Uncertainties in TCSP = Evidential reasoning approach

Abstract

Evidential reasoning is a collection of irgerence techniques for reasoning with uncertain information. These techniques are useful in developing improved frameworks of set based representation and reasoning. The current study considers interval algebra (IA), which is argubaly the most injuential language proposed to represent and reason about temporal relations. The IA representation is often referred to as Temporal Constraint Network or Temporal Constraint Satisfaction Problem (TCSP). In TCSE a constraint is a subset of the set of atomic relations between interval events, and the reasoning tasks are performed using a collection of set operations over the sets of atmomic relations. This motivated us to investigate the usefulness of evidnetial reasoning TCSPs. This paper presents a new TCSP framework by augmenting the standard IA framework with evidential reasoning techniques. It is demonstrated that the new framework is more expressive than IA; and heuristic algorithm can be employed to efficiently solve TCSPs in the new framework. © 2003 ISlF.