The ProB Animator and Model Checker

ProB is an animator, constraint solver and model checker for the B-Method (see the B-Method site of Clearsy). The constraint-solving capabilities of ProB can also be used for model finding, constraint-based symbolic checking and test-case generation.

The B language is rooted in predicate logic, arithmetic and set theory and provides support for data structures such as (higher-order) relations, functions and sequences. In addition to the B language, ProB also supports Event-B, CSP-M, TLA+, and Z. ProB can be installed within Rodin. ProB can also be used as a Jupyter kernel to generate interactive notebooks.

ProB is being used within Siemens, Alstom, Thales and several other companies for data validation of complicated properties for safety critical systems. It was used, e.g., for Paris Line1, Sao Paulo line 4, Barcelona line 9 and many more. ProB is certified T2 SIL4 according to the Cenelec EN 50128 standard for use at Thales. In this video from the Deutsche Bahn you can see ProB animating a formal B model of the ETCS hybrid-level 3 principles in real-time, controlling two trains. For commercial support contact Michael Leuschel.

Versions of ProB

Several versions of ProB are available. They all make use of the same Prolog core (see below).

A command-line tool called probcli. It is distributed with the standard download of the latest release.
The original graphical user interface ProB Tcl/Tk. It is distributed with the standard download of the latest release.
The new graphical user interface ProB2-UI based on Java FX and the ProB2-Java-API of ProB. It is available as a separate download. It contains VisB for SVG-based visualizations (as a successor to BMotionStudio).
ProB for use within Rodin, the toolset for Event-B. It is available as a plugin.
An API for using ProB from Java, called ProB2-Java-API (aka ProB2). It is available via Maven Central.
A kernel for Jupyter, allowing to use B and ProB from within Jupyter notebooks. It is available via a seperate installation procedure.

News

26/1/2021 ProB2-UI 1.1.0 is available, contains VisB.

15/12/2020 ProB 1.10.0 is available. Highlights: well-definedness prover, REAL datatype, -lint comand for VSCode and Atom, improved unsat core and error messages.

19/2/2020 ProB 1.9.3 is available. Highlights: performance improvements, new external functions, performance monitoring, notarized for macOS.

12/7/2019 ProB 1.9.0 is available. Highlights: Unicode support, regular expression library, memoization. New ProB2 UI.

1/10/2018 ProB 1.8.2 is available. Highlights: Jupyter kernel, first support for Alloy models.

11/7/2017 ProB 1.7.0 highlights: improved Latex document generation, improved XML/CSV support, RULE DSL language.

20/10/2016 ProB 1.6.1 highlights: LET and IF-THEN-ELSE for expressions and predicates.

22/4/2016 ProB 1.6.0 highlights: directed model checking, Z3 as backend.

18/8/2014 ProB 1.4.0 is available. Highlights: CLP(FD)-based constraint solver enabled by default, integration of the TLC model checker.

4/3/2013 ProB 1.3.6 is available. Highlights: improved constraint propagation, model checking progress bar, improved Kodkod backend and use within REPL.

30/03/2012 ProB Logic Calculator available.

Implementation

The core of ProB is implemented in SICStus Prolog (but can be run without a SICStus Prolog license). The ProB constraint solver is implemented using co-routining and the CLP(FD) finite domain library of SICStus. An alternate constraint solver based on Kodkod (and thus SAT) is also available within ProB, as is an integration with the SMT solver Z3. An alternate model checking engine (using TLC) well-suited for lower level B specifications is available as is an integration with LTSmin as model checking backend. The ProB Licence can be found here. Automatically generated test coverage reports are also available.

Features

ProB covers a large part of B, and we are striving towards full coverage of Atelier B constructs. ProB supports B features such as non-deterministic operations, arbitrary quantification, sets, sequences, functions, lambda abstractions, set comprehensions, records, and many more. ProB does support multiple machines, refinements, and implementations. ProB can also be used for automated refinement checking and LTL model checking. It also supports CSP-M process descriptions, to be used on their own or to guide B machines for specification and property validation. The state space of the specifications can be graphically visualized. ProB supports Z specifications (ProB in this context is sometimes called ProZ) as well as TLA+ specifications. ProB can be used within Rodin and Atelier-B as a disprover and prover.