Finding Closed Quasigeodesics on Convex Polyhedra
Erik D. Demaine, Adam C. Hesterberg, Jason S. Ku
TL;DR
This work resolves the long-standing open problem of constructively finding a closed quasigeodesic on a convex polyhedron by presenting the first finite algorithm with provable bounds. The method centers on a Real RAM traversal of a state-graph whose edges encode feasible quasigeodesic progress and on controlled unfolding arguments to bound the number of face traversals; the authors prove a pseudopolynomial upper bound on the total number of face visits and provide a near-linear-time result on the Real RAM. To bridge theory with practical finite-precision computation, they introduce the expression RAM, enabling rigorous cost analyses of algorithms manipulating radical expressions and showing how to translate the Real RAM algorithm to the Word RAM with acceptable slowdown. The paper also develops a comprehensive cost framework for radical-expression computations, including a recursive and simple cost model, and demonstrates how to extend these results to multiple input representations of polyhedra via Alexandrov gluing. Overall, the work delivers a foundational, verifiable algorithm for a classical geometric problem and a versatile computational model for exact geometric computation.
Abstract
A closed quasigeodesic is a closed curve on the surface of a polyhedron with at most $180^\circ$ of surface on both sides at all points; such curves can be locally unfolded straight. In 1949, Pogorelov proved that every convex polyhedron has at least three (non-self-intersecting) closed quasigeodesics, but the proof relies on a nonconstructive topological argument. We present the first finite algorithm to find a closed quasigeodesic on a given convex polyhedron, which is the first positive progress on a 1990 open problem by O'Rourke and Wyman. The algorithm also establishes a pseudopolynomial upper bound on the total number of visits to faces (number of line segments), namely, $O\left(\frac{n \, L^2}{ε^2 \, \ell^2}\right)$ where $n$ is the number of vertices of the polyhedron, $ε$ is the minimum curvature of a vertex, $L$ is the length of the longest edge, and $\ell$ is the smallest distance within a face between a vertex and a nonincident edge (minimum feature size of any face). On the real RAM, the algorithm's running time is also pseudopolynomial, namely $O\left(\frac{L^2}{ε^2 \, \ell^2} \, n \lg n\right)$. On a word RAM, the running time grows to $O\left(\frac{b^2 \, Δ^{36} \, L^{146}}{ε^{98} \, \ell^{146}} \, n \lg n \cdot 2^{O(|R|)}\right)$, where $Δ\leq n$ is the polyhedron's maximum vertex degree, assuming the polyhedron's intrinsic geometry is given by constant-size radical expressions with $b$-bit integers and at most $|R|$ distinct square-roots. Along the way, we introduce the expression RAM model of computation, formalizing a connection between the real RAM and word RAM hinted at by past work on exact geometric computation.
