A graph algorithm library based on compact data structures by Joana Modesto Hrotko MSc thesis presentation and discussion. Date: 2021-Jan-13 Time: 18:00 Room: Zoom Abstract: We address the problem of representing dynamic graphs using k2-trees. The k2-tree data structure is one of the succinct data structures proposed for representing static graphs, and binary relations in general. It relies on compact representations of static bit vectors. By adding dynamism to the static compact data structures, we can also represent dynamic graphs. However, this approach suffers from a well known bottleneck in compressed dynamic indexing, the problem of maintaining a changing collection so that we can query the data structure efficiently. In this work we present a k2-tree based implementation which follows instead the ideas by Munro to circumvent this bottleneck. We refactored and extended the work of Coimbra by building a C++library. The library includes efficient edge and neighborhood iterators, as well as some illustrative algorithms. We also included a study on the add operation first proposed by Munro. Our experimental results show that our implementation is competitive in practice.