I am planning on characterizing the performance of an algorithm (specifically the DIRECT algorithm) with respect to a particular parameter mentioned in the original paper but for which I have found no usable characterization of the performance impacts of. This algorithm isn't the most popular in its field (nonlinear numerical optimization), but there have been several papers associated with it and it does have applications in other fields.

To be more specific, I am planning on characterizing the performance of the algorithm with respect to its number of iterations in a more specific way than "increasing the number of iterations increases the quality of the results," including a description of how exactly the performance impact scales.

Is this sort of characterization novel enough to be worth publication?

  • You may get better input if you post on a discipline-specific site (eg or.stackexchange.com). Best wishes! – E. Tucker Nov 25 '19 at 19:56
  • Is this a purely mathematical analysis, or did you do conduct an experiment to verify your hypotheses? – nick012000 Nov 26 '19 at 2:23

It may be thin, but you won't know until you write it up formally and submit it to some appropriate journal. The real feedback will come from reviewers and/or the editor.

The question here is just a hypothetical. Make it concrete and ask someone (the editor) who is in a position to make a determination. You will have to make your case there.

In general, to answer the topline question, it may be publishable or not, but the details matter.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.