Optimizing computationally expensive black-box functions is challenging because each function evaluation incurs a high computational cost. Bayesian optimization (BO) addresses this problem by using Gaussian process surrogate models to guide the selection of new evaluation points. Within the Bayesian optimization with Pareto-based selection (BOPS) framework, the predictive mean and standard deviation define a bi-objective optimization problem that balances exploitation and exploration. A batch of solutions is then selected from its approximate Pareto set. This paper presents a systematic empirical study of Pareto front approximation and Pareto-based solution selection within the BOPS framework. We investigated three space-filling designs (Sobol sequences, Latin hypercube sampling, and uniform random sampling) in combination with five Pareto-based solution selection strategies. Experiments were conducted on the 20-dimensional CEC2022 benchmark suite (12 functions) and a real-world robotic manipulation problem using a batch size of $ q = 4 $, an evaluation budget of 800 function evaluations, and 30 independent runs. BOPS variants were compared with standard acquisition-based batch BO methods, including $ q $-expected improvement, $ q $-upper confidence bound, and $ q $-knowledge gradient. The results showed that optimization performance depends primarily on the solution selection mechanism rather than the Pareto approximation strategy once a sufficiently accurate Pareto set is obtained. Moreover, several BOPS variants using simple space-filling-based Pareto approximation achieve optimization performance comparable to or better than standard acquisition-based BO. These findings provide empirical insights into Pareto-based batch Bayesian optimization, demonstrate the effectiveness of simple space-filling-based Pareto approximation, and provide a unified framework for developing and evaluating future Pareto-based solution selection strategies.
Citation: Kittisak Chaiyotha, Tipaluck Krityakierne. Pareto-based selection strategies for batch Bayesian optimization of expensive black-box functions[J]. AIMS Mathematics, 2026, 11(7): 22166-22204. doi: 10.3934/math.2026898
Optimizing computationally expensive black-box functions is challenging because each function evaluation incurs a high computational cost. Bayesian optimization (BO) addresses this problem by using Gaussian process surrogate models to guide the selection of new evaluation points. Within the Bayesian optimization with Pareto-based selection (BOPS) framework, the predictive mean and standard deviation define a bi-objective optimization problem that balances exploitation and exploration. A batch of solutions is then selected from its approximate Pareto set. This paper presents a systematic empirical study of Pareto front approximation and Pareto-based solution selection within the BOPS framework. We investigated three space-filling designs (Sobol sequences, Latin hypercube sampling, and uniform random sampling) in combination with five Pareto-based solution selection strategies. Experiments were conducted on the 20-dimensional CEC2022 benchmark suite (12 functions) and a real-world robotic manipulation problem using a batch size of $ q = 4 $, an evaluation budget of 800 function evaluations, and 30 independent runs. BOPS variants were compared with standard acquisition-based batch BO methods, including $ q $-expected improvement, $ q $-upper confidence bound, and $ q $-knowledge gradient. The results showed that optimization performance depends primarily on the solution selection mechanism rather than the Pareto approximation strategy once a sufficiently accurate Pareto set is obtained. Moreover, several BOPS variants using simple space-filling-based Pareto approximation achieve optimization performance comparable to or better than standard acquisition-based BO. These findings provide empirical insights into Pareto-based batch Bayesian optimization, demonstrate the effectiveness of simple space-filling-based Pareto approximation, and provide a unified framework for developing and evaluating future Pareto-based solution selection strategies.
