Energy-Efficient Data Transmission With Non-FIFO Packets With Processing Cost

In this paper, we consider the energy-efficient transmission problem with non-first-in-first-out (FIFO) packets for a point-to-point additive white Gaussian noise channel with processing cost. This work can be considered as a generation of previous work in terms of the assumption of the FIFO transmission order, which means the earlier arrived packet must have an earlier deadline. Assuming non-causal data arrival information, we first formulate the objective as a non-convex optimization problem, and then investigate the necessary and sufficient conditions of the optimal offline transmission schedule. Based on the optimal properties of these conditions, an efficient offline algorithm that finds the optimal transmission schedule minimizing the total energy consumption is proposed. Next, we identify the optimality of the proposed offline algorithm by proving it satisfies the sufficient conditions of optimality, and further, respectively, analyze the computational complexity in the best case:
$\mathcal {O}(N^{2})$
and in the worse case:
$\mathcal {O}(N^{3})$
, where
$N$
is the packet number of the sequence. Finally, based on the insights obtained from the offline transmission schedule, an efficient heuristic online algorithm performing close to the offline transmission schedule is proposed under the assumption that the data arrival information is known causally at the transmitter.