Skip to main content

Dual Simplex Stucks when Barrier solves the model in C++



1 comment

  • Jaromił Najman
    Gurobi Staff Gurobi Staff

    Hi Giorgos,

    I have formulate the CVRPTW problem which is solved by the barrier algorithm but when the dual simplex method crossover comes to optimize the solution it stucks forever showing me logs... Any idea why is this happening? 

    The Barrier method followed by crossover solves the root node relaxation in about ~140 seconds. After that the B&B phase of the algorithm begins. VRPs are known to be hard to solve, thus it is not surprising that the algorithm proceeds slowly. In your particular case, it looks like finding a feasible solution point is problematic. You might want to experiment with the NoRelHeurTime parameter (set it to ~1200 seconds to start with). Additionally, you might want to have a look at Most important parameters for MIPs.

    Best regards, 


Please sign in to leave a comment.