Combining symbolic and numerical solvers to simplify indecomposable systems solving

  • Authors:
  • Arnaud Fabre;Pascal Schreck

  • Affiliations:
  • University Louis Pasteur, Strasbourg, France;University Louis Pasteur, Strasbourg, France

  • Venue:
  • Proceedings of the 2008 ACM symposium on Applied computing
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

In Computer-Aided Design, solvers always attempt to decompose geometric constraint systems into smaller ones in order to make faster the resolution. However, this scheme often fails in the case of 3D geometric constraint systems since they are hardly decomposable. We have studied a new method which uses jointly two solvers, a symbolic one and a numerical one, in order to solve a system S: system S is transformed into a parametric system S' "almost" equivalent to S and such that system S" is symbolically solvable and the numerical solver computes solutions of S from solutions of S".