| [1] | A. S. Alvi, B. B. Ru, J.-P. Calliess, S. J. Roberts, M. A. Osborne, Asynchronous batch Bayesian optimisation with improved local penalisation, Proceedings of the 36th International Conference on Machine Learning, PMLR, 97 (2019), 253–262. |
| [2] | M. Awad, R. Khanna, Support vector regression, In: Efficient learning machines, Berkeley: Apress, 2015, 67–80. http://doi.org/10.1007/978-1-4302-5990-9_4 |
| [3] |
F. Berkenkamp, A. Krause, A. P. Schoellig, Bayesian optimization with safety constraints: safe and automatic parameter tuning in robotics, Mach. Learn., 112 (2023), 3713–3747. http://doi.org/10.1007/s10994-021-06019-1 doi: 10.1007/s10994-021-06019-1
|
| [4] |
M. Binois, N. Wycoff, A survey on high-dimensional gaussian process modeling with application to bayesian optimization, ACM Transactions on Evolutionary Learning and Optimization, 2 (2022), 1–26. http://doi.org/10.1145/3545611 doi: 10.1145/3545611
|
| [5] |
J. Blank, K. Deb, pymoo: Multi-objective optimization in Python, IEEE Access, 8 (2020), 89497–89509. http://doi.org/10.1109/ACCESS.2020.2990567 doi: 10.1109/ACCESS.2020.2990567
|
| [6] | A. D. Bull, Convergence rates of efficient global optimization algorithms, J. Mach. Learn. Res., 12 (2011), 2879–2904. |
| [7] | A. Candelieri, A gentle introduction to bayesian optimization, 2021 Winter Simulation Conference (WSC), Phoenix, AZ, USA, 2021, 1–16. http://doi.org/10.1109/WSC52266.2021.9715413 |
| [8] |
A. Candelieri, A. Ponti, I. Giordani, F. Archetti, On the use of Wasserstein distance in the distributional analysis of human decision making under uncertainty, Ann. Math. Artif. Intell., 91 (2023), 217–238. http://doi.org/10.1007/s10472-022-09807-0 doi: 10.1007/s10472-022-09807-0
|
| [9] |
K. Chaiyotha, T. Krityakierne, A comparative study of infill sampling criteria for computationally expensive constrained optimization problems, Symmetry, 12 (2020), 1631. http://doi.org/10.3390/sym12101631 doi: 10.3390/sym12101631
|
| [10] |
S. Chen, C. F. N. Cowan, P. M. Grant, Orthogonal least squares learning algorithm for radial basis function networks, IEEE T. Neural Networ., 2 (1991), 302–309. http://doi.org/10.1109/72.80341 doi: 10.1109/72.80341
|
| [11] |
I. Couckuyt, D. Deschrijver, T. Dhaene, Fast calculation of multiobjective probability of improvement and expected improvement criteria for pareto optimization, J. Glob. Optim., 60 (2014), 575–594. http://doi.org/10.1007/s10898-013-0118-2 doi: 10.1007/s10898-013-0118-2
|
| [12] | D. D. Cox, S. John, A statistical method for global optimization, 1992 IEEE International Conference on Systems, Man, and Cybernetics, Chicago, IL, USA, 1992, 1241–1246. http://doi.org/10.1109/ICSMC.1992.271617 |
| [13] |
R. Dai, J. Jie, Z. Wang, H. Zheng, W. L. Wang, Automated surrogate-assisted particle swarm optimizer with an adaptive parental guidance strategy for expensive engineering optimization problems, J. Comput. Des. Eng., 12 (2025), 145–183. http://doi.org/10.1093/jcde/qwaf023 doi: 10.1093/jcde/qwaf023
|
| [14] |
G. Damblin, M. Couplet, B. Iooss, Numerical studies of space-filling designs: Optimization of latin hypercube samples and subprojection properties, J. Simul., 7 (2013), 276–289. http://doi.org/10.1057/jos.2013.16 doi: 10.1057/jos.2013.16
|
| [15] | G. De Ath, R. M. Everson, J. E. Fieldsend, Asynchronous $\varepsilon$-greedy Bayesian optimisation, Proceedings of the Thirty-Seventh Conference on Uncertainty in Artificial Intelligence, PMLR, 161 (2021), 578–588. |
