Scholarlink

<p>This research delves into the integration of multi-agent systems and Explainable Reinforcement Learning (XRL) to enhance threat detection mechanisms within mobile edge network traffic. The study aims to address the complexities introduced by the dynamic and distributed nature of mobile edge environments, which challenge conventional threat detection methods. By leveraging the collective intelligence of multi-agent systems and the interpretability of XRL, the proposed framework seeks to improve detection accuracy and provide comprehensible insights into decision-making processes. The research involves a comprehensive experimental setup, evaluating the system’s performance across various threat scenarios, including normal traffic, high loads, and sophisticated attacks. Experimental results demonstrate the efficacy of the approach in identifying diverse threat patterns with a high detection accuracy of 97% and low rates of false positives (3.2%) and false negatives (4.1%). The findings underscore the potential of XRL-enhanced multi-agent systems to offer a robust, adaptive, and transparent solution for bolstering network security within mobile edge infrastructures, paving the way for more effective and interpretable threat detection solutions in complex network 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

Collaborative multi-agent XRL for threat detection in mobile edge network traffic

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年9月10日

查看全文
  • Muhammad Yousaf Saeed,
  • Jingsha He,
  • Nafei Zhu,
  • Muhammad Farhan,
  • Ahmad Almadhor,
  • Thippa Reddy Gadekallu,
  • Saira Iqbal

摘要

This research delves into the integration of multi-agent systems and Explainable Reinforcement Learning (XRL) to enhance threat detection mechanisms within mobile edge network traffic. The study aims to address the complexities introduced by the dynamic and distributed nature of mobile edge environments, which challenge conventional threat detection methods. By leveraging the collective intelligence of multi-agent systems and the interpretability of XRL, the proposed framework seeks to improve detection accuracy and provide comprehensible insights into decision-making processes. The research involves a comprehensive experimental setup, evaluating the system’s performance across various threat scenarios, including normal traffic, high loads, and sophisticated attacks. Experimental results demonstrate the efficacy of the approach in identifying diverse threat patterns with a high detection accuracy of 97% and low rates of false positives (3.2%) and false negatives (4.1%). The findings underscore the potential of XRL-enhanced multi-agent systems to offer a robust, adaptive, and transparent solution for bolstering network security within mobile edge infrastructures, paving the way for more effective and interpretable threat detection solutions in complex network environments.