Scholarlink

<p>The rise of Industrial Internet of Things (IIoT) systems has revolutionized modern industries by enabling seamless integration between physical devices and digital frameworks. However, ensuring efficient device selection and optimizing resource utilization in such complex environments remains a significant challenge. This work proposes the development of an Optimized Reinforcement Learning-based Deep Q Network (ORL-DQN) for the device selection process in IIoT systems to enhance the efficiency of Digital Twin (DT) frameworks. Intelligent industrial scenarios are implemented using the IIoT concept. The DT helps to monitor the complicated network environment and model training state in this architecture. The suggested framework links virtual systems at the DT layer with IIoT devices at the physical layer. ORL-DQN is used to investigate and optimize actions in order to solve the device selection issue in IIoT with DT systems. The ORL-DQN combines the Deep Q Network (DQN) with the Election Optimizer Algorithm (EOA), where the optimal Q value is determined using EOA. The proposed approach achieves optimal device selection, mitigates malicious attacks, avoids inefficient devices, and reduces the straggler effect. This methodology enhances resource consumption efficiency and improves sample distribution. The proposed ORL-DQN method is implemented in Python, and its performance is evaluated. The suggested ORL-DQN framework reduces the straggler impact by about 10&#xa0;ms while improving efficiency by more than 15% when compared to conventional techniques. Additionally, it outperforms DRL-FL (0.91) and AFSGD-VP (0.92) with an Area under the Curve (AUC) of 0.94, confirming its superiority in resilience and device selection.</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

Optimized reinforcement learning-based deep q network for efficient device selection in IIoT systems with digital twin integration

  • 期刊论文

  • 发表日期：

    2026年6月20日

查看全文
  • D. Francina Sophiya,
  • V. Raja

摘要

The rise of Industrial Internet of Things (IIoT) systems has revolutionized modern industries by enabling seamless integration between physical devices and digital frameworks. However, ensuring efficient device selection and optimizing resource utilization in such complex environments remains a significant challenge. This work proposes the development of an Optimized Reinforcement Learning-based Deep Q Network (ORL-DQN) for the device selection process in IIoT systems to enhance the efficiency of Digital Twin (DT) frameworks. Intelligent industrial scenarios are implemented using the IIoT concept. The DT helps to monitor the complicated network environment and model training state in this architecture. The suggested framework links virtual systems at the DT layer with IIoT devices at the physical layer. ORL-DQN is used to investigate and optimize actions in order to solve the device selection issue in IIoT with DT systems. The ORL-DQN combines the Deep Q Network (DQN) with the Election Optimizer Algorithm (EOA), where the optimal Q value is determined using EOA. The proposed approach achieves optimal device selection, mitigates malicious attacks, avoids inefficient devices, and reduces the straggler effect. This methodology enhances resource consumption efficiency and improves sample distribution. The proposed ORL-DQN method is implemented in Python, and its performance is evaluated. The suggested ORL-DQN framework reduces the straggler impact by about 10 ms while improving efficiency by more than 15% when compared to conventional techniques. Additionally, it outperforms DRL-FL (0.91) and AFSGD-VP (0.92) with an Area under the Curve (AUC) of 0.94, confirming its superiority in resilience and device selection.