Data-driven environment modeling for adaptive system-of-systems

Since a System-of-Systems (SoS) is constructed and managed under a complex and dynamic environment, self-adaptability has become one of the key capabilities that SoSs should have. To design an adaptive SoS, analyzing and modeling the environment are important. Studies on self-adaptive systems (SAS) have proposed various analysis and design approaches to deal with dynamic environment and operating conditions. However, most existing approaches require a considerable amount of domain experts' knowledge about the operating environment without specific and practical guidelines, so there still remain many challenges for engineers to analyze and design an adaptive SoS. In this study, we propose a data-driven method of generating environment models for adaptive SoS. To guide the analysis and understanding of the environment, we propose a metamodel that encompasses characteristics of the dynamic environment. Based on the metamodel, an environment model is generated from historical data for effective analysis of the SoS's complex environment. As a case study, we apply our method to a traffic environment modeling with real data. We show that our proposed method can practically help engineers generate environment models with concrete features that are necessary for adaptive SoS modeling by considering the environment as a major entity for SAS analysis and design.