Coluna.jl
Coluna.jl copied to clipboard
atoptima
Branch-and-Price-and-Cut in Julia
- Previous storage was embedding a "manual" garbage collector. I simplified the code to remove the manual GC because Julia GC bound memory performances. - UnitPermissions are still present but...
TODO (regressions): - [x] fix test 550 - [x] fix test 591 - [x] fix broken tests - [x] old - node finalizer - [ ] branching tree file -...
@artalvpes it will fix the error you sent. Implementing these two methods is the first step to support representatives in Coluna.
**Describe the bug** The solution is not correct when using continuous variables _x_ instead of binary variables in the Generalized Assignment demo. **To Reproduce** ``` using Coluna, ColunaDemos, GLPK, JuMP,...
- [ ] remove subproblem `Formulation` Currently, when the dev creates a custom model/optimizer, Coluna maintains a `Formulation` to handle subproblem variables. This formulation is useless because we can leave...
- [x] Create a `AbstractCustomData` type to make clear what is the super type of the field `custom_data` in `Variable` & `Constraint` & also the type of `custom_data` in some...
In preprocessing tests, one tries to access Coluna's internal `Formulation`s through the JuMP model to test how the preprocessing modifies elements of `Formulation`s. I think it's a bad pratice and...
It would be nice if we indicate this only if at least one lazy cut callback is defined. For this, I suggest to transform CutCallbacks to CutCallback which separates only...
Hello, In the current version of Coluna, is it possible to take control of the branching decisons in the B&P tree through a callback? If not, where are encoded the...
