Clause Replication and Reuse in Incremental Temporal Induction

Temporal induction is one of the most popular SAT-based model checking techniques. It consists of two parts, the base case and the induction step. With the search length increment, both parts generate a sequence of SAT problems. This paper focuses on learnt clause replication and reuse in incremental temporal induction. Firstly, with the aid of assumption literals, we present an alternative clause replication scheme, which is much easier to implement than existing works. Secondly, based on our clause replication scheme, we present several clause reuse schemes to maximally explore the learnt clauses and their replications in temporal induction. Based on above ideas, we propose two new incremental temporal induction algorithms. Experimental results on a large number of benchmarks show significant performance improvement of our technique.