Hi all,

I am trying optimize a  code, and I am solving one LP subproblem iteratively in my algorithm. The constraints of the LP are in the inequality form, I was wondering if it will be more efficient to add the slack variables to the constraints myself, and feed them as equality constraint to CPLEX. Is there any general rule that feeding inequality constraints as equality is better?

Thanks,

------------------------------
Amin
------------------------------

#DecisionOptimization

I would be extremely surprised that there is any visible difference.  Considering that CPLEX has to check for how the constraints look like anyway, and that such checks are very quick compared to all the computations that CPLEX will do on the problem, I suggest you don't care about that.

------------------------------
Xavier
------------------------------

Original Message:
Sent: Sat October 03, 2020 01:03 PM
From: Amin
Subject: Feeding CPLEX with inequality constraint vs equality

Hi all,

I am trying optimize a  code, and I am solving one LP subproblem iteratively in my algorithm. The constraints of the LP are in the inequality form, I was wondering if it will be more efficient to add the slack variables to the constraints myself, and feed them as equality constraint to CPLEX. Is there any general rule that feeding inequality constraints as equality is better?

Thanks,

------------------------------
Amin
------------------------------

#DecisionOptimization