Benchmark-Driven Algorithm Configuration Applied to Parallel Model-Based Optimization

Supplementary Codes and Data for "Benchmark-Driven Algorithm Configuration Applied to Parallel Model-Based Optimization"

Most of the codes are meant to be run through SLURM (a Linux task scheduler).
Yet, they can also be easily adapted to be run on any other machine.

The following files are included:

* 01_experiments.R starter script that will generate many SLURM jobs, one for each algorithm/setup/function/instance combination. 
* 01_plotsAnalysis.R Takes the results of those experiments and generates the figures which are shown in the paper. Additionally, a large amount of supplementary figures is generated.
* The folder imgResults contains all these supplementary figures.
* algorithm_xx: Each algorithm that was used in the experiments is given one R script. The scripts expect to be passed command line parameters such as which function to run on and how much budget to use. 
* analysis_xx files are additional rmd files that were used for the supplementary analysis of data etc. 
* expResults includes all raw results of the experiments as well as the preprocessed data stored into data frames.
* helperScriptsFunctions includes an assortement of additional scripts used for plotting and statistical analysis of the data
* simulatedFunctions includes the original functions generated by Kriging Simulations. Additional functions can be generated with the analysis_04 script.