Backward reachability of Colored Petri Nets for systems diagnosis

Embedded systems development creates a need of new design, verification and validation technics. Formal methods appear as a very interesting approach for embedded systems analysis, especially for dependability studies. The chosen formalism for this work is based on Colored Petri Net (CPN)for two main reasons : the expressivity and the formal nature. Also, they model easily the static and the dynamic natures of the studied systems. The main challenge of this work is to use existing models, which describe the system structure and/or behavior, to extract the dependability information in a most general case and failure diagnosis information in a particular case. The proposed approach is a CPN structural backward reachability analysis. It can be split in two parts. The first one is to perform the proposed analysis : inverse CPN. It is obtained thanks to structural transformations applied on the original CPN. The second part is the analysis implementation. This part needs some complementary concepts of which the most important is the marking enhancement. The proposed approach is studied under two complementary aspects : algorithmic and theoretic aspects. The first one proposes transformations for the CPN inversion and the analysis implementation. The second aspect (the theoretical one) aims to offer a formal proof for the approach by applying two methods which are linear algebra and linear logic.