Interval Constraints: Results and Perspectives

  • Authors:
  • Frédéric Benhamou;Laurent Granvilliers;Frédéric Goualard

  • Affiliations:
  • -;-;-

  • Venue:
  • Selected papers from the Joint ERCIM/Compulog Net Workshop on New Trends in Contraints
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

Reliably solving non-linear real constraints on computer is a challenging task due to the approximation induced by the resort to floating-point numbers. Interval constraints have consequently gained some interest from the scientific community since they are at the heart of complete algorithms that permit enclosing all solutions with an arbitrary accuracy. Yet, soundness is beyond reach of present-day interval constraint-based solvers, while it is sometimes a strong requirement. What is more, many applications involve constraint systems with some quantified variables these solvers are unable to handle. Basic facts on interval constraints and local consistency algorithms are first surveyed in this paper; then, symbolic and numerical methods used to compute inner approximations of real relations and to solve constraints with quantified variables are briefly presented, and directions for extending interval constraint techniques to solve these problems are pointed out.