Effective reasoning systems for ASP related paradigms

Date

2019-08

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

This dissertation focuses on improving the effectiveness of two Answer Set Programming (ASP) related paradigms. The first paradigm is ALM, a recent action language that allows the modular specification of ontologies and state transition diagrams to model actions and their effects in finite dynamic domains. Currently there is no publicly available compiler or reasoning system which accepts ALM System Descriptions as input. In this dissertation we describe and implement CALM, a compiler and reasoning system that translates ALM System Descriptions to SPARC, a variant of ASP. CALM enables the development of knowledge libraries and the investigation of best practices in modeling with ALM.

One future extension of ALM is to incorporate elements of action language H which enable reasoning about continuous change over time. In order to reason in continuous domains, programs in H are translated to AC(C), an extension of ASP to include constraint logic programming (CLP) reasoning and query answering in continuous domains. Our second area of focus is improving the effectiveness of the CLP algorithm used in AC(C) solvers. AC(C) solvers extend ASP solvers with an incrementally changing query to the CLP program derived from the ACC program. Current solvers restart CLP reasoning from scratch when the query changes, leading to redundant computation in the search for answers to the portions of the query that did not change. In this dissertation we formalize the incremental query problem and provide an incremental algorithm that reuses the solutions of previous queries in the search for answers to the modified query.

Description

Keywords

ASP, ALM, CALM, ICLP, IQTD

Citation