Local Search for SMT on Linear Integer Arithmetic

Abstract Satisfiability Modulo Linear Integer Arithmetic, SMT (LIA) for short, has significant applications in many domains. In this paper, we develop the first local search algorithm by directly operating on variables, breaking through traditional framework. We propose a framework considering distinctions between Boolean and integer variables. Moreover, design novel operator scoring functions tailored LIA, two-level operation selection heuristic. Putting these together, solver called LS-LIA. Experiments are carried out to evaluate LS-LIA benchmarks from SMTLIB two benchmark sets generated job shop scheduling data race detection. The results show that is competitive complementary with state-of-the-art solvers, performs particularly well those formulae only A simple sequential portfolio Z3 improves satisfiable of LIA IDL SMT-LIB. also solves Job Shop Scheduling substantially faster than complete solvers.