SHML: Stochastic Hybrid Modeling Language for CPS Behavior

Cyber-Physical Systems (CPS) connect the cyberworld with physical world with a network of interrelated el-ements, such as sensors and actuators. It is always runningin an open environment and the main characteristics of CPSis hybrid and stochastic. Domain-Specific Modeling Language(DSML) offers a tailor-made solution for modeling a specific field. However, there still lacks of DSML to model hybrid and stochasticbehavior in CPS. To address these issues, we propose a StochasticHybrid Modeling Language (SHML) based on domain modellanguage engineering, which supports modeling stochastic andhybrid behaviors in CPS. The abstract syntax, concrete syntax, and operational semantics of SHML are presented. The SHMLis implemented based on the GEMOC studio. With the help ofthe GEMOC execution engine and the Scilab plugin, the SHMLmodels can be executed to generate simulation traces of thesystem. These traces are fed into a statistical model checker whichsupports simulation-based verification to enable the qualitativeand quantitative analysis. The novelty of our work is that aDSML is proposed to model the behavior of CPS. Moreover, the tool prototype is implemented based on the model-drivenarchitecture. We illustrate the feasibility of our approach withan energy-aware building.