Hi Erica,
I don't have a clear answer why the data is different. Do you mean that the variable level values are completely different? Are you only running one solve statement, or do you have multiple math programs?
Regarding the violated flag, which version of AIMMS are you using? In versions of AIMMS prior to 3.11 PR - SU2 / SU3, it could happen that the bound tolerance used by the solver was different than the bound tolerance used within the Math Program Inspector. Please see the help for the options MPI_bound_tolerance and MPI_feasibility_tolerance for more information.
Guido Diepen
AIMMS Specialist