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

arXiv:2505.06300

ARDNS-FN-Quantum: A Quantum-Enhanced Reinforcement Learning Framework with Cognitive-Inspired Adaptive Exploration for Dynamic Environments

Umberto Gonçalves de Sousa

Published May 12, 2025• 2 citations
arXivHTMLPDF

Abstract

Reinforcement learning (RL) has transformed sequential decisionmaking, yet traditional algorithms like Deep Q-Networks (DQNs) and Proximal Policy Optimization (PPO) often struggle with efficient exploration, stability, and adaptability in dynamic environments. This study presents ARDNS-FN-Quantum (Adaptive Reward-Driven Neural Simulator with Quantum enhancement), a novel framework that integrates a 2-qubit quantum circuit for action selection, a dual-memory system inspired by human cognition, and adaptive exploration strategies modulated by reward variance and curiosity. Evaluated in a 10X10 grid-world over 20,000 episodes, ARDNS-FN-Quantum achieves a 99.5% success rate (versus 81.3% for DQN and 97.0% for PPO), a mean reward of 9.0528 across all episodes (versus 1.2941 for DQN and 7.6196 for PPO), and an average of 46.7 steps to goal (versus 135.9 for DQN and 62.5 for PPO). In the last 100 episodes, it records a mean reward of 9.1652 (versus 7.0916 for DQN and 9.0310 for PPO) and 37.2 steps to goal (versus 52.7 for DQN and 53.4 for PPO). Graphical analyses, including learning curves, steps-to-goal trends, reward variance, and reward distributions, demonstrate ARDNS-FN-Quantum's superior stability (reward variance 5.424 across all episodes versus 252.262 for DQN and 76.583 for PPO) and efficiency. By bridging quantum computing, cognitive science, and RL, ARDNS-FN-Quantum offers a scalable, human-like approach to adaptive learning in uncertain environments, with potential applications in robotics, autonomous systems, and decision-making under uncertainty.

DOI
10.36227/techrxiv.174702508.85583877/v1

Related papers

Q-ARDNS-Multi: A Multi-Agent Quantum Reinforcement Learning Framework with Meta-Cognitive Adaptation for Complex 3D Environments

Umberto Gonçalves de Sousa

This paper presents Q-ARDNS-Multi, an advanced multi-agent quantum reinforcement learning (QRL) framework that extends the ARDNS-FN-Quantum model, where Q-ARDNS-Multi stands for "Quantum Adaptive Reward-Driven Neural Simulator - Multi-Agent". It integrates quantum circuits with RY gates, meta-cognitive adaptation, and multi-agent coordination mechanisms for complex 3D environments. Q-ARDNS-Multi leverages a 2-qubit quantum circuit for action selection, a dual-memory system inspired by human cognition, a shared memory module for agent cooperation, and adaptive exploration strategies modulated by reward variance and intrinsic motivation. Evaluated in a $10 \times 10 \times 3$ GridWorld environment with two agents over 5000 episodes, Q-ARDNS-Multi achieves success rates of 99.6\% and 99.5\% for Agents 0 and 1, respectively, outperforming Multi-Agent Deep Deterministic Policy Gradient (MADDPG) and Soft Actor-Critic (SAC) in terms of success rate, stability, navigation efficiency, and collision avoidance. The framework records mean rewards of $-304.2891 \pm 756.4636$ and $-295.7622 \pm 752.7103$, averaging 210 steps to goal, demonstrating its robustness in dynamic settings. Comprehensive analyses, including learning curves, reward distributions, statistical tests, and computational efficiency evaluations, highlight the contributions of quantum circuits and meta-cognitive adaptation. By bridging quantum computing, cognitive science, and multi-agent RL, Q-ARDNS-Multi offers a scalable, human-like approach for applications in robotics, autonomous navigation, and decision-making under uncertainty.

arXiv:2506.03205•Jun 2, 2025•1 citation

A Quantum-Enhanced Framework for Human-Like Reinforcement Learning: ARDNS-P-Quantum with Piagetian Stages

Umberto Gonçalves de Sousa

Traditional reinforcement learning (RL) models like Deep Q-Networks (DQNs) often lack human-like capabilities such as multi-timescale memory, adaptive plasticity, and cognitive developmental stages. We introduce ARDNS-P-Quantum, an enhanced version of the Adaptive Reward-Driven Neural Simulator with Piagetian Developmental Stages (ARDNS-P), integrating quantum computing to improve action selection. Using a 2-qubit quantum circuit with RY rotations and 16 shots, ARDNS-P-Quantum leverages a dual-memory system and Piaget's developmental stages to achieve superior performance in a 10×10 grid-world. Over 20,000 episodes, ARDNS-P Quantum attains a 99.2% goal-reaching success rate (vs. 91.9% for ARDNS-P and 84.5% for DQN), reduces steps to goal to 33.3 (vs. 149.2 for ARDNS-P and 72.1 for DQN), and increases mean rewards to 9.1169 (vs. 9.12 for ARDNS-P and 3.2207 for DQN) in the last 100 episodes. The quantum circuit design, with fewer shots and optimized depth, enhances exploration efficiency. This framework bridges RL, neuroscience, developmental psychology, and quantum computing, offering a scalable approach for human-like learning in dynamic environments.

