Solving Non convex MIQCP in reasonable time

Awaiting user input

Ritesh Bhalerao

• June 28, 2024 06:02

Hello, I am trying to solve a fleet management problem which involves many bilinear terms. This non convex MIQCP has been running on my other machine since 7hrs and the gap is still 24%. Is there any ways to speed up the convergence of model ? 

I have attached the log output for the model

Related to

• Nonconvex
• MIQCP

0

• 

  Jaromił Najman

  Gurobi Staff

  • July 01, 2024 07:01

  Hi Ritesh,

  Do you think it is the dual or the primal bound that are really slow? To improve the primal bound (best feasible point), you might want to try to experiment with the NoRelHeurTime parameter. For the dual bound, you might want to experiment with the Presolve, MIPFocus, PreSparsify, OBBT parameters.

  Non-convex models require tight variable bounds for good convergence. You might want to try to provide as tight bounds as possible for any variable participating in a bilinear term.

  Often, reformulating a non-convex model is the best way to gain performance. You could share the model. Note that uploading files in the Community Forum is not possible but we discuss an alternative in Posting to the Community Forum.

  Best regards, 
  Jaromił

  0

Please sign in to leave a comment.