一种语义为中心的分布构件系统故障诊断建模方法

A Semantic Centric Fault Diagnostic Model Compiling Approach of Distributed Components Based System

随着分布构件技术的发展 ,复杂的分布构件系统故障诊断问题越来越受到人们的重视 .然而 ,现有的大多数分布构件系统故障诊断的研究没有充分考虑系统语义这一重要因素 .文章提出了一种新颖的语义为中心的分布构件系统故障诊断思想 .系统的语义主要由系统正常或者异常时的结构和行为特征构成 .在给出基于模型的故障诊断问题定义之后 ,提出了具有复杂时序语义表达能力的系统逻辑模型编辑方法 ,通过在Horn短句中增加时序描述来增强诊断模型的表达能力 .这种模型不仅对分布软构件系统具有较强的表达能力 ,还能够很好地利用现有的一阶逻辑定理证明器求解极小诊断问题 ,它并不需要具有高级时序逻辑处理能力的定理证明器的支持 .另外 ,通过哲学家就餐问题还从直观上阐明了这种新颖的模型编辑方法 .StarCCM的诊断子系统由诊断引擎、诊断代理、诊断回调接口构成 .它证明这种模型编辑方法能够有效地解决分布构件系统的语义诊断问题 .

With the development of distributed component technology,people set more and more focus on the diagnosis problem in such systems.Whereas,the current research on the diagnosis problem about distributed component systems often omits an important factor the system semantics.The system semantics are composed of the structure and behavior information under the normal and abnormal conditions.The paperproposes a novel ideal of diagnosis,which is a semantic centric method.After introducing principle of model based diagnosis,the paper puts forward a novel model compiling approach by providing the time series information in the Horn clauses,which can express complicated temporal semantics and compute the minimal diagnosis by the theorem provers of first order logic.Further more,the paper sets forth the model compiling method by a philosopher dinner problem.In the end,the diagnosis subsystem implemented in StarCCM is introduced,which consists of the diagnosis engines,diagnosis agents and diagnosis callback interfaces.The novel diagnosis model compiling approach is proved to be applied in the distributed component system effectively.