不完备信息的航空电子装备诊断规则提取方法

针对不完备信息引发的不确定性给航空电子装备的诊断规则提取带来的挑战，分别从广义狭义两个角度对故障诊断决策系统的不完备性进行定义，设计了一致性优先的相似度及属性值期望最大的缺失信息补齐算法，解决间接补齐算法存在不一致性问题；构建征兆属性概念格及诊断决策属性概念格，生成不完备诊断决策信息系统的扩充辨识矩阵，引入征兆属性概念等价关系计算最大一致征兆概念集，求解最大一致征兆概念辨识函数的析取范式获取最优约简属性集，根据约简后的诊断决策信息系统获取诊断规则。以某型航空装备的武器系统发射系统为例对方法验证，诊断结果准确率达到83.3%，高于现有典型方法，该方法在不完备信息处理、精确度及对象描述的直观简洁性方面具有显著优势。

A Fault Diagnostic Rules Gaining Method for Aviation Electronic Equipment under Incomplete Information Condition

Uncertainty caused by incomplete information brings great challenges to fault diagnostic rules gaining method for aviation electronic equipment. In order to solve the problem, by defining the incompleteness from the two angles of narrow sense and broad sense respectively, the consistency first completer algorithm is designed to solve the inconsistency problem caused by incomplete information based on maximal confidence and attribute value expectation, which exists in indirect completing algorithms. The symptom attributes concept lattice and diagnostic decision-making attributes concept lattice are constructed, and the equivalent relationship on the symptom attributes concept is introduced. Based on this, the disjunctive normal form discernibility function of the maximum inconsistent symptom attributes concept set is computed, then the optimal reduced attribute set is obtained based on which diagnostic rules are gained. Taking the weapon launching system as an example, the approach is validated with a precision of 83.3%. It can be concluded that the approach is better than the existing representative approach in dealing with incomplete information, accuracy and intuitionism of the diagnosis knowledge display.