Gamspy outputs

GAMSPy
1

Hi, I would like to ask about GAMSPy solutions (Cplex solver). In the GAMSPy output, some decision variables appear with values such as 0.99999 instead of exactly 1, whereas in GAMS they appear as 1. This also affects the value of the objective function. Is this behavior normal, and how should it be interpreted?

2

Also, there is a difference in the solution as showing in the following pictures.

IMG_20251225_163424
IMG_20251225_1634241920×1440 372 KB

IMG_20251225_163502
IMG_20251225_1635021920×2560 552 KB

3

This is totally normal. All solvers work with tolerances and they usually do not impact the optimization very much. For the integrality tolerance epintyou can actually set the tolerance to 0 (see CPLEX ). You can still get different solutions (with the same objective) since there can be multiple solutions with the same objective value (primal degenerate problem).

-Michael

4

Thank you for your response.