Data structure for sparse coefficient matrices

Answered

Haakon Grov

• October 23, 2019 11:59
• Edited

I am new to Gurobi and I am trying to understand which data structure to use for coefficients. I want to model a large sparse coefficient matrix A as seen below. I am wondering if it should be done using a tuplelist, a tupledict or some other structure?  In the statement below, A is the coefficient matrix and z is a decision variable. 

0

• 

  Silke Horn

  Gurobi Staff

  • October 25, 2019 05:36

  Welcome to Gurobi!

  Which API are you going to use? (From your mention of tuplelist and tupledict, I am going to assume Python.)

  As of version 8.1.1, Gurobi does not support matrix-based modeling in Python. So right now the recommendation is to model your constraints using the addConstr method (or in bulk using addConstrs). You can have a look at the code examples here.

  With Gurobi 9 (scheduled for later this fall), you will be able to do matrix-oriented modeling, using NumPy or SciPy matrices.

  1

Please sign in to leave a comment.