Degenerating orbits of the Longest Edge Bisection process
Abstract
We study the Longest Edge Bisection (LEB) process as a dynamical system on the projective shape space of simplices. A long-standing conjecture going back to Adler and Rivara-Levin and motivated by finite-element mesh refinement, often taken as a standing assumption, is that this procedure is non-degenerate and, in fact, in a certain way periodic. We prove: • There are 3-dimensional simplices such that the longest edge-bisection algorithm degenerates. • There is an open set of 4-dimensional simplices on which the longest edge-bisection algorithm degenerates. • If parametrizing the space of d-dimensional simplices by independent standard Gaussian vectors, then as d increases, a random simplex degenerates asymptotically almost surely. This is realized through exhibiting hyperbolic behaviour of the LEB process. We also exhibit elliptic behaviour that is nonperiodic.
Related
- Closure complexity of longest-edge bisection for triangular meshes
- Multiform Longest Edge Bisection of Tetrahedra via Sextuple Permutations
- Longest increasing subsequences of dyadic-type chaotic orbits
- On degeneration of tetrahedra under longest-edge n-section refinement
- Projective embedding of degenerating family of Kähler-Einstein manifolds with normal crossing limit
BibTeX
@misc{adiprasito2026degenerating,
title = {{Degenerating orbits of the Longest Edge Bisection process}},
author = {Karim Adiprasito and Daniel Kalmanovich and Yaar Solomon},
howpublished = {HAL},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.08846},
eprint = {2609.08846},
archivePrefix = {arXiv},
url = {https://hal.science/hal-05742991v1},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - Degenerating orbits of the Longest Edge Bisection process
AU - Karim Adiprasito
AU - Daniel Kalmanovich
AU - Yaar Solomon
PY - 2026
DA - 2026/09/08
PB - HAL
DO - 10.48550/arxiv.2609.08846
UR - https://hal.science/hal-05742991v1
AB - We study the Longest Edge Bisection (LEB) process as a dynamical system on the projective shape space of simplices. A long-standing conjecture going back to Adler and Rivara-Levin and motivated by finite-element mesh refinement, often taken as a standing assumption, is that this procedure is non-degenerate and, in fact, in a certain way periodic. We prove: • There are 3-dimensional simplices such that the longest edge-bisection algorithm degenerates. • There is an open set of 4-dimensional simplices on which the longest edge-bisection algorithm degenerates. • If parametrizing the space of d-dimensional simplices by independent standard Gaussian vectors, then as d increases, a random simplex degenerates asymptotically almost surely. This is realized through exhibiting hyperbolic behaviour of the LEB process. We also exhibit elliptic behaviour that is nonperiodic.
ER -