Possible Models Computation and Revision – A Practical Approach

This paper describes a method of computing plausible states of a system as a logical model. The problem of analyzing state-based systems as they evolve over time has been studied widely in the automated reasoning community (and others). This paper proposes a specific approach, one that is tailored to situational awareness applications. The main contribution is a calculus for a novel specification language that is built around disjunctive logic programming under a possible models semantics, stratification in terms of event times, default negation, and a model revision operator for dealing with incomplete or erroneous events – a typical problem in realistic applications. The paper proves the calculus correct wrt. a formal semantics of the specification language and it describes the calculus’ implementation via embedding in Scala. This enables immediate access to rich data structures and external systems, which is important in practice.