Project-Team Coprin

Bio: Project-Team Coprin is an academic researcher. The author has contributed to research in topics: Interval arithmetic. The author has an hindex of 1, co-authored 1 publications receiving 2 citations.

A polynomial time local propagation algorithm for general dataflow constraint problems

Abstract: The multi-way dataflow constraint model allows a user to describe interactive applications whose consistency is maintained by a local propagation algorithm. Local propagation applies a sequence of methods that solve the constraints individually. The local aspect of this solving process makes this model sensitive to cycles in the constraint graph. We use a formalism which overcomes this major limitation by allowing the definition of general methods that can solve several constraints simultaneously. This paper presents an algorithm called General-PDOF to deal with these methods which has a polynomial worst case time complexity. This algorithm therefore has the potential to tackle numerous real-life applications where cycles make local propagation unfeasible. Especially, general methods can implement ruler and compass rules to solve geometric constraints.

A constraint programming approach for Solving Rigid Geometric Systems

Abstract: This paper introduces a new rigidification method -using interval constraint programming techniques- to solve geometric constraint systems. Standard rigidification techniques are graph-constructive methods exploiting the degrees of freedom of geometric objects. They work in two steps: a planning phase which identifies rigid clusters, and a solving phase which computes the coordinates of the geometric objects in every cluster. We propose here a new heuristic for the planning algorithm that yields in general small systems of equations. We also show that interval constraint techniques can be used not only to efficiently implement the solving phase, but also generalize former ad-hoc solving techniques. First experimental results show that this approach is more efficient than systems based on equational decomposition techniques.