The principal aim of this study was to provide more realistic output data based on imprecise measurements of product quality. The real‐world problems in fuzzy testing and selecting better processes performance are considered.
The Taguchi index, which provides numerical measures on process performance, has been widely used in the industry. In practice, the Taghchi index is estimated by sample data, thus it is of interest to obtain the confidence limits of the estimate Cpm for assessing processes. In addition, it is much more realistic, because in general the output quality characteristics of continuous quantities are more or less imprecise. Using the approach taken by Buckley, with some extensions, a general method is used to combine the vector of fuzzy numbers to produce the membership function of a fuzzy estimator of Cpm for further fuzzy testing and selection of better process performances.
As the rapid advancement of manufacturing technology occurs, current firms are increasing their levels of out sourcing and are relying more heavily on their supply chain as a source of their competitive advantage. Supplier selection decisions have become an important component of production and logistics management. Those decisions have a significant impact on manufacturers' ability to compete as purchases from outside suppliers may account for a large proportion of a product's costs.
The authors assume that measurements are taken from normally distributed populations in this research. Using fuzzy inference to assess manufacturing process capability processed using imprecise data under mild and severe departures from normality would be an interesting issue for further research.
From a managerial standpoint, considering stochastic uncertainty and fuzziness of data during testing and selecting the better supplier often provides a strong incentive to suppliers to adhere to the conscious gathering of data and variance reductions, as well as to quality requirements and standards.
An obvious advantage of process capability analysis over traditional classical approaches, which use binomial distribution for estimating low fractions of NC, is that reviewing a smaller sample reduces time, effort, and expenses.
