4
$\begingroup$

With SetHint() from the OR-tools linear solver I am trying to pass a feasible solution to my problem. But I receive the following warning:

WARNING: ignore partial solution (0) because unknown rate is too large (0.862233 > 0.85)

What is the unknown rate and how can I reduce it?

I firstly thought, "unknown rate" refers to the proportion of variables that I provide a value for. But even when providing a value for all variables, I receive the warning above.

Background: I am dealing with a VRP and its split delivery generalization. First, I calculate a solution without splits passing it with SetHint(variables, values) to the generalized model. I have done this before with Gurobi (warm start), but wanted to try the same with the OR-Tools SCIP solver. I am working with Python.

Improve this question
$\endgroup$
0

2 Answers 2

Reset to default
7
$\begingroup$

Try:

SetSolverSpecificParametersAsString("heuristics/completesol/maxunknownrate = 0.9")

References:

Improve this answer
$\endgroup$
1
  • $\begingroup$ This is it, thank you! $\endgroup$
    – Dav
    Commented May 11, 2021 at 5:54
2
$\begingroup$

I believe the warning comes from SCIP. So check the documentation there.

Improve this answer
$\endgroup$
3
  • 1
    $\begingroup$ This may be better suited as a comment (as is). $\endgroup$
    – TheSimpliFire ♦
    Commented May 10, 2021 at 20:10
  • $\begingroup$ I think this answer is likely to be a downvote magnet. As pointed out, it should be a comment since it doesn't really answer the question and simply directs the OP to "the documentation." $\endgroup$
    – SecretAgentMan
    Commented May 10, 2021 at 20:27
  • 1
    $\begingroup$ OK. How do you reply when the warning comes from a dependency? $\endgroup$
    – Laurent Perron
    Commented May 11, 2021 at 9:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.