Closed Tomalwo closed 4 years ago
Indded thanks for the report. In the docs the algortihm should be marked as M-U not M-U-I @bluescarni has a PR opened that may include this change maybe?
Thanks for the report, I'll fold the changes in #35 and esa/pagmo2#410.
Closing this as the docs have been update in both PRs.
In my tests, NSPSO doesn't appear to work correctly with integer variables:
After the first population, which works fine, subsequent populations have non-integer values even for "integer" variables.
The same is true for MOEA/D, but, according to the documentation, it doesn't handle integers. So perhaps the issue is not with the implementation of NSPO, but with the documentation?
Thanks! Thomas
PS: With NSGA-II, integer handling works fine, which is why I think that the issue isn't on my end.