I am looking for a good software system to teach students on response surface designs with. Few things I can think of are SAS Proc RSReg or SAS ADX. Is there any other software?
R is free and has Russell V. Lenth's package RSM. Which:
Provides functions to generate response-surface designs, fit first- and second-order response-surface models, make surface plots, obtain the path of steepest ascent, and do canonical analysis
R is fantastic, but it has a steep learning curve. Minitab is probably the easiest to use, appears to be widely used in both industry and classrooms and has some nice tools built in for response surface optimization. Though not free, students can buy Minitab for a semester for the price of a cheap textbook.
An alternative with R would be the R Commander (or Rcmdr) library with the Rcmdr DOE plugin. Nearly the same same functionality as Minitab in a very similar, though not so polished, interface.
If you have access to MATLAB, I'll recommend that to you for two primary reasons:
- Elaborate library for generating design of experiments such as Factorial designs, Latin hypercube designs, quasi random designs like sobol sequences
- MATLAB module UQLab which has an elaborate library of advanced techniques for developing response surfaces: like Low Rank Tensor Approximations, Polynomial Chaos Expansions, Kriging (Gaussian processes) etc. It also has a very good support community.
Another option is Python.
- Use pyDOE for experimental design; or sobol_seq to generate a sobol sequence
- Scikit-learn for the response surface
The advantage here is that Python is completely free and there are loads of examples available as demonstrations on the internet.