Resource Allocation in Multi-Cellular DF Relayed OFDMA Systems

This paper considers the downlink of a multi- cellular decode-and-forward (DF) relayed OFDMA system, in which multiple base stations (BSs) and relay stations (RSs) are coordinated by a centralized resource allocation (RA) scheme. We consider the problem of maximizing the sum rate of all cells subject to a total power constraint. Three iterative RA schemes are proposed to optimize subcarrier allocation, mode selection (decision whether a relay should help and which one) and power allocation alternatively with the sum rate increasing until convergence. During each iteration, the subcarrier allocation and mode selection (SAMS) is either initialized by a conditioned subcarrier allocation and mode selection (CSAMS) algorithm or updated by a joint subcarrier allocation and mode selection (JSAMS) algorithm with the tentative power allocation result. The power allocation (PA) is updated by either single condensation based geometric programming power allocation (SC-GPPA) or conditioned water filling power allocation (CWFPA). The performances of the three RA schemes are compared by means of numerical experiments and the convergence is also illustrated.