Average AoI Minimization for Energy Harvesting Relay-aided Status Update Network Using Deep Reinforcement Learning

A dual-hop status update system aided by energy-harvesting (EH) relays with finite data and energy buffers is studied in this work. To achieve timely updates, the best should be selected to minimize average age of information (AoI), which a recently proposed metric evaluate freshness. The AoI minimization can formulated as Markov decision process (MDP), but state space for capturing channel buffer evolution grows exponentially number relays, leading high solution complexity. We propose relay selection (RS) scheme based on deep reinforcement learning (DRL) according instantaneous packet freshness each relay. Simulation results show significant improvement DRL-based RS over state-of-art approaches.