Hybrid Reinforcement Learning-based Sustainable Multi-User Computation Offloading for Mobile Edge-Quantum Computing
Minrui Xu, Dusit Niyato, Fellow, IEEE, Jiawen Kang, Zehui Xiong, Mingzhe Chen · 2025
Read Hybrid Reinforcement Learning-based Sustainable Multi-User Computation Offloading for Mobile Edge-Quantum Computing by Minrui Xu, Dusit Niyato, Fellow, IEEE, Jiawen Kang, Zehui Xiong, Mingzhe Chen (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-08134.txt.
Introduction
Xu and colleagues study how phones can send hard tasks to nearby servers that mix normal and quantum chips, in a paper from twenty twenty-five. Quantum chips can speed up some big jobs, like planning networks or guarding data. The team designs a system where a phone can split a task and send parts to the edge. Their goal is to cut both wait time and energy use.
They teach a learning program to pick the best split for each task. The program learns from long runs, not just one choice at a time. In tests, it cut total cost by up to thirty percent next to older methods. The work points to a greener way to run mixed tasks at the edge.
Catalog record
| CREATOR | Minrui Xu, Dusit Niyato, Fellow, IEEE, Jiawen Kang, Zehui Xiong, Mingzhe Chen |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-08134
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Hybrid Reinforcement Learning-based Sustainable Multi-User Computation Offloading for Mobile Edge-Quantum Computing Thanks: Part of this article was presented at the IEEE International Conference on Communications 2023 [ 1 ] . Thanks: M. Xu and D. Niyato are with the College of Computing and Data Science, Nanyang Technological University, Singapore (e-mail: [email protected]; [email protected]). J. Kang is with the School of Automation, Guangdong University of Technology, China (e-mail: [email protected]). Z. Xiong is with the Pillar of Information Systems Technology and Design, Singapore University of Technology and Design, Singapore 487372, Singapore (e-mail: [email protected]). M. Chen is with the Department of Electrical and Computer Engineering and Institute for Data Science and Computing, University of Miami, Coral Gables, FL, 33146 USA (e-mail: [email protected]). D. I. Kim is with the Department of Electrical and Computer Engineering, Sungkyunkwan University, Suwon 16419, South Korea (email: [email protected]). X. Shen is with the Department of Electrical and Computer Engineering, University of Waterloo, Waterloo, ON N2L 3G1, Canada (e-mail: [email protected]). Minrui Xu, Dusit Niyato, Fellow, IEEE, Jiawen Kang, Zehui Xiong, Mingzhe Chen 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. I Introduction Quantum computing and communication are envisioned as strategic technologies across academic and industrial sectors, since they will introduce significant advantages in extant technological fields, including artificial intelligence (AI) [ 2 ] , security [ 3 ] , and finance [ 4 ] . For instance, the Quantum Internet can leverage quantum optimization algorithms, which have the potential to provide faster and more efficient processing of large amounts of data [ 5 , 6 , 7 ] . Moreover, with quantum computing, resource management in the Quantum Internet can be optimized by improving the allocation and utilization of resources at the edge of networks, leading to more efficient use of available resources. Additionally, quantum cryptography can enhance security in mobile edge networks by providing a higher level of protection against cyberattacks and security threats [ 4 ] . Finally, by enabling faster and more efficient data processing, quantum computing can help streamline data management in mobile edge networks, making it easier to manage and analyze large-scale data in real-time. Despite these promising applications, executing such quantum computation tasks necessitates scalable quantum computers endowed with an approximate capacity of 10 6 qubits. Diverging from classical computing [ 8 , 9 ] , quantum computing faces unique challenges as the number of qubits, quantum gates, and measurement operations in scalable quantum computers [ 10 ] . Among these challenges, the most significant one is the inherent qubit noise that might compromise the fidelity of quantum computing [ 11 ] . To achieve fault-tolerant quantum computing, scalable quantum computers can operate quantum processing units (QPUs) with advanced cryogenic components and fault-tolerant schemes [ 12 ] . On the one hand, quantum computers work at extremely low temperatures to cool quantum devices to a low-entropy state. On the other hand, quantum noise is combated through fault-tolerant schemes, including concatenated codes, surface codes, and bosonic qubits [ 11 ] . For instance, using error-correction codes allows quantum information to be maintained across several physical qubits, constituting one logical qubit. Overall, the tremendous computing and energy advantages can be achieved only with fault-tolerant quantum computing operating at cryogenic conditions. As scalable quantum computers meet the requisite scale and quality parameters, mobile edge-quantum computing (MEQC) [ 13 ] , coupled with the remote accessibility offered by edge servers, may extend the reach of quantum advantages into mobile edge networks. By offloading computation tasks to quantum computers in edge servers, users can gain significant benefits from cloud/edge quantum computer providers, such as Amazon Braket [ 14 ] , IBM Quantum [ 15 ] , and Azure Quantum [ 16 ] . This extends the potential of discovering innovative applications for the Quantum Internet and tackling existing issues, thus driving the practical implementation of scalable quantum computers. MEQC can provide users with a range of potentially lethal applications, such as quantum ray tracing [ 17 ] , by significantly increasing the appeal of quantum computing to mobile consumers. Unlike classical edge computing [ 8 ] , MEQC has unique differences in the measurement of computing power and energy consumption. First, quantum computing uses the superposition, interference, and entanglement of qubits to accelerate the execution of computation tasks. Second, quantum computers operate in an extremely low-temperature environment. Therefore, compared to conventional computers, most of the energy consumption of quantum computers is used to maintain the ultra-low temperature container. Third, for the reliability of execution results, quantum computers choose appropriate error correction codes and the concatenation degree of error correction according to their latency and energy constraints. Nonetheless, the integration of quantum computers into mobile edge networks to execute quantum computation tasks introduces complexities that go beyond those that are inherent in classical MEC. Finally, to determine offloading strategies for mobile devices to perform hybrid classical-quantum computation tasks in edge servers efficiently is still challenging. Fig. 1: An illustration of mobile edge-quantum computing in the Quantum Internet. In the Quantum Internet, the problem of determining sustainable strategies in the hybrid classical-quantum computation offloading problem involves continuous and discrete optimization variables in both classical and quantum computing. Therefore, deciding on how much computation and which server to offload is a complex mixed-integer programming problem with the non-convex and non-linear objective of minimizing computational cost in terms of latency and energy consumption. Fortunately, deep reinforcement learning (RL) [ 18 , 19 , 20 ] is used by mobile devices to learn a sustainable offloading and partitioning strategy without prior knowledge of the computation tasks, edge server status, and local environment. To minimize long-term system cost, deep RL can let mobile devices act as agents to learn and obtain sustainable offloading and partitioning strategies under dynamically changing quantum noise and the uncertain requests of computation tasks. By leveraging hybrid discrete-continuous policies to interact with the MEQC environment, mobile devices can learn to determine the discrete offloading and the continuous partitioning decisions for classical-quantum computation tasks to maximize the sustainability of the MEQC system. Finally, we leverage variational quantum circuits (VQCs) to parameterize the actor-critic networks of agents, which can accelerate the convergence process of learning-based algorithms without the loss of performance, and thus achieving a higher sustainability. The main contributions can be summarized as follows. • We propose a novel paradigm of mobile edge-quantum computing in the Quantum Internet