Mario Ruthmair

Hi Keerthi, The Gurobi solver uses multiple algorithms to solve a model involving presolving, branch-and-bound, adding cutting planes, heuristics, etc.In your case, the dual bound obtained from the...

Hi Matthijs, The barrier algorithm itself rarely ends up with a basic solution (in a corner of the polytope). The first crossover phase (basically simplex) tries to accomplish this, i.e., move the ...

Hi Hany, If you use an academic WLS license, you can run up to two Gurobi Compute Servers in parallel (and optionally a Cluster Manager to handle them comfortably). You can find instructions on how...

A reasonable value for NoRelHeurTime is unfortunately not easy to guess, it depends heavily on your model. But this heuristic is usually quite good in finding a feasible solution. In the case you m...

Hi Lorenz, Are you comparing the performance in terms of solution quality or in terms of solver runtime?If you have models that give different optimal solutions since they model the shift changes i...

Hi Maura, VRPs with time windows in general have solutions where the vehicles need to wait at some customer locations until the time window starts. So, even in an optimal solution, there could be w...

Hi Richard, The branching on binary variables in Gurobi's branch-and-bound algorithm follows a set of complicated rules that depend on many criteria, i.e., set of fractional variables, current boun...

In the formulation from your initial screenshot, constraints (5.6) ensure that the time windows are satisfied. But in this type of formulation, those constraints are quadratic. You could also add q...

Here is an article on how to deal with infeasible models. In general, there could be several reasons why your model is infeasible: The VRP instance itself could be infeasible. There are issues in ...

Hi Iason, 1) If the solver did not find a solution yet but it is still going on, then there is a chance that your model is infeasible. But the solver does not know yet. It could also be that findin...