Optimizing agents with genetic programming - An evaluation of hyper-heuristics in dynamic real-time logistics - datasets and results

This directory contains the data and results that we used and obtained during the realization of the following paper:

Optimizing agents with genetic programming - An evaluation of hyper-heuristics in dynamic real-time logistics. Rinde R.S. van Lon, Juergen Branke, and Tom Holvoet. Genetic Programming and Evolvable Machines (2017).

The code that has been used can be found on GitHub (https://github.com/rinde/vanLon17-GPEM-code), an archive is available at https://doi.org/10.5281/zenodo.260130.

This repository contains the following files:

• evo.zip (expands to 21.2 GB) A zip file containing all results of the training experiment and tuning experiment.
• JAAMAS.zip (expands to 227.6 MB) A zip file containing the main results of the JAAMAS experiment, also available at https://doi.org/10.5281/zenodo.209760
• realtime.zip (expands to 913.5 MB) A zip file containing the main results
• dataset10k.zip (expands to 4.9 GB) Contains the scenarios used for the training and tuning experiments.
• overview.zip (expands to 42 MB) Contains the following files:
  • overview/experiment-overview.csv - table containing basic overview of the 40 main evolutionary runs that were performed for this paper.
  • overview/heuristics/ - folder with a tree visualization for every evolved heuristic.
  • overview/makefile - executes the analysis scripts that were used in the paper (requires the R programming language, https://www.r-project.org/). All tables and figures will be placed in results/generated/.
  • readme.txt - this file.
  • overview/results/ - contains the scripts
  • overview/results/data/ - empty directory in which evo.zip, JAAMAS.zip, and realtime.zip have to be unpacked for the analysis scripts to work.