Skip to content

Latest commit

 

History

History
52 lines (35 loc) · 2.46 KB

README.md

File metadata and controls

52 lines (35 loc) · 2.46 KB

IOHdata

IOHinspector test datasets

In the folder 'IOHinspector', two new example datasets (generated using IOHinspector 0.3.18) can be found: one for single-objective and one for multi-objective optimization

'iohprofiler' folder

This repository contains several data sets of 12 reference algorithms tested on the 23 Pseudo-Boolean Optimization (PBO) problems. PBO is ones of the benchmark suite compiled under the framework of IOHexperimenter. The 12 reference algorithms are:

  • (1+(λ,λ)) GA
  • (1+1) EA_>0
  • (1+1) fGA
  • (1+10) EA_>0
  • (1+10) EA_logNormal
  • (1+10) EA_normalized
  • (1+10) EA_var_ctrl
  • (1+10) EA_{r/2,2r}
  • (30,30) vGA
  • RLS
  • UMDA (previously called EDA)
  • gHC

The detailed description of those algorithms can be found here.

When performing independent runs of those algorithms, one could decide whether to use exactly the same problem instance from all the runs, or to use different problem instances for every run. Hence, two data sets are produced:

  • 2019gecco-ins1-11run.rds: one one problem instance (ins1) is used for all 11 repetitions, and
  • 2019gecco-ins11-1run.rds: 11 problems instances are created and each algorithm is executed only once (1run) on each instance.

The prefix 2019gecco- says these data sets are used for the following GECCO '19 papers published by the authors (and collaborators):

  • Carola Doerr, Furong Ye, Naama Horesh, Hao Wang, Ofer M. Shir, and Thomas Bäck. Benchmarking Discrete Optimization Heuristics with IOHprofiler. GECCO ‘19
  • Borja Calvo, Ofer M. Shir, Josu Ceberio, Carola Doerr, Hao Wang, Thomas Bäck, and Jose A. Lozano. Bayesian Performance Analysis for Black-Box Optimization Benchmarking. GECCO ‘19

For the complete and up-to-date list of papers that use IOHprofiler, please look here.

Data sets are stored as the .rds files (one of the compressed data type in R). Internally, the data are organized in a S3 structure DataSetList which is implemented in IOHnanalyzer. To load the data set into you R session, please use the following commands:

library(IOHanalyzer)
dsl <- readRDS('path/to/the/downloaded/rds/file')
dsl

'bbob' folder

This contains some examples of BBOB-files processed using the R-interface of IOHanalzyer

'nevergrad' folder

This contains some examples of nevergrad-data processed using the R-interface of IOHanalyzer