A comparison of optimisation algorithms for high-dimensional particle and astrophysics applications
Abstract
Optimisation problems are ubiquitous in particle and astrophysics, and involve locating the optimum of a complicated function of many parameters that may be computationally expensive to evaluate. We describe a number of global optimisation algorithms that are not yet widely used in particle astrophysics, benchmark them against random sampling and existing techniques, and perform a detailed comparison of their performance on a range of test functions. These include four analytic test functions of varying dimensionality, and a realistic example derived from a recent global fit of weak-scale supersymmetry. Although the best algorithm to use depends on the function being investigated, we are able to present general conclusions about the relative merits of random sampling, Differential Evolution, Particle Swarm Optimisation, the Covariance Matrix Adaptation Evolution Strategy, Bayesian Optimisation, Grey Wolf Optimisation, and the PyGMO Artificial Bee Colony, Gaussian Particle Filter and Adaptive Memory Programming for Global Optimisation algorithms.
Cite
@article{arxiv.2101.04525,
title = {A comparison of optimisation algorithms for high-dimensional particle and astrophysics applications},
author = {The DarkMachines High Dimensional Sampling Group and Csaba Balázs and Melissa van Beekveld and Sascha Caron and Barry M. Dillon and Ben Farmer and Andrew Fowlie and Eduardo C. Garrido-Merchán and Will Handley and Luc Hendriks and Guðlaugur Jóhannesson and Adam Leinweber and Judita Mamužić and Gregory D. Martinez and Sydney Otten and Pat Scott and Roberto Ruiz de Austri and Zachary Searle and Bob Stienen and Joaquin Vanschoren and Martin White},
journal= {arXiv preprint arXiv:2101.04525},
year = {2021}
}
Comments
Experimental framework publicly available at http://www.github.com/darkmachines/high-dimensional-sampling