Mechanism reduction optimization question

[wandadars]

From your perspectives, would a nonlinear optimization approach to producing a reduced mechanism be a useful capability compared to the capabilities that pyMARS already has? Is there any value in a method such as this that isn't capture by the other methods already in pyMARS?

Reference for my question: https://pdfs.semanticscholar.org/e02b/c29c7ec93e183c6ccf91e9064f6a5fd381bd.pdf

https://arxiv.org/pdf/1712.06281.pdf

[kyleniemeyer]

@wandadars Potentially, although it isn't clear to me whether that sort of method would be more effective—unfortunately many articles in this area that demonstrate new methods don't actually compare against prior methods. In part this is because most of the codes that implement those are not released openly, which is one motivation for our development of pyMARS.

I'd certainly be interested in seeing a module for pyMARS that includes such an approach, so that we could critically evaluate it against the existing methods.