A SYSTEM TO MAINTAIN KNOWLEDGE BASES OF INTELLIGENT AGENTS IN DLV'S WEAK CONSTRAINTS

摘要

Updating agents' knowledge bases is an important topic in Artificial Intelligence and a key problem of knowledge representation and reasoning. One of the latest ideas to update logic programs is to use abduction by means of Minimal Generalised Answer Sets (MGAS). MGAS provide a more general and flexible semantics than standard ASP, and we propose it to overcome disadvantages of previous approaches as well as to simplify both the semantics definition and its implementation. Moreover, the author provides a link to a modest online prototype of the semantics here introduced. This prototype models updates in a well-known Disjunctive Datalog, DLV which is an implementation (solver) for Answer Set Programming. Accordingly, the two-fold contribution of this paper is showing a simpler theoretical framework characterised in DLVs weak constraints, as well as an outline of the basic structure of its implementation, description of the employed technology.