CPNsim

CPNsim is a Rust library and command-line tool for simulating Colored Petri Nets (CPNs). It parses CPN models in the .ocpn JSON format and executes simulations step by step — handling token consumption, guard evaluation, and token production based on arc inscriptions and variable bindings.

Note: This library is a work in progress. Feel free to leave an issue with your feature request or bug report — your feedback is greatly appreciated!

CPNsim powers the simulation engine behind OCPN Tools and compiles to WebAssembly for in-browser execution.

Features

Core Simulation

• CPN model parsing — Reads .ocpn JSON format with support for places, transitions, and arcs
• Color sets — Basic types (INT, STRING, BOOL, UNIT), product types, record types, enumerated types, indexed types, subsets, lists, and INT with ranges
• Variables & bindings — Automatic variable binding during transition firing with support for multi-variable arcs
• Guard evaluation — Transition guards evaluated using the Rhai scripting engine
• Arc inscriptions — Input and output arc expressions evaluated via Rhai for flexible token manipulation
• Timed simulation — Time expressions, arc delays, and automatic time advancement for timed CPN models
• Hierarchical nets — Substitution transitions with port/socket bindings and fusion places
• Priority levels — Transition priority support for controlling firing order
• Distribution functions — 14 CPN Tools-compatible stochastic functions: bernoulli, beta, binomial, chisq, discrete, erlang, exponential, gamma, normal, poisson, rayleigh, student, uniform, weibull

Monitors

• Marking-size monitors — Track token counts on watched places over simulation steps
• Transition-count monitors — Count firings of specific transitions
• Breakpoint monitors — Stop simulation when a place-based or transition-based predicate is met (e.g., place empty, transition fires)
• Data collector monitors — User-defined monitors with custom Rhai observation and predicate scripts
• Statistics — Each monitor computes running statistics: count, sum, average, min, max, and standard deviation

State Space Analysis

• BFS exploration — Breadth-first state space construction with configurable limits (max states, max arcs)
• Canonical markings — Deterministic state deduplication; timed models include simulation time in the state key
• SCC decomposition — Tarjan's algorithm for strongly connected component detection
• Report metrics — Dead markings, home markings, terminal SCCs, place bounds (min/max tokens), dead/live transitions, and transition fairness (arc occurrence counts)
• Timed state spaces — Supports timed CPN models with configurable time bounds
• Deterministic overrides — Replace stochastic distribution functions and IntRange random bindings with constant values for finite, deterministic exploration

Integration

• Library interface — Use as a Rust crate in other projects
• WebAssembly — Compiles to WASM via wasm-pack for browser-based simulation
• Debug CLI — Command-line executable for running simulations and inspecting model behavior

Building

Debug (CLI)

cargo run -- examples/petrinet-2.ocpn

WebAssembly

wasm-pack build --scope rwth-pads --target web --release -- --features wasm

License

This project is licensed under the MIT License. See the LICENSE file for details.