Scholarlink

As new IoT technologies evolve quickly, digital twins (DT) are being suggested for a number of uses. It is anticipated that DT and a mobile network would combine to create a DTMN. Nevertheless, DTMN is vulnerable to serious security risks, and the defenses in place are mostly inert and only react when an assault takes place. In this study, we present a collaborative mutation-based MTD (CM-MTD) for DTMN, which uses two techniques to modify network attributes and interrupt several stages of the cyber death chain: HAM and RM. We use a SMDP to represent dynamic deployment of MTD schemes and time-varying security events. LSTM predicts security occurrences and eliminates impractical activities from the action space. Lastly, we develop a deep reinforcement learning system for collaborative scheduling that is hierarchical. In comparison to baseline solutions, simulation results show how successful CM-MTD is.

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 Moving Target Defense for DTMN: Collaborative Mutation-Based Moving Target Defense Based on Hierarchical Reinforcement Learning

  • 章节

  • 发表日期：

    2026年

查看全文
  • Tao Zhang,
  • Xiangyun Tang,
  • Jiawen Kang,
  • Changqiao Xu

摘要

As new IoT technologies evolve quickly, digital twins (DT) are being suggested for a number of uses. It is anticipated that DT and a mobile network would combine to create a DTMN. Nevertheless, DTMN is vulnerable to serious security risks, and the defenses in place are mostly inert and only react when an assault takes place. In this study, we present a collaborative mutation-based MTD (CM-MTD) for DTMN, which uses two techniques to modify network attributes and interrupt several stages of the cyber death chain: HAM and RM. We use a SMDP to represent dynamic deployment of MTD schemes and time-varying security events. LSTM predicts security occurrences and eliminates impractical activities from the action space. Lastly, we develop a deep reinforcement learning system for collaborative scheduling that is hierarchical. In comparison to baseline solutions, simulation results show how successful CM-MTD is.