I am solving some LPs using Gurobi and the interior point method and without crossover. I notice that it returns the solution to be suboptimal and terminates with status LOCALLY_SOLVED
, which, according to documentation, means the solver found a locally optimal solution (which may also be globally optimal, but it could not prove so). Why is Gurobi unable to prove optimality (being a global solver)? Note that it doesn’t hit the time limit. I am probably missing something trivial here, but any help is appreciated. Optimization problems are part of the subproblems in an algorithm, so I am unable to provide a MWE here. Thanks!
We’re just working on a change for this:
- Strange reporting with respect to feasibility · Issue #546 · jump-dev/Gurobi.jl · GitHub
- Check feasibility of solutions in PrimalStatus and DualStatus by odow · Pull Request #548 · jump-dev/Gurobi.jl · GitHub
But I haven’t made a new release yet.
LOCALLY_SOLVED
maps to Gurobi’s SUBOPTIMAL
status:
1 Like
Sounds good. Thanks!
1 Like