Spatial Reasoning in Constraint Logic Programming based on Numerical Optimisatio

May 23, 2016 at 9:48:09 AM
New CS Department,
Engineering Dr,
11794 Stony Brook, NY
From Jul 8, 2016, 11:30 AM
To Jul 8, 2016, 12:00 PM
A paper by Carl Schultz (The DesignSpace Group, University of Münster) and Mehul Bhatt (University of Bremen)
Abstract. In this paper we formulate (qualitative) spatial reasoning within the framework of numerical optimisation, and develop an approach that extends Constraint Logic Programming (CLP) to support variables that range over spatial domains. The crux of our approach is extending standard unification by the mechanism of attributed variables, and using specialised numerical optimisation algorithms in a stateless manner. We prove that our approach is fully consistent with the semantics of CLP, and demonstrate a range of powerful spatial reasoning features.