Task Offloading and Resource Allocation for Cloud-Edge Collaboration in Low Earth Orbit Satellite Networks

With the rapid advancement of satellite communication technology, integrating decentralized satellite computing capabilities through low earth orbit (LEO) satellite edge computing has emerged as a promising solution. However, computation offloading and resource allocation in the mobile edge computing environment face significant challenges due to dynamic system states and evolving user demands. This paper presents a cloud-edge collaborative computation offloading model that determines optimal offloading strategies and resource allocation based on task features. Initially, a network model is established with well-defined task and resource parameters, which includes ground terminals, satellites, and clouds. Then, a joint latency and energy consumption optimization problem is formulated and described using the Markov decision process (MDP). Finally, we propose a soft actor-critic-based optimal task offloading and resource allocation (SAC-OTORA) algorithm to minimize system la-tency and energy consumption simultaneously. Simulation results demonstrate that the proposed scheme outperforms the existing schemes, achieving lower latency and energy consumption.