The optimal algorithm for dynamic support of the Voronoi Diagram for a set of points

Abstract

The article is devoted to the development of a dynamic data structure for solving proximity problems based on the dynamic Voronoi Diagram. This data structure can be used as the core of the common algorithmic space model for solving a set of visualization and computer modeling problems. The data structure is based on the strategy of "divide and rule" for Voronoi diagram construction. Similar to the original algorithm, we store a binary tree that represents the Voronoi diagram, but define three new operations: insert, delete, and balance. To ensure the efficiency of operations, it is proposed to use red-black tree. In general, the proposed data structure shows much better results than the original static algorithm. Compared to existing algorithms, this data structure is both simple and efficient.