Le Thi Hoai An, Pham Dinh Tao: "Large Scale Molecular Conformation via the exact distance geometry problem".

Abstract: We develop in this paper a method based on a d.c. (difference of convex functions) optimization approach called DCA for solving large-scale exact distance geometry problem. Requiring only matrix-vector products and one Cholesky factorization, the DCA seems to be robust and efficient in large scale problems. Moreover it allows exploiting sparsity of the given distance matrix. A technique using the triangle inequality to generate a complete approximate distance matrix was investigated in order to compute a good starting point for the DCA. Numerical simulations of the molecular conformation problems with up to 12288 variables are reported which prove the robustness, the efficiency, and the globality of our algorithms.


Keywords: d.c. programming, d.c. algorithm (DCA), exact distance geometry problem, molecular conformation.

Citation: Le Thi Hoai An, Pham Dinh Tao, Large Scale Molecular Conformation via the exact distance geometry problem, in ”Optimization”, Lecture Notes in Economics and Mathematical Systems, volume 481, pp. 260-277, Springer-Verlag, Heidelberg 2000.