Discrete uniformization of polyhedral surfaces with non-positive curvature and branched covers over the sphere via hyper-ideal circle patterns (Q517458): Difference between revisions

This paper studies a discrete version of the classical uniformization theorem for Riemann surfaces. The key ingredient is to consider hyper-ideal circle patterns, which is generalization of Delaunay circle patterns, on closed orientable polyhedral surfaces with non-positive Gaussian curvature. The main result is to show that discrete uniformization via hyper-ideal circle patterns always exists and is unique. A similar result is also obtained for surfaces represented as finite branched covers over the Riemann sphere. Let \(S\) be a closed orientable surface which admits a metric \(d\) of constant non-positive Gaussian curvature with finitely many cone singularities. We take a finite set, say \(V\), of points on \(S\) which contains all the cone singularities, and then take a geodesic cell complex \(\mathcal{C}\) obtained from \(V\) which means that the interior of each edge is geodesic. A hyper-ideal circle pattern on \(S\) consists of (i) a circle for each vertex \(i\) with the center at \(i\) and radius of \(r_i\) such that any two circles bound mutually disjoint disks, and (ii) a circle for each face which is orthogonal to all the circles for vertices of the face and admits intersection angles less than \(\pi/2\) with nearby vertex circles if their interiors intersect. Given a hyper-ideal circle pattern, we can extract its combinatorial angle data \((\mathcal{C},\theta,\Theta)\), where \(\theta\) is the data of intersection angles for adjacent face circles and \(\Theta\) is the data of cone angles for each point of \(V\). The data \((\mathcal{C},\theta,2\pi)\) means that there are no cone singularities. We note that if \(r_i\) is equal to zero for each vertex \(i\), then it is the classical Delaunay circle pattern. For a given \((S,d,V)\) as above, consider the unique Delaunay circle pattern and let \((\mathcal{C},\theta,\Theta)\) be its combinatorial angle data. Then the authors prove that there is a unique (up to isometry isotopic to identity) hyper-ideal circle pattern on \(S\) with underlying hyperbolic metric \(h\) whose combinatorial angle data is \((\mathcal{C},\theta,2\pi)\). They also obtain a similar result for surfaces represented as finite branched covers over the Riemann sphere by well considering the \textit{covering data}.