nresults that the distributed RA with aggressive frequency reuse\nis able to approach the performance of the centralized RA when\nthe number of users is large, while preserving the same level of\nfairness.