Intelligent Task Offloading and Resource Allocation for NOMA-based Multi-beam Satellite Mobile Edge Computing Systems

In this paper, we investigate the problem of task offloading and resource allocation in non-orthogonal multiple access (NOMA) based multi-beam satellite mobile edge computing (SMEC) systems, aiming to minimize the average task completion latency over a finite number of time steps. Such a problem is formulated as a sequential mixed-integer programming problem, which cannot be solved by standard optimization techniques. Thus, we convert the problem into a Decentralized Partially Observable Markov Decision Process (Dec-POMDP) with hybrid action space and propose a deep reinforcement learning-based solution. In particular, a multi-agent hybrid Proximal Policy Optimization algorithm (MAHPPO) is designed to address the hybrid action space. Simulation results show that the proposed scheme outperforms several benchmark schemes.