用于混合整数编程（MIP）的PULP，CPLEX或GUROBI

Question

I am interested to develop a code who uses PulP. I Have some questions and I Will be very gratful if you can help me.

• Does PulP have a restriction in the number of linear constraint or integer variables? • If I have a problem with many constraintes or integer variables, Have I to buy a solver like CPLEX or Gurobi

I really thank you for your time.

Answer 1

1. I don't think there are artificial limits on the size of models you can generate with PuLP

2. For larger, more difficult problems, commercial solvers like Cplex or Gurobi typically are much faster and more reliable than open source solvers. Of course you can use an open source solver like glpk or CBC for prototyping, even if the final model is large. Note also that Cplex and Gurobi come with their own Python based modeling interfaces (these may offer access to the more esoteric aspects of the solvers). One advantage of Pulp is that you can develop the model with an open source solver and then switch to a commercial solver without changing the model code.