Circle patterns with hyperbolic background via topological degree method (Q6974773)

Circle patterns are a natural approximation of conformal mappings and an important tool in three-dimensional geometric topology. They also play an essential role in studying the existence of hyperbolic polyhedra and the discretization of conformal geometry.\N\NIn spherical geometry circle pattern are connected to triangular polyhedra, therefore the authors study an oriented closed surface \(S\) with genus \(g \geq 2\), and a triangulation \(T=T(V,E,F)\) on \(S\). The principle idea is to connect the circle pattern \(P\) with a contact graph \(G(P)\) such that\N\begin{itemize}\N\item [(1)] Each disk \(D_i \in P\) is associated with a vertex \(v_i \in G(P)\);\N\item [(2)] Disks \(D_i\) and \(D_j\) intersect if and only if there is an edge \(e \in G(P)\) that connects vertices \(v_i\) and \(v_j\).\N\end{itemize}\N\NFurthermore the connection of \(P\) in hyperbolic geometry to the metric \(\mu\) is given by the angles of intersection, resp. their evaluation \(\Theta\). E.g. this connects the angles of \(P\) with the angles of the triangulation \(T\).\N\NThe main theorem studies the following conditions:\N\begin{itemize}\N\item [(C1)] If \(e_1 + e_2 + e_3\) forms the boundary of a face \(f\) , and \(\sum_{i=1}^3 \Theta(e_i) > \pi\), then \(\Theta(e_1) + \Theta(e_2) < \pi + \Theta(e_3)\), \(\Theta(e_2) + \Theta(e_3) < \pi + \Theta(e_1)\), \(\Theta(e_3) + \Theta(e_1) < \pi + \Theta(e_2)\).\N\item [(C2)] If \(e_1 + e_2 + \dots + e_s\) is a non-vacant pseudo-Jordan \(\gamma\) curve on \(S\), i.e. it is a closed curve where \(S \setminus \gamma\) contains a simply-connected component \(K\) with boundary \(\partial K = \gamma\) and where one of its enclosing vertex sets is non-empty, then \( \sum_{i=1}{3} \Theta(e_i) < (s-2) \cdot \pi. \)\N\item [(C3)] If \(e_1 + e_2\) are homologically non-adjacent, then \(\Theta(e_1) + \Theta(e_2) \leq \pi\).\N\item [(C4)] If \(e_1 + e_2 + e_3\) forms a closed loop not bounding a face, then \(\sum_{i=1}^3 \Theta(e_i) < \pi\).\N\end{itemize}\NThen there exists a constant curvature metric \(\mu\) on \(S\) with curvature \(-1\), such that \((S,\mu)\) supports a circle pattern \(P\) such that the disks in \(P\) cover \(S\), each disk plays an irreplaceable role in this covering, and the intersection angle functions are \(\Theta\).\N\NThis establishes the existence of at least one 'good' circle pattern. \N\NThe next theorem shows that the numbers of such patterns is finite, and that the pattern is unique if condition (C1) is replaced by the stronger:\N\begin{itemize}\N\item [(S1)] If \(e_1 + e_2 + e_3\) form the boundary of a face \(f\), then \(\cos \Theta_i + \cos \Theta_j \cdot \cos \Theta_k \geq 0\), \(\cos \Theta_j + \cos \Theta_k \cdot \cos \Theta_i \geq 0\), \(\cos \Theta_k + \cos \Theta_i \cdot \cos \Theta_j \geq 0\).\N\end{itemize}\NFinally the authors, building upon Thurston's observation of the duality between polyhedra and circle patterns, propose a conjecture concerning three-dimensional Fuchsian convex polyhedra:\N\N\smallskip Let \(P\) be an abstract trivalent Fuchsian polyhedron. Then a simple closed curve \(\Gamma\) is a \(k\)-circuit if it consists of \(k\) edges from the dual complex. It is a prismatic \(k\)-circuit, if all endpoints of the edges of \(P\) intersected by \(\Gamma\) are distinct. Furthermore suppose \(\Theta: E \rightarrow (0, \pi)\), i.e. on the edges, is a function satisfying the following conditions, where \(\Gamma\) is always prismatic circuit:\N\begin{itemize}\N\item[(F1)] If \(e_1\), \(e_2\), \(e_3\) are three distinct edges meeting at the same vertex, then \(\sum_{i=1}^n \Theta(e_i) > \pi\), and \(\Theta(e_1) + \Theta(e_2) < \pi + \Theta(e_3)\), \(\Theta(e_2) + \Theta(e_3) < \pi + \Theta(e_1)\), \(\Theta(e_3) + \Theta(e_1) < \pi + \Theta(e_2)\).\N\N\item[(F2)] If \(\Gamma\) intersects the edges \(e_1\), \(e_2\), \(\dots\) , \(e_s\), then \(\sum_{i=1}^n \Theta(e_i) > (s-2) \cdot \pi\).\N\N\item[(F3)] If \(\Gamma\) intersects edges \(e_1\), \(e_2\), where \(e_1\), \(e_2\) are homologically nonadjacent, then \(\Theta(e_1) +\Theta(e_2) \leq \pi\).\N\N\item[(F4)] If \(\Gamma\) intersects edges \(e_1\), \(e_2\), \(e_3\), then \(\sum_{i=1}^3 \Theta(e_i) < \pi\).\N\end{itemize}\N\NThen for any oriented closed surface \(S\) of genus \(g \geq 2\), there should exist a Fuchsian manifold \(F\) and a compact Fuchsian convex polyhedron \(Q \in F\) such that \(Q\) is isomorphic to \(P\), and the dihedral angle function of the compact Fuchsian convex polyhedron \(Q\) equals \(\Theta\). Moreover, \((F, P)\) is unique up to isometry.