0
$\begingroup$

I'm sure there are standard test cases of hard, combinatorial, NP problems, ... (like knapsack, bin packing, partitioning, ...) that practitioners or algorithm developers use when they compare a new algorithm against existing ones (as I often see in papers), but where can I find them ? (a single repository is more preferable than many scattered)

$\endgroup$
  • $\begingroup$ Complexity Zoo and it's companions Complexity Garden + Complexity Dojo might be of interest to you. $\endgroup$ – Aristu Sep 30 '18 at 22:20
  • $\begingroup$ Hi, I mean test cases with numbers. Files in txt or csv or json or some format with numbers that describe a real problem to be solved and that serves as a kind of benchmark. For example for subset sum, the file would include all numbers, the target sum, the result subset (if known), an indication if it is proved to be optimal (and perhaps some performance numbers from runs on various cpus and with various languages) ... $\endgroup$ – dsnz Oct 1 '18 at 10:48
  • $\begingroup$ This depends entirely on the problem. Typically, there is no central repository for problem instances, but they are scattered here and there (for example, some are on the homepage of individual researchers, some might be somewhere on a research project homepage, etc.) What problem(s) are you interested in? $\endgroup$ – Juho Oct 31 '18 at 21:32
0
$\begingroup$

somehow I found the following

OR-Library at http://people.brunel.ac.uk/~mastjjb/jeb/info.html (many datasets for many problems)

Bin Packing page at https://www2.wiwi.uni-jena.de/Entscheidung/binpp/index.htm (datasets for bin packing)

$\endgroup$

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.