DEVELOPING ONE-SIDED SPECIFICATION SIX-SIGMA FUZZY QUALITY INDEX AND TESTING MODEL TO MEASURE THE PROCESS PERFORMANCE OF FUZZY INFORMATION

Depending on the quality characteristic, a process capability index (PCI) can be used for one-sided specifications or for bilateral specifications. They are widely employed in the evaluation and improvement of process performance and are also good tools for communication between sales departments and vendors. The earliest PCIs for one-sided specifications were proposed by Kane (1986) and have one-to-one relationships with yield. A number of researchers have investigated the statistical properties of one-sided specification indices and proposed methods for applications. The later introduction of the Six Sigma approach also assisted many firms in effectively enhancing their production capacities, reducing waste, and increasing effectiveness. Like the PCIs, the Six Sigma quality levels have also become good quality control methods and communication tools in the industry. As a result, many researchers have examined the relationships between PCIs and the Six Sigma and their applications so that both approaches can be easily used to solve engineering problems in practice. In view of this, Chen et al. (2017a) modified the PCI for one-sided specifications and proposed the Six Sigma Quality Index (SSQI), which coincidently equals the quality level and has a one-to-one relationship with yield. However, uncertainty in quality characteristic measurements is common in practice, which can lead to judgment errors in conventional process capability assessment methods. In these cases, fuzzy calculation methods can determine the actual process capability more precisely. This study therefore developed an SSQI for one-sided specifications based on the fuzzy testing method created by Buckley (2005) and developed a Six Sigma fuzzy evaluation index and testing model. In addition to having a simpler calculation procedure, the model takes the process capability and Six Sigma quality level into consideration and can process the uncertainties in the data to make it more convenient for the industry to solve engineering issues. Finally, we presented a practical example to demonstrate the applications. The model proposed in this study can provide the industry with a practical approach to assess process quality in a fuzzy environment.