A New SMT Algorithm Based on Lazy Framework Combining Clause Weight and Reward Mechanism

Abstract

Satisfiability Modulo Theories (SMT) is an extension of the Boolean Satisfiability problem(SAT). It is a theory for satisfiability judgment of multi-type first-order logic formulas. It is being recognized as increasingly important due to its applications in different communities, particularly in compilation optimization, scheduling, static program analysis, software and hardware verification and other fields. Firstly, according to the VSIDS branch selection strategy, this paper proposes the branch selection algorithm WR-SAT based on the “clause weights-reward mechanism”, which gives priority to selecting variables in short clauses for assignment and reward variables in the conflict learning clauses appropriately, in order to reduce unless search in SAT during branch selection. Then, using OpenSMT, an open-source SMT solver, this paper combines WR-SAT with SMT solving in Lazy framework to form a new SMT solver, Modified-OpenSMT. Finally, by testing some instances in benchmarks of 2017 SMT-COMP, OpenSMT is compared with Modified-OpenSMT on the efficiency of solving and it can be found that Modified-OpenSMT is feasible for solving SMT problems. Meanwhile, Modified-OpenSMT has a good effect in solving SMT, especially on the eq_diamond benchmark family and qg5 benchmark family, which is better than OpenSMT.