Articles | Volume 5
30 May 2024
 | 30 May 2024

Land Evaluation Configuration using Answer Set Programming

Mina Karamesouti and Etienne Tignon

Keywords: declarative knowledge representation, logic programming, transparency, model semantics, problem instance

Abstract. In the realm of Land Evaluation (LE) interdisciplinary and transdisciplinary knowledge exchange is critical for land use preservation. Geographic Information Systems (GIS) are powerful tools for real-world Knowledge Representation (KR), facilitating inter- and transdisciplinary communication. In such knowledge exchange contexts, heterogeneity, ambiguity, abstraction are only indicative issues, underscoring the necessity for a rigorous commitment to broader transparency in KR. Answer Set Programming (ASP), a declarative, human-readable, logic-based formalism, could serve this objective and facilitate productive, case-relevant dialogues. Similarly to the fundamental GIS knowledge organization structures, ASP formalizes knowledge as entities and relations between them. In current work, leveraging Rossiter’s theoretical framework for LE, and employing ASP, we aim for greater transparency in the epistemological and ontological assumptions underpinning the complex LE problem. ASP-based system configuration is used to formalize the LE Problem Instance as Components (C) with Properties (P) and Values (V ). Fact-type specifications in predicate format materialize relations between problem components. Over 40 concepts, corresponding to distinct domains, 30 mereological relations and relational requirements between components, and 60 requirements on component properties have been described. We showcase the Problem Instance formalization of the non-spatial, single-area LE, based on Land Characteristics (LC), model type. The clear separation between domain knowledge (Problem Instance) and high-level theories (Problem Encoding) enables the consistent LE problem formalization using the ASP-based system configuration paradigm. A declarative Problem Instance formalization provides insight into the problem’s nature and assumptions. Modular knowledge formalization using ASP, among others, enhances flexibility, scalability, and adaptability, given new knowledge becomes available.