How does Gurobi recognize/solve simple stochastic programs?

Ongoing

Alexander Torgovitsky

• June 03, 2019 21:10

I watched this recent Gurobi webinar about solving stochastic programs with Gurobi.

It wasn't clear to me from the video whether Gurobi is actually implementing a decomposition method (or another approach) for solving stochastic programs, or whether it was just solving large deterministic equivalents.

So, does Gurobi has special algorithmic capability for stochastic programs? If so, how does it recognize the problem structure? Is there any documentation for this? (I looked but wasn't able to find any...)

0

• 

  Daniel Espinoza

  • June 04, 2019 14:29

  Hi Alexander,

  For now, Gurobi does not have any special algorithm for dealing with stochastic optimization problems. In the webinar we used the deterministic equivalent formulation, i.e. the extensive formulation of the problem, to solve the stochastic problems.

  Now, for many practical problems, this approach is more than able to solve the stochastic problem, specially for two-stage stochastic problems without chance constraints.

  Daniel

  0
• 

  shahla rezaei

  • May 20, 2020 18:05
  • Edited

  

  

  Hi Alexander

  is there any approach to equip gurobi to cover optimization problem with uncertainty and clearly multi-stage stochastic programming. as you know these type of problems these days are the most demanding optimization problems in industry and academia.

  regards

  Shahla

  0
• 

  Joe Xing

  • May 27, 2022 23:26

  

  

  Hello Daniel,

  Could you share the entire Python code associated with the webinar?  Thank you.

  Regards,

  Joe

   

  0

Please sign in to leave a comment.