grapherator-package {grapherator}

grapherator: A modular multi-step graph generator

Description

Due to lack of real world graphs, e.g., the optimization community often relies on artificial graphs to benchmark algorithms. The grapherator package implements a multi-step approach for the generation of weighted graphs. A set of predefined node, edge and weight generators allows for fast and convenient graph generation. Furthermore, the modular structure of the package enables writing user-defined generators and use them within the framework in a plug-and-play style.

Generation philosophy

The graph generation follows a three step procedure. A bare graph (see graph), i.e., an empty graph object, the following three serves as a staring point for several iterations of the following steps. Note that once edges have been added, no furhter nodes may be added. Likewise, after weights have been attached to edges, no further edges may be added. 1) Node generation via addNodes: nodes are generated by placing node coordinates in the two-dimensional Euclidean plane using different node generators. 2) Edge generation via addEdges: links between nodes are established via one or multiple edge generators. 3) Weight generation via addWeights: One or more weights are attached to each edge by different weight generators.