| [16] | G. De Ath, R. M. Everson, J. E. Fieldsend, A. A. M. Rahat, $\varepsilon$-shotgun: $\varepsilon$-greedy batch bayesian optimisation, Proceedings of the 2020 Genetic and Evolutionary Computation Conference, New York: Association for Computing Machinery, 2020,787–795. http://doi.org/10.1145/3377930.3390154 |
| [17] |
G. De Ath, R. M. Everson, A. A. M. Rahat, J. E. Fieldsend, Greed is good: Exploration and exploitation trade-offs in bayesian optimisation, ACM Trans. Evol. Learn. Optim., 1 (2021), 1–22. http://doi.org/10.1145/3425501 doi: 10.1145/3425501
|
| [18] | G. De Ath, J. E. Fieldsend, R. M. Everson, What do you mean? The role of the mean function in bayesian optimisation, Proceedings of the 2020 Genetic and Evolutionary Computation Conference Companion, New York: Association for Computing Machinery, 2020, 1623–1631. http://doi.org/10.1145/3377929.3398118 |
| [19] | J. Dem{š}ar, Statistical comparisons of classifiers over multiple data sets, J. Mach. Learn. Res., 7 (2006), 1–30. |
| [20] |
B. Do, T. Adebiyi, R. D. Zhang, Epsilon-greedy thompson sampling to Bayesian optimization, J. Comput. Inf. Sci. Eng., 24 (2024), 121006. http://doi.org/10.1115/1.4066858 doi: 10.1115/1.4066858
|
| [21] |
L. Du, R. B. Gao, P. N. Suganthan, D. Z. W. Wang, Bayesian optimization based dynamic ensemble for time series forecasting, Inform. Sciences, 591 (2022), 155–175. http://doi.org/10.1016/j.ins.2022.01.010 doi: 10.1016/j.ins.2022.01.010
|
| [22] | D. Eriksson, M. Poloczek, Scalable constrained bayesian optimization, International Conference on Artificial Intelligence and Statistics, PMLR, 130 (2021), 730–738. |
| [23] | M. Ester, H.-P. Kriegel, J. Sander, X. W. Xu, A density-based algorithm for discovering clusters in large spatial databases with noise, Proceedings of the Second International Conference on Knowledge Discovery and Data Mining, Oregon: AAAI Press, 1996,226–231. |
| [24] | P. I. Frazier, Bayesian optimization, Recent Advances in Optimization and Modeling of Contemporary Problems, 2018 2018,255–278. http://doi.org/10.1287/educ.2018.0188 |
| [25] | M. A. Gelbart, J. Snoek, R. P. Adams, Bayesian optimization with unknown constraints, Proceedings of the Thirtieth Conference on Uncertainty in Artificial Intelligence, Quebec City, Canada, 2014,250–259. |
| [26] | GPy: A Gaussian process framework in Python, 2012–2024. Available from: http://github.com/SheffieldML/GPy. |
| [27] |
F. Häse, M. Aldeghi, R. J. Hickman, L. M. Roch, A. Aspuru-Guzik, Gryffin: An algorithm for bayesian optimization of categorical variables informed by expert knowledge, Appl. Phys. Rev., 8 (2021), 031406. http://doi.org/10.1063/5.0048164 doi: 10.1063/5.0048164
|
| [28] |
C. L. He, Y. Zhang, D. W. Gong, X. F. Ji, A review of surrogate-assisted evolutionary algorithms for expensive optimization problems, Expert Syst. Appl., 217 (2023), 119495. http://doi.org/10.1016/j.eswa.2022.119495 doi: 10.1016/j.eswa.2022.119495
|
| [29] |
J. D. Huang, D. W. Zhan, An adaptive dropout approach for high-dimensional Bayesian optimization, Optim. Eng., 2026 (2026), 1–24. http://doi.org/10.1007/s11081-026-10087-4 doi: 10.1007/s11081-026-10087-4
|
| [30] |
C. Jiang, M. Q. Li, Multi-objectivising acquisition functions in bayesian optimisation, ACM Transactions on Evolutionary Learning and Optimization, 5 (2025), 1–33. http://doi.org/10.1145/3716504 doi: 10.1145/3716504
|
| [31] |
C. Jiang, M. Q. Li, Trading off quality and uncertainty through multi-objective optimisation in batch bayesian optimisation, Proceedings of the AAAI Conference on Artificial Intelligence, 39 (2025), 27027–27035. http://doi.org/10.1609/aaai.v39i25.34909 doi: 10.1609/aaai.v39i25.34909
