Investigation of Variation and Uncertainty in Six Sigma

ABSTRACT Six Sigma is a widely used moniker to represent a set of design and manufacturing techniques, the aggregate of which should effectively improve quality and lower cost. Six Sigma implies that six standard deviations of performance should be maintained between the performance mean and the closest specification limit. When multiple performance attributes must be satisfied, aggregation techniques may be applied for evaluation and optimization. This paper evaluates two common aggregation techniques, the multi-process capability index and the process desirability index, and investigates their appropriateness relative to variation and uncertainty. These methods are applied to a data set for DVD manufacturing generated according to a Central Composite Design for 6 processing variables and 8 quality attributes. The results indicate that both indices may lead to erroneous design and manufacturing decisions with poor estimation of quality and cost.