Automatic Modular Abstractions for Template Numerical Constraints

David Monniaux
DOI: https://doi.org/10.2168/lmcs-6(3:4)2010
2010-07-20
Logical Methods in Computer Science
Abstract:<p>We propose a method for automatically generating abstract transformers forstatic analysis by abstract interpretation. The method focuses on linearconstraints on programs operating on rational, real or floating-point variablesand containing linear assignments and tests. Given the specification of anabstract domain, and a program block, our method automatically outputs animplementation of the corresponding abstract transformer. It is thus a form ofprogram transformation. In addition to loop-free code, the same method alsoapplies for obtaining least fixed points as functions of the precondition,which permits the analysis of loops and recursive functions. The motivation ofour work is data-flow synchronous programming languages, used for buildingcontrol-command embedded systems, but it also applies to imperative andfunctional programming. Our algorithms are based on quantifier elimination andsymbolic manipulation techniques over linear arithmetic formulas. We also giveless general results for nonlinear constraints and nonlinear programconstructs.</p>
computer science, theory & methods,logic
What problem does this paper attempt to address?