With the motivation of optimizing the power consumption, in the past few years, much attention has been devoted to the smart grid system. As we know, one significant feature of the smart grid is that electric appliances can obtain the real-time power price, and power suppliers can implement the power allocation via networks. Through smart grid, the appliances can adjust their power consumptions according to their practical needs, and hence the efficiency of the power utilization is enhanced. Usually, the appliances' practical needs rely on the customer satisfaction degree which belongs to the field of quality of experience (QoE). Indeed, QoE-driven power scheduling is important for smart grid: Firstly, the current quality of service (QoS)-based power optimization only makes sense in the case of static or single quality model whereas smart grid uses a dynamic or heterogeneous model based on quality guarantees. In such a case, the system operators should maximize customer satisfaction, irrespective of power requirements. Secondly, customer satisfaction is also very important to the control centers who realize that unsatisfied users would quit the system, resulting in severe loss of system revenues. Thirdly, QoE-based scheduling potentially allows, at least in theory, more customers to be served simultaneously without a severe perceived quality loss.
In this article, the authors propose a QoE-driven power scheduling in the context of smart grid from the perspectives of architecture, strategy, and methodology. Specifically, it takes into account the QoE requirement when designing the power allocation scheme. To obtain the QoE requirement, the authors analyze the fluctuation of the power load and the impact of the transmission delay. In particular, the power allocation is formulated as an optimization problem that maximizes the system performance. Based on the derived QoE model, an efficient power scheduling scheme is proposed by jointly considering the admission control and QoE expectation. Extensive simulation results indicate that the proposed scheme can efficiently allocate the power according to the dynamic QoE requirements in a practical smart grid system.