Concurrent transmission versus time sharing in Gaussian interference channels

A key to efficient mobile communication networks is an efficient utilization of scarce resources, which includes mechanisms for resource allocation and interference management. In practice, time sharing methods such as Time Division Multiple Access (TDMA) are often exploited to create orthogonal connections. TDMA schemes however are inflexible and not suitable for decentralized implementation. In this context, approaches based on concurrent transmissions in the same frequency spectrum admitting the interference offer a promising and attractive alternative. Yet their performances deteriorate dramatically if the underlying interference coefficients are relatively large. Assuming a Gaussian Interference Channel (GIC), this paper characterizes the so-called TDMA region which is defined as a set of all interference coefficient matrices for which TDMA outperforms concurrent transmissions. We use the dimensional lifting method to prove the convexity and monotonicity of the TDMA region, which provides some insights into the scheduling design under uncertainty with respect to the interference coefficients. Finally, the analysis is used to develop strategies for resource allocation in STDMA-based wireless networks.