Martin Bromberger

Hi Ben,OR-Tools indeed uses a different parameter configuration than Gurobi. OR-Tools in Python at least changes OptimalityTol, IntFeasTol, and FeasTol to 1e-7, which can increase the computation t...

Hi Enrico,This might be a bit late, but NODCNT is a good choice for a unique identifier. Just to let you know, there might be multiple callbacks for the same node, as processing a node involves num...

Hi Marian,Thanks for the explanation. I think I now understand what the expected behavior should be.In the current formulation, the constraints \( (p − p ′ ) \cdot v \geq x \) force the variable \(...

Hi Marian,Gurobi does not have to satisfy every constraint exactly, but only up to some tolerance (see FeasibilityTol). That way, Gurobi prevents minor rounding errors from causing infeasibility. I...

Hi Matthew,One advantage of simple general constraints like Abs/Max is that Gurobi will recognize most well-known cases for which a simpler formulation (e.g., continuous without introducing binary ...

Hi Sam,Exactly. If you want it based on a buffer, you'd need a custom parser and use the Start attribute.Pretty much the same is true if you want to write a solution to a buffer/string. Gurobi does...

Hi Sam,There is no direct way for Gurobi to read a buffer. However, there are some workarounds without I/O operations that give you similar behavior.If your problem is an LP, then you can give Guro...

There are three competing factors to consider: Likelihood and magnitude of rounding errors Strength of your formulation, i.e., how close is the relaxed problem to the convex hull of the integer fea...

Hello Junseo,Yes, you can improve the search for "all" solutions with the best objective (i.e., using PoolSearchMode=2, PoolGapAbs=0, and a sufficiently large value for PoolSolutions) by providing ...

Hi Adrian,In the blended case, the objectives are combined into a single linear objective, and Gurobi only has to optimize towards this one objective. Therefore, blended multi-objectives are typica...