Control Charts Applications For Multivariate Attribute Processes

摘要

Two procedures were proposed for the construction of control charts to monitor multivariate process based on multi-dimenstional linguistic data. The first is based on probability theory and the second on the fuzzy theory. The performance of resulting control charts and their sensitivity to process shift are presented. These procedures are applied to a real data example. This real data application allow us to illustrate the implementation and the use of the proposed charts. Comparison of these charts is based on the ARL following a predefined process shift. A new procedure for interpreting out-of-control signals is also developed.