贝叶斯网络仿真元模型及其在装备战场损伤仿真中的应用研究

The Study of Simulation Metamodel Based on Bayesian Network and Its Application in Equipment Battlefield Damage Simulation

结合装备战场损伤仿真系统,研究了贝叶斯网络仿真元模型的构建方法.从条件概率角度描述了仿真模型输入参数与输出参数之间的映射关系,研究了构建贝叶斯网络仿真元模型的可行性,分析了贝叶斯网络仿真元模型的优点;研究了贝叶斯网络仿真元模型构建过程中的关键问题,包括:元模型参数的确定、原始模型参数向贝叶斯网络节点的转化、联结强度的计算、衍生元模型的构建;针对不完全信息条件下装备战场损伤快速定位问题,研究了基于K2算法的贝叶斯网络仿真元模型构建方法;构建了某型高炮的战场损伤贝叶斯网络仿真元模型.

The construction method of simulation metamodel based on Bayesian network is studied for the equipment battlefield damage simulation system. The mapping relation be- tween input parameters and output parameters of simulation model is described from the po- sition of condition probability, the feasibility of construction of metamodel based on Bayesian network is studied, the merits of metamodel based on Bayesian network are analyzed. The key problems are studied in the process of construction of metamodel based on Bayesian network, including the ensuring of metamodel parameters, the translation from original model param- eters to Bayesian network nodes, the calculation of joining intensity, and the construction of derived metamodel. Aiming at equipment battlefield damage fast location without complete information, the construction method of metamodel based on Bayesian network is studied based on the K2 method. The metamodel based on Bayesian network is constructed for one type of antiaircraft artillery.