•Apr 18, 2025•1 citation

Quantum-Enhanced Reinforcement Learning for Adaptive Decision-Making in Dynamic Environments

Priyanka Singh, Sonam Dubey

Many challenges that occur with traditional reinforcement learning in uncertain environments, such as slow convergence and scalability, can be addressed with Quantum-Enhanced Reinforcement Learning (QERL). Specifically, this research is for adaptive decision-making in environments with rapid change and great uncertainty. A QERL framework that combines variational quantum circuits and deep Q-networks is introduced. This framework optimizes decision-making processes in real-time and is applicable in areas such as autonomous systems, smart manufacturing, and intelligent resource allocation. Training for the framework took place in simulated dynamic environments during 50,000 training episodes and was then evaluated alongside Deep Q-Learning and Proximal Policy Optimization models. The QERL self-model outperformed the baseline in policy accuracy (+18.6%) and cumulative reward loss (-21.4%). Decision latency decreased 16.2%, and convergence improved 23.8% with the use of a hybrid quantum architecture. The proposed analysis demonstrated improved average reward (78.4 to 94.7) with 95% CI, indicating model consistency. Ultimately, QERL improves decision-making in complex scenarios and provides robust adaptive learning. Also noted, QERL frameworks with hybrid quantum architecture are useful for real-world systems. Furthermore, the learning framework proved to be more stable and adaptable to rapid and stochastic environmental changes compared to classical agents, in environments with incomplete dynamic and rapidly changing environments.

•Aug 1, 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-Inspired Algorithms for Reinforcement Learning

Jincheng Zhang

This paper investigates the application of quantum-inspired algorithms to the field of reinforcement learning (RL). Traditional RL methods often struggle with exploration challenges and can be computationally expensive, particularly in complex environments. We propose a novel RL algorithm that leverages concepts from quantum mechanics, specifically superposition and entanglement, to address these limitations. The core idea is to represent states, actions, and rewards using quantum-inspired representations, allowing the agent to explore the environment more efficiently and potentially converge to optimal policies faster. This work presents a detailed formulation of the algorithm, including the mathematical framework and key considerations for implementation. The potential benefits of this approach, including improved exploration and reduced computational complexity, are discussed. The algorithm is designed to iteratively refine its policy based on simulated quantum measurements, aiming to minimize the expected return. The exploration strategy is guided by the inherent probabilistic nature of quantum states, leading to a more diverse and potentially faster learning process. Further research is warranted to fully explore the capabilities and limitations of this approach across various RL benchmark problems.

•Sep 1, 2026•0 citations

Quantum-Inspired Neural Networks for Reinforcement Learning with Delayed Rewards

Jincheng Zhang

Reinforcement learning (RL) faces significant challenges when dealing with environments characterized by delayed rewards and sparse feedback. Traditional RL algorithms often struggle to efficiently learn optimal policies in such scenarios due to the difficulty in associating actions with their eventual consequences. This research proposes a novel approach to RL leveraging quantum-inspired concepts – superposition and entanglement – to improve exploration and exploitation within the state space. The core idea involves designing a neural network architecture where nodes exist in a superposition of states, allowing the agent to concurrently evaluate multiple actions. Furthermore, entangled node interactions are implemented to facilitate the rapid and efficient transfer of information across the network, effectively addressing the problem of delayed rewards. We demonstrate, theoretically and through a conceptual framework, the potential benefits of this approach, highlighting its capacity to improve learning speed and robustness in complex RL environments. The network architecture is defined as follows: Let $N = (n_1, n_2, ..., n_K)$ represent the neural network, where $n_i$ denotes the i-th node. Each node $n_i$ is associated with a state vector $s_i = (s_{i1}, s_{i2}, ..., s_{id})$ and an associated potential $V_i$. The superposition state of node $n_i$ is denoted as $|n_i\rangle = \alpha_i |0\rangle + \beta_i |1\rangle$, where $\alpha_i$ and $\beta_i$ are complex amplitudes. Entanglement is represented by a correlation function $C(n_i, n_j)$ which quantifies the degree of interconnectedness between nodes $n_i$ and $n_j$. The learning process involves updating the node potentials based on the expected future reward, considering the superposition state and entanglement interactions. The central focus is on the theoretical framework and the design principles, rather than a fully realized experimental implementation.

