On the representation of temporal knowledge

Abstract

The growing interest in an adequate modelling of time in Artificial Intelligence has given rise to the research discipline of Temporal Reasoning (TR). Due to different views, different approaches towards TR such as PL1, modal logics or Allen's intervall logic have been investigated. It was realized at an early stage that each of this approaches has some strong points whereas it suffers from certain drawbacks. Thus recently, a number of research activities have emerged aiming at a combination of the classical paradigms for representing time. In the first part of this paper, we present an overview of the most important approaches to the integration of temporal knowledge into logic programming. In the second part, we present the CRONOLOG temporal logic programming language which has been developed to cover the quintessence of the approaches presented before. The third part of the paper describes TRAM, which it is an extension of CRONOLOG to a temporal knowledge representation system. Using TRAM it is possible to represent knowledge depending on time and to reason about this knowledge. TRAM has been conceptually based on a combination of modal logics with Allen's interval logic. We present the Extended Modal Logics (EML) which establishes the theoretical framework for TRAM. We define an operational semantics and a horizontal compilation scheme for TRAM.