|
| [32] |
P. Y. Jiang, Y. S. Cheng, J. X. Yi, J. Liu, An efficient constrained global optimization algorithm with a clustering-assisted multiobjective infill criterion using gaussian process regression for expensive problems, Inform. Sciences, 569 (2021), 728–745. http://doi.org/10.1016/j.ins.2021.05.015 doi: 10.1016/j.ins.2021.05.015
|
| [33] |
R. W. Jiao, S. Y. Zeng, C. H. Li, Y. H. Jiang, Y. C. Jin, A complete expected improvement criterion for gaussian process assisted highly constrained expensive optimization, Inform. Sciences, 471 (2019), 80–96. http://doi.org/10.1016/j.ins.2018.09.003 doi: 10.1016/j.ins.2018.09.003
|
| [34] |
S. Joe, F. Y. Kuo, Constructing sobol sequences with better two-dimensional projections, SIAM J. Sci. Comput., 30 (2008), 2635–2654. http://doi.org/10.1137/070709359 doi: 10.1137/070709359
|
| [35] |
D. R. Jones, M. Schonlau, W. J. Welch, Efficient global optimization of expensive black-box functions, J. Global Optim., 13 (1998), 455–492. http://doi.org/10.1023/A:1008306431147 doi: 10.1023/A:1008306431147
|
| [36] | T. Kathuria, A. Deshpande, P. Kohli, Batched Gaussian process bandit optimization via determinantal point processes, Proceedings of the 30th International Conference on Neural Information Processing Systems, Red Hook: Curran Associates Inc., 2016, 4213–4221. |
| [37] |
T. Krityakierne, D. Baowan, Aggregated GP-based optimization for contaminant source localization, Operations Research Perspectives, 7 (2020), 100151. http://doi.org/10.1016/j.orp.2020.100151 doi: 10.1016/j.orp.2020.100151
|
| [38] | T. Krityakierne, D. Ginsbourger, Global optimization with sparse and local gaussian process models, International Workshop on Machine Learning, Optimization and Big Data, Cham: Springer, 2015,185–196. http://doi.org/10.1007/978-3-319-27926-8_16 |
| [39] |
A. Kulesza, B. Taskar, Determinantal point processes for machine learning, Found. Trends Mach. Le., 5 (2012), 123–286. http://doi.org/10.1561/2200000044 doi: 10.1561/2200000044
|
| [40] | A. Kumar, K. V. Price, A. W. Mohamed, A. A. Hadi, P. N. Suganthan, Problem Definitions and Evaluation Criteria for the 2022 Special Session and Competition on Single Objective Bound Constrained Numerical Optimization, Technical report, Nanyang Technological University, Singapore, 2021. Available from: https://github.com/P-N-Suganthan/2022-SO-BO. |
| [41] |
J. Lehman, K. O. Stanley, Abandoning objectives: Evolution through the search for novelty alone, Evol. Comput., 19 (2011), 189–223. http://doi.org/10.1162/EVCO_a_00025 doi: 10.1162/EVCO_a_00025
|
| [42] |
B. Liu, Q. F. Zhang, G. G. E. Gielen, A Gaussian process surrogate model assisted evolutionary algorithm for medium scale expensive optimization problems, IEEE T. Evolut. Comput., 18 (2014), 180–192. http://doi.org/10.1109/TEVC.2013.2248012 doi: 10.1109/TEVC.2013.2248012
|
| [43] | J. MacQueen, Some methods for classification and analysis of multivariate observations, Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, 1 (1967), 281–297. |
| [44] | J. Mockus, V. Tiesis, A. Zilinskas, The application of Bayesian methods for seeking the extremum, In: Towards global optimization, Amsterdam: North-Holland, 1978,117–129. |
| [45] | F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, et al., Scikit-learn: Machine learning in Python, J. Mach. Learn. Res., 12 (2011), 2825–2830. |
| [46] | V. Perrone, M. Donini, M. B. Zafar, R. Schmucker, K. Kenthapadi, C. Archambeau, Fair bayesian optimization, Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, New York: Association for Computing Machinery, 2021,854–863. http://doi.org/10.1145/3461702.3462629 |
| [47] | V. Picheny, A stepwise uncertainty reduction approach to constrained global optimization, Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics, PMLR, 33 (2014), 787–795. |
| [48] |
G. Pirot, T. Krityakierne, D. Ginsbourger, P. Renard, Contaminant source localization via bayesian global optimization, Hydrol. Earth Syst. Sci., 23 (2019), 351–369. http://doi.org/10.5194/hess-23-351-2019 doi: 10.5194/hess-23-351-2019
|
| [49] | pyDOE2 Contributors, pyDOE2: Design of experiments for Python, 2018–2024. Available from: https://github.com/clicumu/pyDOE2. |
| [50] |
J. C. Qian, Y. S. Cheng, J. L. Zhang, J. Liu, D. W. Zhan, A parallel constrained efficient global optimization algorithm for expensive constrained optimization problems, Eng. Optimiz., 53 (2021), 300–320. http://doi.org/10.1080/0305215X.2020.1722118 doi: 10.1080/0305215X.2020.1722118
|
| [51] | C. E. Rasmussen, C. K. I. Williams, Gaussian processes for machine learning, Cambridge: MIT Press, 2005. http://doi.org/10.7551/mitpress/3206.001.0001 |
| [52] |
F. Rehbach, M. Zaefferer, A. Fischbach, G. Rudolph, T. Bartz-Beielstein, Benchmark-driven configuration of a parallel model-based optimization algorithm, IEEE T. Evolut. Comput., 26 (2022), 1365–1379. http://doi.org/10.1109/TEVC.2022.3163843 doi: 10.1109/TEVC.2022.3163843
|
| [53] | F. L. Richardson, G. De Ath, T. Chugh, Is greed still good in multi-objective bayesian optimisation, Proceedings of the Genetic and Evolutionary Computation Conference Companion, New York: Association for Computing Machinery, 2024, 2103–2106. http://doi.org/10.1145/3638530.3664189 |
| [54] |
O. Roustant, D. Ginsbourger, Y. Deville, DiceKriging, DiceOptim: Two R packages for the analysis of computer experiments by kriging-based metamodeling and optimization, J. Stat. Softw., 51 (2012), 1–55. http://doi.org/10.18637/jss.v051.i01 doi: 10.18637/jss.v051.i01
|
| [55] |
M. Schonlau, W. J. Welch, D. R. Jones, Global versus local search in constrained optimization of computer models, IMS Lecture Notes Monogr. Ser., 34 (1998), 11–25. http://doi.org/10.1214/lnms/1215456182 doi: 10.1214/lnms/1215456182
|
| [56] |
E. Schulz, M. Speekenbrink, A. Krause, A tutorial on gaussian process regression: Modelling, exploring, and exploiting functions, J. Math. Psychol., 85 (2018), 1–16. http://doi.org/10.1016/j.jmp.2018.03.001 doi: 10.1016/j.jmp.2018.03.001
|
| [57] | N. Srinivas, A. Krause, S. M. Kakade, M. W. Seeger, Gaussian process optimization in the bandit setting: No regret and experimental design, Proceedings of the 27th International Conference on Machine Learning (ICML), Madison: Omnipress, 2010, 1015–1022. |
| [58] |
N. V. Thieu, Opfunu: An open-source Python library for optimization benchmark functions, Journal of Open Source Software, 12 (2024), 8. http://doi.org/10.5334/jors.508 doi: 10.5334/jors.508
|
| [59] |
J. Ungredda, J. Branke, Bayesian optimisation for constrained problems, ACM T. Model. Comput. S., 34 (2024), 1–26. http://doi.org/10.1145/3641544 doi: 10.1145/3641544
|
| [60] |
N. S. Upadhye, R. Chowdhury, Constrained Bayesian optimization with lower confidence bound, Technometrics, 66 (2024), 561–574. http://doi.org/10.1080/00401706.2024.2336535 doi: 10.1080/00401706.2024.2336535
|
| [61] |
A. H. Victoria, G. Maragatham, Automatic tuning of hyperparameters using bayesian optimization, Evol. Syst., 12 (2021), 217–223. http://doi.org/10.1007/s12530-020-09345-2 doi: 10.1007/s12530-020-09345-2
|
| [62] |
