Energy-Efficient Adaptive Modulation and Data Schedule for Delay-Sensitive Wireless Communications

In this paper, targeting at improving the energy efficiency (EE) for Quality-of-Service (QoS)-guaranteed wireless communications, we develop new adaptive modulation and data scheduling algorithms for delay-sensitive bursty data. Assuming a-priori knowledge on data arrivals and latency requirements, the problem is formulated as a mix-integer programming that minimizes the total energy consumption at the transmitter with a non-linear Doherty power amplifier (PA) and non-negligible circuit power. According to the different properties of the PA in different output power regions, we decouple the formulated problem and solve it in two stages. In the first stage, assuming the PA has a linear efficiency, we develop an optimal modulation and data scheduling scheme (MDS) relying on convex relaxation and the resultant optimality conditions. The MDS is able to reveal the specific structure of the optimal policy in a computationally efficient and graphical manner. On top of that, a heuristic MDS scheme (HMDS) is proposed to adjust the MDS when the PA works in the non-linear region in the second stage, where a quadratic function is obtained to approximate the non-linear PA model. The offline HMDS algorithm is further extended to practical online scenarios in a well-structured way, where the modulation and data scheduling policy is produced on-the-fly. Simulation corroborates that the proposed offline algorithm can achieve the exactly same performance as the standard CVX solver, while requiring only 0.69% of its computational time.