This paper presents a numerically robust algorithm to construct a Voronoi diagram of circles in the plane. The circles are allowed to have intersections among them, but one circle cannot fully contain another circle. The Voronoi diagram is a tessellation of the plane into Voronoi regions of given circles. Each circle has its Voronoi region which is defined by a set of points in the plane closer to the circle than any other circles. The distance from a point p to a circle ci of center pi and radius ri is ||p-pi||-ri, which is the closest Euclidean distance from p to the circle boundary. The proposed algorithm first constructs the point Voronoi diagram of centers of given circles, then it enlarges each point to the circle and expands its Voronoi region accordingly. This region-expansion process is done by local modifications and after completing this process for the whole circles the desired circle Voronoi diagram can be obtained. The proposed algorithm is numerically robust and we provide with a few examples to show its robustness. The algorithm runs in O(n2) time in the worst case and O(n) time on average where n is the number of the circles. The experiment shows that the region-expansion algorithm is robust and runs fast with strong linear time behavior.

Presented in this paper is an algorithm to compute a Voronoi diagram of circles in a circle, where circles are located in a large circle. Given circles in a large circle, the region in the plane is divided into regions associated with the given circles. The proposed algorithm uses point Voronoi diagram, and then some topological remedies are applied so that we obtain proper initial topology including enclosing circle. From this initial topology, we can obtain the correct topology by a series of edge-flip operations. After getting the correct topology, the equations of edges are computed and represented in a rational quadratic Bézier curve form.