Next: Methodology Up: Programme and Methodology Previous: Programme and Methodology

Overall Aims and Objectives

The overall aim of this project is to explore the possibility of augmenting static analysis with quantitative features by developing an abstract interpretation methodology in a probabilistic setting. This will allow us to replace, in cases like the above presented example, ad hoc methods by mathematically sound reasoning.

The main expected outcome of this investigation is a framework in which one can estimate -- in a way similar to statistical methods -- probabilities associated to the properties of programs, instead of asserting the properties in a conservative way.

The introduction of probability will be studied on two levels. On an operational level, this will result in a statistical data-flow analysis; on a more abstract level, we will re-define the Cousot framework to accommodate within it probabilistic information.

The methodology we develop will apply to both probabilistic and deterministic programs. Although the framework will be general enough to include a wide range of programming languages -- functional, imperative, logic, etc. -- we will concentrate on constraint languages, and in particular on Concurrent Constraint Programming languages.

Next: Methodology Up: Programme and Methodology Previous: Programme and Methodology