1: \begin{abstract}

2:       Bayesian optimi\sz{}ation (BO) is a popular method for optimi\sz{}ing expensive black-box functions. BO has several well-documented shortcomings, including computational slowdown with longer optimi\sz{}ation runs, poor suitability for non-stationary or ill-conditioned objective functions, and poor convergence characteristics. Several algorithms have been proposed that incorporate local strategies, such as trust regions, into BO to mitigate these limitations; however, none address all of them satisfactorily. To address these shortcomings, we propose the \ALGPA{} algorithm, which extends \revision{trust-region-based} BO by adding a \revision{rotation aligning the trust region with the weighted principal components} and an adaptive rescaling strategy based on the length-scales of a local Gaussian process surrogate model with automatic relevance determination. Through extensive numerical experiments using a set of synthetic test functions and the well-known COCO benchmarking software, we show that the \ALGPA{} algorithm outperforms several state-of-the-art BO and other black-box optimi\sz{}ation algorithms.

3:   \end{abstract}

4: