Mono-Monostatic Convex Bodies

• Mono-monostatic convex bodies are convex objects with exactly one stable and one unstable equilibrium point when placed on a flat surface.
• Research investigates combinatorial constraints using quadratic inequalities and semidefinite programming to determine minimal vertex and face bounds.
• Symmetry classification by dihedral and cyclic groups underpins explicit constructions and connects equilibrium theory to optimization methods.

A mono-monostatic convex body is a convex object possessing exactly one stable and one unstable static equilibrium point when placed on a flat surface, with no degenerate equilibria. This property may be defined for both smooth convex solids and convex polyhedra (potentially with vertex-only, or "0-skeleton," mass distributions). The classification, construction, and combinatorial minima of such bodies are the subject of deep investigation in three-dimensional geometry and computational mathematics, with connections to rigidity theory, convex optimization, and symmetry classification.

1. Definitions and Equilibrium Theory

Let $K \subset \mathbb{R}^3$ be a convex body with center of mass $p \in K$. Critical points of the distance-to-mass-center function

$$\delta_K: \partial K \rightarrow \mathbb{R}, \quad \delta_K(q) = \|q-p\|$$

correspond to equilibrium points—stable (local minima), unstable (local maxima), and saddle points. For a nondegenerate (Morse) function, their counts $S, U, H$ satisfy the Poincaré–Hopf relation

$S - H + U = 2.$

A mono-stable body has $S = 1$, a mono-unstable body has $U = 1$, and a mono-monostatic body has $S = U = 1$. For the 0-skeleton model, mass is concentrated at the vertices, and equilibrium notions are inherited from the geometry of the supporting planes at vertices or faces (Bozóki et al., 2021).

2. Combinatorial and Geometric Constraints

The search for mono-monostatic convex bodies is governed by strict combinatorial lower and upper bounds. In the 0-skeleton (unit vertex mass) case, every convex mono-monostatic polyhedron must have at least 8 vertices and, by Euler-type relations, at least 6 faces (Bozóki et al., 2021). These lower bounds were established using reductions of the shadowing criterion to systems of quadratic inequalities, where a vertex is an unstable equilibrium precisely when it is not shadowed by any other vertex; the existence problem then becomes a feasibility problem for such systems. Specifically, for a convex polyhedron with $V$ vertices, no mono-monostatic 0-skeleton exists for $V < 8$.

For homogeneous (volume mass) polyhedra, bounds are less sharp but relate closely: mono-stable or mono-unstable homogeneous polyhedra cannot be realized by tetrahedra, setting $V, F \geq 5$, and the best explicit constructions for the homogeneous case involve polyhedra with $V \leq 18$ vertices and $F \leq 14$ faces (Bozóki et al., 2021, Papp et al., 2023).

3. Explicit Constructions and Upper Bounds

The first explicit mono-monostatic 0-skeleton construction was achieved by Domokos and Kovács via a "Conway spiral" method, yielding a 21-vertex, 21-face convex polyhedron with each vertex carrying a unit mass and exactly one stable and one unstable equilibrium (Domokos et al., 2021). The construction proceeds by:

• Generating a planar "spiral" sequence, parameterized by turning angles $(\alpha_1, ..., \alpha_n)$.
• Rotating to produce $k$-fold symmetry, building $kn+1$ vertices.
• Filling in faces using polygons and quadrilaterals, ensuring combinatorial convexity.
• Rigorously verifying monotonicity and Morse conditions to ensure that only the topmost vertex and the lowest face host equilibria.

This establishes $V_0^* \leq 21$ for the 0-skeleton case. For mono-unstable 0-skeletons, the known minimal configuration features $V=11$ vertices and $F=8$ faces, established by exhaustive semidefinite feasibility certification over all possible shadowing patterns, exploiting SDP solvers and dimension-free certificates (Papp et al., 2023). The summary of current known bounds for 0-skeletons is:

Type	Minimal Vertices (V)	Minimal Faces (F)
Mono-unstable	11	8
Mono-monostatic	8 ≤ V ≤ 21	6 ≤ F ≤ 21

The exact minimal values for mono-monostatic 0-skeletons remain between these bounds.

