Scholarlink

<p>The rapid growth of IoT-driven applications such as smart cities, smart agriculture, and healthcare systems, is generating a large amount of data that needs efficient processing. To cover a wider area, and also for some limitations, battery-powered devices are part of an IoT-Fog network. These devices need energy-efficient task processing and offloading strategies to expand their operational lifespan. Furthermore, due to the large-scale nature of IoT, network traffic and connection failures can result in network congestion, task loss, and data degradation. In order to address these challenges, a decentralized AI-based approach has been proposed here, for task-offloading problems that consider energy consumption and network traffic. Experimental results demonstrate that this approach outperforms other methods, by effectively balancing energy efficiency, network traffic, and response time, ensuring reliable task execution and offloading in the IoT-Fog-Cloud continuum. The proposed method reduced deadline violations to near zero across all workload scenarios, improved response time by up to 50%, and achieved up to 60% battery savings compared to the No-Offload method. The statistical analysis using an ANOVA test confirms the improvements are statistically significant (p-value &lt; 0.05).</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

Ai-driven energy-aware task offloading with network traffic considerations in fog-cloud environments

  • 期刊论文

  • 发表日期：

    2025年9月3日

查看全文
  • Mohammad Zolghadri,
  • Parvaneh Asghari,
  • SeyedEbrahim Dashti,
  • Alireza Hedayati

摘要

The rapid growth of IoT-driven applications such as smart cities, smart agriculture, and healthcare systems, is generating a large amount of data that needs efficient processing. To cover a wider area, and also for some limitations, battery-powered devices are part of an IoT-Fog network. These devices need energy-efficient task processing and offloading strategies to expand their operational lifespan. Furthermore, due to the large-scale nature of IoT, network traffic and connection failures can result in network congestion, task loss, and data degradation. In order to address these challenges, a decentralized AI-based approach has been proposed here, for task-offloading problems that consider energy consumption and network traffic. Experimental results demonstrate that this approach outperforms other methods, by effectively balancing energy efficiency, network traffic, and response time, ensuring reliable task execution and offloading in the IoT-Fog-Cloud continuum. The proposed method reduced deadline violations to near zero across all workload scenarios, improved response time by up to 50%, and achieved up to 60% battery savings compared to the No-Offload method. The statistical analysis using an ANOVA test confirms the improvements are statistically significant (p-value < 0.05).