Fuzzy logic based assignable causes ranking system for control chart abnormity diagnosis

When using control chart patterns as signals to identify the cause for faster and easier process diagnosis, tradition method is hard to handle with the uncertainties, ambiguities and vagueness associated with the problem. Based on fuzzy logic, this paper develops a fuzzy inference system (FIS), composed by six sub modules. Each determines the intensity of corresponding causes based on degree of presence of each pattern. All the evidence supporting each cause from the unnatural patterns are aggregated using fuzzy connective operators and causes are prioritized according to the final aggregating results. The search can be done from the cause having highest priority when process goes out of control.