Scholarlink

<p>In today’s rapidly changing business environment, effective strategic and risk management is essential for enterprises to remain competitive and resilient. Reinforcement learning (RL) has attracted significant attention due to its ability to support decision-making in uncertain, complex, and dynamic environments. This study explores the role of RL in enterprise risk management by proposing a Skill-Optimized Efficient Deep Q-Network (SO-EDQN) framework for dynamic strategic decision-making. The proposed model utilizes historical and real-time enterprise data collected from multiple sources, including market trends, financial reports, operational logs, customer records, and external economic and regulatory indicators. Data preprocessing techniques, such as data cleaning and normalization, are applied to enhance input quality and ensure reliable learning behavior. The skill optimization (SO) mechanism acts as an adaptive control layer that improves the balance between exploration and exploitation, enabling the model to learn continuously from past experiences and adjust to changing risk conditions. Simulation-based experiments are conducted to evaluate the model’s ability to handle large, high-dimensional state spaces and support real-time risk mitigation strategies. The SO-EDQN framework achieves a precision of 98.51%, recall of 97.65%, accuracy of 98.81%, and an F1-score of 97.26%, demonstrating strong performance in managing dynamic enterprise risks. These results indicate that SO-EDQN offers a data-driven and adaptive approach for enterprise risk management in volatile and uncertain business environments.</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

The role of reinforcement learning algorithms in dynamic strategic management of enterprises

  • 期刊论文

  • 发表日期：

    2026年3月24日

查看全文
  • Qinling Li

摘要

In today’s rapidly changing business environment, effective strategic and risk management is essential for enterprises to remain competitive and resilient. Reinforcement learning (RL) has attracted significant attention due to its ability to support decision-making in uncertain, complex, and dynamic environments. This study explores the role of RL in enterprise risk management by proposing a Skill-Optimized Efficient Deep Q-Network (SO-EDQN) framework for dynamic strategic decision-making. The proposed model utilizes historical and real-time enterprise data collected from multiple sources, including market trends, financial reports, operational logs, customer records, and external economic and regulatory indicators. Data preprocessing techniques, such as data cleaning and normalization, are applied to enhance input quality and ensure reliable learning behavior. The skill optimization (SO) mechanism acts as an adaptive control layer that improves the balance between exploration and exploitation, enabling the model to learn continuously from past experiences and adjust to changing risk conditions. Simulation-based experiments are conducted to evaluate the model’s ability to handle large, high-dimensional state spaces and support real-time risk mitigation strategies. The SO-EDQN framework achieves a precision of 98.51%, recall of 97.65%, accuracy of 98.81%, and an F1-score of 97.26%, demonstrating strong performance in managing dynamic enterprise risks. These results indicate that SO-EDQN offers a data-driven and adaptive approach for enterprise risk management in volatile and uncertain business environments.