Feedback-Aware AoI Policies in Finite-Horizon Multiuser IoT Networks

This work considers finite-horizon status-update systems for real-time IoT monitoring, where multiple devices contend for a shared slotted random-access channel and each device has a strict transmission budget (e.g., due to duty-cycle/energy limits). Unlike a recently proposed single-device perspective, we quantify how cost-effective feedback interacts with multiuser congestion in Age of Information (AoI) optimization. We compare an Offline (OFF) policy that follows a precomputed schedule and requires no feedback, and a Feedback-Aware (FB) policy that combines an optimal single-user dynamic-programming rule with an AoI-aware randomized back-off to mitigate collisions. By means of Monte Carlo simulations, we characterize the operating regimes in terms of network load, channel reliability, and feedback cost. The results show that feedback is not universally beneficial: when feedback is expensive or contention is mild, OFF can be preferable because it preserves transmission opportunities. In contrast, as the network becomes dense and links become more reliable, collisions dominate update failures, and the FB policy yields substantial AoI gains, making feedback increasingly valuable in multi-device deployments.