strange behavior - explainable by numerical issues?

Alicia Buhk

• December 04, 2019 15:11

I am seeking assistance with a network design problem (including Big-M-Constraints) that I modeled as an IP via the java api.

The gurobi log warns of a large RHS range and a large Matrix range.

I am confused by the following observations and appreciate your opinion on whether or not they can be caused by numerical issues :

1) changing the optimization parameters changes the value of the optimal solution by a factor of 2, i.e. changing the Cut parameter from the default parameter to 0 (corresponding to not making any cuts) improves the solution. However, in both runs, the solution was indicated to be optimal (mip gap = 0.0000%), in particular in the run with the worse solution!

2) extending the feasible set (changing some equality constraints to inequality constraints) worsens the solution significantly (by more than 60%), while both runs still indicate that the solutions found are optimal (mip gap = 0.0000%).

Any ideas?

Kind regards!

Alicia

1

• Official comment

  

  Simranjit Kaur

  Gurobi Staff

  • December 06, 2024 00:51

  This post is more than three years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum. Or why not try our AI Gurobot?.

Post is closed for comments.