# Age of Information Minimization in Multicarrier-Based Wireless Powered Sensor Networks

## TL;DR

This paper proposes a new method to reduce information delay in wireless sensor networks by combining Lyapunov optimization and deep reinforcement learning.

## Contribution

A novel approach combining Lyapunov optimization and model-free deep reinforcement learning to minimize Age of Information in WPSNs.

## Key findings

- The proposed algorithm achieves significantly lower weighted average Age of Information than existing methods.
- It effectively reduces excessive instantaneous AoI for individual sensors compared to DQN.
- Simulation results validate the effectiveness of the method in real-world scenarios.

## Abstract

This study investigates the challenge of ensuring timely information delivery in wireless powered sensor networks (WPSNs), where multiple sensors forward status-update packets to a base station (BS). Time is partitioned to multiple time blocks, with each time block dedicated to either data packet transmission or energy transfer. Our objective is to minimize the long-term average weighted sum of the Age of Information (WAoI) for physical processes monitored by sensors. We formulate this optimization problem as a multi-stage stochastic optimization program. To tackle this intricate problem, we propose a novel approach that leverages Lyapunov optimization to transform the complex original problem into a sequence of per-time-bock deterministic problems. These deterministic problems are then solved using model-free deep reinforcement learning (DRL). Simulation results demonstrate that our proposed algorithm achieves significantly lower WAoI compared to the DQN, AoI-based greedy, and energy-based greedy algorithms. Furthermore, our method effectively mitigates the issue of excessive instantaneous AoI experienced by individual sensors compared to the DQN.