**1. Introduction**

Recently, industry is experiencing a new revolution. In the past, product design, as well as their associated manufacturing processes, were based on the use of nominal models, nominal loadings (in their broadest sense), and a small amount of data for calibrating those models, with the product performance as a design target.

Very recently, predictions enabling real-time decision-making targeting zero defects in processing and zero unexpected faults in operation, were needed everywhere within the Internet of Things (IoT) paradigm, on the work-floor (smart processes), in the city (autonomous systems and smart-city), at the nation level (e.g., smart nation), etc., i.e., anywhere where engineering designs operate.

In those circumstances, the use of traditional simulation-based engineering (SBE) that was the major protagonist of 20th century engineering, is not anymore a valuable option due to three main reasons: (i) models become sometimes crude approximations of the observed reality; (ii) assimilating data enabling the continuous calibration of the models in operation remains difficult to perform under the stringent real-time constraint; and (iii) the real-time simulation of those extremely complex mathematical models needs alternative techniques to those commonly employed in traditional SBE.

It was at the beginning of the XXI century that two new revolutions in the domain of digital engineering emerged.
