A comparison of optimisation algorithms for high-dimensional particle and astrophysics applications

• Published in: The journal of high energy physics Vol. 2021; no. 5; pp. 1 - 46
• Main Authors: Balázs, Csaba, van Beekveld, Melissa, Caron, Sascha, Dillon, Barry M., Farmer, Ben, Fowlie, Andrew, Garrido-Merchán, Eduardo C., Handley, Will, Hendriks, Luc, Jóhannesson, Guðlaugur, Leinweber, Adam, Mamužić, Judita, Martinez, Gregory D., Otten, Sydney, de Austri, Roberto Ruiz, Scott, Pat, Searle, Zachary, Stienen, Bob, Vanschoren, Joaquin, White, Martin
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.05.2021; Springer Nature B.V; SpringerOpen
• Subjects: Algorithms; Astrophysics; Classical and Quantum Gravitation; Covariance matrix; Elementary Particles; Evolutionary computation; Global optimization; High energy physics; Mathematical analysis; Particle swarm optimization; Phenomenology of Field Theories in Higher Dimensions; Physics; Physics and Astronomy; Quantum Field Theories; Quantum Field Theory; Quantum Physics; Random sampling; Regular Article - Theoretical Physics; Relativity Theory; String Theory; Supersymmetry; Supersymmetry Phenomenology; Phenomenology of Field Theories in Higher Dimensions; Supersymmetry Phenomenology
• ISSN: 1029-8479; 1126-6708; 1029-8479
• DOI: 10.1007/JHEP05(2021)108

A bstract 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.