DOE help needed
Six Sigma – iSixSigma › Forums › Old Forums › General › DOE help needed
 This topic has 6 replies, 5 voices, and was last updated 15 years, 11 months ago by Belgus.

AuthorPosts

August 31, 2006 at 10:11 am #44484
Hello,
I am trying to do a DOE on abandonment rate where i have considered volumes,AHT and staffing as the factors with 3 levels each.I was able to do the Analyze factorial design and come up with significant factors main effects and interactions. However minitab is not giving me option of response optimiser!! Any idea why this has happened? did i do anything wrong?
pst: Also it didnt give me the option of pareto in analyze factorial design as well as i didnt get the option of cube plots in factorial plots
I have pasted the results below from the analyze factorial design.plz advise.
Thanks and Regards
Dee
Analysis of Variance for AB%, using Adjusted SS for Tests
Source DF Seq SS Adj SS Adj MS F P
Volumes 2 244.610 244.610 122.305 271.22 0.000
AHT 2 143.936 143.936 71.968 159.59 0.000
Staffing 2 46.593 46.593 23.296 51.66 0.000
Volumes*AHT 4 69.472 69.472 17.368 38.51 0.000
Volumes*Staffing 4 24.902 24.902 6.226 13.81 0.000
Error 12 5.411 5.411 0.451
Total 26 534.924
S = 0.671529 RSq = 98.99% RSq(adj) = 97.81%
Unusual Observations for AB%
Obs AB% Fit SE Fit Residual St Resid
9 18.9500 17.9856 0.5005 0.9644 2.15 R
13 2.8000 1.8522 0.5005 0.9478 2.12 R
R denotes an observation with a large standardized residual.0August 31, 2006 at 12:17 pm #142564
Robert ButlerParticipant@rbutler Include @rbutler in your post and this person will
be notified via email.I don’t know the mechanics of the Minitab optimizer but the output you’ve posted is for an ANOVA. This method will tell you the degree of significance of a term but it won’t tell you magnitude and direction – in short ANOVA won’t give you a predictive equation complete with coefficients and estimates of their error.
To get this you will have to run a regression. The output from a regression will be a correlation (predictive) equation which, as far as I know, is what any kind of optimizing routine for any statistics package needs in order to do “what if” analysis.0September 1, 2006 at 4:15 am #142633
Eric MaassParticipant@poetengineer Include @poetengineer in your post and this person will
be notified via email.Dee,
Minitab should be giving you the option of Response Optimizer – I would suggest contacting Minitab technical support and sharing your Minitab project with them. They should be able to get it to work.Alternatively, if you want to try something yourself, you could use the Stat/DOE/Factorial PullDown Menu and then select “Define Customer Factorial Design”. That will enable you to respecify your factors and responses, and after you reanalyze, the Response Optimizer might be available for you.
Best of luck!0September 1, 2006 at 9:53 am #142642Thanks a lot. I shall try doing both the ideas.
Regards
Dee0September 1, 2006 at 10:39 am #142644Hi Dee,
If i’m not wrong, you can download some good calculators available for free on the net to understand the relationship b/w abandonment rate, AHT and Staffing. These calculators are based on the queue theories i guess…I dont really understand the objective behind the DOE that you’re talking about0September 1, 2006 at 10:59 am #142645Hello VP,
I have used erlang calculator to determine the optimum staffing in my project. However i wanted to use DOE to experiment on various factors and understand its interdependancy and to look how the equation looks like..
Dee0September 1, 2006 at 12:28 pm #142647If you have run your DOE out of Minitab, but used Minitab to analyse the data, it won’t give you the option of using the pareto, and posibly this is why you can’t use the optimiser. If you try to let Minitab create a new design then cut and paste the results in then you should be able to use both the pareto and the optimiser. Once you reduce the model down you can take the coded or uncoded units as your final model.
0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.