基于随机决策准则的改进多级影响图机动决策方法

针对传统影响图机动决策方法需要由飞行员预先给出空战态势评估难于准确确定的先验似然函数,将随机决策理论引入到影响图机动决策当中,提出了基于随机决策准则的改进多级影响图机动决策方法. 该方法仅需在决策时刻评估潜在空战态势,依据决策准则便可以产生机动决策. 为了优化该方法,基于典型随机决策准则,分别设计了5种机动决策准则,在典型空战想定条件下仿真分析了各机动决策准则的特点. 仿真结果表明:在态势劣势时,后悔值准则决策效果较好;在态势优势时,悲观准则、乐观准则、折中准则、等概率准则决策效果较好. 

Improved Multistage Influence Diagram Maneuvering Decision Method Based on Stochastic Decision Criterions

The traditional influence diagram maneuvering decision method requires pilot's determination of the prior likelihood function for situation assessment, but the prior likelihood function is hard to be determined precisely. In order to solve this problem, this work introduces the stochastic decision theory into the influence diagram and presents an improved multistage influence diagram maneuvering decision method. The new method only needs to assess potential air combat situations at decision stage and the maneuvers decision could be generated according to the decision criterions. To optimize the presented method, this research designs five types of maneuvering decision criterion and, under typical air combat situations, numerically analyzes their characters respectively.