Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

IEEE Internet of Things Journal

Quantum Reinforcement Learning With Classical Policy Deployment for Resource Allocation in Multibeam GEO–LEO Satellite Networks

Quynh Tu Ngo, Ying He, Beeshanga Abewardana Jayawickrama, Eryk Dutkiewicz, Shiva Raj Pokhrel

Published Feb 10, 2026• 1 citation
Publisher

Abstract

Satellite communications (SatCom) are envisioned as a critical enabler of 6G networks, enabling seamless global coverage by integrating terrestrial infrastructures with multi-layered satellite constellations. Among these, the integration between geostationary (GEO) and low Earth orbit (LEO) satellite networks is particularly attractive, as they combine the broad coverage of GEO satellites with the low latency and high capacity of LEO systems. Within this context, we address the resource allocation problem for LEO satellite through a joint design of beam size and transmit power, while accounting for GEO interference constraints, residual Doppler frequency offsets, and frequency reuse strategies. The objective is to maximize the spectral efficiency of LEO system operating in multi-beam GEO-LEO networks. Motivated by the limitations of classical deep reinforcement learning (RL) in such dynamic orbital settings and the potential of quantum RL for accelerated convergence, we propose a hybrid solution that exploits quantum acceleration during offline training and subsequently exports the learned policy into a classical representational format for onboard LEO satellite deployment. A fully quantum deep deterministic policy gradient framework with variational quantum circuit-based actor and critic is developed, along with a neural network-based policy translator for classical inference. To the best of our knowledge, this is the first deployment-ready quantum RL framework in SatCom, offering efficient offline training, reduced retraining latency, and practical deployment compatibility with existing LEO satellite hardware.

DOI
10.1109/jiot.2026.3663363

Related papers

Quantum-Enhanced Attention for Federated Resource Allocation in LEO Satellite Networks

Muhammad Ahsan Nazir, Ming Chen, Abdul Aziz +2 more

Dynamic Radio Resource Allocation (RRA) in Low Earth Orbit-Space Integrated Ground Networks (LEOSIGNs) presents unprecedented challenges due to extreme dynamism, sparse resource availability, and heterogeneous quality-of-service requirements across satellite constellations like Starlink, OneWeb, and Kuiper. This paper introduces a paradigmshifting Quantum-Optimized Deep Learning with Federated Edge Intelligence (QO-FEDI) framework that synergistically integrates: (1) Variational Quantum Circuits (VQC) for decomposing the NP-hard RRA problem into parametric quantum optimization, (2) federated attention-based predictive models across satellite clusters for privacy-preserving demand forecasting, and (3) adaptive edge-compute resource coordination for sub- 100 ms allocation latency. Novel contributions include: mathematical formulation of quantum cost Hamiltonians capturing interference constraints as entanglement patterns, federated learning topology with Byzantine-robust aggregation, and spatio-temporal attention mechanisms for orbital trajectory-aware predictions. Extensive evaluations across 15 realistic LEO scenarios with varying satellite configurations (32-256 satellites), traffic patterns ($1-25 \text{Gbps}$), and environmental conditions demonstrate: (1$) 71 \%$spectrum efficiency improvement (reaching$6.2 \text{bps} / \text{Hz}$), (2) 58% latency reduction (achieving 8.4 ms average), (3) 82 % handover optimization with & 96 % success rates, and (4) superior scalability to$\mathbf{1 0 0 0 +}$satellite constellations. Comparative analysis with 8 state-of-the-art methods validates near-optimal solution quality (96.1 % vs. theoretical bounds). Real-world testbed deployment on constellation simulator validates practical applicability and deployment readiness for commercial satellite operators.

