Learning-based Sustainable Multi-User Computation Offloading for Mobile Edge-Quantum Computing

Abstract

In this paper, a novel paradigm of mobile edge-quantum computing (MEQC) is proposed, which brings quantum computing capacities to mobile edge networks that are closer to mobile users (i.e., edge devices). First, we propose an MEQC system model where mobile users can offload computational tasks to scalable quantum computers via edge servers with cryogenic components and fault-tolerant schemes. Second, we show that it is NP-hard to obtain a centralized solution to the partial offloading problem in MEQC in terms of the optimal latency and energy cost of classical and quantum computing. Third, we propose a multi-agent hybrid discrete-continuous deep reinforcement learning using proximal policy optimization to learn the long-term sustainable offloading strategy without prior knowledge. Finally, experimental results demonstrate that the proposed algorithm can reduce at least 30% of the cost compared with the existing baseline solutions under different system settings.