Linear Refutation and Clause Splitting

Abstract

Linear resolution is one of the ancient methods for first-order theorem proving. We extend linear resolution with clause splitting, producing subgoals dispatched independently. An incremental SAT solver keeps track of refutations and thus provides a “lemma” mechanism. We describe some implementation considerations, present some initial experimental results, and discuss future directions for this approach.