TY - JOUR
T1 - Benchmarking surrogate-based optimisation algorithms on expensive black-box functions
AU - Bliek, Laurens
AU - Guijt, Arthur
AU - Karlsson, Rickard
AU - Verwer, Sicco
AU - de Weerdt, Mathijs
PY - 2023
Y1 - 2023
N2 - Surrogate algorithms such as Bayesian optimisation are especially designed for black-box optimisation problems with expensive objectives, such as hyperparameter tuning or simulation-based optimisation. In the literature, these algorithms are usually evaluated with synthetic benchmarks which are well established but have no expensive objective, and only on one or two real-life applications which vary wildly between papers. There is a clear lack of standardisation when it comes to benchmarking surrogate algorithms on real-life, expensive, black-box objective functions. This makes it very difficult to draw conclusions on the effect of algorithmic contributions and to give substantial advice on which method to use when. A new benchmark library, EXPObench, provides first steps towards such a standardisation. The library is used to provide an extensive comparison of six different surrogate algorithms on four expensive optimisation problems from different real-life applications. This has led to new insights regarding the relative importance of exploration, the evaluation time of the objective, and the used model. We also provide rules of thumb for which surrogate algorithm to use in which situation. A further contribution is that we make the algorithms and benchmark problem instances publicly available, contributing to more uniform analysis of surrogate algorithms. Most importantly, we include the results of the six algorithms on all evaluated problem instances. This unique new dataset lowers the bar for researching new methods as the number of expensive evaluations required for comparison and for the creation of new surrogate models is significantly reduced.
AB - Surrogate algorithms such as Bayesian optimisation are especially designed for black-box optimisation problems with expensive objectives, such as hyperparameter tuning or simulation-based optimisation. In the literature, these algorithms are usually evaluated with synthetic benchmarks which are well established but have no expensive objective, and only on one or two real-life applications which vary wildly between papers. There is a clear lack of standardisation when it comes to benchmarking surrogate algorithms on real-life, expensive, black-box objective functions. This makes it very difficult to draw conclusions on the effect of algorithmic contributions and to give substantial advice on which method to use when. A new benchmark library, EXPObench, provides first steps towards such a standardisation. The library is used to provide an extensive comparison of six different surrogate algorithms on four expensive optimisation problems from different real-life applications. This has led to new insights regarding the relative importance of exploration, the evaluation time of the objective, and the used model. We also provide rules of thumb for which surrogate algorithm to use in which situation. A further contribution is that we make the algorithms and benchmark problem instances publicly available, contributing to more uniform analysis of surrogate algorithms. Most importantly, we include the results of the six algorithms on all evaluated problem instances. This unique new dataset lowers the bar for researching new methods as the number of expensive evaluations required for comparison and for the creation of new surrogate models is significantly reduced.
KW - Bayesian optimisation
KW - Benchmarking
KW - Expensive optimisation
KW - Surrogate-based optimisation
UR - http://www.scopus.com/inward/record.url?scp=85170226208&partnerID=8YFLogxK
U2 - 10.1016/j.asoc.2023.110744
DO - 10.1016/j.asoc.2023.110744
M3 - Article
AN - SCOPUS:85170226208
SN - 1568-4946
VL - 147
JO - Applied Soft Computing
JF - Applied Soft Computing
M1 - 110744
ER -