一种支持实验设计的智能体推演仿真系统设计  

A combat simulation system developed based on agent theory

作　　者：张恒[1] 刘佳[1] 王晓路[1] ZHANG Heng;LIU Jia;WANG XiaoLu(China Academy of Vehicle Technology,Beijing 100076,China)

机构地区：[1]中国运载火箭技术研究院,中国北京100076

出　　处：《计算机仿真》2020年第12期16-19,71,共5页Computer Simulation

摘　　要：为实现基于蒙特卡洛的作战推演仿真,基于Agent智能体建模理论和仿真技术定制开发了一套具有实验设计与批量仿真功能的态势推演系统。系统由Agent仿真实体、态势推演引擎、用户界面组成,支持二维战场态势、三维电子沙盘展示功能。通过引入实验设计算法,针对模型的初始状态参数、组件参数和任务参数三类参数进行实验设计并生成了满足仿真要求的随机参数样本。设计了想定文件的结构,将初始想定文件和多次运行参数文件分离,实现了高效率的批量仿真运行。系统具备了作战概念设计、作战模式验证以及武器装备的作战效能评估能力。In order to carry out combat deduction simulation based on Monte Carlo,we developed a system based on agent modeling theory and simulation technology to support design of experiments,multi-run simulation,2D bat-tlefield situation,and 3D electronic sand table display.The system is composed of agent entities,situation engine and user interface.By introducing the experimental design algorithm,the experimental design was carried out for the initial state parameters,component parameters and task parameters of the model.We customized a type of scenario file to realize the high efficiency batch simulation operation.The system can be used to design operational concepts,verify combat mode,and evaluate the effectiveness of weapons.

关 键 词：智能体 批量 作战推演 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]