Warm start Non-Convex Quadratic Optimization

Answered

Shuvomoy Das Gupta

June 08, 2021 14:03

• 
• 
• 

Dear All,

I am trying to solve a non-convex quadratic optimization problem in Gurobi, and Gurobi can solve it without any issue though it takes a while to solve it. However, for this particular problem, I do have a very good feasible point (that is suboptimal but close to the final objective value). I was wondering if there is a way to use this feasible solution to speed up the solver? 

Any tips/suggestions will be much appreciated. 

0

• Official comment

  

  Simranjit Kaur

  • Gurobi Staff

  June 23, 2025 02:53 Edited

  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?.
• 

  Jaromił Najman

  • Gurobi Staff

  June 10, 2021 09:20

  Hi,

  You can provide an initial point for a (nonconvex) quadratic problem by setting the Start attribute of your optimization variables or by generating an MST file and using the model.read() function if using an API or the InputFile parameter if using the command line.

  Best regards,
  Jaromił

  1
• 

  Shuvomoy Das Gupta

  July 13, 2021 10:08 Edited

  • 
  • 
  • 

  Thanks for your info, Jaromił! I am using Gurobi with JuMP in Julia programming language, and your suggestion works. I found out that there are two ways to provide an initial point in JuMP. They are:

  1. set_start_value(variable_name, variable_start_value) function

  or 

  2. MOI.set(nonlinear_model, Gurobi.VariableAttribute("Start"), variable_name, variable_start_value)

   

   

  1

Post is closed for comments.