Graph-based weighted sum throughput maximization in OFDMA cellular networks

This paper addresses resource allocation for weighted sum throughput maximization (WSTM) in OFDMA multi-cellular networks. The WSTM problem is a trade-off between maximizing the sum data rate, and balancing the users´ queue lengths. It may therefore be seen as a cross-layer resource allocation objective suited for Best Effort users. A graph-based subcarrier allocation for WSTM, deduced from the pairwise capacity region, is determined. Joint transmission of two interfering links is allowed depending on the convexity of the capacity region, and on the relative links´ weights. Subcarrier allocation may be followed by an iterative power control process. Numerical results show that graph-based subcarrier allocation is necessary to avoid interference-limited situations at medium to high load, and that this method followed by power control is more effective than binary allocation to maximize the weighted sum data rate and balance the queue lengths.