Data Sets for the study "Leveraging Benchmarking Data for Informed One-Shot Dynamic Algorithm Selection"

This is the result of the study "Leveraging Benchmarking Data for Informed One-Shot Dynamic Algorithm Selection". A One-shot informed dynamic genetic algorithm selection (DAS) is promoted. We present the theoretical performance and experimental results of the DAS on IOHprofiler problems.

This Data set consists of 2 parts:

1. The theoretical performance of the DAS policies for 25 pseudo-Boolean problems defined in IOHprofiler (https://iohprofiler.github.io/). The DAS policies are combinations over 80 GAs and 38 switch targets.

2. The experimental result of 100 DAS polices for the IOHprofiler problems and result of the DAS polices selected with constraints for F8. The 100 DAS polices are the best ones of the theoretical performance, and the DAS polices tested for F8 constrains the number of GAs being selected for A1 and A2 not exceeding 20 respectively. 

Contact: if you have any questions or suggestions, please feel free to contact Furong Ye .