Modeling Adaptive Socio-Cyber-Physical Systems with Goals and SysML

Abstract

Socio-cyber-physical systems (SCPSs) are cyber-physical systems (CPSs) with a socio-technical system (STS) aspect. Several SCPSs need to adapt dynamically to changing situations in order to reach an optimal symbiosis with users in their contexts. Tailoring requirements engineering activities and modeling techniques is needed for developing SCPSs and supporting their runtime adaptability. The proposed thesis aims to combine a common way of modeling STSs (i.e., goal modeling with the Goal-oriented Requirement Language - GRL) to a conventional way of modeling CPSs (i.e., SysML and feature models) in order to integrate social concepts early in SCPS requirements, design, simulation, and implementation activities. To help guarantee system quality and compliance during both design time and runtime adaptations, the thesis proposes to translate goal and feature models to mathematical functions used to validate possible design and adaptation alternatives both during simulations at design time and adaptations at runtime. These functions can be used outside goal modeling tools and be combined to SysML models, simulations, problem solvers, and implementation tools. Furthermore, an integration between GRL and SysML models via a third-party requirements management system is proposed in order to strengthen system traceability and help ensure that stakeholder goals were considered properly during the SCPS development process.