about solve.status

[wagjixiang]

Hellow,
I had create a MIP model, and I used warm-start to solve and limited the solve time, but I found a problem whenever my solution's status is optimal, i know it is not optimal, just is feasible, so can you tell me about this problem?
Thank you very much!

[Pasuukki]

Hi @wagjixiang! I think you may be talking about something that I struggled a long time with. I eventually discovered that the pulp.LpProblem class has 2 different statuses:

• status = status of the problem. If this equals pulp.constants.LpStatusOptimal, it does not yet mean that the solution is optimal, but that a feasible solution has been found.
• sol_status = status of the solution. If this equals pulp.constants.LpSolutionOptimal, I think then the solution should be optimal (at least inside the tolerances that you have defined).