A Principle for Global Optimization with Gradients
Optimization and Control
2023-08-21 v1 Numerical Analysis
Neural and Evolutionary Computing
Numerical Analysis
Machine Learning
Abstract
This work demonstrates the utility of gradients for the global optimization of certain differentiable functions with many suboptimal local minima. To this end, a principle for generating search directions from non-local quadratic approximants based on gradients of the objective function is analyzed. Experiments measure the quality of non-local search directions as well as the performance of a proposed simplistic algorithm, of the covariance matrix adaptation evolution strategy (CMA-ES), and of a randomly reinitialized Broyden-Fletcher-Goldfarb-Shanno (BFGS) method.
Cite
@article{arxiv.2308.09556,
title = {A Principle for Global Optimization with Gradients},
author = {Nils Müller},
journal= {arXiv preprint arXiv:2308.09556},
year = {2023}
}
Comments
16 pages, 3 figures