A Robust Sampling-Based Method for Batch Bayesian Optimization

Abstract
This paper presents a sampling-based method called Sampling-Calculation-Optimization (SCO) for the experimental design of batch Bayesian optimization. SCO does not construct new multi-point acquisition functions but samples from the existing one-point acquisition function to obtain candidate designs. The rejection sampling is modified to address different shape of acquisition function. Besides, general discrepancy is computed to compare different designs. To reduce the uncertainty of designs, the genetic algorithm is applied to optimize the designs. Several strategies are proposed to reduce the burden of calculation in the SCO. After calculation and optimization, the SCO designs are less uncertain than other sampling-based methods. Numerical results shows that SCO is not only comparable with other BBO methods in many analytic problem, but also robust when the smoothness of objective function and kernel function are inconsistent.

Keywords
Batch Bayesian optimization, Design of experiment, Acquisition function, General discrepancy

Cite this paper
Yiming Ye, Xin Qin, Lening Ma, A Robust Sampling-Based Method for Batch Bayesian Optimization , SCIREA Journal of Mathematics. Volume 11, Issue 1, February 2026 | PP. 11-31. 10.54647/mathematics110575

References

[ 1 ]	Bishop, C.M. (2006). Pattern recognition and machine learning.
[ 2 ]	Briffoteaux, G., Gobert, M., Ragonnet, R., Gmys, J., Mezmaz, M.-S., Melab, N., Tuyttens, D. (2020). Parallel surrogate-assisted optimization: Batched Bayesian neural network-assisted GA versus q-EGO. Swarm Evol. Comput., 57 , 100717.
[ 3 ]	Cai, X., Qiu, H., Gao, L., Yang, P., Shao, X. (2017). A multi-point sampling method based on kriging for global optimization. Structural and Multidisciplinary Optimization, 56 (1), 71–88.
[ 4 ]	Chen, R.-B., Hsu, Y.-W., Hung, Y., Wang, W. (2014). Discrete particle swarm optimization for constructing uniform design on irregular regions. Computational Statistics and Data Analysis, 72 , 282–297.
[ 5 ]	Cheng, J., Jiang, P., Zhou, Q., Hu, J., Shu, L. (2021). A parallel constrained lower confidence bounding approach for computationally expensive constrained optimization problems. Appl. Soft Comput., 106 , 107276.
[ 6 ]	Dai, R., & Chen, W. (2026). Research on stock price forecasting models based on lstm and Bayesian optimization algorithms. Proceedings of the 2025 5th international conference on big data, artificial intelligence and risk management (p. 853–860). New York, NY, USA: Association for Computing Machinery.
[ 7 ]	Desautels, T., Krause, A., Burdick, J.W. (2012). Parallelizing exploration-exploitation tradeoffs with gaussian process bandit optimization. J. Mach. Learn. Res., 15 , 3873-3923.
[ 8 ]	de Villiers, D.I.L., Couckuyt, I., Dhaene, T. (2017). Multi-objective optimization of reflector antennas using kriging and probability of improvement. 2017 IEEE International Symposium on Antennas and Propagation & USNC/URSI National Radio Science Meeting , 985-986.
[ 9 ]	Fang, K.T., Lin, D.K., Winker, P., Zhang, Y. (2000). Uniform design: Theory and application. Technometrics, 42 (3), 237–248.
[ 10 ]	Fang, K.-T., Liu, M.-Q., Qin, H., Zhou, Y.-D. (2018). Theory and application of uniform experimental designs.
[ 11 ]	Ginsbourger, D., Riche, R.L., Carraro, L. (2010). Kriging is well-suited to parallelize optimization. (pp. 131–162).
