Benchmarking different models is not trivial, however it is something I frequently do. It may be fitting, to provide a nice example on how to do a proper evaluation fitting for a publication.
As example, we could use different TSP models. (1) The circuit-constraint (2) the MTZ formulation (3) the classical Dantzig formulation that will require multiple iterations because CP-SAT does not support lazy constraints.
Show all the steps if implementing and evaluating this via Git tags. Can be used for generally teaching about empirical evaluations of optimization algorithms.
d-krupke
commented
8 months ago
Benchmarking different models is not trivial, however it is something I frequently do. It may be fitting, to provide a nice example on how to do a proper evaluation fitting for a publication.
As example, we could use different TSP models. (1) The circuit-constraint (2) the MTZ formulation (3) the classical Dantzig formulation that will require multiple iterations because CP-SAT does not support lazy constraints.
Show all the steps if implementing and evaluating this via Git tags. Can be used for generally teaching about empirical evaluations of optimization algorithms.