Scholarlink

<p>The increasing integration of renewable energy resources into modern power systems has intensified the need for intelligent control strategies capable of addressing operational uncertainty, environmental constraints, and competing performance objectives. In this context, this study proposes a Quantum-Inspired Multi-Agent Reinforcement Learning (QI-MARL) framework for the decentralized and multi-objective optimization of large-scale wind farms operating under stochastic environmental conditions. The proposed framework combines amplitude-based probabilistic policy representations, federated constraint negotiation, entropy-regularized learning, and Monte Carlo–driven uncertainty modeling to simultaneously address energy production, acoustic compliance, and environmental considerations. A multi-fidelity simulation environment is developed to capture wind variability, turbine dynamics, and agent-level adaptive decision-making within a realistic coastal wind farm setting. Simulation results demonstrate that the proposed approach enhances energy generation while maintaining compliance with acoustic and environmental constraints and exhibits stable learning behavior across diverse stochastic scenarios. The principal contribution of this work lies in the integration of decentralized multi-agent learning, uncertainty-aware optimization, and constraint-sensitive decision-making within a unified control framework for wind energy systems. The findings provide a simulation-based proof-of-concept for intelligent wind farm management and highlight the potential of learning-driven optimization approaches for renewable energy applications. Nevertheless, further validation using real-world operational data and comprehensive benchmarking against established control and reinforcement learning methods remains necessary to assess practical deployment potential.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimizing quantum-inspired federated multi-agent reinforcement architectures for sustainable wind energy governance under uncertainty

  • 期刊论文

  • 发表日期：

    2026年8月3日

查看全文
  • Khamiss Cheikh,
  • EL Mostapha Boudi,
  • Rabi Rabi,
  • Hamza Mokhliss

摘要

The increasing integration of renewable energy resources into modern power systems has intensified the need for intelligent control strategies capable of addressing operational uncertainty, environmental constraints, and competing performance objectives. In this context, this study proposes a Quantum-Inspired Multi-Agent Reinforcement Learning (QI-MARL) framework for the decentralized and multi-objective optimization of large-scale wind farms operating under stochastic environmental conditions. The proposed framework combines amplitude-based probabilistic policy representations, federated constraint negotiation, entropy-regularized learning, and Monte Carlo–driven uncertainty modeling to simultaneously address energy production, acoustic compliance, and environmental considerations. A multi-fidelity simulation environment is developed to capture wind variability, turbine dynamics, and agent-level adaptive decision-making within a realistic coastal wind farm setting. Simulation results demonstrate that the proposed approach enhances energy generation while maintaining compliance with acoustic and environmental constraints and exhibits stable learning behavior across diverse stochastic scenarios. The principal contribution of this work lies in the integration of decentralized multi-agent learning, uncertainty-aware optimization, and constraint-sensitive decision-making within a unified control framework for wind energy systems. The findings provide a simulation-based proof-of-concept for intelligent wind farm management and highlight the potential of learning-driven optimization approaches for renewable energy applications. Nevertheless, further validation using real-world operational data and comprehensive benchmarking against established control and reinforcement learning methods remains necessary to assess practical deployment potential.