Joint Freshness and Channel Aware Scheduling for Multi-User Wireless Communications

Age of Information (AoI) has attracted much attention recently due to its capability of characterizing the freshness of information in Industrial Internet of Things (IIoT). To improve the information freshness over a fading channel shared by multiple users, efficient scheduling methods are highly desired for IIoT. In this paper, we are interested in a cross-layer scheduling policy, in which power and rate-adaptive transmissions are enabled in the physical layer. More specifically, we adopt a probabilistic scheduling method to minimize AoI while satisfying an average power constraint. The optimization of the probabilistic scheduling policy is formulated as a Constrained Markov Decision Process (CMDP), the dimension of which grows exponentially with the number of users and the buffer size. To overcome the prohibitive computational cost, we reveal the threshold-based structure of the AoI-optimal scheduling. Then we conceive a low complexity iterative algorithm that gives us the optimal freshness-aware scheduling policy.