Active Systems

Abstract

An active system is a network of components. Each component is endowed with input and output terminals , which are connected with terminals of other components of the system through proper connection links. Interactions among components are represented by events produced and received on terminals through connection links. System components may also produce observable outputs. These are modeled as events on a specific terminal, which are assumed to be collected in some way by an external observer. The behavioral model of a component is described by a communicating finite automaton, which incorporates both nominal and faulty behavior. In the transition function, the current state and the input event at a given time instant give rise to the production of some output events and to an instantaneous state transition. The behavior of system components in the presence of a relevant input event coming from the external world is the reaction of the system. The observable outputs produced by the system during its reaction represent the observation of the system. The system observation is always associated with a relevant observer, that is, the mode in which the observable outputs can be actually seen from the external world. In fact, only a subset of the set of observable outputs are actually visible by an observer. Furthermore, the total ordering of the observed events can be lost and transformed into partial ordering, thereby making the system observation temporally underconstrained. A diagnostic problem is the association among an observer, a system observation, and the state of the system prior to the reaction.