Abstract
A global optimization problem is considered where the objective functions are assumed “black box” and “expensive”. An algorithm is theoretically substantiated using a statistical model of objective functions and the theory of rational decision making under uncertainty. The search process is defined as a sequence of bi-objective selections of sites for the computation of the objective function values. It is shown that two well known (the maximum average improvement, and the maximum improvement probability) algorithms are special cases of the proposed general approach.
Similar content being viewed by others
References
Calvin, J.: Probability models in global optimization. Informatica 27(2), 323–334 (2016)
Calvin, J., Žilinskas, A.: A one-dimensional P-algorithm with convergence rate \(o(n^{-3+\delta })\) for smooth functions. J. Optim. Theory Appl. 106, 297–307 (2000)
Emmerich, M., Yang, K., Deutz, A., Wang, H., Fonseca, C.: A multicriteria generalization of Bayesian global optimization. In: Pardalos, P.M., Zhigljavsky, A., Žilinskas, J. (eds.) Advances in Stochastic and Deterministic Global Optimization, pp. 229–242. Springer, Berlin (2016)
Gimbutas, A., Žilinskas, A.: An algorithm of simplicial Lipschitz optimization with the bi-criteria selection of simplices for the bi-section. J. Global Optim. (2018). https://doi.org/10.1007/s10898-017-0550-9
Huang, D., Allen, T., Notz, W., Miller, R.: Sequential kriging optimization using multiple-fidelity evaluations. Struct. Multidiscip. Optim. 32, 369–382 (2006)
Jones, D.: A taxonomy of global optimization methods based on response surfaces. J. Glob. Optim. 21, 345–383 (2001)
Kleijnen, J., van Beers, W., van Nieuwenhuyse, I.: Expected improvement in efficient global optimization through bootstrapped kriging. J. Glob. Optim. 54, 59–73 (2012)
Knowles, J.: ParEGO: a hybrid algorithm with on-line landscape approximation for expensive multiobjective optimization problems. IEEE Trans. Evolut. Comput. 10(1), 50–66 (2006)
Knowles, J., Corne, D., Reynolds, A.: Noisy multiobjective optimization on a budget of 250 evaluations. In: Ehrgott, M., et al. (eds.) Lecture Notes in Computer Science, vol. 5467, pp. 36–50. Springer (2009)
Kushner, H.: A versatile stochastic model of a function of unknown and time-varying form. J. Math. Anal. Appl. 5, 150–167 (1962)
Mockus, J.: On Bayes methods for seeking an extremum. Avtomatika i Vychislitelnaja Technika 3, 53–62 (1972). in Russian
Pepelyshev, A.: Fixed-domain asymtotics of the maximum likelihood estiomator and the gaussian process approach for deterministic models. Stat. Methodol. 8(4), 356–362 (2011)
Picheny, V.: Multiobjective optimization using gaussian process emulators via stepwise uncertainty reduction. Stat. Comput. 25, 1265–1280 (2015)
Sasena, M.: Dissertation: Flexibility and Efficiency Enhancements for Constrained Global Design Optimization with Kriging Approximations. Michigan University (2002)
Strongin, R.: Information method of global minimization in the presence of noise. Eng. Cybern. 6, 118–126 (1969). in Russian
Strongin, R.G., Sergeyev, Y.D.: Global Optimization with Non-convex Constraints: Sequential and Parallel Algorithms. Kluwer Academic Publishers, Dordrecht (2000)
Zhigljavsky, A., Žilinskas, A.: Stochastic Global Optimization. Springer, Berlin (2008)
Žilinskas, A.: One-step Bayesian method for the search of the optimum of one-variable functions. Cybernetics 1, 139–144 (1975). in Russian
Žilinskas, A.: Axiomatic characterization of a global optimization algorithm and investigation of its search strategies. Oper. Res. Lett. 4, 35–39 (1985)
Žilinskas, A.: A statistical model-based algorithm for black-box multi-objective optimisation. Int. J. Syst. Sci. 45(1), 82–92 (2014)
Žilinskas, A.: Global search as a sequence of rational decisions under uncertainty. In: AIP Conference Proceedings, vol. 1776, No. 020001, pp. 1–8 (2016)
Žilinskas, A., Zhigljavsky, A.: Stochastic global optimization: a review on the occasion of 25 years of Informatica. Informatica 27(2), 229–256 (2016)
Žilinskas, A., Žilinskas, J.: A hybrid global optimization algorithm for non-linear least squares regression. J. Glob. Optim. 56, 265–277 (2013)
Acknowledgements
The research of A. Žilinskas was supported by the Research Council of Lithuania under Grant No. P-MIP-17-61. The research of James Calvin was supported by the National Science Foundation under Grant No. CMMI-1562466.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Žilinskas, A., Calvin, J. Bi-objective decision making in global optimization based on statistical models. J Glob Optim 74, 599–609 (2019). https://doi.org/10.1007/s10898-018-0622-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10898-018-0622-5