Supplemental material for the paper: "Subfunction Structure Matters: a New Perspective on Local Optima Networks"

Supplemental material for the paper: "Subfunction Structure Matters: a New Perspective on Local Optima Networks" This repository contains code, data, and plots. In the construct-lon folder is all code [construct-lon-x.py, where x is the algorithm shortname] and instance information needed to construct LONs using the three different algorithms and for the considered problems. The dist-and-corr-plots folder contains distribution plots for the LON subfunction metrics and also correlation plots for LON metrics for all algorithms and problems. The subfunction-metrics-and-lon-viz folder contains the code used to compute LON subfunction metrics [analyse-subfuncs.py inside the subfolders]. Finally, the traditional-lon-metrics folder contains code to parse LONs and then compute a set of metrics from the literature on them. Pipeline: from within a problem folder [e.g. "dec"]: python3 parse-lons.py -> ./construct-lon.R -> ./cmlon-construct.R -> cd lons -> ./lon-metrics-funnel.R