Question about constraint modelling

[Christopher-Wang]

Sorry if this is an obvious question, however I was reviewing the code for the constraint algorithm and stumbled across this line. I was wondering about the logic behind checking the lowerbound for the upper constraint. Why is this not checking the upper bound instead?

BayesianOptimization/bayes_opt/constraint.py

Line 206 in dc6f251

norm(loc=y_mean, scale=y_std).cdf(self._ub[0]) if self._lb[0] != np.inf else np.array([1])

[till-m]

Hi @Christopher-Wang,

first of all, great thing that you're checking the code!

You're completely right, this should be checking the upper constraint (frankly, quite an embarrassing thing to mess up on my part). The good news is that this just makes the code slightly less efficient by making an unnecessary call to .cdf, but doesn't break anything -- probably also why no one noticed it before. However, given that .cdf is about 1000x slower than calling np.array([1]) this should definitely be corrected.

Would you be interested in making a PR to fix this? :)