International Conference on Communications and Broadband Networking (ICCBN)•Nov 27, 2025•0 citations

Hybrid quantum-classical neural networks for low earth orbit satellite communications in 6G

Ki-Chang Tong, Yeong-Gyun Jung, Hyeon-Su Yu +4 more

Low Earth Orbit (LEO) satellite communications are pivotal for next-generation 6G wireless systems but suffer from severe channel impairments, including large Doppler shifts and time-varying Shadowed-Rician fading. These dynamic conditions, coupled with channel aging, significantly degrade the reliability of classical receivers. While deep learning-based solutions offer improved performance, their high computational complexity poses a challenge for resource-constrained satellite payloads. This paper proposes a parameter-efficient hybrid quantum-classical receiver based on a Quantum Convolutional Neural Network. In the remainder of this paper, the receiver is referred to as the Hybrid QCNN. The model is enhanced with an iterative data re-uploading mechanism to maximize the expressivity of limited qubits. To improve optimization stability in hybrid quantum-classical training, this paper introduces a Dual Learning Rate optimization strategy. In this setting, gradient magnitudes can be small and can differ across parameter groups, so the strategy assigns distinct learning rates to the quantum and classical parameters. This paper validates the proposed model using a high-fidelity channel simulator that integrates the 3GPP TR 38.811 elevation-dependent K-factor model with Loo's Shadowed-Rician fading model, dynamically capturing realistic fading characteristics across different elevation angles. Simulation results demonstrate that the proposed Hybrid QCNN achieves bit error rate performance that is nearly identical to that of the classical CNN baseline. The performance gap remains below 1% across varying channel conditions, while the number of parameters is reduced by approximately 89%. Furthermore, in an ablation study, the iterative re-processing mechanism achieves up to a 26.9% relative BER reduction at an SNR of 10 dB compared with a standard quantum pooling baseline on the considered test set. This result indicates improved feature preservation during quantum pooling.

Scientific Reports•Jul 21, 2026•0 citations

Quantum-Inspired Reinforcement Learning for Edge-Driven Cyber-Physical Systems

Jagjit Singh Dhatterwal, Sandeep Kumar, Shilpa Choudhary

The integration of Quantum-Inspired Reinforcement Learning (QiRL) into Edge-Driven CyberPhysical Systems (CPS) presents a transformative approach to enhance real-time decision-making in distributed environments. While classical reinforcement learning (RL) struggles with scalability and computational complexity in edge-CPS, quantum-inspired Quantum-Inspired Reinforcement Learning for Edge-Driven Cyber-Physical Systems algorithms leverage superposition and entanglement principles to optimize policy exploration and convergence. This paper proposes a novel QiRL framework that combines quantum-inspired optimization with deep RL, enabling efficient resource allocation and adaptive control in latency-sensitive applications (e.g., smart grids, autonomous drones). We formalize the mathematical foundations of QiRL, design a hybrid neural-quantum architecture for edge deployment, and evaluate its performance under dynamic CPS constraints. Simulations on IoT-based case studies demonstrate $\mathbf{2 0}-\mathbf{3 5 \%}$ faster convergence and $\mathbf{1 5 \%}$ higher energy efficiency compared to classical Deep Q-Networks (DQN). Key challenges, such as quantum noise simulation and edge-hardware limitations, are discussed to guide future research. This work bridges quantum computing principles with edge intelligence, offering a scalable pathway for nextgeneration CPS.

International Conference on Mobile Networks and Wireless Communications (ICMNWC)•Dec 10, 2025•0 citations

Quantum Adaptive Learning for Coverage Optimization in LEO Satellite Network

Silvirianti Silvirianti, Georges Kaddoum

A broader coverage of low earth orbit (LEO) satellite networks has been an object of research interest in recent years and will remain to be of keen interest to both start-ups and established companies in the future. However, achieving optimal coverage remains a challenge today, as the satellites dynamically move along their orbit, requiring frequent constellation size and beamsteering adjustments. Such frequent adjustments and the number of variables that need to be optimized result in a high computational complexity. Aiming to attend to these concerns, the present study proposes a quantum adaptive learning (QAL) as a potential solution for coverage optimization of stochastic geometry-based LEO satellite networks with low computational complexity by taking advantage of quantum computing and adaptive learning. The QAL scheme utilizes quantum computing and a feedback mechanism to improve the learning network design and parameters for higher accuracy with low computational complexity. As a study case, considering the binomial point process (BPP) distribution of contact distance between satellites and user terminals, we use the QAL scheme to optimize the satellites' constellation size and their corresponding beamsteering at each time step to achieve maximum coverage probability. To evaluate performance, the simulation results of the QAL scheme are compared with those of quantum machine learning (QML), which lacks feedback and adaptive mechanisms. The results show that the proposed QAL outperforms QML by achieving higher coverage probability, higher accuracy, and faster convergence.

