New Bayesian method for multiextremal optimization

Natalija Pozniak, Leonidas Sakalauskas

Abstract


This paper is focused on the Bayes approach to multiextremal optimization problems, based on modelling the objective function by Gaussian random field (GRF) and using the Euclidean distance matrices with fractional degrees for presenting GRF covariances. A recursive optimization algorithm has been developed aimed at maximizing the expected improvement of the objective function at each step, using the results of the optimization steps already performed. Conditional mean and conditional variance expressions, derived by modelling GRF with covariances expressed by fractional Euclidean distance matrices, are used to calculate the expected improvement in the objective function. The efficiency of the developed algorithm was investigated by computer modelling, solving the test tasks, and comparing the developed algorithm with the known heuristic multi-extremal optimization algorithms.

Keywords: Bayesian optimization, Gaussian random fields.

DOI:https://doi.org/10.15181/csat.v7i0.1956


References


Abrahamsen P., 1997, A Review of Gaussian random fields and correlation functions, Norsk Regnesentral/Norwegian Computing Center, p. 64.

Gower J. C., 1984, Multivariate analysis: Ordination, multidimensional scaling and allied topics, Handbook of applicable mathematics Vol. VI: Statistics. (E. H. Lloyd, Ed.), J. Wiley and Sons, Chichester, P. 727-781.

Mockus J., 1989, Bayesian approach to global optimization, Mathematics and its Applications, Springer Netherlands, Kluwer Academic Publishers.

Pozniak N., Sakalauskas L., Saltyte L., 2019, Kriging Model with Fractional Euclidean Distance Matrices, Informatica, vol. 30, nr. 2, p. 367-390.

Pozniak, N., Sakalauskas, L., 2017, Fractional Euclidean distance matrices extrapolator for scattered data. Journal of Young Scientists. 47(2), 56–61.

Schoenberg I. J., 1935, Remarks to Maurice Frechet’s article: Sur la definition axiomatique d’une classe d’espaces vectoriels distancies applicables vectoriellement sur l’espace de Hilbert,” Ann. Math., Vol. 36, P. 724-732.

Weinberger K.Q, Sha F., Saul L.K., 2004, Learning a Kernel Matrix for Nonlinear Dimensionality Reduction, Proceedings of the twenty-first international conference on Machine learning (ICML '04), P. 106-113.


Full Text: PDF

Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.

eISSN: 2029-9966

Creative Commons License