You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When I'm calling cobyla through Optimization.jl (i.e. using COBYLA()), the constraints seem to be handled incorrectly: for comparison, using IPNewton() gives a totally different, better, solution. Calling cobyla directly in PRIMA.jl (i.e. using cobyla()), the constraints (when properly modified) do provide the expected results.
Expected behavior
Results from calling solve with COBYLA() or IPNewton() should provide the same results.
@ChrisRackauckas You are right; however, the conditions should be equivalent here vis-a-vis the domain restrictions (both variables are non-negative, so both constraint functions cannot become negative either; any non-positive lower bound thus suffices. I've changed the above code nonetheless to make the distinction more clear: lcons is now negative Inf; the problem persists.
Describe the bug 🐞
When I'm calling cobyla through Optimization.jl (i.e. using COBYLA()), the constraints seem to be handled incorrectly: for comparison, using IPNewton() gives a totally different, better, solution. Calling cobyla directly in PRIMA.jl (i.e. using cobyla()), the constraints (when properly modified) do provide the expected results.
Expected behavior
Results from calling
solve
with COBYLA() or IPNewton() should provide the same results.Minimal Reproducible Example 👇
Error & Stacktrace⚠️
Environment (please complete the following information):
using Pkg; Pkg.status()
using Pkg; Pkg.status(; mode = PKGMODE_MANIFEST)
versioninfo()
Additional context
Add any other context about the problem here.
The text was updated successfully, but these errors were encountered: