Scholarlink

<p>The Internet of Things (IoT) is revolutionizing various sectors by connecting numerous devices, enabling them to collect and exchange data seamlessly. However, IoT networks’ vast scale and diverse nature present significant challenges, especially in constrained environments where resources such as bandwidth, computational power, and energy are limited. Efficient parallel processing is essential to manage the massive influx of data and maintain optimal system performance. Parallel processing in Constrained IoT Environments faces numerous challenges. To address this limitation in this research, we develop innovative methods to improve the performance and data retrieval efficiency of IoT systems. Initially, the Elliptic curve cryptography-based authentication is used to secure the overall communication. Integrating the Improved Differential Evolution Algorithm (IDEA) with the Concurrent Federated Reinforcement Learning (CFRL) method improves resource allocation and ensures efficient resource utilization. To handle the delay constraints efficiently, we integrate the Lyapunov Optimization and the Whale Optimization algorithm with Deep Reinforcement Learning (DRL-LOWOA) for optimal task offloading. After optimizing the task offloading, we focus on improving the accuracy of parallel data processing using the Deep Neutral Network (DNN). Finally, we propose a Markov Decision Process (MDP) algorithm to optimize the decision-making process under uncertain conditions. By using these methods, we can efficiently perform parallel processing in a constrained IoT environment. The performance of the proposed method is evaluated using key metrics of communication overhead, response time, accuracy, and latency.</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

Innovative approaches to parallel processing in constrained IoT environments

  • 期刊论文

  • 发表日期：

    2025年6月26日

查看全文
  • Nasser S. Albalawi

摘要

The Internet of Things (IoT) is revolutionizing various sectors by connecting numerous devices, enabling them to collect and exchange data seamlessly. However, IoT networks’ vast scale and diverse nature present significant challenges, especially in constrained environments where resources such as bandwidth, computational power, and energy are limited. Efficient parallel processing is essential to manage the massive influx of data and maintain optimal system performance. Parallel processing in Constrained IoT Environments faces numerous challenges. To address this limitation in this research, we develop innovative methods to improve the performance and data retrieval efficiency of IoT systems. Initially, the Elliptic curve cryptography-based authentication is used to secure the overall communication. Integrating the Improved Differential Evolution Algorithm (IDEA) with the Concurrent Federated Reinforcement Learning (CFRL) method improves resource allocation and ensures efficient resource utilization. To handle the delay constraints efficiently, we integrate the Lyapunov Optimization and the Whale Optimization algorithm with Deep Reinforcement Learning (DRL-LOWOA) for optimal task offloading. After optimizing the task offloading, we focus on improving the accuracy of parallel data processing using the Deep Neutral Network (DNN). Finally, we propose a Markov Decision Process (MDP) algorithm to optimize the decision-making process under uncertain conditions. By using these methods, we can efficiently perform parallel processing in a constrained IoT environment. The performance of the proposed method is evaluated using key metrics of communication overhead, response time, accuracy, and latency.