These slides were presented at the NAFEMS World Congress 2025, held in Salzburg, Austria from May 19–22, 2025.

Abstract

The design of complex engineering systems, such as aircraft, gas turbines, or modern cars amongst others presents significant challenges on both technical and project management fronts. Finding an optimal design solution to meet the demanding performance, cost, and safety requirements of such complex engineering systems is not trivial. Traditional monolithic design optimization approaches often struggle to handle the complexity of complex systems and their inherent interdependencies, where various disciplines interact in intricate ways (such as aerodynamics, structures, propulsion, and controls to only name the most obvious systems in case of an aircraft). An appealing approach to solve this challenge is to apply a multidisciplinary design optimization technique, which enables the integration and consideration of multiple disciplines into the design process. However, in practice, complex engineering systems are often decomposed into siloed design teams based on their specific functionality, regulatory requirements, and role in ensuring safety and performance of the system. These teams focus on optimizing their subsystems using isolated tools, making it difficult for systems engineers to integrate these optimizations cohesively and efficiently. This poses an additional challenge that monolithic multidisciplinary design optimization techniques cannot handle. To tackle this problem, the authors propose a distributed multidisciplinary design optimization solution. The proposed solution enables an interactive decomposition of the global problem into separate subproblem studies which can be run in a distributed manner in parallel, while still considering the interactions between the different disciplines involved in the multidisciplinary design optimization problem. The two main elements in the proposed methodology rely on applying Nonhierarchical Analytic Target Cascading for solving the multidisciplinary design optimization problem and a client-server type framework that enables distributing the different disciplinary studies to different client applications, which can be located anywhere around the world. To represent the complete system, including all subsystems and their interdependencies concisely, the authors employ a Design Structure Matrix (DSM), also known as an N-squared (N2) matrix. This approach enables not only to tackle the design challenge of complex engineering systems, but also enables and facilitates the collaboration of separate design teams involved around the world. In this publication, the authors introduce the new solution for distributed multidisciplinary design optimization and demonstrate it on a few selected application cases.