Lukas Kühne

Generating matroids using HPC-GAP and ArangoDB

A matroid is a fundamental structure occurring in various fields of mathematics, such as combinatorics, algebra and optimization. It generalizes the notion of linear dependence in linear algebra. In this talk, we will describe a general framework of parallelized iterators which can be applied to generate matroids parallely. Other applications include e.g. the generation of phylogenetic trees. This approach is implemented in HPC-GAP and we will also show experimental results and run times. Finally using the new ArangoDB GAP package, we have established a database of matroids which can be used for future experiments quickly. This is joint work with Mohamed Barakat, Reimer Behrends, and Chris Jefferson.