Delay-optimal mode selection in device-to-device communications for smart grid

The smart grid communication network adopts a hierarchical structure which consists of three kinds of networks including Neighborhood Area Networks (NANs). The smart grid NANs comprise of the communication infrastructure used to manage the electricity distribution to the end users. Cellular technology with LTE-based standards is a widely-used and forward-looking technology hence it is a promising technology that can meet the requirements of different applications in NANs. However, the LTE has limitations to cope with the data traffic characteristics of smart grid applications and cannot guarantee the delay requirement of real-time smart grid applications, thus requires enhancements. Device-to-Device (D2D) communications enable direct data transmissions between devices by exploiting the cellular resources, which could guarantee the improvement of LTE performances. Delay is one of the important communication requirements for the real-time smart grid applications. In this paper, the application of D2D communications for the smart grid NANs is investigated to improve the average end-to-end delay of the system. A relay selection algorithm that considers both the queue state and the channel state of nodes is proposed. The optimization problem is formulated as a constrained Markov decision process (CMDP) and a linear programming method is used to find the optimal policy for the CMDP problem. Simulation results are presented to prove the effectiveness of the proposed scheme.