摘要
武器装备体系的脆性是其固有特性,直接关系到武器装备体系作战效能的稳定发挥.武器装备体系的脆性分析主要研究体系的子系统在遭受内外部干扰或破坏后,体系出现整体损伤或崩溃的现象,并分析导致体系崩溃的关键因素.本文利用贝叶斯网络对体系内部子系统之间复杂的相互作用和因果关系进行建模,从而将体系的脆性分析问题转化为体系崩溃的后验概率计算问题,进而可通过贝叶斯网络的信念传播算法和MCMC随机采样来有效计算.本文还提出将体系内部的正反馈回路建模为动态贝叶斯网络,从而可通过网络的概率推断,动态分析体系脆性的内部传导机制.本文提出的脆性分析方法可为武器装备体系的设计与定量分析提供一种有效的技术手段.
The brittleness of a weapon equipment system of systems is an inherent characteristic of the system and directly affects its stable exertion in combat effectiveness.A brittleness analysis of a weapon equipment system of systems mainly studies the phenomenon of damage or collapse of the system after the subsystems of the system suffer from internal and external interference or destruction and analyzes the key factors leading to the collapse of the system.In this paper,Bayesian networks are used to model the complex interactions and causal relationships between subsystems within a system.This allows the brittleness analysis problem of the system to be converted into a posterior probability calculation problem of system collapse,which can be effectively analyzed through the belief propagation algorithm of Bayesian networks or MCMC random sampling.This paper also proposes to model the positive feedback loop inside the system as a dynamic Bayesian network so that the internal transmission mechanism of system brittleness can be dynamically analyzed through the probability inference of the network.The brittleness analysis method proposed in this paper provides an effective technical means for designing and quantitatively analyzing weapon system of systems.
作者
高亮
缪文民
姜晓辉
陈龙
叶军
张永乐
GAO Liang;MIAO WenMin;JIANG XiaoHui;CHEN Long;YE Jun;ZHANG YongLe(China Research and Development Academy of Machinery Equipment,Beijing 100089,China;Beijing Special Mechatronics Research Institute,Beijing 100012,China;Troops 78092 of PLA,Chengdu 610000,China)
出处
《中国科学：技术科学》
EI
CSCD
北大核心
2023年第9期1522-1532,共11页
Scientia Sinica(Technologica)
关键词
武器装备体系
脆性分析
贝叶斯网络
概率推断
信念传播算法
MCMC采样
weapon equipment system of systems
brittleness analysis
Bayesian networks
probability inference
belief propagation
MCMC sampling
