In the classical Petri net model, tokens are indistinguishable. If one wishes to distinguish processes, one must separately construct places and transitions for each process.
In the present paper, the author generalizes the usual Petri net model to one in which tokens can be distinguished. The major part of the paper is devoted to a study of “S- and T-invariants” of these generalized nets. An S-invariant is a function that assigns to each place s an integer (“weight”) ms in such a way that for any given initial marking the weighted sum of tokens is invariant for all markings reachable from the initial marking. A T-invariant is a function that assigns to each transition t an integer mt in such a way that for any transition t, any firing sequence that fires t mt times restores the initial marking. Invariants provide a tool for analysis.
There are other known Petri net models with individual tokens that have previously been studied: colored nets and predicate/transition nets. The author indicates how his results can be applied to these two models as well.
The paper is carefully written, but it cannot be appreciated without a familiarity with previous work in this rather specialized area of Petri net theory.