An improved algorithm for subdivision traversal without extra storage

Abstract

We describe an algorithm for enumerating all vertices, edges and faces of a planar subdivision stored in any of the usual pointer-based representations, while using only a constant amount of memory beyond that required to store the subdivision. The algorithm is a refinement of a method introduced by de Berg et al (1997), that reduces the worst case running time from O(n2) to O(n log n). We also give experimental results that show that our modified algorithm runs faster not only in the worst case, but also in many realistic cases.