Robust resource management for energy efficient cellular networks

Cross layer scheduling techniques rely on accurate system state information, in particular the channel state. Since effective link gains are inherently time-varying, scheduling decisions have to be made every time the channel changes and hence accurate channel state information must be collected. In this paper we propose scheduling techniques that are robust to channel variations, allowing less frequent schedule updates and reducing in that sense the communication overhead and improving the system performance. Lower bounds on the achieved goodput when robust scheduling is used, in single and multi hop networks, are derived and through simulation we study the trade-offs between data rates, robustness and energy consumption. Finally, it is shown that the price of robustness in relay-aided networks is less when compared to traditional architectures.