Verifying χ Models of Industrial Systems with Spin

Abstract

The language χ has been developed for modeling of industrial systems. To obtain performance measures, its simulator has been successfully used in many industrial areas. However, for functional analysis, simulation is less applicable. Such analysis can be done in other environments. In this paper we present guidelines and techniques for translating χ specifications to Promela, the input language of the well known model checker Spin. We highlight the differences between the two languages and show, in a step by step manner, how some of them can be resolved. We conclude by giving the translation scheme and apply it to a small industrial case study.

A Model-based Integration and Testing Method to Reduce System Development Effort

“…However, TorX supports Trojka [10], a slightly modified version of Promela, the specification formalism for the model checker Spin [15]. Spin is also used in the model verification back end of the χ toolset [5], for which a translation scheme from χ to Promela is developed [22]. By combining the translation of χ to Promela and the model-based testing capabilities of TorX, the conformance of a component realization with respect to the χ model used for model-based integration can be determined.…”

Stuttering Congruence for χ

“…We have illustrated the use of stuttering congruence in correctness proofs of syntactic transformations by indicating how (a part of) the preprocessing phase of the translation from χ to Promela can be proved correct. It is explained in detail in [16] that if a χ model satisfies a few general restrictions, then the preprocessing phase yields a χ model that can be straightforwardly translated into Promela. The resulting Promela specification can then be verified with Spin.…”

“…In [16], the translation from χ to Promela is discussed informally. It is pointed out that the translation is straightforward for some constructs of χ (e.g., for assignments and alternative composition), since they also exist in Promela.…”

“…In [16], the translation of χ specifications into Promela is discussed in more generality. The translation proceeds in two phases.…”

“…The resulting notion we call stuttering congruence and we prove that it is indeed a congruence with respect to the syntactic constructs of the discrete-event, untimed part of χ. In Section 5 we briefly discuss how our notion of stuttering congruence can be used to establish part of the correctness of the translation proposed in [16]. The paper ends with a conclusion.…”