基于深度强化学习的防空反导智能任务分配  

Intelligent Task Assignment Research for Air Defense and Anti-missiles Based on Deep Reinforcement Learning

作　　者：刘家义 王刚[2] 夏智权 王思远 付强[2] LIU Jiayi;WANG Gang;XIA Zhiquan;WANG Siyuan;FU Qiang(Joint Operations College,National Defense University,Shijiazhuang 050000,China;Air and Missile Defense College,Air Force Engineering University,Xi'an 710051,China;Unit 93126 of PLA,Beijing 100000,China)

机构地区：[1]国防大学联合作战学院,石家庄050000 [2]空军工程大学防空反导学院,西安710051 [3]解放军93126部队,北京100000

出　　处：《火力与指挥控制》2024年第1期43-48,55,共7页Fire Control & Command Control

基　　金：国家自然科学基金资助项目(62106283)。

摘　　要：随着作战双方不断采用新技术,信息时代的战争呈现出强博弈对抗性。在分析防空反导任务分配过程和决策的本质基础上,从敌我两个角度深入探讨了强博弈对抗环境下防空反导任务分配所面临的挑战。讨论了基于深度强化学习的防空反导智能任务分配方法的优势,提出了其实际应用所面临的问题,有望解决相关问题的技术途径和方法评价指标,为防空反导智能任务分配提供新思路。With the continuous adoption of new technologies by both combatants,warfare in the infor-mation age has taken on a strongly game-based adversarial nature.Based on the analysis of the process of air defence and anti-missile mission assignment and the nature of decision-making,the challenges faced by air defense and anti-missile mission assignment in a strong game confrontation environment are ex-plored in depth from both the perspectives of enemy and us.The advantages of the deep reinforcement learning-based intelligent task assignment method for air defense and anti-missile defense are discussed,and the key problems faced by the practical application of intelligent task allocation,the promising techni-cal ways to solve the relevant problems and the method evaluation indexes are proposed to provide new ideas for intelligent task assignment for air defense and anti-missile defense.

关 键 词：强博弈对抗 防空反导 深度强化学习 任务分配 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术] TJ013[自动化与计算机技术—计算机科学与技术]