International Conference on Communications•Jun 8, 2025•1 citation

Variational Quantum Circuit-Based Reinforcement Learning for Dynamic Portfolio Optimization

Vincent Gurgul, Ying Chen, Stefan Lessmann

This paper presents a Quantum Reinforcement Learning (QRL) solution to the dynamic portfolio optimization problem based on Variational Quantum Circuits. The implemented QRL approaches are quantum analogues of the classical neural-network-based Deep Deterministic Policy Gradient and Deep Q-Network algorithms. Through an empirical evaluation on real-world financial data, we show that our quantum agents achieve risk-adjusted performance comparable to, and in some cases exceeding, that of classical Deep RL models with several orders of magnitude more parameters. However, while quantum circuit execution is inherently fast at the hardware level, practical deployment on cloud-based quantum systems introduces substantial latency, making end-to-end runtime currently dominated by infrastructural overhead and limiting practical applicability. Taken together, our results suggest that QRL is theoretically competitive with state-of-the-art classical reinforcement learning and may become practically advantageous as deployment overheads diminish. This positions QRL as a promising paradigm for dynamic decision-making in complex, high-dimensional, and non-stationary environments such as financial markets. The complete codebase is released as open source at: https://github.com/VincentGurgul/qrl-dpo-public

cs.LGCross-listedarXiv:2601.18811•Jan 21, 2026•0 citations

Quantum-Enhanced DRL Optimization for DoA Estimation and Task Offloading in ISAC Systems

Anal Paul, Keshav Kumar Singh, Aryan Kaushik +4 more

This work proposes a quantum-aided deep reinforcement learning (DRL) framework designed to enhance the accuracy of direction-of-arrival (DoA) estimation and the efficiency of computational task offloading in integrated sensing and communication systems. Traditional DRL approaches face challenges in handling high-dimensional state spaces and ensuring convergence to optimal policies within complex operational environments. The proposed quantum-aided DRL framework that operates in a military surveillance system exploits quantum computing’s parallel processing capabilities to encode operational states and actions into quantum states, significantly reducing the dimensionality of the decision space. For the very first time in literature, we propose a quantum-enhanced actor-critic method, utilizing quantum circuits for policy representation and optimization. Through comprehensive simulations, we demonstrate that our framework improves DoA estimation accuracy by 91.66% and 82.61% over existing DRL algorithms with faster convergence rate, and effectively manages the trade-off between sensing and communication and optimizing task offloading decisions under stringent ultra-reliable low-latency communication requirements. Comparative analysis also reveals that our approach reduces the overall task offloading latency by 43.09% and 32.35% compared to the DRL-based deep deterministic policy gradient and proximal policy optimization algorithms, respectively.

