Scholarlink

<p>This paper proposed a novel approach integrating dual linear programming and a greedy algorithm for resource allocation in decentralized federated learning within Internet of Things (IoT) environments. The approach, (Privacy-Preserving Federated Learning with Adaptive Resource Allocation), seeks to optimize computational resource utilization among IoT devices engaged in federated learning while ensuring the preservation of privacy. The dual linear programming technique serves to model the resource allocation problem, addressing the heterogeneous computational capabilities and energy constraints inherent to IoT devices. A greedy algorithm is then applied to facilitate efficient, decentralized resource allocation, aimed at maximizing both convergence speed and model accuracy. The contributions of this research are twofold: it enhances the efficiency of federated learning systems operating in resource-constrained environments and introduces a robust mechanism for privacy preservation. Experimental results conducted on various datasets demonstrate the superiority of the proposed method in comparison to state-of-the-art techniques.</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

Privacy-preserving federated learning with adaptive resource allocation in internet of things

  • 期刊论文

  • 发表日期：

    2025年9月15日

查看全文
  • Majid Mohammadpour,
  • Seyedakbar Mostafavi

摘要

This paper proposed a novel approach integrating dual linear programming and a greedy algorithm for resource allocation in decentralized federated learning within Internet of Things (IoT) environments. The approach, (Privacy-Preserving Federated Learning with Adaptive Resource Allocation), seeks to optimize computational resource utilization among IoT devices engaged in federated learning while ensuring the preservation of privacy. The dual linear programming technique serves to model the resource allocation problem, addressing the heterogeneous computational capabilities and energy constraints inherent to IoT devices. A greedy algorithm is then applied to facilitate efficient, decentralized resource allocation, aimed at maximizing both convergence speed and model accuracy. The contributions of this research are twofold: it enhances the efficiency of federated learning systems operating in resource-constrained environments and introduces a robust mechanism for privacy preservation. Experimental results conducted on various datasets demonstrate the superiority of the proposed method in comparison to state-of-the-art techniques.