Petri nets (PNs) are a central formalism for distributed systems, though, PNs turn out to be inadequate to model changes in real systems' layout. To cope with such a major issue, a framework based on Reisig's algebraic PNs has been recently proposed. It consists of an algebraic net emulating a P/T system encoded in net's inscriptions, and a set of rewriting rules defined at the net level. This approach permits the reuse of classical techniques and is provided with a sound initial semantics. In this paper the framework extends to Stochastic PNs, to make performance analysis possible. The ability to set up state-dependent transition rates is retained.