1

I'm a CS student and I'm using Gurobi for a project.

I am here because I have encountered a little problem with the Gurobi solver used along with the MiniZinc driver. In particular I have noticed that Gurobi returns two different solutions on the same optimisation problem (modelled with MiniZinc) but with a different domains for all the float problem variables. The first problem uses a domain equal to -2^31.0 .. 2^31.0, while the second uses a domain equal to -3.402823e+38..3.402823e+38.

In the first case Gurobi returns a solution equal to 1.0, in the second case it returns UNSATorUNBOUNDED.

Francesco Contaldo
  • 35
  • 2
  • 1
    What is the difference between the two runs? Can you tell if one of them is incorrect? – Dekker1 Feb 18 '19 at 23:18
  • @Dekker1 The solver is called with the same way on both the two instance problems. The correct result is 1.0 since the problem comes from a known benchmark-set and other solver return that value as solution. The two instance problems are available at this links: [bigger domain variables](https://pastebin.com/Bua7ZK9u) [smaller domain variables](https://pastebin.com/TGa3eLBF) – Francesco Contaldo Feb 19 '19 at 11:01

0 Answers0