Delay-Power Tradeoff of Fixed-Rate Wireless Transmission With Arbitrarily Bursty Traffics

Wireless communication system is expected to provide service with low latency and high energy efficiency. To improve the energy efficiency, the transceiver prefers sending packets, when the channel states are good. However, such opportunistic transmission may induce undesirably large latency. Therefore, a fundamental tradeoff exists between the average transmission power and average queuing delay, and is studied in this paper via cross-layer probabilistic scheduling. In particular, we consider the delay-power tradeoff when the packet arrivals have arbitrary probabilistic distributions. A Markov reward model is adopted to model the queue of the backlogged packets. Based on that, we formulate a nonlinear optimization problem and convert it into a linear programming (LP) problem by using variable substitution. The optimal solution to the LP problem allows us to derive the optimal scheduling parameters. Based on the optimal solution, we can derive the optimal scheduling policy, which turns out to be threshold-based. Besides, we consider the source scheduling with the specific packet arrival distribution being unknown. Adaptive algorithms are proposed to achieve the corresponding delay-power tradeoff.