Quantum gated recurrent neural networks for continuous dynamic temporal decision-making
Physica Scripta · 28.08.2026 · 101 (34) · 346001
DOI: 10.1088/1402-4896/ae99cd ↗Kurz gefasst
Fachartikel in der Zeitschrift Physica Scripta, online erschienen am 24.08.2026. Verfasst von Jiazhao Shen und 4 weiteren Autor:innen. Die Arbeit nennt 49 Literaturangaben. Gefördert von: Joint Fund of Shandong Provincial Natural Science Foundation, Shandong Provincial Natural Science Foundation, China, Key R&D Program of Shandong Province, China.
Abstract (Originalsprache)
Performance breakthroughs in dynamic temporal decision-making tasks have long been a core research objective in reinforcement learning. Yet traditional policy networks, including long short-term memory (LSTM) networks, gated recurrent unit (GRU) and feedforward neural network (FNN), have long been hampered by inherent bottlenecks: low reward efficiency and training instability, which renders them less capable of meeting the high-precision decision-making demands of complex scenarios. This paper introduces a novel policy network, quantum gated recurrent neural network (QGRNN), which incorporates quantum superposition encoding and cyclic entanglement structures to achieve more efficient state-space exploration. Furthermore, the network employs a quantum–classical hybrid residual design that effectively integrates quantum measurement outcomes with classical inputs, thereby strengthening the representational capacity for temporal decision-making features. Trained over 800 episodes via proximal policy optimization in CartPole-v1, the QGRNN achieves an average reward of 380.3 ± 128.6 in the last 100 episodes-significantly outperforming GRU (275.9 ± 140.7), LSTM (239.7 ± 131.8), and FNN (219.2 ± 109.5). As a 4-qubit proof-of-concept, it achieves 45% high-performance episodes (rewards > 450) and demonstrates low-resource, high-performance characteristics for temporal decision-making.
Förderung
- Joint Fund of Shandong Provincial Natural Science Foundation
- Shandong Provincial Natural Science Foundation, China
- Key R&D Program of Shandong Province, China
Zitieren
APA
Shen, J., Guo, Z., Ding, W., Shentu, G., & Ma, H. (2026). Quantum gated recurrent neural networks for continuous dynamic temporal decision-making. Physica Scripta, 101(34), 346001. https://doi.org/10.1088/1402-4896/ae99cd
Vancouver
Shen J, Guo Z, Ding W, Shentu G, Ma H. Quantum gated recurrent neural networks for continuous dynamic temporal decision-making. Physica Scripta. 2026;101(34):346001. doi:10.1088/1402-4896/ae99cd
BibTeX
@article{Shen2026,
author = {Shen, Jiazhao and Guo, Zijun and Ding, Wei and Shentu, Guoliang and Ma, Hongyang},
title = {Quantum gated recurrent neural networks for continuous dynamic temporal decision-making},
journal = {Physica Scripta},
year = {2026},
volume = {101},
number = {34},
pages = {346001},
doi = {10.1088/1402-4896/ae99cd}
}
EUVIMED