I'm working on a MINLP model where small changes in input constants (e.g., 1e-1 vs 1e-2) are leading to inconsistent results, suggesting possible model sensitivity

New to AMPL
1

I’m working on a MINLP optimization model and noticing unexpected results when I change a constant input by some factor. For example, adjusting the input by 1e-1 gives incorrect results, but a smaller change like 1e-2 yields the correct output. I’m trying to understand if this is due to model sensitivity or formulation issues.

2

To get some help with this issue, please give a more detailed explanation. Specifically, explain how you can tell that the results are incorrect, when the input is adjusted by 1e-1. If possible, attach copies of your files.