基于强化学习的对抗意图识别  

Adversarial Intention Recognition Based on Reinforcement Learning

作　　者：白亮[1] 肖延东[1] 齐景涛 BAI Liang;XIAO Yandong;QI Jingtao(College of Systems Engineering,National University of Defense Technology,Changsha 410073,China)

机构地区：[1]国防科技大学系统工程学院,长沙410073

出　　处：《指挥与控制学报》2024年第1期112-116,共5页Journal of Command and Control

基　　金：国家自然科学基金(61902418)资助。

摘　　要：未来智能化战争复杂多变,敌我双方往往以对抗博弈情况出现,当我方作为攻击者时,如何有效隐藏我方意图实施有效打击,以及我方作为防御者时如何识别敌方意图实施有效防御,对战争局势的走向起到决定性作用。为把握战争局势,做到“知己知彼,百战不殆”,基于强化学习方法针对对抗意图识别问题提出了解决方案。仿真结果表明,基于强化学习被识别者可以有效地进行意图欺骗,而识别者可以通过意图识别设计方法有效地迫使被识别者暴露意图,从而采取攻防措施。In the future,intelligent warfare is complex and changeable.Both sides in warfare appear as adversarial roles.When our side acts as an attacker,how we can hide our intentions efficiently to carry out effective strike,and when our side acts as a defender,how we can recognize the enemy’s intentions and carry out effective defenses,which plays a decisive role in warfare.In order to control the situation of the war and achieve“know the enemy and know yourself,and you need not fear the result of a hundred battles”,a solution for the problem of adversarial intention recognition is proposed based on the reinforcement learning.The simulation results show that the attacker can effectively deceive the defender to hide their intentions based on reinforcement learning,and the defender can effectively force the attacker to expose their intention through the intention recognition design method,so as to take offensive and defensive measures.

关 键 词：强化学习 对抗意图识别 对抗博弈 智能化作战 

分 类 号：E91[军事]