A Novel Deadline-/Interference-Aware Cooperative Data Transmission Scheduling Scheme for Optimizing AoI in Wireless Networks

In knowledge-freshness-sensitive wireless networking scenarios, a major identified gap lies in overlooking the notable impact of physical link interference and the validity duration (i.e. deadline) of newly acquired knowledge on the network-wide achieved Age of Information (AoI). This paper aims at closing this gap not only through accounting for deadline and interference constraints but also through proposing efficient mechanisms leveraging inter-nodal cooperativeness to work around them and, indeed, migrate the achieved network-wide average AoI to its next lower level. First, a Deadline-constrained Interference-aware Cooperative data transmission Scheduling (DICS) framework is presented. This framework encloses a mathematical formulation of the cooperative data transmission scheduling optimization problem as an Integer Linear Program (ILP) that accounts for deadline and interference constraints and has the two-fold objective of:
i
) investigating the impact of integrating auxiliary Secondary Nodes (SNs) in the network to assist Primary Nodes (PNs) in improving their overall achieved AoI, and
ii
) evaluating the effect of altering various network-wide parameter values on the network's AoI performance. Owing to the ILP's remarkable complexity, a GReedy Algorithm (GRA) is proposed, which generates acceptable sub-optimal schedules (error of 5.222%) but is found to be non-scalable. To address GRA's scalability issue, a Reinforcement Learning Actor-Critic Algorithm (RL-ACA) is developed, which, is found to outperform GRA by 2.849% in terms of accuracy. Thorough simulations and numerical analyses are conducted to gauge the merit of the proposed DICS's RL-ACA and highlight its AoI performance improvement as compared to existing benchmarks.