Randomized Scheduling for Periodic Multi-Source Systems with PAoI Violation Guarantees

Abstract: The Age of Information (AoI) has been recognized as a critical metric for assessing the freshness of information in modern communication systems. In this work, we examine an information update system where multiple information sources transmit updates to their respective destinations via a shared base station. Our main contribution is the proposal of a randomized scheduling algorithm that offers distinct statistical AoI guarantees for heterogeneous sources. Specifically, we rigorously derive an analytical upper bound on peak age of information (PAoI) violation probability by leveraging properties of the multivariate noncentral hypergeometric Wallenius distribution. Building on these analytical results, two designs of coefficients for the randomized policy are proposed to meet the outage constraints for all sources, tailored to the long and short sampling delay cases, respectively. Simulation results demonstrate the accuracy of our analysis on PAoI violation probability and also show that our proposed design always provides a feasible solution in most cases.