search

numbbo / coco

Numerical Black-Box Optimization Benchmarking Framework
https://numbbo.github.io/coco
Other
262 stars 87 forks source link

[DATA SUBMISSION] BORG MOEA on BBOB-BiObj #2203

Closed Slickytail closed 1 year ago

Slickytail commented 1 year ago

Reference

Submitted to GECCO 2023 Workshop, will update once accepted/rejected

Description of the Algorithm(s)

BORG-MOEA: Borg MOEA with epsilon = 10^-4. BORG-MOEA-adapt: Borg MOEA with epsilon set dynamically per objective, with eta = 10^-2.. See paper and code for details.

Link to Data

https://slickytail.github.io/borg-adaptive-epsilon/data-archives

Source Code of Experiment

slickytail/borg-adaptive-epsilon

brockho commented 1 year ago

Dear @Slickytail,

Many thanks for your submission of the BORG data. Unfortunately, I cannot download it because the above link does not work for me. The link to the source code, though, is working. Is it possible that you forgot to push your changes?

Slickytail commented 1 year ago

@brockho hello, the link I gave is to the root folder, which doesn't have any content. But the following links should work:

https://slickytail.github.io/borg-adaptive-epsilon/data-archives/coco_archive_definition.txt https://slickytail.github.io/borg-adaptive-epsilon/data-archives/BORG-MOEA.zip https://slickytail.github.io/borg-adaptive-epsilon/data-archives/BORG-MOEA-adapt.zip

brockho commented 1 year ago

Many thanks again for your data submission. Both Borg data sets should be now available via the postprocessing of COCO. If you don't see the entries in cocopp.bbob_biobj, you will have to type import cocopp; cocopp.archives.update_all() in your favorite python shell/notebook first (once should be enough).