A Curated Corpus of Simulink Models for Model-Based Empirical Studies

Abstract

Recent years have seen many empirical studies of model-based cyber-physical systems and commercial CPS development tool chains such as Matlab/Simulink. To benefit such research, this paper presents the by-far largest corpus of freely available Simulink models to date, containing over 1,000 models. Surprising findings based on this corpus include that (a) tool support for metric collection is not adequate and (b) users do not reuse model components as they would in object-oriented programs. The paper both confirms and contradicts earlier findings that are based on significantly fewer models, suggesting the utility of the corpus for future research. While others have not yet leveraged this model corpus, we hope that our freely available corpus and infrastructure will benefit future model-based empirical research and tool development efforts, by reducing the model-collection overhead and thus easing evaluation.