A Knowledge based method for Data Flow Analysis

Abstract

Data Flow Analysis (DFA) is essential in software engineering tasks such as program verification, compiler optimization, and program understanding. Knowledge based methods for problem solving admit highly efficient and understandable problem description. This paper proposes a novel Answer Set Programming (ASP) based solution to the DFA of imperative programs. Given a program, all knowledge necessary for DFA is produced automatically. With answer set solvers such as DLV, SMODELS and CMODELS, the analysis results could be generated automatically. Several experiments are conducted to show the applicability and efficiency of our method. It is shown that the ASP based DF A method is very suitable for quick formulation and solving of DFA problem.