Nondeterminism bug in parmest

[blnicho]

Summary

@djlaky discovered an issue in parmest with constraints being generated in a nondeterministic order. We think we've tracked the issue to this line in parmest where we are using a Python set to make sure the list of theta names does not have duplicates:

pyomo/pyomo/contrib/parmest/parmest.py

Line 312 in df51e99

self.estimator_theta_names = list(set(theta_names))

The fix suggested by @jsiirola is to use the following to create a list without duplicates:

my_list = list(dict.fromkeys(list_with_duplicates))

[djlaky]

@blnicho Wanted to follow up on this issue from our conversation at the PSE stakeholder meeting last week.