Energy optimal scheduling under average throughput constraint

By adapting radio setting, such as the modulation or error code, in response to changes in the wireless channel, the energy consumption of a communication system can be greatly reduced. In paper, we investigate the problem of minimizing the energy for a desired average throughput. When the option of shutting down the radio completely is also considered, deciding on the optimal setting becomes a complex scheduling problem. In this case, we have to take power of both the radio electronics and the power amplifier into account, the relative importance of which has a critical impact on the nature of the adaptation strategy, as we show in this paper. We present a new algorithm that does not require any knowledge of the channel statistics, but instead quickly learns the desired adaptation strategy. It is easy to implement, yet is optimal in terms of energy savings.