## Introduction

Complex networks have become increasingly popular for modeling various real-world phenomena. Realistic generative network models are important in this context as they simplify complex network research regarding data sharing, reproducibility, and scalability studies. Random hyperbolic graphs are a very promising family of geometric graphs with unit-disk neighborhood in the hyperbolic plane. Previous work provided empirical and theoretical evidence that this generative graph model creates networks with many realistic features. In our recent work we provide the first generation algorithm for random hyperbolic graphs with subquadratic running time. We prove a time complexity of \(O((n^{3/2}+m) \log n)\) with high probability for the generation process.## Generation Algorithm

In the generative model, vertices are generated as points in polar coordinates \((\phi, r)\) on a disk of radius \(R\) in the hyperbolic plane with curvature \(-\zeta^2\). Any two vertices \(u\) and \(v\) are connected by an edge if their hyperbolic distance \(\mathrm{dist}_{\mathcal{H}}(u,v)\) is below \(R\). (A more general model with non-deterministic edges exists, but is not considered in this implementation. An extended implementation for the general model is available on request.) A straightforward implementation would iterate over all point pairs to generate the graph, leading to a quadratic time complexity. Instead, we use the Poincaré disk model to map the hyperbolic plane into the Euclidean unit disk. This model of hyperbolic space maps hyperbolic circles to Euclidean circles, the neighborhood of a vertex is then an Euclidean circle with a moved center: Getting all points that are in an Euclidean circle can be modeled as a range query and we can use a polar quadtree on the Poincaré disk to reduce the time complexity of this query.## Publication and Implementation

The corresponding article is due to appear in the proceedings of*26th Int'l Symp. on Algorithms and Computation*(ISAAC 2015). More material is available in the preprint on arxiv. The implementation is available on the Dev branch of NetworKit, our network analysis framework.