Abstract
Traditional Petri nets lack specific features to conveniently describe systems with an evolving structure. A model based on the Symmetric Net formalism has been recently introduced. It is composed of an emulator reproducing the behaviour of a Place/Transition net (encoded as a marking) and a basic set of net-transformation primitives to specify evolutionary behaviour. In this paper, we discuss the adoption of the stochastic extension of Symmetric Nets for performance analysis, considering important issues related to time specification and analysis complexity. We put into place theoretical aspects by using a running example consisting in a self-healing manufacturing system.