[ 12 ]	Goldberg, D.E. (1988). Genetic algorithms in search, optimization, and machine learning.
[ 13 ]	Gonz´alez, J.I., Dai, Z., Hennig, P., Lawrence, N.D. (2016). Batch Bayesian optimization via local penalization. Aistats.
[ 14 ]	Jones, D.R., Schonlau, M., Welch, W.J. (1998). Efficient global optimization of expensive black-box functions. Journal of Global Optimization, 13 (4), 455–492.
[ 15 ]	Kushner, H.J. (1963). A new method of locating the maximum point of an arbitrary multipeak curve in the presence of noise. Journal of Basic Engineering , 86 , 97-106.
[ 16 ]	Li, M., Yi, B., Yang, Y. (2021). An efficient global optimization method with multi-point infill sampling based on kriging. Engineering Optimization, 0 (0), 1-18.
[ 17 ]	Li, Y., Kang, L., Hickernell, F.J. (2020). Is a transformed low discrepancy design also low discrepancy? In: Fan, J., Pan, J. (eds) Contemporary Experimental Design, Multivariate Analysis and Data Mining. Springer, Cham. https://doi.org/10.1007/978-3-030-46161-4_5
[ 18 ]	Liu, J., Jiang, C., Zheng, J. (2021). Batch Bayesian optimization via adaptive local search. Appl. Intell., 51 , 1280-1295.
[ 19 ]	Nguyen, V., Rana, S., Gupta, S., Li, C., Venkatesh, S. (2017). Budgeted batch Bayesian optimization with unknown batch sizes. ArXiv , abs/1703.04842 .
[ 20 ]	Xiao, Y., Ning, J., Xiong, Z. et al. (2022). Batch sequential adaptive designs for global optimization. J. Korean Stat. Soc. 51, 780–802.
[ 21 ]	Rasmussen, C.E., & Williams, C.K.I. (2009). Gaussian processes for machine learning. Adaptive computation and machine learning.
[ 22 ]	Srinivas, N., Krause, A., Kakade, S.M., Seeger, M.W. (2010). Gaussian process optimization in the bandit setting: No regret and experimental design. Icml.
[ 23 ]	Sun, G., Li, L., Fang, J., Li, Q. (2021). On lower confidence bound improvement matrix-based approaches for multi-objective Bayesian optimization and its applications to thin-walled structures. Thin-walled Structures, 161 , 107248.
[ 24 ]	Tang, C., Zhang, F., Zhang, J., Lv, Y., Wang, G. (2022). Novel reliability evaluation method combining active learning kriging and adaptive weighted importance sampling. Structural and Multidisciplinary Optimization.
[ 25 ]	Tao, T., Zhao, G., Ren, S. (2019, 11). An Efficient Kriging-Based Constrained Optimization Algorithm by Global and Local Sampling in Feasible Region. Journal of Mechanical Design, 142 (5), 1-48.
[ 26 ]	Victoria, A.H., & Maragatham, G. (2021). Automatic tuning of hyperparameters using Bayesian optimization. Evolving Systems, 12 , 217-223.
[ 27 ]	Wang, Z., Dahl, G.E., Swersky, K., Lee, C., Nado, Z., Gilmer, J., . . . Ghahramani, Z. (2024). Pre-trained gaussian processes for Bayesian optimization. Journal of Machine Learning Research, 25 (1), 83.
[ 28 ]	Yu, R., Picard, C., Ahmed, F. (2025). Fast and accurate Bayesian optimization with pre-trained transformers for constrained engineering problems: Fast and accurate Bayesian optimization with pre-trained transformers. R. yu et al. Structural & Multidisciplinary Optimization, 68 (3).
[ 29 ]	Zhan, D., Qian, J., Cheng, Y. (2017). Pseudo expected improvement criterion for parallel EGO algorithm. Journal of Global Optimization, 68 , 641-662.
[ 30 ]	Zhan, D., & Xing, H. (2020). Expected improvement for expensive optimization: a review. Journal of Global Optimization, 1-38.
[ 31 ]	Zhou, Y.-D., & Fang, K.-T. (2013). An efficient method for constructing uniform designs with large size. Computational Statistics, 28 (3), 1319– 1331.