•Sep 5, 2026•0 citations
Reinforcement learning (RL) has transformed sequential decisionmaking, yet traditional algorithms like Deep Q-Networks (DQNs) and Proximal Policy Optimization (PPO) often struggle with efficient exploration, stability, and adaptability in dynamic environments. This study presents ARDNS-FN-Quantum (Adaptive Reward-Driven Neural Simulator with Quantum enhancement), a novel framework that integrates a 2-qubit quantum circuit for action selection, a dual-memory system inspired by human cognition, and adaptive exploration strategies modulated by reward variance and curiosity. Evaluated in a 10X10 grid-world over 20,000 episodes, ARDNS-FN-Quantum achieves a 99.5% success rate (versus 81.3% for DQN and 97.0% for PPO), a mean reward of 9.0528 across all episodes (versus 1.2941 for DQN and 7.6196 for PPO), and an average of 46.7 steps to goal (versus 135.9 for DQN and 62.5 for PPO). In the last 100 episodes, it records a mean reward of 9.1652 (versus 7.0916 for DQN and 9.0310 for PPO) and 37.2 steps to goal (versus 52.7 for DQN and 53.4 for PPO). Graphical analyses, including learning curves, steps-to-goal trends, reward variance, and reward distributions, demonstrate ARDNS-FN-Quantum's superior stability (reward variance 5.424 across all episodes versus 252.262 for DQN and 76.583 for PPO) and efficiency. By bridging quantum computing, cognitive science, and RL, ARDNS-FN-Quantum offers a scalable, human-like approach to adaptive learning in uncertain environments, with potential applications in robotics, autonomous systems, and decision-making under uncertainty.
This paper presents Q-ARDNS-Multi, an advanced multi-agent quantum reinforcement learning (QRL) framework that extends the ARDNS-FN-Quantum model, where Q-ARDNS-Multi stands for "Quantum Adaptive Reward-Driven Neural Simulator - Multi-Agent". It integrates quantum circuits with RY gates, meta-cognitive adaptation, and multi-agent coordination mechanisms for complex 3D environments. Q-ARDNS-Multi leverages a 2-qubit quantum circuit for action selection, a dual-memory system inspired by human cognition, a shared memory module for agent cooperation, and adaptive exploration strategies modulated by reward variance and intrinsic motivation. Evaluated in a 10×10×310 \times 10 \times 310×10×3 GridWorld environment with two agents over 5000 episodes, Q-ARDNS-Multi achieves success rates of 99.6\% and 99.5\% for Agents 0 and 1, respectively, outperforming Multi-Agent Deep Deterministic Policy Gradient (MADDPG) and Soft Actor-Critic (SAC) in terms of success rate, stability, navigation efficiency, and collision avoidance. The framework records mean rewards of −304.2891±756.4636-304.2891 \pm 756.4636−304.2891±756.4636 and −295.7622±752.7103-295.7622 \pm 752.7103−295.7622±752.7103, averaging 210 steps to goal, demonstrating its robustness in dynamic settings. Comprehensive analyses, including learning curves, reward distributions, statistical tests, and computational efficiency evaluations, highlight the contributions of quantum circuits and meta-cognitive adaptation. By bridging quantum computing, cognitive science, and multi-agent RL, Q-ARDNS-Multi offers a scalable, human-like approach for applications in robotics, autonomous navigation, and decision-making under uncertainty.
Traditional reinforcement learning (RL) models like Deep Q-Networks (DQNs) often lack human-like capabilities such as multi-timescale memory, adaptive plasticity, and cognitive developmental stages. We introduce ARDNS-P-Quantum, an enhanced version of the Adaptive Reward-Driven Neural Simulator with Piagetian Developmental Stages (ARDNS-P), integrating quantum computing to improve action selection. Using a 2-qubit quantum circuit with RY rotations and 16 shots, ARDNS-P-Quantum leverages a dual-memory system and Piaget's developmental stages to achieve superior performance in a 10×10 grid-world. Over 20,000 episodes, ARDNS-P Quantum attains a 99.2% goal-reaching success rate (vs. 91.9% for ARDNS-P and 84.5% for DQN), reduces steps to goal to 33.3 (vs. 149.2 for ARDNS-P and 72.1 for DQN), and increases mean rewards to 9.1169 (vs. 9.12 for ARDNS-P and 3.2207 for DQN) in the last 100 episodes. The quantum circuit design, with fewer shots and optimized depth, enhances exploration efficiency. This framework bridges RL, neuroscience, developmental psychology, and quantum computing, offering a scalable approach for human-like learning in dynamic environments.
