An Abstraction-Refinement Framework for Reasoning with Large Theories

摘要

In this paper we present an approach to reasoning with large theories which is based on the abstraction-refinement framework. The proposed approach consists of the following approximations: the over-approximation, the under-approximation and their combination. We present several concrete abstractions based on subsumption, signature grouping and argument filtering. We implemented our approach in a theorem prover for first-order logic iProver and evaluated over the TPTP library.