A Data-Driven Bayesian Nonparametric Approach for Black-Box Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Haowei, Zhang, Xun, Ng, Szu Hui, Wang, Songhao
Format: Preprint
Published: 2020
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910344372289536
author Wang, Haowei
Zhang, Xun
Ng, Szu Hui
Wang, Songhao
author_facet Wang, Haowei
Zhang, Xun
Ng, Szu Hui
Wang, Songhao
contents We present a data-driven Bayesian nonparametric approach for global optimization (DaBNO) of stochastic black-box function. The function value depends on the distribution of a random vector. However, this distribution is usually complex and hardly known in practice, and is often inferred from data (realizations of random vectors). The DaBNO accounts for the finite-data error that arises when estimating the distribution and relaxes the commonly-used parametric assumption to reduce the distribution-misspecified error. We show that the DaBNO objective formulation can converge to the true objective asymptotically. We further develop a surrogate-assisted algorithm DaBNO-K to efficiently optimize the proposed objective function based on a carefully designed kernel. Numerical experiments are conducted with several synthetic and practical problems, demonstrating the empirical global convergence of this algorithm and its finite-sample performance.
format Preprint
id arxiv_https___arxiv_org_abs_2008_02154
institution arXiv
publishDate 2020
record_format arxiv
spellingShingle A Data-Driven Bayesian Nonparametric Approach for Black-Box Optimization
Wang, Haowei
Zhang, Xun
Ng, Szu Hui
Wang, Songhao
Optimization and Control
We present a data-driven Bayesian nonparametric approach for global optimization (DaBNO) of stochastic black-box function. The function value depends on the distribution of a random vector. However, this distribution is usually complex and hardly known in practice, and is often inferred from data (realizations of random vectors). The DaBNO accounts for the finite-data error that arises when estimating the distribution and relaxes the commonly-used parametric assumption to reduce the distribution-misspecified error. We show that the DaBNO objective formulation can converge to the true objective asymptotically. We further develop a surrogate-assisted algorithm DaBNO-K to efficiently optimize the proposed objective function based on a carefully designed kernel. Numerical experiments are conducted with several synthetic and practical problems, demonstrating the empirical global convergence of this algorithm and its finite-sample performance.
title A Data-Driven Bayesian Nonparametric Approach for Black-Box Optimization
topic Optimization and Control
url https://arxiv.org/abs/2008.02154