Quantitative analysis for symbolic heap bounds of CPS software

One important quantitative property of CPS (Cyber-Physical Systems) software is its heap bound for which a precise analysis result needs to combine shape analysis and numeric reasoning. In this paper, we present a framework for statically finding symbolic heap bounds of CPS software. The basic idea is to separate numeric reasoning from shape analysis by first constructing an ASTG (Abstract State Transition Graph) and then extracting a pure numeric representation which can be analyzed for the heap bounds. A quantitative shape analysis method based on symbolic execution is defined in the framework to generate the ASTG. The numeric representation is extracted based on program slicing technique and inputted into an abstract interpretation tool for computing the heap bounds. We take list manipulating programs as an example to explain how to instantiate the framework for important data structures and to exhibit its practicability. A novel list abstraction method is also presented to support the instantiation of the framework.