Nogoods in Qualitative Constraint-Based Reasoning

Abstract

Abstract. The prevalent method of increasing reasoning efficiency in the domain of qualitative constraint-based spatial and temporal reasoning is to use domain splitting based on so-called tractable subclasses. In this paper we analyze the application of nogood learning with restarts in combination with domain splitting. Previous results on nogood recording in the constraint satisfaction field feature learnt nogoods as a global constraint that allows for enforcing generalized arc consistency. We present an extension of such a technique capable of handling domain splitting, evaluate its benefits for qualitative constraint-based reasoning, and compare it with alternative approaches.

Qualitative Spatial and Temporal Reasoning with Answer Set Programming

“…At each step of the search, a-closure is enforced to prune the labels of the edges, and powerful heuristics using tractable subsets reduces both the depth and branching factor of the search. The approach has been compared favorably against other existing solvers [26], and delivered the fastest average solving time for most instances. However, it has been observed that for certain instances in the most difficult phase transition region the algorithm may get stuck for hours without finding a solution [14].…”

Incomplete information in RDF databases with an emphasis on the geospatial domain

“…Due to this the implementation of the path-consistency algorithm is not as efficient as that of Renz and Nebel presented above. However, when it comes to checking of consistency using backtracking search, the modern techniques of GQR based on nogood recording [39] and restarting of the backtracking search process make it very powerful [175].…”

“…3. We show that our partitioning-based techniques can achieve scalability for very large real-world networks and in general for networks of low average degree, but for networks of high average degree or complex spatial relations, such as the ones from the so-called "hard" set of relations N P 8 , the state of the art reasoners, such as GQR [53,175] or the more recent reasoners by Sioutis et al [158,156] should be preferred. We also juxtapose the evaluation of this chapter with the evaluation carried out in our paper [131] showing that in all cases the performance can be even better, except for two cases in which performance is worse but still better than the reasoners considered in [131].…”