IEEE Journal on Selected Areas in Communications•Sep 13, 2024•32 citations
Satellite communications (SatCom) are envisioned as a critical enabler of 6G networks, enabling seamless global coverage by integrating terrestrial infrastructures with multi-layered satellite constellations. Among these, the integration between geostationary (GEO) and low Earth orbit (LEO) satellite networks is particularly attractive, as they combine the broad coverage of GEO satellites with the low latency and high capacity of LEO systems. Within this context, we address the resource allocation problem for LEO satellite through a joint design of beam size and transmit power, while accounting for GEO interference constraints, residual Doppler frequency offsets, and frequency reuse strategies. The objective is to maximize the spectral efficiency of LEO system operating in multi-beam GEO-LEO networks. Motivated by the limitations of classical deep reinforcement learning (RL) in such dynamic orbital settings and the potential of quantum RL for accelerated convergence, we propose a hybrid solution that exploits quantum acceleration during offline training and subsequently exports the learned policy into a classical representational format for onboard LEO satellite deployment. A fully quantum deep deterministic policy gradient framework with variational quantum circuit-based actor and critic is developed, along with a neural network-based policy translator for classical inference. To the best of our knowledge, this is the first deployment-ready quantum RL framework in SatCom, offering efficient offline training, reduced retraining latency, and practical deployment compatibility with existing LEO satellite hardware.
Dynamic Radio Resource Allocation (RRA) in Low Earth Orbit-Space Integrated Ground Networks (LEOSIGNs) presents unprecedented challenges due to extreme dynamism, sparse resource availability, and heterogeneous quality-of-service requirements across satellite constellations like Starlink, OneWeb, and Kuiper. This paper introduces a paradigmshifting Quantum-Optimized Deep Learning with Federated Edge Intelligence (QO-FEDI) framework that synergistically integrates: (1) Variational Quantum Circuits (VQC) for decomposing the NP-hard RRA problem into parametric quantum optimization, (2) federated attention-based predictive models across satellite clusters for privacy-preserving demand forecasting, and (3) adaptive edge-compute resource coordination for sub- 100 ms allocation latency. Novel contributions include: mathematical formulation of quantum cost Hamiltonians capturing interference constraints as entanglement patterns, federated learning topology with Byzantine-robust aggregation, and spatio-temporal attention mechanisms for orbital trajectory-aware predictions. Extensive evaluations across 15 realistic LEO scenarios with varying satellite configurations (32-256 satellites), traffic patterns (1−25Gbps1-25 \text{Gbps}1−25Gbps), and environmental conditions demonstrate: (1)71%) 71 \%)71%spectrum efficiency improvement (reaching6.2bps/Hz6.2 \text{bps} / \text{Hz}6.2bps/Hz), (2) 58% latency reduction (achieving 8.4 ms average), (3) 82 % handover optimization with & 96 % success rates, and (4) superior scalability to1000+\mathbf{1 0 0 0 +}1000+satellite constellations. Comparative analysis with 8 state-of-the-art methods validates near-optimal solution quality (96.1 % vs. theoretical bounds). Real-world testbed deployment on constellation simulator validates practical applicability and deployment readiness for commercial satellite operators.
Low Earth Orbit (LEO) satellite communications are pivotal for next-generation 6G wireless systems but suffer from severe channel impairments, including large Doppler shifts and time-varying Shadowed-Rician fading. These dynamic conditions, coupled with channel aging, significantly degrade the reliability of classical receivers. While deep learning-based solutions offer improved performance, their high computational complexity poses a challenge for resource-constrained satellite payloads. This paper proposes a parameter-efficient hybrid quantum-classical receiver based on a Quantum Convolutional Neural Network. In the remainder of this paper, the receiver is referred to as the Hybrid QCNN. The model is enhanced with an iterative data re-uploading mechanism to maximize the expressivity of limited qubits. To improve optimization stability in hybrid quantum-classical training, this paper introduces a Dual Learning Rate optimization strategy. In this setting, gradient magnitudes can be small and can differ across parameter groups, so the strategy assigns distinct learning rates to the quantum and classical parameters. This paper validates the proposed model using a high-fidelity channel simulator that integrates the 3GPP TR 38.811 elevation-dependent K-factor model with Loo's Shadowed-Rician fading model, dynamically capturing realistic fading characteristics across different elevation angles. Simulation results demonstrate that the proposed Hybrid QCNN achieves bit error rate performance that is nearly identical to that of the classical CNN baseline. The performance gap remains below 1% across varying channel conditions, while the number of parameters is reduced by approximately 89%. Furthermore, in an ablation study, the iterative re-processing mechanism achieves up to a 26.9% relative BER reduction at an SNR of 10 dB compared with a standard quantum pooling baseline on the considered test set. This result indicates improved feature preservation during quantum pooling.
