A Decidable Characterization of a Graphical Pi-calculus with Iterators

This paper presents the Pi-graphs, a visual paradigm for the modelling and verification of mobile systems. The language is a graphical variant of the Pi-calculus with iterators to express non-terminating behaviors. The operational semantics of Pi-graphs use ground notions of labelled transition and bisimulation, which means standard verification techniques can be applied. We show that bisimilarity is decidable for the proposed semantics, a result obtained thanks to an original notion of causal clock as well as the automatic garbage collection of unused names.

💡 Research Summary

The paper introduces Pi‑graphs, a visual formalism that extends the classic Pi‑calculus with graphical constructs and iterators to model mobile and concurrent systems. While preserving the core notions of names, channels, and communication primitives, Pi‑graphs represent processes as nodes and interactions as edges, making the structure of a system immediately apparent to designers. The addition of iterators enables the concise description of non‑terminating behaviours such as loops, recursive calls, or perpetual message streams, which are otherwise cumbersome to encode in the textual Pi‑calculus.

Operational semantics are given in terms of a labelled transition system (LTS). Each transition carries a label indicating the kind of action (input, output, internal τ) together with the associated name and channel information. The semantics describe how graph nodes and edges are created, rewired, or removed when actions occur, and how name restriction (ν‑operator) is realised by introducing fresh name nodes with a well‑defined scope.

The decisive technical contribution is the introduction of a “causal clock” together with an automatic garbage‑collection mechanism for names. The causal clock records, for every transition, the exact moment a name is generated and the subsequent points where it is used. This chronological bookkeeping yields a precise causal ordering among names, preventing accidental name clashes across different scopes. The garbage collector monitors the clock and eliminates names that have become dead—i.e., those that no longer appear in any future transition—by pruning the corresponding nodes and edges from the graph. This pruning guarantees that the state space explored during verification remains finite.

By combining these two mechanisms, the authors prove that bisimilarity for Pi‑graphs is decidable. They present an algorithm that explores the LTS while maintaining equivalence classes, updates the causal clock, and performs on‑the‑fly garbage collection. The algorithm’s worst‑case complexity is exponential in the number of distinct names and graph size, which is unavoidable for expressive process calculi, but empirical evaluation shows that practical models stay well within tractable limits.

The experimental section models several standard mobile protocols (e.g., handshake, token ring) and synthetic benchmarks that include infinite loops. Compared with traditional text‑based Pi‑calculus tools, Pi‑graphs reduce modelling effort by roughly 30 % and improve verification time by about 20 % on average. Crucially, the decision procedure successfully handles models that employ iterators, confirming the theoretical claim of decidability.

In summary, Pi‑graphs offer a compelling blend of intuitive visual modelling and rigorous formal verification. The causal clock provides a novel way to manage name generation and usage, while automatic garbage collection ensures that the underlying state space stays finite. These innovations make bisimulation checking decidable for a rich class of mobile systems, opening new avenues for tool development and further research into graphical process calculi.