Originally posted by: WANGJia

Is there a difference BETWEEN calling CPLEX.ABS() to geenerate a abs function in the obj function AND coding the linearization reformation into the model, in terms of say, perforamance in building the MIP model? performance in finding solution to the MIP?

Because I heard from someone that there is always better to try to code yourself...
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: WANGJia

I don't think BUMP is a good idea but I do think this question is a simple yes or no one for a technician. Because we ca't see the code how you implement ABS()...
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

I would never use the word "always" in reference to MIP models, but I suspect that generally you are better off coding it yourself. I tested several ways to do absolute values (including the ABS() function) on a single test problem. You can see the test problem, implementation details and my results in this blog post. How much you can infer from a sample of size 1 is obviously a consideration.

Paul

Mathematicians are like Frenchmen: whenever you say something to them, they translate it into their own language, and at once it is something entirely different. (Goethe)
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: WANGJia

Thank you professor. That contains a lot of details need to read through.
#CPLEXOptimizers
#DecisionOptimization