Abstract:

Analysis and simulation of complex systems is a hard task that requires 
the use of proper modelling formalism and tools. In many cases, no single 
analysis and modelling method can successfully cope with the growing 
complexity of a real system. A multi-formalism multi-solution (MFMS) 
approach is very appealing, since it allows to cope with the complexity 
of the problem by using different formalisms to model and analyze different
formalisms to model and analyze different subsystems and also to define 
reusable building blocks.Nevertheless problems arise at many levels.
The major concerns are: the interoperability of differnet formalisms
and analysis/simulation tools, the definition and the implementation of 
mechanisms to guarantee the flexibility and the scalability of the 
modelling frameworks and the development of proper strategies for the 
analysis of MFMS models.
This paper describes a MFMS approach to the construction of dependability
models based on the integration of different graph-based formalisms. 
The proposed approach is supported by the DrawNET++ framework through 
(a) an object-oriented construction method that enables the definition
and the usage of model classes (including formalism definition and 
inheritance), model objects (including templates), model instances, 
weak and strong aggregation and (b) a proper interface to the external 
solvers (analysis/simulation engines) realized by means of the XML and 
XSL technologies. A simple application example is used through the 
paper to describe the DrawNET++ modelling process and the possibility
of analyzing a system by integrating two different formalisms:
Fault Trees (FT) and Generalized stochastic Petri Nets (GSPN).