Research And Implement Of Measurement In Model-based Diagnosis

Model-based diagnosis is a tributary of artificial intelligence which has been called the revolution of the diagnostic theory and technology in recent years. Model-based diagnosis constructs the system model (backgroud theory) and its behavior expected. If the observed system behavior is not the same as the expected one, then we can use the system model to generate all candidates of the diagnosis. However, the candidates may be not just one, so a series of tests must be done to find out the real faulty components from the candidates. This is the measure of the diagnosis. The main purpose to study the method of measures is to help people find out real faulty components faster with the full results of the diagnoses.Reiter generated diagnoses based upon the conflict sets. They concluded that the candidates of diagnoses can be not only one and they can be generated by the given faulty system. One way to recognize the real diagnoses is to make measurements. HouAimin provided an efficient incremental method for computing new diagnoses given a new measurement, based on the previous diagnoses predicting the opposite. The method for computing new diagnoses resulting from a new measurement was also based on the concept of a conflict set and chose first-order logic as a language for representing such a system. Whatever one's choice of representation logic is, the description of a system will specify how that system normally behaves on the assumption that all its components are functioning correctly. If we have available an observation of the system's actual behavior and if this observation conflicts with (namely, is logically inconsistent with) the way the system is meant to behave, then there is a diagnostic problem. The problem is to determine those system components which, when assumed to be functioning...