Benchmarking Gaussian Processes and Random Forests Surrogate Models on the BBOB Noiseless Testbed

0446912 - ÚI 2016 RIV US eng C - Konferenční příspěvek (zahraniční konf.)
Bajer, Lukáš - Pitra, Z. - Holeňa, Martin
Benchmarking Gaussian Processes and Random Forests Surrogate Models on the BBOB Noiseless Testbed.
GECCO Companion '15. Genetic and Evolutionary Computation Conference. Companion Material Proceedings. New York: ACM, 2015 - (Silva, S.), s. 1143-1150. ISBN 978-1-4503-3488-4.
[GECCO Companion '15. Genetic and Evolutionary Computation Conference. Madrid (ES), 11.07.2015-15.07.2015]
Grant CEP: GA ČR GA13-17187S
Grant ostatní: ČVUT(CZ) SGS14/205/OHK4/3T/14; GA MŠk(CZ) ED2.1.00/03.0078; GA MŠk(CZ) LM2010005
Institucionální podpora: RVO:67985807
Klíčová slova: benchmarking * black-box optimization * surrogate model * Gaussian process * random forest
Kód oboru RIV: IN - Informatika

Speeding-up black-box optimization algorithms via learning and using a surrogate model is a heavily studied topic. This paper evaluates two different surrogate models: Gaussian processes and random forests which are interconnected with the state-of-the art optimization algorithm CMA-ES. Results on the BBOB testing set show that considerable amount of fitness evaluations can be saved especially during the initial phase of the algorithm's progress.
Trvalý link: http://hdl.handle.net/11104/0248874