absolute value constraint on vectorised variable setting

Answered

afsaneh mastouri

December 04, 2020 19:36

• 
• 
• 

Hi, 

I am using AddMVar to define a matrix/vectorised setting for a quadratic optimisation problem (linear objective with quadratic and linear constraint). for one of the constraint I need to access to absolute value of variables. I used "addGenConstrAbs" but it doesn't look that it works on vector level. 

I read in other posts that we dont have access to individual variable at vector level (i.e. v[i]) to use addGenConstrAbs for individual variable. 

Do you have any solutions\suggestions?

Thank you.

0

• Official comment

  

  Simranjit Kaur

  • Gurobi Staff

  June 23, 2025 03:03 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?.
• 

  Eli Towle

  • Gurobi Staff

  December 04, 2020 19:51

  Unfortunately, MVar objects cannot currently be used with general constraints. The development team is aware of the limitation. For now, you can use MVar.tolist() (new in Gurobi 9.1) to obtain the list of Var objects associated with the matrix variable(s). These Var objects can be used with Model.addGenConstrAbs()/abs_().

  0
• 

  afsaneh mastouri

  December 05, 2020 10:46

  • 
  • 
  • 

  Thank you Eli for your prompt response. 

  How would you suggest that i implement the abs constraint at matrix setting? 

  Does it mean that I should use AddVar instead of AddMVar?

  Thank you, 

  0
• 

  Eli Towle

  • Gurobi Staff

  December 07, 2020 21:09

  You can use the MVar.tolist() method I mentioned to obtain the list of Var objects associated with the matrix variables. These Var objects can be used in general constraint functions. E.g.:

  import gurobipy as gp
  
  m = gp.Model()
  x = m.addMVar(3)
  y = m.addMVar(3)
  
  # MVar objects are not currently integrated with general constraints
  # m.addConstr(y == gp.abs_(x)) does not work
  
  # Use corresponding Var objects instead
  for vx, vy in zip(x.tolist(), y.tolist()):
      m.addConstr(vy == gp.abs_(vx))

  Alternatively, you could just use Var objects throughout by defining your variables with Model.addVar() or Model.addVars(). However, this would require you to rewrite any code that uses matrix semantics.

  0

Post is closed for comments.