Automatic Verification by Abstract Interpretation

We discuss the use of abstract interpretation in the context of automatic program veri cation requiring precise abstractions. We compare entirely manual versus user-guided abstractions ranging from program-speci c abstractions including predicate abstraction to the systematic design of abstract domains and iteration strategies. 1 Abstract Interpretation Theory Abstract interpretation theory [1,2,3,4,5,6] formalizes the notion of abstraction for mathematical constructs involved in the speci cation of computer systems. Applications range from static program analysis [2,3,4,6] (including dataow analysis [3], set-based analysis [7], etc), typing [8], model-checking [9], parsing [10] to the design of semantics [11] and program transformations [12]. In this invited tutorial we discuss applications to automatic program verification.interpretation theory [1,2,3,4,5,6] formalizes the notion of abstraction for mathematical constructs involved in the speci cation of computer systems. Applications range from static program analysis [2,3,4,6] (including dataow analysis [3], set-based analysis [7], etc), typing [8], model-checking [9], parsing [10] to the design of semantics [11] and program transformations [12]. In this invited tutorial we discuss applications to automatic program verification.