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

TL;DR

This work introduces mobile edge quantum computing (MEQC) by integrating edge-mounted QPUs with classical edge servers to enable hybrid classical-quantum task offloading from mobile devices. It formulates a non-convex, mixed-integer objective to minimize system cost (latency and energy) and recasts the problem as a partially observable Markov decision process (POMDP). A hybrid discrete-continuous multi-agent reinforcement learning (HMADRL) framework is proposed, featuring VQC-based quantum-hybrid policies to learn sustainable offloading and partitioning strategies under dynamic quantum noise and system states. Experimental results show the approach can reduce MEQC costs by up to 30% relative to baselines and accelerates convergence, demonstrating practical potential for edge-assisted quantum acceleration in next-generation networks.

Abstract

Exploiting quantum computing at the mobile edge holds immense potential for facilitating large-scale network design, processing multimodal data, optimizing resource management, and enhancing network security. In this paper, we propose a pioneering paradigm of mobile edge quantum computing (MEQC) that integrates quantum computing capabilities into classical edge computing servers that are proximate to mobile devices. To conceptualize the MEQC, we first design an MEQC system, where mobile devices can offload classical and quantum computation tasks to edge servers equipped with classical and quantum computers. We then formulate the hybrid classical-quantum computation offloading problem whose goal is to minimize system cost in terms of latency and energy consumption. To solve the offloading problem efficiently, we propose a hybrid discrete-continuous multi-agent reinforcement learning algorithm to learn long-term sustainable offloading and partitioning strategies. Finally, numerical results demonstrate that the proposed algorithm can reduce the MEQC system cost by up to 30% compared to existing baselines.

Figures (9)

• Figure 1: An illustration of mobile edge-quantum computing in the Quantum Internet.
• Figure 2: The workflow of hybrid classical-quantum computation offloading for mobile edge-quantum computing in the Quantum Internet.
• Figure 3: The proposed hybrid discrete-continuous multi-agent reinforcement algorithms. In the algorithm, each learning agent consists of the continuous agent and the discrete agent, whose actor-critic networks can be parameterized by classic/quantum neural networks.
• Figure 4: System cost v.s. training epochs, U=10 and E=10.
• Figure 5: System cost v.s. training epochs, U=20 and E=20.