电子学报 ›› 2021, Vol. 49 ›› Issue (3): 527-535.DOI: 10.12263/DZXB.20191070

• 学术论文 • 上一篇    下一篇

基于多阶段Markov信号博弈的移动目标防御最优决策方法

蒋侣1, 张恒巍1,2, 王晋东1   

  1. 1. 战略支援部队信息工程大学三院, 河南郑州 450001;
    2. 河南省信息安全重点实验室, 河南郑州 450001
  • 收稿日期:2019-09-19 修回日期:2020-10-26 出版日期:2021-03-25
    • 通讯作者:
    • 张恒巍
    • 作者简介:
    • 蒋侣 男,1995年出生,四川广安人,战略支援部队信息工程大学硕士,助理工程师,主要研究方向为移动目标防御、网络安全与攻防对抗;王晋东 男,1966年出生,山西洪桐人,战略支援部队信息工程大学教授,主要研究方向为网络与信息安全、云资源管理.
    • 基金资助:
    • 河南省科技攻关计划基金 (No.182102210144)

A Markov Signaling Game-theoretic Approach to Moving Target Defense Strategy Selection

JIANG Lü1, ZHANG Heng-wei1,2, WANG Jin-dong1   

  1. 1. The Third Institute, PLA SSF Information Engineering University, Zhengzhou, Henan, 450001, China;
    2. Henan Key Laboratory of Information Security, Zhengzhou, Henan 450001, China
  • Received:2019-09-19 Revised:2020-10-26 Online:2021-03-25 Published:2021-03-25
    • Corresponding author:
    • ZHANG Heng-wei
    • Supported by:
    • Technology Research and Development Program Fund of Henan Province (No.182102210144)

摘要: 随着移动目标防御技术研究的不断深入,移动目标防御策略选取问题成为当前研究的热点问题之一,本文提出一种基于多阶段Markov信号博弈模型的移动目标防御最优策略选取方法.首先,结合攻防实际,提出实施攻击所需构建的攻击链模型.其次,在考虑状态随机跳变的基础上,将多阶段信号博弈模型与Markov决策过程相结合,构建基于多阶段Markov信号博弈的移动目标防御模型.同时,引入Logistic映射刻画攻防博弈系统中可能造成概率更新过程失真的随机干扰因素.在形式化建模的基础上,设计折扣收益目标函数,并提出均衡求解算法,给出最优防御策略选取算法.最后,通过仿真实验验证模型和方法的有效性.

关键词: 移动目标防御, Markov决策, 多阶段信号博弈, 最优策略选取, logistic映射

Abstract: With the development of the research on Moving Target Defense (MTD) technique, how to effectively select the optimal strategy of moving target defense has become an urgent issue in the current research. To solve this problem, we propose a MTD optimal strategy selection method based on multi-stage Markov signaling game model. Firstly, combined with the actual attack-defense process, we construct an attack chain model that attackers need to build to carry out the attack. Secondly, due to the random jump between states, we combine multi-stage signaling with Markov Decision Process (MDP) to construct the corresponding MTD model. Meanwhile, we adopt Logistic mapping to characterize the stochastic interference factors that may cause the distortion of the probability updating in the attack-defense process. Additionally, on the basis of formally modeling, we design an objective function with discounted total payoff. Besides, we give a solution method for multi-stage signaling game equilibrium and design an optimal defense strategy selection algorithm for MTD. Finally, the simulation demonstrates the effectiveness and feasibility of the proposed model and method.

Key words: moving target defense, Markov decision, multi-stage signaling game, optimal strategy selection, logistic mapping

中图分类号: