AoI-Aware D2D Communication in 5G-Enabled Smart Grids: A Multi-Armed Bandit Approach

Device-to-device (D2D) communication has been introduced as an innovative approach in 5G networks to enhance network performance and coverage. D2D-assisted relay helps to further boost transmission performance where the relay nodes assist in wireless data offloading for users with compromised direct cellular links. Typically, D2D links are established between the source users and the relay nodes in a deterministic manner which is inefficient considering the dynamic nature of the wireless channel. Particularly, the selection of D2D relay is critical for time-sensitive applications where data freshness of a significant importance. In this work, we propose an intelligent relay selection mechanism to improve the Age of Information (AoI) in 5G-enabled smart grids. The problem of selecting the best relay is structured within a Multi-Arm Bandit (MAB) framework and solved using the Upper-Confidence-Bound (UCB) algorithm. Each source node compiles a set of possible relay options according to established criteria. Subsequently, the learning mechanism employed in the UCB algorithm progressively gravitates towards choosing a relay node that enhances a combined AoI metric that integrates both the average AoI and the probability of AoI violation probability. The performance of the proposed intelligent relay selection method is evaluated using extensive discrete-event simulations, and the results affirm the effectiveness of the proposed method over the deterministic relay selection method.