Artificial Intelligence Capabilities for Effective Model‐Based Systems Engineering: A Vision Paper

Abstract

Both Model-Based Systems Engineering (MBSE) and Artificial Intelligence (AI) have been challenged for their deployment in real-world applications. Although MBSE remains the focal point of any systems engineering activities, its adoption still faces significant hurdles to demonstrate its return on investment. Recently, AI has received intensive attention, and its applications made their way into our daily life products. From an industrial perspective, within the context of the design and development of mechatronic systems, there is a lack of coherent foundation to enable the application of AI in MBSE. This vision paper discusses the role of AI in solving a set of MBSE challenges. As a result, we contribute by describing the actual MBSE adoption challenges and follow up with the characterization of the capabilities of AI in solving these challenges. With this initial work, we aim to trigger both AI and MBSE communities for further research discussions and industrial applications to help in achieving an intelligent design and development environment. Motivation and BackgroundDuring the last decades, technology has been enormously revolutionized for products we use in our daily lives and people's expectations increased. Indeed, competition between companies got more intense and brought new challenges to deliver smarter, safer, adaptable, and sustainable products in a faster and cheaper way. Companies designing and developing mechatronic products, for instance in transportation, aerospace and automotive, regularly face huge difficulties due to the multidisciplinary nature and complexity of their products. In most cases, these difficulties could be traced back to a set of technological, financial, or human factors (Chami & Bruel, 2018). In some cases, these difficulties can lead to massive unforeseen costs and catastrophic failures.In order to maintain a profitable business, employees perform diverse technical, administrative and cognitive activities to bridge their stakeholder needs with most of their products' features. Although these activities might sound trivial, their evolving nature triggers new challenges for keeping them up-to-date, efficient and optimized. Therefore, instead of focusing solely on delivering intelligent products, we ask ourselves why not supporting as well designing and developing them with the help of some intelligent environment?