Towards compositional CSL model checking

The Continuous Stochastic Logic (CSL) is a powerful means to state properties which refer to Continuous Time Markov Chains (CTMCs). The verification of such properties on a model can be achieved through a suitable algorithm. In this doctoral thesis, the CSL logic has been considered and two major aspects have been addressed: the analysis of its expressiveness and the study of methods for a decomposed verification of formuale. Concerning expressiveness, we have observed that the CSL syntax can lead to formulae with a trivial semantics. As a consequence the idea of well-formed CSL formula has been introduced. Furthermore, a simpler and equivalent syntax for referring to ergodic CMTCs have defined as well as a brand new event-bounded Until operator. With respect to compositionality, we have referred our study to a specific type of decomposed CTMCs, namely the bidimensional Boucherie framework. A number of basic properties concerning the Boucherie framework have been demonstrated and, relying on this, a compositional semantics for a subset of the CSL syntax has been derived. The considered subset is obtained by disallowing nesting of probabilistic path-formulae, something whose impact on the ability to state useful properties is low.