Generate DoE matrix and test plan
27 ビュー (過去 30 日間)
古いコメントを表示
I would like to generate both a test matrix and a test plan using DoE methodology. I already found these functions: hadamard (x); ff2n(x) and fracfact(...) But, the 1st function doesn't give me a clear vision of the experimental sequence and of the existing aliase. I can obtain this vision by using the function fracfact, where I can define the aliases. But the output of this function give just the combinaison of -1, +1 of the parameters. I don't have the complete DoE matrix for te further DoE analysis. Moreover, I didn't find the way to generate a full factorial design. The fullfact function doesn't do it.
Additionally, for experimental data extraction purpose, I would like to replace in the matrix, the -1 and +1 values by the respective extrema of parameters values.
Thanks for your help
0 件のコメント
採用された回答
Tom Lane
2011 年 7 月 26 日
Phillippe, you are right that the Statistics Toolbox offers a collection of DOE tools rather than a beginning-to-end DOE application. Here's how to use some of those tools. You can generate a full factorial design for four factors each taking two values as follows:
design = fullfact([2 2 2 2])
For historical reasons, this codes the two levels as 1 and 2. The fracfact function codes them as -1 and 1. Here's how to use that function to get the equivalent design and its confounding pattern:
[design,confounding] = fracfact('a b c d')
Here's how to generate a 2^(4-1) design:
[design,confounding] = fracfact('a b c abc')
Here's how to change the limits in the first column to other limits that you specify (you could repeat for other columns or loop over them):
limits = [5 10];
design(:,1) = limits(1) + (limits(2)-limits(1))*(1+design(:,1))/2
Finally, you may want to randomize the order of runs when you carry out the experiment. Here's how to do that:
order = randperm(8)
design = design(order,:)
0 件のコメント
その他の回答 (2 件)
Rick Rosson
2011 年 7 月 22 日
Do you have access to the Statistics Toolbox? If so, there are some functions provided there related to DoE and SPC.
HTH.
Rick
0 件のコメント
参考
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!