Model Finding Strategies in Semantically Guided Instance-based Theorem Proving

Semantic hyper-linking has recently been proposed [1] to use semantics in an instance-based theorem prover. The basic procedure is to generate ground instances of the input clauses until the ground clause set is unsatisfiable. Models for the satisfiable ground set are constructed periodically to guide generation of the new ground instances to change the models, until no model can be constructed. In this paper we discuss some model finding strategies that can generate useful ground instances, without using semantics, to change the ground models. We show that such strategies are helpful and will not increase the search space of the semantic hyper-linking. In addition, using semantics is often expensive. Since semantics is not used in those model finding strategies, they help to find the proofs earlier and faster.