One needs to set the randomization seed for the study object (trial) to make sure the random sequence of hyperparameters are repeatable. The "random_state" in Random Forest controls the procedures for Random Forest but not the imputed hyperparameters.