Multiobjective Real-coded Bayesian Optimization Algorithm Revisited: Diversity Preservation Abstract This paper provides empirical studies on MrBOA, which have been designed for strengthening diversity of nondominated solutions. The studies lead to modified sharing. A new selection scheme has been suggested for improving. diversity performance. Empirical tests validate their effectiveness on uniformity and front-spread (i.e., diversity) of nondominated set. A diversity-preserving MrBOA (dp-MrBOA) has been designed by carefully combining all the promising components; i.e., modified sharing, dynamic crowding, and diversity-preserving selection. Experiments demonstrate that dp-MrBOA is able to significantly improve diversity performance (for the scaling problems), without weakening proximity of nondominated set.