Modeling and Simulation Based Engineering in the Context of Cyber-Physical Systems

TL;DR

This paper introduces MSBE, a methodology that explicitly models execution semantics to improve the verification and validation of Cyber-Physical Systems, bridging the gap between model behavior and physical execution.

Contribution

It formalizes execution conditions as first-class entities and proposes an iterative cycle integrating formal and experimental execution for CPS engineering.

Findings

MSBE formalizes execution semantics, activity, constraints, and properties.

The methodology applies to four CPS classes, demonstrating generality.

MSBE enhances the alignment between verified models and actual system behavior.

Abstract

Cyber-Physical Systems (CPS) produce behavior through execution on substrates coupling computation with physical processes. However, usual engineering approaches do not treat execution semantics as first-class engineering entities. Formal verification reasons about model behaviors under fixed semantic assumptions that are not revisable and do not account for physical execution constraints. Simulation-based validation explores scenarios under execution semantics that are implicitly determined by the simulation engine. In both cases, physical constraints of the execution substrate are addressed as implementation details rather than as semantic boundary conditions. In this article, it is hypothesized that making execution semantics explicit as first-class engineering entities is necessary and sufficient to bridge the gap between verified model behaviors and validated executed behaviors in CPS. To test this hypothesis, Modeling and Simulation Based Engineering (MSBE) is proposed: a methodology grounded in the Theory of Modeling and Simulation. MSBE formalizes execution conditions as four components: execution semantics, activity (behaviorally meaningful changes), admissibility constraints (physical bounds), and specified properties (behavioral guarantees). MSBE organizes engineering around an iterative cycle alternating formal execution, experimental execution, verification, and activity-mediated validation. Executability is defined as stabilization of execution conditions and the induced admissible model space. The cycle is applied to four CPS classes (human-centric, biophysical, technological, and digital twins). These applications show that the framework generalizes beyond CPS to any system whose behavior depends on explicitly defined execution conditions. Modeling and Simulation-Based Engineering