Type-2 fuzzy process capability indices for non-normal processes

Abstract

We developed type-2 fuzzy percentile-based standard PCIs for non-normal data via Clements' method. For modeling uncertainty and imprecision of data, interval type-2 fuzzy sets are utilized. To prove usefulness and applicability of the proposed type-2 fuzzy percentile-based standard PCIs, a numerical illustration is performed for the data randomly generated from Log-normal distribution. The results show that in comparison with their crisp types, the proposed type-2 fuzzy percentile-based standard PCIs for non-normal processes are more informative, sensitive and flexible to evaluate the process performance of the industrial processes.