基于可计算任务树的人机混合智能任务规划新范式  

A New Paradigmof Human-machine Hybrid Intelligent Mission Planning Based on Computational Mission Tree

作　　者：刘东红 LIU Donghong(Academy of Military Sciences,Beijing 100091,China)

机构地区：[1]军事科学院,北京100091

出　　处：《指挥与控制学报》2023年第1期85-92,共8页Journal of Command and Control

摘　　要：为加速智能化技术在任务规划中的应用,提出一种基于可计算任务树的人机混合智能任务规划新范式.以可计算任务树作为人机混合智能任务规划的桥梁,机器配合指挥员对任务进行分解,通过智能算法在虚拟对抗环境中开展海量训练,支持方案的自动生成.同时,给出了3种不同的使用模式与典型运行流程,并对关键支撑技术进行了分析.该范式为未来智能化任务规划提供了一种可行的技术途径,对提升任务规划能力和效率具有重要意义.To promote the application of intelligent techniques in mission planning,a new paradigm of human-machine hybrid intelligent mission planning based on computational mission tree is proposed.With the computational mission tree as the bridge for the planning of human-machine hybrid intelligent mission,the commander could decompose the mission into many sub-missions with the machine.The support schemes of carrying out lots of trainings could be generated automatically with intelligent algorithms in the virtual confrontation environment.Meanwhile,three different usage patterns,a typical running process are provided and some key support technologies are analyzed.The proposed paradigm provides a feasible technical approach for future intelligent mission planning,and is of great significance for improving the capability and efficiency of current mission planning.

关 键 词：智能任务规划 人机混合智能 可计算任务树 智能算法 虚拟对抗环境 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] E91[自动化与计算机技术—控制科学与工程]