fix empty optimization result + test of infeasible ip restricted master #173
Conversation
Codecov Report
@@ Coverage Diff @@
## master #173 +/- ##
==========================================
+ Coverage 76.87% 77.06% +0.18%
==========================================
Files 38 38
Lines 2171 2193 +22
==========================================
+ Hits 1669 1690 +21
- Misses 502 503 +1
Continue to review full report at Codecov.
|
|
@vitornesello GLPK throws an exception when you give it a model without variables. |
| @@ -50,8 +50,8 @@ getprimalsols(res::OptimizationResult) = res.primal_sols | |||
| getdualsols(res::OptimizationResult) = res.dual_sols | |||
| nbprimalsols(res::OptimizationResult) = length(res.primal_sols) | |||
| nbdualsols(res::OptimizationResult) = length(res.dual_sols) | |||
| getbestprimalsol(res::OptimizationResult) = res.primal_sols[1] | |||
| getbestdualsol(res::OptimizationResult) = res.dual_sols[1] | |||
| getbestprimalsol(res::OptimizationResult) = get(res.primal_sols, 1, nothing) | |||
There was a problem hiding this comment.
Maybe we should leave it to the user to ask if there is a solution available in case he is not sure.
We could add a function issolavailable
Anyway after calling getbestsol he will need to check if the return value was nothing
There was a problem hiding this comment.
I can add nbprimalsol() and nbdualsol()
There was a problem hiding this comment.
They are already there..
There was a problem hiding this comment.
So what is the best behavior? In any case, the user should assure himself that there is at least one solution available, it he uses directly the result of the new getbestsol he might also get in trouble
There was a problem hiding this comment.
I think it's fine for now if it returns nothing. I added new methods to handle this case e.g.
set_lp_primal_sol!(inc::Incumbents, ::Nothing) = false. If the user works on the solution, he will have to test that solution != nothing.
There was a problem hiding this comment.
It means we will have to add methods for functions that expect a solution all over the place.
| problem, x, dec = CLD.GeneralizedAssignment.model(data, coluna) | ||
|
|
||
| JuMP.optimize!(problem) | ||
| @test JuMP.objective_value(problem) == Inf |
There was a problem hiding this comment.
Can you also check the feasibility status?
There was a problem hiding this comment.
It returns INFEASIBLE. I think it should be NODE_LIMIT. We should do that in another PR and tests all possible termination statuses.
Fix #167
Raise and fix #174