Surrogate models with logarithmic terms (ln) do not evaluate properly

[tarnold17]

I trained a surrogate and obtained the following expression for the model:

The "ln" term in the surrogate is problematic. When attempting to evaluate the surrogate model against input data, the following traceback is obtained:

This surrogate model was trained using alamopy; I am not sure if the same can occur for surrogates trained using PySMO.

[andrewlee94]

@tarnold17 Thanks for finding this - the issue is that ALAMO uses ln to represent natural logs but Pyomo uses log, thus it is unable to parse the expression it receives from ALAMO. We need to add a check in the parsing code to replace the ln with a log - this is only related to ALAMO and does not affect anything else.