Define an index set that is dependent on another two index sets - in Python

Answered

Zeyu Wang

June 19, 2022 00:04

• 
• 
• 

Hi,

I'm transferring my model from AMPL to python now, but I could not figure out how to build the index sets that rely on other indices.

s.t. c1{a in A}: sum{j in J, f in 1..F[a,j]} x[j,f+1] >= 1;

where J={0,1,2}, A is a non-continuous set ({0,1,5} for example) and F gives the max value that f can take under each (a,j) pair (for example, F=[ 1 1 2; 1 2 3; 1 1 3])

I've found a solution posted here but did not work it out when the dependency is on two indices.  https://support.gurobi.com/hc/en-us/community/posts/5989957537553-Defining-a-set-dependent-on-an-other-set?sort_by=created_at 

Thank you so much.

 

0

• Official comment

  

  Simranjit Kaur

  • Gurobi Staff

  November 16, 2025 23:34

  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 try Gurobot, our chatbot interface offering instant, expert-level support.
• 

  David Torres Sanchez

  • Gurobi Staff

  June 20, 2022 11:10

  Hi,
  
  Assuming your variables are appropriately defined, the syntax is a bit similar to AMPL.
  You just need to ensure that the order of the indices is correct.

  m.addConstrs(
      (gp.quicksum(x[j, f + 1] for j in J for f in range(1, F[a][j])) >= 1 for a in A),
      name="set_cov",
  )

  In this case, the outer for-loop defines \(\texttt{a}\), the first in-line for loop of the sum defines \(\texttt{j}\) and the second defines \(\texttt{f}\) (which uses both \(\texttt{a}\) and \(\texttt{j}\)).

  I suggest you take a look at the examples, there are some simple cases when this is done: e.g. diet.py  or workforce1.py. The latter uses tupledict.sum().
  You will need some basic Python knowledge to understand these, particularly Python Generators and list comprehensions.

  Cheers, 
  David

  0

Post is closed for comments.