Noisy test suite

There are two ways to run experiments on a noisy test suite.

  1. Using the bbob-noisy suite (Hansen et al. 2009), suite = cocoex.Suite('bbob-noisy', '', ''), see also the function visualizations. This suite features 30 functions, f101–f130, with three different noise models in two different magnitudes based on 8 of the 24 functions from the bbob test suite.
  2. Using the cocoex.noiser.Noisifier (since 2025, only available in Python) as described in more detail in this issue.1 The noisifier can be used with any COCO test suite. The resulting data can be post-processed as usual. However, the post-processing does not recognize whether a noisifier was applied during the experiment. Hence the figures themselves won’t give any hints on the applied noise parameters.

References

Hansen, N., Finck, S., Ros, R., and Auger, A. (2009), Real-parameter black-box optimization benchmarking 2009: Noisy functions definitions, research report RR-6869, Inria.

Footnotes

  1. [Feature description] A Noisifier for objective functions. Issue #36 of the numbbo/coco-experiment Github repository.↩︎