Many challenges that occur with traditional reinforcement learning in uncertain environments, such as slow convergence and scalability, can be addressed with Quantum-Enhanced Reinforcement Learning (QERL). Specifically, this research is for adaptive decision-making in environments with rapid change and great uncertainty. A QERL framework that combines variational quantum circuits and deep Q-networks is introduced. This framework optimizes decision-making processes in real-time and is applicable in areas such as autonomous systems, smart manufacturing, and intelligent resource allocation. Training for the framework took place in simulated dynamic environments during 50,000 training episodes and was then evaluated alongside Deep Q-Learning and Proximal Policy Optimization models. The QERL self-model outperformed the baseline in policy accuracy (+18.6%) and cumulative reward loss (-21.4%). Decision latency decreased 16.2%, and convergence improved 23.8% with the use of a hybrid quantum architecture. The proposed analysis demonstrated improved average reward (78.4 to 94.7) with 95% CI, indicating model consistency. Ultimately, QERL improves decision-making in complex scenarios and provides robust adaptive learning. Also noted, QERL frameworks with hybrid quantum architecture are useful for real-world systems. Furthermore, the learning framework proved to be more stable and adaptable to rapid and stochastic environmental changes compared to classical agents, in environments with incomplete dynamic and rapidly changing environments.
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.
This paper investigates the application of quantum-inspired algorithms to the field of reinforcement learning (RL). Traditional RL methods often struggle with exploration challenges and can be computationally expensive, particularly in complex environments. We propose a novel RL algorithm that leverages concepts from quantum mechanics, specifically superposition and entanglement, to address these limitations. The core idea is to represent states, actions, and rewards using quantum-inspired representations, allowing the agent to explore the environment more efficiently and potentially converge to optimal policies faster. This work presents a detailed formulation of the algorithm, including the mathematical framework and key considerations for implementation. The potential benefits of this approach, including improved exploration and reduced computational complexity, are discussed. The algorithm is designed to iteratively refine its policy based on simulated quantum measurements, aiming to minimize the expected return. The exploration strategy is guided by the inherent probabilistic nature of quantum states, leading to a more diverse and potentially faster learning process. Further research is warranted to fully explore the capabilities and limitations of this approach across various RL benchmark problems.
Reinforcement learning (RL) faces significant challenges when dealing with environments characterized by delayed rewards and sparse feedback. Traditional RL algorithms often struggle to efficiently learn optimal policies in such scenarios due to the difficulty in associating actions with their eventual consequences. This research proposes a novel approach to RL leveraging quantum-inspired concepts – superposition and entanglement – to improve exploration and exploitation within the state space. The core idea involves designing a neural network architecture where nodes exist in a superposition of states, allowing the agent to concurrently evaluate multiple actions. Furthermore, entangled node interactions are implemented to facilitate the rapid and efficient transfer of information across the network, effectively addressing the problem of delayed rewards. We demonstrate, theoretically and through a conceptual framework, the potential benefits of this approach, highlighting its capacity to improve learning speed and robustness in complex RL environments. The network architecture is defined as follows: Let N=(n1,n2,...,nK)N = (n_1, n_2, ..., n_K)N=(n1​,n2​,...,nK​) represent the neural network, where nin_ini​ denotes the i-th node. Each node nin_ini​ is associated with a state vector si=(si1,si2,...,sid)s_i = (s_{i1}, s_{i2}, ..., s_{id})si​=(si1​,si2​,...,sid​) and an associated potential ViV_iVi​. The superposition state of node nin_ini​ is denoted as ∣ni⟩=αi∣0⟩+βi∣1⟩|n_i\rangle = \alpha_i |0\rangle + \beta_i |1\rangle∣ni​⟩=αi​∣0⟩+βi​∣1⟩, where αi\alpha_iαi​ and βi\beta_iβi​ are complex amplitudes. Entanglement is represented by a correlation function C(ni,nj)C(n_i, n_j)C(ni​,nj​) which quantifies the degree of interconnectedness between nodes nin_ini​ and njn_jnj​. The learning process involves updating the node potentials based on the expected future reward, considering the superposition state and entanglement interactions. The central focus is on the theoretical framework and the design principles, rather than a fully realized experimental implementation.