Quasigeodesics on the Cube

MIT CompGeom Group, Hugo A. Akitaya, Erik D. Demaine, Adam Hesterberg, Thomas C. Hull, Anna Lubiw, Jayson Lynch, Klara Mundilova, Chie Nara, Joseph O’Rourke, Frederick Stock, Josef Tkadlec, Ryuhei Uehara · 2025

Read Quasigeodesics on the Cube by MIT CompGeom Group, Hugo A. Akitaya, Erik D. Demaine, Adam Hesterberg, Thomas C. Hull, Anna Lubiw, Jayson Lynch, Klara Mundilova, Chie Nara, Joseph O’Rourke, Frederick Stock, Josef Tkadlec, Ryuhei Uehara (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-10376.txt.

Introduction

The MIT CompGeom Group with Hugo Akitaya and co-authors count in a 2025 paper the closed quasigeodesics on a cube. Such a curve keeps at most a half plane of surface on each side at every point. Unlike true geodesics, these may pass through corners of the solid.

Every convex solid holds at least three such closed loops, a known result. Beyond the tetrahedron, counts were open until this cube study. The cube holds exactly 15 simple closed quasigeodesics past the three known geodesics. Fifteen examples are built by hand for the lower bound in the paper. A matching upper bound, the main feat, closes the count for the cube.

Catalog record

CREATORMIT CompGeom Group, Hugo A. Akitaya, Erik D. Demaine, Adam Hesterberg, Thomas C. Hull, Anna Lubiw, Jayson Lynch, Klara Mundilova, Chie Nara, Joseph O’Rourke, Frederick Stock, Josef Tkadlec, Ryuhei Uehara
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE17 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-10376

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Quasigeodesics on the Cube MIT CompGeom Group, Hugo A. Akitaya, Erik D. Demaine, Adam Hesterberg, Thomas C. Hull, Anna Lubiw, Jayson Lynch, Klara Mundilova, Chie Nara, Joseph O’Rourke, Frederick Stock, Josef Tkadlec, Ryuhei Uehara Abstract A quasigeodesic is a curve on the surface of a convex polyhedron that has ≤ π \leq\pi surface to each side at every point. In contrast, a geodesic has exactly π \pi to each side and so can never pass through a vertex, whereas quasigeodesics can. Although it is known that every convex polyhedron has at least three simple closed quasigeodesics, little else is known. Only tetrahedra have been thoroughly studied. In this paper we explore the quasigeodesics on a cube, which have not been previously enumerated. We prove that the cube has exactly 15 15 simple closed quasigeodesics (beyond the three known simple closed geodesics). For the lower bound we detail 15 15 simple closed quasigeodesics. Our main contribution is establishing a matching upper bound. For general convex polyhedra, there is no known upper bound. 1 Introduction 1.1 Quasigeodesics A quasigeodesic is a curve on the surface of a convex polyhedron that has \leq\pi surface to each side at every point. In contrast, a geodesic has exactly \pi to each side. Because a vertex is a point with less than 2\pi surface, a geodesic can never pass through a vertex. Quasigeodesics can pass through vertices. Ever since since Poincaré’s investigations more than a century ago, closed geodesics have played an important role in the topology of Riemannian manifolds [ Ber03 , p. 433] . It is a famous 1929 theorem of Lyusternik-Schnirelmann that every smooth genus- 0 surface has at least three simple (non-self-intersecting) closed geodesics [ LS29 ] . Pogorelov proved in 1949 a natural analog: Every convex surface has at least three simple closed quasigeodesics [ Pog49 ] . Pogorelov’s existence proof does not suggest a way to identify the three quasigeodesics, and it is only recently that finite algorithms have been proposed [ DHK20 ] [ CdM24 ] . Aside from these algorithms, simple closed quasigeodesics have only been systematically studied on tetrahedra. Two results in [ OV22 ] are: (1) On any non-isosceles tetrahedron, there is at least one 1 -vertex, one 2 -vertex, and one 3 -vertex simple closed quasigeodesic. (2) There is an open set in the space of all tetrahedra, each element of which has at least 34 simple closed quasigeodesics. In contrast to (1), it is known from [ DDTY17 ] that the cube does not have a 1 -vertex simple closed quasigeodesic. Simple closed quasigeodesics play central roles in [ HLM+22 ] and [ OV24 ] , and are of interest in their own right. But beyond their existence, much remains unknown. There is no known upper bound on the number of simple closed quasigeodesics on a given polyhedron, and there is an n -vertex polyhedron with 2^{\Omega(n)} distinct simple closed quasigeodesics [ DO07 , Sec. 24.4] . In contrast, it is known that isosceles tetrahedra 1 1 1 Also called disphenoids, tetramonohedra, isotetrahedra, and several other names. All faces are congruent acute triangles. have arbitrarily long “spiraling” simple closed geodesics [ Pro07 ] [ AP18 ] . In this paper we make a complete inventory of simple closed quasigeodesics on a cube. It was known that there are precisely three simple closed geodesics on the cube. We identify a further 15 simple closed quasigeodesics (up to symmetries), and prove that this list is complete. We consider this proof to be our most significant contribution. 1.2 Three simple closed geodesics To describe geodesics and quasigeodesics explicitly, we adopt the notation for faces and vertices displayed in Fig. 1 . Note that we label vertices in figures by their index i , but refer to them in the text as v_{i} . Figure 1: F,R,T,K,L,B = Front, Right, Top, bacK, Left, Bottom. B vertices indexed 1,2,3,4 ; T vertices indexed 5,6,7,8 . v_{1} is marked white. It has long been known that there are precisely three simple closed geodesics on the cube [ FF07 ] , displayed in Fig. 2 . 2 2 2 Note these three are not the three from Pogorelov’s theorem. Figure 2: The three simple closed geodesics on a cube. The first is an equatorial band. The other two are as depicted. Note that each of the three geodesics can slide within a range, maintaining parallelism. This is because each geodesic lies on a cylinder, with 2\pi curvature (four vertices, each with \pi/2 curvature) to each side. 2 Outline of Argument We mentioned that simple closed geodesics can spiral around isosceles tetrahedra. A simple closed quasigeodesic also may spiral around other convex polyhedra, as shown in Fig. 3 below. A central aspect of our proof is to show that quasigeodesics cannot spiral on a cube. Define a geodesic segment as a non-self-intersecting vertex-to-vertex geodesic. 3 3 3 In some literature, a geodesic segment is a shortest path between its endpoints. In this paper, our geodesic segments may or may not be shortest. A simple closed quasigeodesic is composed of a sequence of geodesic segments, satisfying the \leq\pi condition to both sides at each vertex. An instructive example was identified in [ DHK20 ] : a long box with a spiraling simple closed quasigeodesic. See Fig. 3 . Each of the four marked vertices has \pi angle to one side and \pi/2 to the other side. Since there is freedom to partition the 3\pi/2 surface angle differently (while maintaining \leq\pi to each side), the number of spiraling simple closed quasigeodesics of a long box grows with the length of the long side of the box. A crucial property of spiraling is that some geodesic segment re-enters its initial face. For example, the blue geodesic segment from v_{1} to v_{2} in the figure starts on the long front-side face and later re-enters that face. We will prove that this cannot happen on a cube: a geodesic segment cannot return to its initial face, and in fact, cannot cross any face more than once. Figure 3: (v_{1},v_{2},v_{3},v_{3},v_{4}) is a simple closed quasigeodesic. Based on Fig. 2 in [ DHK20 ] . 3 Fifteen Simple Closed Quasigeodesics Here is our main result: Theorem 1 There are exactly 15 simple closed quasigeodesics on the cube (beyond the three simple closed geodesics noted above). These are displayed in Fig. 4 and described in Table 1. As our sole focus in the remainder is on “simple closed quasigeodesics,” we often simplify that term to quasigeos . Figure 4: The 15 simple closed quasigeodesics. A_{i} : Four quasigeos using only {0/1} segments (each of length 1 ). B_{i} : Five quasigeos using at least one {1/1} segment (length \sqrt{2} ), and none longer. C_{i} : Four quasigeos using at least one {1/2} segment (length \sqrt{5} ), and none longer. D_{1} : One quasigeo using a single {1/3} (length \sqrt{10} ), and none longer. E_{1} : One quasigeo using a single {2/3} (length \sqrt{13} ). Table 1: Description of the five categories of quasigeos. The quasigeos are listed in order of the length of the geodesic segments comprising them, as described in Table 3 . We identify a geodesic segment by

Read the complete text