Weiming Liu

On redundant topological constraints

The Region Connection Calculus (RCC) is a well-known calculus for representing part-whole and topological relations. It plays an important role in qualitative spatial reasoning, geographical information science, and ontology. The computational complexity of reasoning with RCC has been investigated in depth in the literature. Most of these works focus on the consistency of RCC constraint networks. In this paper, we consider the important problem of redundant RCC constraints. For a set Γ of RCC constraints, we say a constraint (x R y) in Γ is redundant if it can be entailed by the rest of Γ. A prime network of Γ is a subset of Γ which contains no redundant constraints but has the same solution set as Γ. It is natural to ask how to compute a prime network, and when it is unique. In this paper, we show that this problem is in general co-NP hard, but becomes tractable if Γ is over a tractable subclass of RCC. If S is a tractable subclass in which weak composition distributes over non-empty intersections, then we can show that Γ has a unique prime network, which is obtained by removing all redundant constraints from Γ. As a byproduct, we identify a sufficient condition for a path-consistent network being minimal.

Solving qualitative constraints involving landmarks

Consistency checking plays a central role in qualitative spatial and temporal reasoning. Given a set of variables V, and a set of constraints Γ taken from a qualitative calculus (e.g. the Interval Algebra (IA) or RCC-8), the aim is to decide if Γ is consistent. The consistency problem has been investigated extensively in the literature. Practical applications e.g. urban planning often impose, in addition to those between undetermined entities (variables), constraints between determined entities (constants or landmarks) and variables. This paper introduces this as a new class of qualitative constraint satisfaction problems, and investigates the new consistency problem in several well-known qualitative calculi, e.g. IA, RCC-5, and RCC-8. We show that the usual local consistency checking algorithm works for IA but fails in RCC-5 and RCC-8. We further show that, if the landmarks are represented by polygons, then the new consistency problem of RCC-5 is tractable but that of RCC-8 is NP-complete.

Solving minimal constraint networks in qualitative spatial and temporal reasoning

The minimal label problem (MLP) (also known as the deductive closure problem) is a fundamental problem in qualitative spatial and temporal reasoning (QSTR). Given a qualitative constraint network Γ, the minimal network of Γ relates each pair of variables (x,y) by the minimal label of (x,y), which is the minimal relation between x,y that is entailed by network Γ. It is well-known that MLP is equivalent to the corresponding consistency problem with respect to polynomial Turing-reductions. This paper further shows, for several qualitative calculi including Interval Algebra and RCC-8 algebra, that deciding the minimality of qualitative constraint networks and computing a solution of a minimal constraint network are both NP-hard problems.

Cardinal directions: a comparison of direction relation matrix and objects interaction matrix

How to express and reason with cardinal directions between extended objects such as lines and regions is an important problem in qualitative spatial reasoning (QSR), a common subfield of geographical information science and Artificial Intelligence (AI). The direction relation matrix (DRM) model, proposed by Goyal and Egenhofer in 1997, is one very expressive relation model for this purpose. Unlike many other relation models in QSR, the set-theoretic converse of a DRM relation is not necessarily representable in DRM. Schneider et al. regard this as a serious shortcoming and propose, in their work published in ACM TODS (2012), the objects interaction matrix (OIM) model for modelling cardinal directions between complex regions. OIM is also a tiling-based model that consists of two phases: the tiling phase and the interpretation phase. Although it was claimed that OIM is a novel concept, we show that it is not so different from DRM if we represent the cardinal direction of two regions a and b by both the DRM of a to b and that of b to a. Under this natural assumption, we give methods for computing DRMs from OIMs and vice versa, and show that OIM is almost the same as DRM in the tiling phase, and becomes less precise after interpretation. Furthermore, exploiting the similarity between the two models, we prove that the consistency of a complete basic OIM network can be decided in cubic time. This answers an open problem raised by Schneider et al. regarding efficient algorithms for reasoning with OIM.

On Redundant Topological Constraints (Extended Abstract).

The Region Connection Calculus (RCC) is a well-known calculus for representing part-whole and topological relations. It plays an important role in qualitative spatial reasoning, geographical information science, and ontology. The computational complexity of reasoning with RCC has been investigated in depth in the literature. Most of these works focus on the consistency of RCC constraint networks. In this paper, we consider the important problem of redundant RCC constraints. For a set Γ of RCC constraints, we say a constraint (xRy) in Γ is redundant if it can be entailed by the rest of Γ. A prime subnetwork of Γ is a subset of Γ which contains no redundant constraints but has the same solution set as Γ. It is natural to ask how to compute a prime subnetwork, and when it is unique. In this paper, we show that this problem is in general intractable, but becomes tractable if Γ is over a tractable subclass of RCC. If S is a tractable subclass in which weak composition distributes over non-empty intersections, then we can show that r has a unique prime network, which is obtained by removing all redundant constraints from Γ. As a byproduct, we identify a sufficient condition for a path-consistent network being minimal.