Essentials of Design Robustness in Design for Six Sigma (DFSS) Methodology. 3″Design for Six Sigma (DFSS) is a systematic process and a disciplined problem preventionapproach to achieve business excellence. Robust design is the heart of DFSS.”In DFSS, it is important to develop transfer function and identify vital few Critical to Qualitycharacteristics (CTQ).
A DFSS project starts with defining CTQ and ends with validating CTQbased on the customer requirements. Applying the axiomatic design framework enables the DFSSteam to focus on the functional requirements and enhances creative thinking. This process resultsin better solutions for robustness and reliability.Six Sigma method aims to reduce waste in the manufacturing process and during the operationsphase mainly focuses on a) Increasing productivity to quickly develop new products at a low rate,and b) Management which is based on value; while the Robust design or Taguchi designcenters on improving engineering productivity.Over the years, Six Sigma has made it possible to reduce cost by discovering problems whichoccur during manufacturing and resolving instant causes. Robust Design on the other hand hasmade it feasible to prevent issues or problems by rigorously developing designs for bothmanufacturing process and product.
Breakdown Point”The breakdown point is a measure of testing the robustness of a statistical procedure.”There are two breakdown points: finite sample breakdown point which is that fraction of thedata that can be given arbitrary values without making the estimator go bad. This is however, alocal breakdown point and asymptotic breakdown point is the limit of the finite samplebreakdown point as n tends to infinity. Greater the value of the breakdown point better is theestimator used.From the analysis 4 of the breakdown point characterization of robustness, it is observed thatthe sample mean is the worst estimator invented and median is the of all three, Pseudo-medianbeing the third one.
Another way 5 of considering Breakdown point(BP) is one which measures the robustnessproperties of a statistic T in a global sense. It is defined as the maximal amount of modelmisspecification an estimator can withstand before it breaks down. If the Gross-error sensitivity(GES, which is measured by the Influence Function IF) of T is infinite, then its BP is nil. But, thismethod could fail if BP is too small.In this definition a term “breakdown point” is introduced to define robustness. System withhigher breakdown point means higher robustness. But do we will measure breakdown point in not