Priority-Oriented Intelligent Resource Management in Space-Air-Ground Integrated IoT Networks

In this paper, we study intelligent multi-domain collaborative computing offloading within the space-air-ground integrated Internet of Things (SAG-IoT). While non-terrestrial transmission alleviates the burden on scarce terrestrial resources, it introduces significant propagation delay, rendering it unsuitable for all tasks. To address this issue, we categorize tasks into priority and general groups and design a dynamic priority resource management (DPRM) framework. This framework strategically pre-allocates resources to priority tasks, ensuring their completion on edge nodes. Within this framework, we formulate an optimization problem focused on offloading path selection and multi-dimensional resource management, to maximize the completion rates of general tasks while meeting the quality of service requirements for priority tasks. We introduce a hierarchical hybrid policy optimization based on DPRM (HHPO-DPRM) algorithm to tackle the aforementioned problem in highly dynamic network environments. Comparative analysis with two traditional algorithms underscores the effectiveness of our approach.