Optimizing Age of Information for Energy Harvesting Systems Using an Energy-Aware Hybrid Preemptive/Non-Preemptive Discipline

Abstract

The growing integration of Internet of Things technologies, alongside limited energy resources of sensor nodes, has driven extensive research into real-time status update systems powered by energy harvesting (EH) modules. The unpredictability of available energy can adversely affect critical timeliness, as measured by the age of information (Aol) metric. Packet management schemes have been developed to enhance AoI under stochastic energy conditions. However, conventional preemptive (PR) and non-premptive (NP) service disciplines operate independently of system dynamics, limiting their adaptability to variations in traffic and energy profiles. In this work, we propose a novel energy-aware hybrid PR/NP discipline for a singlesource status update system. The system is powered by EH technology with a limited-capacity battery and a Poisson energy replenishment process. Service preemptions, which lead to energy depletion, are permitted only when the battery's energy level exceeds a predetermined threshold parameter. We analyze the average AoI using the stochastic hybrid system approach, which leads to deriving a formula for the energy packet loss rate. The numerical study demonstrates that the proposed discipline, by adjusting its threshold parameter, enhances system adaptability and achieves a significant balance between AoI optimization and energy loss reduction.