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

Journal of High Energy Physics

Martin White¹  Zachary Searle¹  Adam Leinweber¹  Ben Farmer²  Will Handley³  Andrew Fowlie⁴  Joaquin Vanschoren⁵  Sydney Otten⁶  Sascha Caron⁷  Bob Stienen⁷  Luc Hendriks⁷  Eduardo C. Garrido-Merchán⁸  Barry M. Dillon⁹  Roberto Ruiz de Austri^1,10  Judita Mamužić^1,10  Gregory D. Martinez^1,11  Melissa van Beekveld^1,12  Pat Scott^1,13  Csaba Balázs^1,14  Guðlaugur Jóhannesson^1,15 

[1] ARC Centre for Dark Matter Particle Physics, Department of Physics, University of Adelaide, 5005, Adelaide, SA, Australia;Bureau of Meteorology, 3001, Melbourne, VIC, Australia;Cavendish Laboratory, University of Cambridge, JJ Thomson Avenue, CB3 0HE, Cambridge, U.K.;Kavli Institute for Cosmology, Madingley Road, CB3 0HA, Cambridge, U.K.;Department of Physics and Institute of Theoretical Physics, Nanjing Normal University, 210023, Nanjing, Jiangsu, China;Eindhoven University of Technology, Groene Loper 5, 5612 AZ, Eindhoven, Netherlands;High Energy Physics, IMAPP, Radboud University Nijmegen, Heyendaalseweg 135, 6525 AJ, Nijmegen, Netherlands;Gravitation Astroparticle Physics Amsterdam (GRAPPA), Institute for Theoretical Physics Amsterdam and Delta Institute for Theoretical Physics, University of Amsterdam, Science Park 904, 1098 XH, Amsterdam, The Netherlands;High Energy Physics, IMAPP, Radboud University Nijmegen, Heyendaalseweg 135, 6525 AJ, Nijmegen, Netherlands;Nikhef, Science Park 105, 1098 XG, Amsterdam, Netherlands;Higher Polytechnic School, Universidad Autonoma de Madrid, Francisco Tomas y Valiente 25, Madrid, Spain;Institut für Theoretische Physik, Universität Heidelberg, Heidelberg, Germany;Instituto de Física Corpuscular, IFIC-UV/CSIC, Carrer del Catedrátic José Beltrán Martinez, 2, Valencia, Spain;Physics and Astronomy Department, University of California, 90095, Los Angeles, CA, U.S.A.;Rudolf Peierls Centre for Theoretical Physics, 20 Parks Road, OX1 3PU, Oxford, U.K.;School of Mathematics and Physics, The University of Queensland, St. Lucia, 4072, Brisbane, QLD, Australia;Department of Physics, Imperial College London, Blackett Laboratory, Prince Consort Road, SW7 2AZ, London, U.K.;School of Physics and Astronomy, Monash University, 3800, Melbourne, VIC, Australia;Science Institute, University of Iceland, Dunhaga 7, IS-107, Reykjavik, Iceland;Nordita, KTH Royal Institute of Technology and Stockholm University, Roslagstullsbacken 23, SE-106 91, Stockholm, Sweden;

关键词: Phenomenology of Field Theories in Higher Dimensions;    Supersymmetry Phenomenology;   

DOI  :  10.1007/JHEP05(2021)108

来源: Springer

【 摘 要 】

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.

【 授权许可】

CC BY