A General Labeling Algorithm for Assumption-Based Truth Maintenance

Assumption-based truth maintenance systems have become a powerful and widely used tool in Artificial Intelligence problem solvers. The basic ATMS is restricted to accepting only horn clause justifications. Although various generalizations have been made and proposed to allow an ATMS to handle more general clauses, they have all involved the addition of complex and difficult to integrate hyperresolution rules. This paper presents an alternative approach based on negated assumptions which integrates simply and cleanly into existing ATMS algorithms and which does not require the use of a hyperresolution rule to ensure label consistency.