Cautious Models for General Logic Programs

摘要

In this paper, cautious models of general logic programs are investigated. Such models are constructed iteratively using a monotonic operator which performs case analysis on total interpretations generated by enumerations of atoms. Consequently, every general logic program has a unique cautious model. The new class of partial models is compared with well-founded and stable models of general logic programs. Various extensions of these models are are also addressed. The time complexity of constructing cautious models is analyzed. Results indicate that the major reasoning task is co-NP-complete. Finally, the connection to cautious autoepistemic logic is explained.