Constructing Dependability Analysis Models of Reconfigurable Production Systems

Model-driven engineering methodologies are often used for the design and integration of the software and hardware components in automated production systems. Architecture models in general purpose or domain specific modeling languages allow the modular and systematic construction of formal analysis models for the evaluation of dependability and performability. The on-line reconfiguration and fault handling actions in the system introduce changes in the architecture and parameters, and thus result in a multi-phased operation. We propose a mission automaton formalism to describe reconfigurations, fault handling and parameter changes over architecture models. By maintaining the architecture model, which is modified by the reconfigurations, along with the corresponding analysis models, a stochastic phased-mission system (PMS) is obtained and solved for dependability.