Originally posted by: MarcoMojana

Hi,
I'm trying to use the CPXuncrushform function (C Callable library). It's not clear to me the precise meaning of the offset parameter. Does it mean that the rhs in the original problem is equal to the rhs of the presolved row minus the offset?

Thank you!

M. Mojana
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

It is probably the easiest to explain this with an example. Suppose the following preprocessing steps have been applied:

x = 3x' + 2
y = 5y' - 4

So, the variables in the presolved model are x' and y'. If you now have a linear sum (like the left hand side of a cut) that is

6x' - 15y'

and you call CPXuncrushform on this, it will result in

6x' - 15y' = 6(x/3 - 2/3) - 15(y/5 - 4/5)
           = 2x - 4 - 3y + 12
           = 2x - 3y + 8

So, the resulting offset will be 8. If you are uncrushing a constraint, then need to subtract this from the right hand side.

Tobias
#CPLEXOptimizers
#DecisionOptimization