The integration of Quantum-Inspired Reinforcement Learning (QiRL) into Edge-Driven CyberPhysical Systems (CPS) presents a transformative approach to enhance real-time decision-making in distributed environments. While classical reinforcement learning (RL) struggles with scalability and computational complexity in edge-CPS, quantum-inspired Quantum-Inspired Reinforcement Learning for Edge-Driven Cyber-Physical Systems algorithms leverage superposition and entanglement principles to optimize policy exploration and convergence. This paper proposes a novel QiRL framework that combines quantum-inspired optimization with deep RL, enabling efficient resource allocation and adaptive control in latency-sensitive applications (e.g., smart grids, autonomous drones). We formalize the mathematical foundations of QiRL, design a hybrid neural-quantum architecture for edge deployment, and evaluate its performance under dynamic CPS constraints. Simulations on IoT-based case studies demonstrate 20−35%\mathbf{2 0}-\mathbf{3 5 \%}20−35% faster convergence and 15%\mathbf{1 5 \%}15% higher energy efficiency compared to classical Deep Q-Networks (DQN). Key challenges, such as quantum noise simulation and edge-hardware limitations, are discussed to guide future research. This work bridges quantum computing principles with edge intelligence, offering a scalable pathway for nextgeneration CPS.
A broader coverage of low earth orbit (LEO) satellite networks has been an object of research interest in recent years and will remain to be of keen interest to both start-ups and established companies in the future. However, achieving optimal coverage remains a challenge today, as the satellites dynamically move along their orbit, requiring frequent constellation size and beamsteering adjustments. Such frequent adjustments and the number of variables that need to be optimized result in a high computational complexity. Aiming to attend to these concerns, the present study proposes a quantum adaptive learning (QAL) as a potential solution for coverage optimization of stochastic geometry-based LEO satellite networks with low computational complexity by taking advantage of quantum computing and adaptive learning. The QAL scheme utilizes quantum computing and a feedback mechanism to improve the learning network design and parameters for higher accuracy with low computational complexity. As a study case, considering the binomial point process (BPP) distribution of contact distance between satellites and user terminals, we use the QAL scheme to optimize the satellites' constellation size and their corresponding beamsteering at each time step to achieve maximum coverage probability. To evaluate performance, the simulation results of the QAL scheme are compared with those of quantum machine learning (QML), which lacks feedback and adaptive mechanisms. The results show that the proposed QAL outperforms QML by achieving higher coverage probability, higher accuracy, and faster convergence.
This paper presents a Quantum Reinforcement Learning (QRL) solution to the dynamic portfolio optimization problem based on Variational Quantum Circuits. The implemented QRL approaches are quantum analogues of the classical neural-network-based Deep Deterministic Policy Gradient and Deep Q-Network algorithms. Through an empirical evaluation on real-world financial data, we show that our quantum agents achieve risk-adjusted performance comparable to, and in some cases exceeding, that of classical Deep RL models with several orders of magnitude more parameters. However, while quantum circuit execution is inherently fast at the hardware level, practical deployment on cloud-based quantum systems introduces substantial latency, making end-to-end runtime currently dominated by infrastructural overhead and limiting practical applicability. Taken together, our results suggest that QRL is theoretically competitive with state-of-the-art classical reinforcement learning and may become practically advantageous as deployment overheads diminish. This positions QRL as a promising paradigm for dynamic decision-making in complex, high-dimensional, and non-stationary environments such as financial markets. The complete codebase is released as open source at: https://github.com/VincentGurgul/qrl-dpo-public
This work proposes a quantum-aided deep reinforcement learning (DRL) framework designed to enhance the accuracy of direction-of-arrival (DoA) estimation and the efficiency of computational task offloading in integrated sensing and communication systems. Traditional DRL approaches face challenges in handling high-dimensional state spaces and ensuring convergence to optimal policies within complex operational environments. The proposed quantum-aided DRL framework that operates in a military surveillance system exploits quantum computing’s parallel processing capabilities to encode operational states and actions into quantum states, significantly reducing the dimensionality of the decision space. For the very first time in literature, we propose a quantum-enhanced actor-critic method, utilizing quantum circuits for policy representation and optimization. Through comprehensive simulations, we demonstrate that our framework improves DoA estimation accuracy by 91.66% and 82.61% over existing DRL algorithms with faster convergence rate, and effectively manages the trade-off between sensing and communication and optimizing task offloading decisions under stringent ultra-reliable low-latency communication requirements. Comparative analysis also reveals that our approach reduces the overall task offloading latency by 43.09% and 32.35% compared to the DRL-based deep deterministic policy gradient and proximal policy optimization algorithms, respectively.