The geometry of Euclidean surfaces with conical singularities (Q343641)

A closed surface \(S\) equipped with a metric \(d\) is said to be a Euclidean surface with conical singularities \(s_1,s_2,\dots,s_n\) if every point \(p\in S\setminus\{s_1,s_2,\dots,s_n\}\) has a neighborhood isometric to a disk in the Euclidean plane, and each \(s_i\) has a neighborhood isometric to a neighborhood of vertex \(v\) of the standard cone \(C(v,\theta(s_i))\), where \(\theta(s_i)\) is the angle of the conical point \(s_i\). An isometric map \(h:[a,b]\to S\) is called a ``geodesic segment'', and if \(h\) is a periodic local isometric map then it is called a ``closed geodesic''. If \(GS\) is the space of all local isometric maps \(\gamma:\mathbb R\to S\) equipped with the compact-open topology, then it is sequentially compact. In this paper, the authors study the geometry of Euclidean surfaces with conical singularities. They prove that if \(\gamma\) is a non-closed geodesic in \(S\), then its image has \(0\) distance from the set \(\{s_1,s_2,\dots,s_n\}\) of conical points, that is, \(d(\text{Im}\gamma,\{s_1,s_2,\dots,s_n\})=0\). In case when \(\gamma\) is a closed geodesic, then the authors show that it is dense in \(GS\). Next, they show that if \(\{s_1,s_2,\dots,s_k,s_{k+1},\dots,s_n\}\) are conical singularities with \(\theta(s_i)\in(0,2\pi)\) if \(i=1,\dots,k\) and \(\theta(s_i)\in(2\pi,\infty)\) if \(i=k+1,\dots,n\), then for any non-closed geodesic \(\gamma\) in \(S\), \(d(\text{Im}\gamma,\{s_1,s_2,\dots,s_n\})=0\). Also, it is shown that every element in \(GS\) can be approximated, in the compact-open topology, either by a sequence of closed geodesics or, by a sequence of closed piece-wise geodesics.