A method for solving the problem of isometric realization of developments (Q950816)

The author describes an algorithm for finding all isometric realizations of a triangulated polyhedral sphere. Let \(K\) be a simplicial complex homeomorphic to a closed 2--surface, and let \(l\) be a positive real-valued function on the edge set of \(K\). If the values of \(l\) on the edges of every triangle of \(K\) satisfy the triangle inequality, then the pair \((K,l)\) is called an (abstract) triangulated polyhedral surface. An isometric realization of \((K,l)\) in \({\mathbb R}^3\) is a map \(p\) of the vertex set of \(K\) to \({\mathbb R}^3\) such that \[ \|p_i - p_j\| = l_{ij} \] for every edge \(ij\) of \(K\). Since finding all realizations is equivalent to solving a system of polynomial equations, methods of computational algebraic geometry can be applied. However, the Gröbner basis method is too expensive, and it is not clear how to use the special structure of the system to save time. There is an algorithm suggested by \textit{I. Kh. Sabitov} [Izv. Math. 66, No. 2, 377--391 (2002; Zbl 1076.51513)] which is more efficient than the Gröbner basis method, but it works only for surfaces in general position. For example, it does not allow to find flexible realizations. The method suggested in the present paper is based on the following idea. Assume that the 1--skeleton of \(K\) contains a Hamiltonian cycle that bounds a subcomplex \(L\) homeomorphic to a disk. Then the isometric realizations of \((L,l)\) are parametrized by the dihedral angles at the diagonals of \(L\). Each edge in \(K \setminus L\) yields an equation in terms of these dihedral angles. The equations can be reduced to a system of polynomial equations. The assumption on the existence of a Hamiltonian cycle that bounds a disk is fulfilled, by a result of \textit{H. Whitney} [Ann. Math. (2) 32, 378--390 (1931; Zbl 0002.16101)], for all triangulations of the sphere such that every 3--edge cycle is spanned by a triangle. If a triangulation of the sphere contains empty triangles, then it can be decomposed into connected sum along these triangles. The problem thus reduces to isometric realizations of the summands. The algorithm is illustrated on a number of examples.