Scholarlink

<p>This paper addresses the critical challenge of modeling coordinated cyberattacks in adversarial environments where attackers must adapt to progressive detection and neutralization. We introduce an augmented Markov game framework that captures multi-agent attack dynamics through a novel dummy state mechanism, maintaining team coordination despite agent losses. The framework’s theoretical contributions include a formal model of coordinated attacks under detection constraints and a symmetry-aware value iteration (SAVI) algorithm that achieves exponential state-space reduction while preserving optimality. Experimental evaluation demonstrates that coordinated strategies significantly outperform non-coordinated approaches, exhibiting superior success rates, enhanced detection evasion through adaptive path planning, and greater resilience to varying defense configurations. The framework provides both theoretical foundations and practical algorithms for analyzing coordinated attack strategies in adversarial environments with dynamic team composition, establishing a comprehensive approach to security analysis under realistic detection constraints.</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

A probabilistic attack graph-based model for coordinated attacks in sensor-defended networks

  • 期刊论文

  • 发表日期：

    2025年12月23日

查看全文
  • Romaric Mofouet,
  • Arnold Kouam,
  • Haoxiang Ma,
  • Jie Fu,
  • Charles Kamhoua,
  • Gabriel Deugoue

摘要

This paper addresses the critical challenge of modeling coordinated cyberattacks in adversarial environments where attackers must adapt to progressive detection and neutralization. We introduce an augmented Markov game framework that captures multi-agent attack dynamics through a novel dummy state mechanism, maintaining team coordination despite agent losses. The framework’s theoretical contributions include a formal model of coordinated attacks under detection constraints and a symmetry-aware value iteration (SAVI) algorithm that achieves exponential state-space reduction while preserving optimality. Experimental evaluation demonstrates that coordinated strategies significantly outperform non-coordinated approaches, exhibiting superior success rates, enhanced detection evasion through adaptive path planning, and greater resilience to varying defense configurations. The framework provides both theoretical foundations and practical algorithms for analyzing coordinated attack strategies in adversarial environments with dynamic team composition, establishing a comprehensive approach to security analysis under realistic detection constraints.