P. Virtanen, R. Gommers, T. E. Oliphant, M. Haberland, T. Reddy, D. Cournapeau, et al., SciPy 1.0: Fundamental algorithms for scientific computing in Python, Nat. Methods, 17 (2020), 261–272. http://doi.org/10.1038/s41592-019-0686-2 doi: 10.1038/s41592-019-0686-2
|
| [63] |
H. Y. Wang, H. Xu, Y. Yuan, Z. Q. Zhang, An adaptive batch bayesian optimization approach for expensive multi-objective problems, Inform. Sciences, 611 (2022), 446–463. http://doi.org/10.1016/j.ins.2022.08.021 doi: 10.1016/j.ins.2022.08.021
|
| [64] |
H. Y. Wang, H. Xu, Z. Q. Zhang, High-dimensional multi-objective bayesian optimization with block coordinate updates: Case studies in intelligent transportation system, IEEE T. Intell. Transp., 25 (2024), 884–895. http://doi.org/10.1109/TITS.2023.3241069 doi: 10.1109/TITS.2023.3241069
|
| [65] |
L. W. Wang, S. Yerramilli, A. Iyer, D. Apley, P. Zhu, W. Chen, Scalable gaussian processes for data-driven design using big data with categorical factors, J. Mech. Des., 144 (2022), 021703. http://doi.org/10.1115/1.4052221 doi: 10.1115/1.4052221
|
| [66] |
W. Z. Wang, H.-L. Liu, K. C. Tan, A surrogate-assisted differential evolution algorithm for high-dimensional expensive optimization problems, IEEE T. Cybernetics, 53 (2023), 2685–2697. http://doi.org/10.1109/TCYB.2022.3175533 doi: 10.1109/TCYB.2022.3175533
|
| [67] |
Z. Y. Wang, G. P. Rangaiah, Application and analysis of methods for selecting an optimal solution from the pareto-optimal front obtained by multiobjective optimization, Ind. Eng. Chem. Res., 56 (2017), 560–574. http://doi.org/10.1021/acs.iecr.6b03453 doi: 10.1021/acs.iecr.6b03453
|
| [68] | Z. Wang, S. Jegelka, Max-value entropy search for efficient Bayesian optimization, Proceedings of the 34th International Conference on Machine Learning, PMLR, 2017, 3627–3635. |
| [69] | C. K. I. Williams, C. E. Rasmussen, Gaussian processes for regression, Advances in Neural Information Processing Systems, 8 (1995), 1–7. |
| [70] | J. T. Wilson, F. Hutter, M. P. Deisenroth, Maximizing acquisition functions for Bayesian optimization, Advances in Neural Information Processing Systems, 31 (2018), 1–12. |
| [71] | J. Wu, P. I. Frazier, The parallel knowledge gradient method for batch Bayesian optimization, Advances in Neural Information Processing Systems, 29 (2016), 1–9. |
| [72] | W. J. Xu, Y. N. Jiang, B. Svetozarevic, C. N. Jones, Constrained efficient global optimization of expensive black-box functions, Proceedings of the 40th International Conference on Machine Learning, Honolulu, Hawaii, USA, 202 (2023), 38485–38498. |
| [73] |
Y. Zeng, Y. S. Cheng, J. Liu, An efficient global optimization algorithm for expensive constrained black-box problems by reducing candidate infilling region, Inform. Sciences, 609 (2022), 1641–1669. http://doi.org/10.1016/j.ins.2022.07.162 doi: 10.1016/j.ins.2022.07.162
|
| [74] |
D. W. Zhan, Expected coordinate improvement for high-dimensional Bayesian optimization, Swarm Evol. Comput., 91 (2024), 101745. http://doi.org/10.1016/j.swevo.2024.101745 doi: 10.1016/j.swevo.2024.101745
|
| [75] |
Y. J. Zhang, Y. Z. Wang, J. P. Wang, Objective attributes weights determining based on shannon information entropy in hesitant fuzzy multiple attribute decision making, Math. Probl. Eng., 2014 (2014), 463930. http://doi.org/10.1155/2014/463930 doi: 10.1155/2014/463930
|
Figures(7) / Tables(10)
Kittisak Chaiyotha, Tipaluck Krityakierne. Pareto-based selection strategies for batch Bayesian optimization of expensive black-box functions[J]. AIMS Mathematics, 2026, 11(7): 22166-22204. doi: 10.3934/math.2026898
DownLoad: