Dangling points in area-minimizing Meissner polyhedra

Abstract: Meissner polyhedra are constant-width bodies obtained from extremal finite sets of unit diameter. Such a generating set may contain dangling points, namely points having exactly two diametric neighbors. This article studies whether these points can play an essential role in surface-area minimization. Given an extremal set we show that the smallest surface area among the Meissner polyhedra based on it cannot increase by deleting a dangling point. Adding a dangling point cannot decrease the smallest achievable surface area. This reduces the search for area-minimizing Meissner polyhedra to generating sets without dangling points.

• The paper proves that deleting a dangling point from an extremal generating set never increases the optimal Meissner surface area, with a strict decrease when the subdivided edge is longer than its dual edge.
• The analysis combines an exact area-additivity identity for edge subdivision with longest-edge switching and a new convexity result for the smoothing functional to compare all geometric cases.
• The result reduces area and volume minimization to dangling-free extremal sets, while leaving broader inclusion monotonicity and the Meissner tetrahedron conjectures open.

Overview

The paper studies the three-dimensional Blaschke–Lebesgue problem through the lens of Meissner polyhedra, the family of constant-width bodies obtained by smoothing extremal finite point sets. Its central object is a dangling point: a vertex of an extremal generating set having exactly two diametric neighbors (diametric valence two). The author proves that dangling points cannot help reduce surface area: for an extremal set $X$ and a dangling point $x \in X$, the optimal Meissner area satisfies

$$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$

with a matching statement for insertion. This reduces the search for area-minimizing Meissner polyhedra to generating sets without dangling points. By Blaschke's relation $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$, the same reduction transfers verbatim to volume minimization.

Background and setup

For a finite set $X \subset \mathbb{R}^3$ of unit diameter with $m \geq 4$ points, the Vázsonyi problem gives $e(X) \leq 2m-2$ diametric pairs; an extremal set attains equality. The associated ball polyhedron $\bigcap_{p \in X} \overline{B}(p,1)$ has $m-1$ dual edge pairs, and choosing one edge of each pair as smoothing centers yields a Meissner polyhedron of constant width one. The surface area admits an explicit formula in terms of the angular lengths $\theta(uv) = 2\arcsin(|u-v|/2)$ of the dual pair:

$x \in X$0

The key structural fact is that deleting a dangling point removes exactly two diametric pairs, preserving extremality: $x \in X$1. Geometrically, the dangling point subdivides a circular edge $x \in X$2 of the smaller ball polyhedron, splitting one dual pair $x \in X$3 into two pairs $x \in X$4 and $x \in X$5.

The area comparison

Writing $x \in X$6 and $x \in X$7, the paper exploits two facts: the longest-edge rule $x \in X$8 for $x \in X$9 (so centers go on the shorter edge, the longer edge is smoothed), and an exact additivity identity $$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$0 for the two subedges, which follows from the linear branch $$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$1 along the circle of radius $$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$2 containing the subdivided edge.

Two cases arise, and their contrast is the analytical core of the paper:

• If $$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$3, deletion gives exact equality: $$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$4, and compatible minimizing choices produce literally the same Meissner body, since the center sets $$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$5 and $$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$6 coincide.
• If $$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$7, deletion strictly decreases the optimal area. The proof splits into three exhaustive subcases depending on whether zero, one, or both subedges $$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$8 exceed $$\min_{M \in \mathcal{M}(X)} \mathcal{H}^2(\partial M) \;\geq\; \min_{M \in \mathcal{M}(X \setminus \{x\})} \mathcal{H}^2(\partial M),$$9. Cases (i) and (ii) use a switching gain $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$0, shown strictly increasing via a mixed-derivative comparison. Case (iii), where both subedges are longer than $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$1, is the technical heart: the author proves that $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$2 is convex on the active range $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$3. This convexity is established through a chain of elementary but delicate estimates involving the auxiliary function $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$4, with $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$5, and lower bounds $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$6 and $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$7 for $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$8. Convexity then forces the maximum of $$|K| = \tfrac{1}{2}\mathcal{H}^2(\partial K) - \pi/3$$9 to occur at an endpoint, where strict monotonicity of $X \subset \mathbb{R}^3$0 yields the strict inequality.

The conclusion is the boxed comparison $X \subset \mathbb{R}^3$1, with equality exactly when $X \subset \mathbb{R}^3$2.

The dangling-free reduction

Iterating the deletion yields the paper's structural corollary: successively deleting dangling points terminates at a dangling-free extremal set (a four-point extremal set has all $X \subset \mathbb{R}^3$3 diametric pairs and hence no dangling points), with no larger optimal area. Consequently, if the global minimum over Meissner polyhedra is attained, it is attained by one whose generating set has no dangling points. This is a genuine reduction of the search space, though the paper is careful to note it is not a strict exclusion: equality can occur, e.g., when insertion merely subdivides an edge already smoothed the same way.

Limitations and open questions

The result covers only the addition or removal of a single dangling point to an existing extremal set. The paper explicitly concedes that adding multiple points simultaneously may produce more complex diameter-graph combinatorics requiring different techniques. Three conjectures are stated: (1) monotonicity of optimal Meissner area under arbitrary inclusion of extremal sets, $X \subset \mathbb{R}^3$4 for $X \subset \mathbb{R}^3$5; (2) that for any extremal set $X \subset \mathbb{R}^3$6, the minimum surface area over constant-width bodies containing $X \subset \mathbb{R}^3$7 is attained by a Meissner polyhedron based on $X \subset \mathbb{R}^3$8; and (3) the special case where $X \subset \mathbb{R}^3$9 is the vertex set of a regular tetrahedron, i.e., that the Meissner tetrahedron minimizes area among constant-width bodies containing the tetrahedron's vertices. All three remain open.

Conclusion

The paper establishes that dangling points are inessential for area minimization among Meissner polyhedra: deleting them never increases the optimal area, with strict decrease precisely when the subdivided edge exceeds its dual in angular length. The proof combines the exact additivity of the local area functional under edge subdivision with a new convexity result for the smoothing functional. The reduction to dangling-free generating sets narrows the finite-dimensional search space relevant to the Bonnesen–Fenchel conjecture, while leaving the monotonicity under general set inclusion and the constrained minimization over bodies containing a fixed extremal set as open problems.