The AltaRica formalism was developed at the LaBRI in Bordeaux jointly with
industrial partners and allows to analyse a given system with several
different methods (fault tree, Petri nets, Markov chains) in order to
perform safety analysis of the system. This talk is concerned with the
model checking of such models, and in particular in the Mec 5 tool which
we developed.
We will describe in this talk a few key points of the AltaRica formalism,
and
then the very expressive specification language we use to compute arbitrary
(finite) relations which allow us to express usual "temporal" properties of
models as well as e.g. bisimulation relations, in a very natural way.
The specification language is first-order logic extended with fixpoints over
relations.
We will also give some insight of the implementation of Mec 5 which
basically
uses binary decision diagrams, and discuss the different ideas we are
investigating to improve the performance of Mec 5.
|