Detect unbounded and infeasible problems#332
Open
BenjaminPINEAU wants to merge 5 commits intoJuliaSmoothOptimizers:masterfrom
Open
Detect unbounded and infeasible problems#332BenjaminPINEAU wants to merge 5 commits intoJuliaSmoothOptimizers:masterfrom
BenjaminPINEAU wants to merge 5 commits intoJuliaSmoothOptimizers:masterfrom
Conversation
dpo
requested changes
May 5, 2026
| diverging_obj_tol::T = -eps(T)^(-1), | ||
| cviol_tol::T = eps(T)^(-1), | ||
| diverging_max_iter::Int = 5, | ||
| cviol_max_iter::Int = 5, |
Member
There was a problem hiding this comment.
The main question is to find out if the criteria you implemented result in a lot of "false positives", i.e., problems that are falsely detected as (locally) infeasible.
| mu *= factor_penalty_up | ||
| if cviol > cviol_tol | ||
| cviol_iter += 1 | ||
| end |
Member
There was a problem hiding this comment.
I think cviol_iter should be reset to zero if cviol ≤ cviol_tol. We want the tolerance to be exceeded for a number of consecutive iterations, right?
Contributor
Author
There was a problem hiding this comment.
Actually, my mistake, this has been implemented in R2 and R2N but not in AL.
| end | ||
| if (fx + hx < diverging_obj_tol) || (norm(solver.x) > diverging_iterates_tol) | ||
| mu *= factor_penalty_up | ||
| diverging_iter = diverging_iter + 1 |
Co-authored-by: Dominique <dominique.orban@gmail.com>
Co-authored-by: Dominique <dominique.orban@gmail.com>
Co-authored-by: Dominique <dominique.orban@gmail.com>
Co-authored-by: Dominique <dominique.orban@gmail.com>
Contributor
Author
|
@dpo I was also wondering if we could add the criterion problem = "elec"
nlp = MathOptNLPModel(OptimizationProblems.PureJuMP.eval(Meta.parse(problem))(), name = problem)
quadratic_model = QuadraticModel(nlp, nlp.meta.x0, name = nlp.meta.name)we can detect that the problem is unbounded much more quickly (considering the number of iteration of the subsolver): |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Allows AL, R2N and R2 to detect unbounded problems (see #331).
Considering the following unbounded problem:
The AL algorithm now return
A problem (or subproblem) is stated as$$f+h$$ or $$\lVert x \rVert$$ is higher than a tolerance for a given number of iterations.
unboundedifThis pull request also allowed
ALto detect infeasible problem. Considering the following infeasible problemThe
ALalgorithm now returnThe algorithm stops when
cviolis hight and the regularization parameter increases for a given number of iterations.@dpo