The way to implement GPyOpt from the existing X and Y

If you have already run this program for iterations and have all the random initial points, you can use

opt_model = GPyOpt.methods.BayesianOptimization(f=f, domain=bounds, X=existing_x, Y=existing_y)

to use the previous data. Caveat: X and Y are all expected to be 2d array.