定性因果与多项式算法的模型诊断方法

MODEL BASED DIAGNOSIS USING QUALITATIVE MODEL AND POLYNOMIAL ALGORITHM

在基于模型的诊断方面，Ｒｅｉｔｅｒ最初提出了使用第一原理的诊断方法，但它应用于复杂物理系统时遇到很多困难。后来，Ｍｉｋｉｔｏ等人提出了使用定性因果与覆盖集的模型诊断方法，此方法有很大改进，但覆盖集的计算复杂费时。提出一种使用定性因果与多项式算法的模型诊断方法，此方法可以明显地减少诊断的计算量，并且对这种新算法的正确性给予了证明。

In the model based diagnosis method,Reiter proposed firstly a method to calculate diagnosis from first principle,but it is difficult toapply to the complex physical domain.Afterward,Mikito Iwamasa et.al.presented a model based diagnosis using qualitative causal model and set covering,Which is improved greatly,but set covering calculating is complicated and needs long time.In the paper,a model based diagnosis using qualitative causal model and polynomial algorithm is presented,which has less computation burden and the correctness of this new algorithm is proved.