## An Efficient Randomized Algorithm for Higher-Order Abstract Voronoi Diagrams

- Cecilia Bohler, Rolf Klein, Chih-Hung Liu
- Symposium on Computational Geometry
- 2016

- Published 2013 in SODA

We investigate the higher-order Voronoi diagrams of n point sites with respect to the geodesic distance in a simple polygon with h > 0 polygonal holes and c corners. Given a set of n point sites, the korder Voronoi diagram partitions the plane into several regions such that all points in a region share the same k nearest sites. The nearest-site (first-order) geodesic Voronoi diagram has already been well-studied, and its total complexity isO(n+c). On the other hand, Bae and Chwa [3] recently proved that the total complexity of the farthest-site ((n− 1)-order) geodesic Voronoi diagram and the number of faces in the diagram are Θ(nc) and Θ(nh), respectively. It is of high interest to know what happens between the first-order and the (n− 1)order geodesic Voronoi diagrams. In this paper we prove that the total complexity of the k-order geodesic Voronoi diagram is Θ(k(n − k) + kc), and the number of faces in the diagram is Θ(k(n− k)+ kh). Our results successfully explain the variation from the nearest-site to the farthest-site geodesic Voronoi diagrams, i.e., from k = 1 to k = n − 1, and also illustrate the formation of a disconnected Voronoi region, which does not occur in many commonly used distance metrics, such as the Euclidean, L1, and city metrics. We show that the k-order geodesic Voronoi diagram can be computed in O(k(n+c) log(n+c)) time using an iterative algorithm.

@inproceedings{Liu2013HigherOrderGV,
title={Higher-Order Geodesic Voronoi Diagrams in a Polygonal Domain with Holes},
author={Chih-Hung Liu and Charles Tzu-Chi Lee},
booktitle={SODA},
year={2013}
}