Transmission Policies for Time Varying Channels with Average Delay Constraints

We consider the problem of transferring data with a xed level of reliability over a discrete time channel whose characteristics vary with time, with a constraint on the average delay su ered by the data and a constraint on peak transmitter power. We assume that the channel state is known to the transmitter at the beginning of each slot, and seek transmission policies that minimize the average power expended by the transmitter. Using a dynamic programming formulation coupled with a duality argument, we nd that good transmission policies will in general take action in accordance with how much data is queued at the transmitter, as well as in accordance with the channel state. In particular, if the backlog is su ciently large the transmitter should use the channel even if it is in a \bad" state, and if the backlog is su ciently small the transmitter should only use the channel when it is in a \good" state. However, we nd that the time scale of the variations in the channel relative the maximum tolerable average delay is a crucial factor in uencing the qualitative features of the optimal transmission policy.