The COCO ppdata Archive — postprocessed data for browsing
Here we provide already postprocessed data from benchmarking experiments for browsing. The bases for these data are the official benchmarking data archives of the COCO platform. For each test suite we provide a separate page, listing the official data sets by year. Data sets for the following test suites are available:
- bbob: 250+ algorithm data sets for unconstrained, noiseless optimization
- bbob-noisy: 40+ algorithm data sets for noisy optization
- bbob-biobj: 40+ algorithm data sets for bi-objective optimization
- bbob-largescale: 10+ algorithm data sets for large-scale optimization (20 < dim < 640)
- bbob-mixint: 8 algorithm data sets for mixed-integer optimization
- bbob-constrained: 9 algorithm data sets for constrained optimization
- bbob-boxed: 3 algorithm data sets for box-constrained problems, also known as
sbox-cost
There is also a machine readable JSON file with data from the experiments for programmatic access. If you are using the cocopp Python module, the data is directly accessible “by name”.