Machine Learning Models for Performance Predictions & Behavior Evolutions in Complex Systems

A complex system is characterized by the emergence of global properties, which are very difficult to anticipate just from complete knowledge of component behaviors. Further, advances in technology have made it easy to integrate multiple modern systems to form complex system-of-systems (SoS) to achieve unparalleled levels of functionality that are otherwise not achievable by the constituent systems in isolation. Recently, there is a significant increase in the adoption of data-driven machine learning models in many engineered physical systems, such as cars and drones, and are being used to govern their functionality and behavior. It is critical to ensure that the users of these complex engineered systems can be certain of the behavior and performance of that system. Hence, there is a need for data-driven models to also be physics aware and leverage the same towards achieving enhanced confidence in systems. This presentation discusses some of the recent approaches and challenges towards using machine learning models for predicting the performance and behavior of complex engineered systems. Further, approaches to inculcate adaptable intelligence in constituent systems to adapt their behaviors in the SoS context are also discussed.