Investigating benchmark correlations when comparing algorithms with parameter tuning.

Christie, Lee A.; Brownlee, Alexander E.I.; Woodward, John R.

Abstract

Benchmarks are important for comparing performance of optimisation algorithms, but we can select instances that present our algorithm favourably, and dismiss those on which our algorithm under-performs. Also related are automated design of algorithms, which use problem instances (benchmarks) to train an algorithm: careful choice of instances is needed for the algorithm to generalise. We sweep parameter settings of differential evolution to applied to the BBOB benchmarks. Several benchmark functions are highly correlated. This may lead to the false conclusion that an algorithm performs well in general, when it performs poorly on a few key instances. These correlations vary with the number of evaluations.

Citation

CHRISTIE, L.A., BROWNLEE, A.E.I. and WOODWARD, J.R. 2018. Investigating benchmark correlations when comparing algorithms with parameter tuning. In Aguirre, H.E. (ed.) Proceedings of the 2018 Genetic and evolutionary computation conference companion (GECCO'18 companion), 15-19 July 2018, Kyoto, Japan. New York: Association for Computing Machinery [online], pages 209-210. Available from: https://doi.org/10.1145/3205651.3205747

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 2018 Genetic and evolutionary computation conference (GECCO'18)
• Start Date: Jul 15, 2018
• End Date: Jul 19, 2018
• Acceptance Date: Mar 24, 2018
• Online Publication Date: Jul 6, 2018
• Publication Date: Jul 31, 2018
• Deposit Date: Mar 9, 2020
• Publicly Available Date: Jul 31, 2018
• Publisher: Association for Computing Machinery (ACM)
• Peer Reviewed: Peer Reviewed
• Pages: 209-210
• Book Title: Proceedings of the 2018 Genetic and evolutionary computation conference companion (GECCO'18 companion)
• ISBN: 9781450357647
• DOI: https://doi.org/10.1145/3205651.3205747
• Keywords: Benchmarking; Benchmarks; Black-Box Optimization Benchmarking (BBOB); Differential evolution; Continuous optimisation; Algorithm automation
• Public URL: https://rgu-repository.worktribe.com/output/876279
• Related Public URLs: https://rgu-repository.worktribe.com/output/876294