4. Symmetry Classification

Every mono-monostatic convex body—whether smooth or polyhedral—admits as its symmetry group only a discrete two-dimensional point group: either a dihedral group $D_n$ or a cyclic group $C_n$, for $n \geq 1$ (Domokos et al., 2021). No higher-order polyhedral or continuous symmetries are possible, as these would necessarily preserve more than two equilibria. The main result asserts full classification:

• For any discrete two-dimensional point group $G \subset O(3)$ and $\varepsilon > 0$, there exists a mono-monostatic convex body with $\operatorname{Sym}(K) = G$ and Hausdorff distance $< \varepsilon$ to the unit ball.
• Examples can be constructed for each $C_n$ and $D_n$ via symmetric perturbations of nearly spherical shapes, ensuring precisely two equilibria.

The prototypical smooth body, the Gömböc ($D_2$ symmetry), is a limiting case of this construction.

5. Methodological Innovations

Key advances in the enumeration and certification of mono-monostatic or mono-unstable convex bodies rely on the translation of geometric equilibrium constraints into systems of quadratic inequalities. The main tools are:

• Shadowing matrices encoding which vertices shadow others (for 0-skeletons), with explicit formulae for equilibrium counting.
• Reduction to quadratic feasibility problems, certified infeasible via construction of positive-definite conic combinations of quadratic forms—operationalized through semidefinite programming (SDP) (Papp et al., 2023, Bozóki et al., 2021).
• Enumeration over all $O(V!)$ patterns, implementable for moderate $V$ using symmetry reduction, parallelism, and SDP solver technology.
• The extension of infeasibility certificates across all dimensions by virtue of the tensor structure of the relevant matrices (Papp et al., 2023).

These methodologies provided the first complete resolution of the minimal-vertex mono-unstable 0-skeleton problem and established new avenues for higher-dimensional and more complex mass distributions.

6. Open Problems and Limitations

Despite major progress, several open questions persist:

• The minimal vertex and face count for mono-monostatic 0-skeletons is known to satisfy $8 \leq V^*_0 \leq 21$, but no explicit 0-skeleton with $V = 8$ and $S = U = 1$ is known (Bozóki et al., 2021, Domokos et al., 2021).
• For homogeneous (volume-mass) polyhedra, the true minimal $V^*_3$ remains open; best explicit constructions use $V \leq 18$.
• Sum-of-squares and SDP-based methods become intractable for larger $V$ or continuous mass, and analogous combinatorial reduction for the mono-stable (as opposed to mono-unstable) case is considerably more difficult due to higher degree and complexity of the constraints.
• Homogeneous bodies analogous to the constructed mono-monostatic 0-skeletons are obstructed by integral balance constraints, e.g., the center-of-mass moment cannot be simultaneously at the required position while maintaining only two equilibria (Domokos et al., 2021).

7. Implications and Broader Context

Mono-monostatic convex bodies constitute extreme and rigid examples in static equilibrium theory, providing benchmarks for combinatorial and geometric rigidity, convex optimization, and applied mechanics. The constructive and symmetry-classification results connect the enumerative geometry of convex polytopes to the rich theory of group actions and Morse functions on manifolds. The methodology of translating geometric constraints into algebraic feasibility problems—reducible to high-dimensional SDP—offers a template for resolving other extremal combinatorial and equilibrium-count problems. The classification by symmetry groups establishes a tight link between equilibrium theory and classical crystallographic classifications.

Further developments may exploit symmetry, moment relaxations, and advances in computational optimization to resolve remaining minimal cases or to generalize to higher equilibrium-number or higher-dimensional analogues (Papp et al., 2023, Domokos et al., 2021, Bozóki et al., 2021, Domokos et al., 2021).