Age of Information-Oriented Link Scheduling in Device-to-Device Networks

This paper focuses on optimizing the long-term average age of information (AoI) in device-to-device (D2D) networks through age-aware link scheduling. The problem is naturally formulated as a Markov decision process (MDP). However, finding the optimal policy for the formulated MDP in its original form is challenging due to the intertwined AoI dynamics of all D2D links. To address this, we employ the Lyapunov optimization framework to develop a dynamic age-aware scheduling policy. Specifically, we explore two scenarios: known statistical channel state information (CSI) and known instantaneous CSI. For the statistical CSI case, we propose a message passing neural network (MPNN)-based policy for real-time scheduling. The MPNN is trained in an unsupervised manner with a loss function designed to minimize per-slot Lyapunov drift. For the instantaneous CSI case, we introduce a Gurobi-based policy, using the solver to minimize per-slot Lyapunov drift for scheduling decisions. To further reduce the computational complexity, we also propose a greedy heuristic policy that approximates drift minimization. Extensive simulation results show that our proposed age-aware scheduling policies have superior performance compared to the baselines, and can be applied to large-scale D2D networks.