Scheduling to Minimize Age of Information with Multiple Sources

Abstract

Finding an optimal/near-optimal scheduling algorithm to minimize the age of information (AoI) in a multi-source G/G/1 system is well-known to be a hard problem. In this paper, we consider this problem for the non-preemptive setting, where an algorithm is free to choose which update to transmit, but an update under transmission is not allowed to be preempted. For this problem, we propose a novel randomized scheduling algorithm and show that its competitive ratio is at most 3 plus the maximum of the ratio of the variance and the mean of the inter-arrival time distribution of sources. Notably, the competitive ratio is independent of the number of sources, or their service time distributions. For several common inter-arrival time distributions such as exponential, uniform and Rayleigh, the competitive ratio is at most 4.