AoI-Aware Computation Offloading and Resource Allocation in MEC-Enabled IIoT Systems

Abstract

For time-sensitive Industrial Internet of Things (IIoT) applications demanding ultra-reliable and low-latency communication (URLLC), it is critical to integrate the age of information (AoI) into computation offloading designs. This paper proposes a joint optimization of task offloading and resource allocation in the mobile edge computing (MEC)-enabled IIoT system, aiming to minimize the long-term weighted sum of overall AoI and system energy consumption. To address the variable coupling in this problem, an asynchronous two-stage deep reinforcement learning (DRL) algorithm, JORA-MADDPG, is proposed. Distributed offload agents make parallel offloading decisions, while a centralized resource allocation (RA) agent allocates computational resources. Simulation results confirm the effectiveness of JORA-MADDPG, showing significant improvements in minimizing weighted sum of overall AoI and system energy consumption compared to other proposed schemes.