Energy-Efficient Wireless System Within Average Delay and With Mixed-Erlang-Distributed Data

Energy efficiency (EE) is a key performance indicator in modern wireless systems. Consider a point-to-point buffer-aided communication system model, where 1) the wireless channel has Shannon capacity and is characterized by the Nakagami-
${m}$
distribution; 2) application data has an average data rate with mixed-Erlang distributed data sizes and a quality-of-service (QoS) constraint on average delay. In this letter, we first develop a numerical expression of average delay in the above system model, and then propose an optimal transmit power allocation scheme to maximize the EE under the QoS constraint. It is shown in the Monte-Carlo simulation that our numerical expression can accurately approximate the average delay. From the numerical simulation, the maximum EE as a function of data rate is shown to have a bell-shaped curve.