基于攻防信号博弈模型的防御策略选取方法.pdfVIP

Vo1 .37 No.5 通信学报 第 37 卷第5 期 Journa1 on Communications May2016 2016 年5 月 doi: 10.11959/j .issn.l 000-436x.20 16092 基于攻防信号博弈模型的防御策略选取方法 张恒巍，余定坤，韩继红，王晋东，李涛 (信息工程大学三院，河南郑州 45∞01) 摘 要:当前基于博弈理论的防御策略选取方法大多采用完全信息或静态博弈模型，为更加符合网络攻防实际， 从动态对抗和有限信息的视角对攻防行为进行研究。构建攻防信号博弈模型，对策略量化计算方法进行改进， 并提出精炼贝叶斯均衡求解算法。在博弈均衡分析的基础上，设计了最优防御策略选取算法。通过实验验证了 模型和算法的有效性，并在分析实验数据的基础上总结了攻防信号博弈的一般性规律，能够指导不同类型防御 者的决策。 关键词z 动态博弈:不完全信息:攻防信号博弈:精炼贝叶斯均衡:均衡分析:策略选取 中图分类号: TP309 文献标识码:A Defense policies selection method based on attack-defense signaling game model ZHANG Heng-wei, YU Ding-kun, HAN Ji-hong, WANG Jin圄dong， Ll Tao (The Third Institu钮， Information E~伊阳ing Universi:纱， Zhen阱ou450侧，China ) Abstract: Currently defense policies selection based on g缸ne theory mostly applied either the complete information game model or the static g缸ne mode l. In order to be more in line wi曲曲e reality of network a忧郁k and defense, at- tack-defense behavior was studied by dynamic rivalry and incomplete information.ηle attack-defense signaling 伊me mode was built, the method ωquantifY policies was improved and an a1gori也m to obtain the 严rfect Bayesian equilib- rium was proposed. On the basis of analyzing equilibrium, the a1gorithm for selecting the optimal def出se policy was proposed. The simulation experiment demonstrates 也at the model and a1gorithms are feasible and effective. By the ex- perimental 也钮， general rules on signaling 甜ack-defense game are summarized, which can guide defenders of different