The difference in the post-processing is caused by the difference in the model status and the solution information given by the solver. Cplex is able to provide a solution even in the iteration interrupt case, and provides a status that is reported by GAMS as Intermediate Infeasible:
**** SOLVER STATUS 2 Iteration Interrupt
**** MODEL STATUS 6 Intermediate Infeasible
**** OBJECTIVE VALUE 73549121.4264
However, Gurobi is not able to handle the interrupt equally gracefully, as it is not able to provide any solution, and GAMS therefore reports the status No Solution Returned:
**** SOLVER STATUS 2 Iteration Interrupt
**** MODEL STATUS 14 No Solution Returned
**** OBJECTIVE VALUE NA
So, the difference in the results processing is caused by what the solvers provide to GAMS. Cplex appears better in handling exceptions. You can of course only get a GDX file of the results only when the solver provides a solution.
**** SOLVER STATUS 2 Iteration Interrupt
**** MODEL STATUS 6 Intermediate Infeasible
**** OBJECTIVE VALUE 73549121.4264
However, Gurobi is not able to handle the interrupt equally gracefully, as it is not able to provide any solution, and GAMS therefore reports the status No Solution Returned:
**** SOLVER STATUS 2 Iteration Interrupt
**** MODEL STATUS 14 No Solution Returned
**** OBJECTIVE VALUE NA
So, the difference in the results processing is caused by what the solvers provide to GAMS. Cplex appears better in handling exceptions. You can of course only get a GDX file of the results only when the solver provides a solution.