DOE and the Quality Loss Function
Six Sigma – iSixSigma › Forums › Old Forums › General › DOE and the Quality Loss Function
 This topic has 4 replies, 3 voices, and was last updated 14 years, 11 months ago by Erik L.

AuthorPosts

October 21, 2007 at 11:29 am #48464
I am trying to solve a problem in a Simulation for Six Sigma Book. The book analyze a DOE and calculate Quality Loss Function Grouping.
How they calculate The Quality Loss Function with the use of the Minitab 15. What is the purpose of the calculation.
Then they perform a regression analysis with the Total QLF versus the Critiacl to Quality Variables?
What they can get with the calculation?.
What is the relationship of the QLF and the Six Sigma calculation ? ( If any)
How can I improve the QLF?
Thanks so much for your help
Jose
I can send the chart at any mail. Thanks0October 22, 2007 at 2:14 am #163450
fake accrington alertParticipant@fakeaccringtonalert Include @fakeaccringtonalert in your post and this person will
be notified via email.First understand the function QLF and then use the Minitab.
good luck0October 22, 2007 at 4:36 am #163452Thanks so much and then the question is What is the Quality Loss Function?.
Thanks so much0October 22, 2007 at 6:58 am #163456
fake accrington alertParticipant@fakeaccringtonalert Include @fakeaccringtonalert in your post and this person will
be notified via email.A continuous “Taguchi” function that measures the cost implications of product quality.A common form is the quadratic loss function:
L=k(ym)^2 Where L is the loss associated with a particular value of the indepedent variable y.The specification nominal value is m,while k is a contrant depending on the cost and width of the specification limits.This type of philosophy encourages,for example,a television manufacturer to continually strive to routinely manufacture products that have a very high quality picture.
Hope that helps0October 22, 2007 at 6:59 pm #163497Jose,
There is a nice linkage between DOE and the QLF that you have sighted. DOE is a key tool in the creation of a causal transform equation that explains the behavior in the response from either its central tendency and/or its variability. DOE points you to the key parameters (xs) that impact the response and through tools like Response Optimizer provides what settings those xs should be at to achieve certain objectives for your response.
As long as you have a believable cost to replace/repair defective output (part of the loss coefficient and remember that this is suppose to capture internal as well as external costs) and a meaningfully derived tolerance you can look at what combination of settings create a predicted average response (and associated s.d.). Ala Monte Carlo simulation you could then come up with a prediction of the average loss for a process if its run at those settings (Assuming target is best is the best model for the response). You would feed your data into the equation L(y)=$[s^2+(average Target)^2]. Iterating through various what if scenarios you could then begin to appreciate the sensitivities of the factors in your model and which settings are most impacting the projected loss of the process.
There is a direct linkage between Six Sigma thinking and the Taguchi loss function. Both measures are positively impacted by achieving the nominal (targeted) response with a minimum of associated variation relative to the Voice of the Customer (VOC). My personal choice in defining quality. Sigma levels of a process are calculated as the number of standard deviations that exist between the average of the process and the closest of either the upper/lower specification limit. Typically if we can count six standard deviations we feel that we have a high quality product which will generate low COPQ and high Customer Sat.
So, how can one improve QLF? 1.) Adjust your processes average so it hits the customers desired target. 2.) Minimize the variation of your response. 3.) Combine both options #1 and #2.
Regards,
Erik0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.