现代防御技术 ›› 2025, Vol. 53 ›› Issue (6): 122-133.DOI: 10.3969/j.issn.1009-086x.2025.06.013

• 指挥控制与通信 • 上一篇    

基于前景理论的网络防御决策三方演化博弈分析

胡航, 冯凯, 张玉臣, 刘鹏程, 谭晶磊   

  1. 信息工程大学,河南 郑州 450001
  • 收稿日期:2024-07-31 修回日期:2024-11-07 出版日期:2025-12-28 发布日期:2025-12-31
  • 作者简介:胡航(1994-),男,河南桐柏人。硕士生,研究方向为网络安全博弈与智能决策。
  • 基金资助:
    国家重点研发计划(2019YFB0801904)

Evolutionary Game Analysis of Network Defense Decision Based on Prospect Theory

Hang HU, Kai FENG, Yuchen ZHANG, Pengcheng LIU, Jinglei TAN   

  1. Information Engineering University,Zhengzhou 450001,China
  • Received:2024-07-31 Revised:2024-11-07 Online:2025-12-28 Published:2025-12-31

摘要:

针对网络攻防对抗中防御策略选取问题,为更好地适应不完全理性网络攻防场景,结合奖惩机制与前景理论,应用演化博弈模型刻画网络攻防博弈过程,通过求解动态演化方程研究了攻击方、防御方及监管方行为策略选取演化规律,对各主体决策行为进行了稳定性分析,并探讨了防御成本、奖惩机制及价值函数对各主体行为策略选取的影响。结果表明,初始决策概率不影响博弈系统总体演化趋势,均衡点稳定且可预测;增加防御投资能降低攻击强度及次数,但成本增加会抑制防御投入;监管方实施奖惩机制能遏制网络攻击行为,但会影响防御方投资防御策略积极性;引入价值函数后,攻防双方决策受到价值感知、风险偏好等影响会扩大感知损益与实际值的差距,进而改变博弈系统演化路径。

关键词: 网络攻防, 三方演化博弈, 前景理论, 演化稳定点, 最优防御策略

Abstract:

This study aims to address defense strategy selection in network attack-defense confrontations and adapt to incomplete rationality scenarios in network attack-defense better. Incorporating reward-punishment mechanisms and prospect theory, the study utilized an evolutionary game model to depict the network attack-defense game process. By solving the dynamic evolution equations, it investigated the evolutionary patterns of behavioral strategy selection among the attacker, defender, and regulator. It conducted a stability analysis of the decision-making behaviors of each game participant and explored the effect of defense costs, reward-punishment mechanisms, and value functions on the selection of behavioral strategies of each party. The results show that initial decision-making probabilities do not affect the overall evolutionary trend of the game system, and the evolutionary equilibrium points possess predictability and stability. Increasing defense investment can reduce the intensity and frequency of attacks, but increased costs may curb defense investment. Implementing reward-punishment mechanisms by the regulator can deter network attacks, but it may affect the defender’s enthusiasm for investing in defense strategies. After the introduction of a value function, the decision-making of both attackers and defenders is influenced by factors such as value perception and risk preference. This can expand the gap between perceived gains and losses and actual values, thereby changing the evolutionary path of the game system.

Key words: network attack-defense, tripartite evolutionary game, prospect theory, evolutionary stable point, optimal defense strategy

中图分类号: