What is it about?

Designing complex systems like airplanes starts with figuring out the overall structure, but traditional methods often explore too few options. This study looks at three different ways to improve this process by linking early design decisions with later-stage optimization. The three approaches tested are: (1) global exploration, which optimizes everything at once and generally works best; (2) nested optimization, which breaks the problem into smaller steps and helps when the full design is not clear; and (3) decision chain optimization, which uses AI but is complicated to set up and does not improve results much. The study recommends using global exploration when possible and nested optimization when the full design is not known. The big test case is the design of an aileron, which demonstrates the effectiveness of the strategies proposed.

Featured Image

Why is it important?

This work is important because designing complex systems like airplanes involves making big decisions early on that affect how well the final design performs. Traditional methods often overlook better options because they don’t explore enough possibilities. By improving how these early decisions are connected to later design steps, this research aims to help engineers find better, more efficient solutions.

Perspectives

This article summarizes my work as a master's student at TU Delft and GKN Fokker. It was a pleasure to carry out this work, as a great supervisory team, including Dr.ir, supported me from start to finish. Gianfranco La Rocca, Dr.ir. Ton van Der Laan, Dr.ir. Tobie van den Berg, Anne-Liza Bruggeman, and Jente Sonneveld. I hope this article gets more people interested in the world of system architecture optimization and opens the door to future collaborations!

Santiago Valencia-Ibáñez
Concordia University

Read the Original

This page is a summary of: Optimization Strategies for System Architecting Problems, January 2025, American Institute of Aeronautics and Astronautics (AIAA),
DOI: 10.2514/6.2025-2847.
You can read the full text:

Read

Resources

Contributors

The following have contributed to this page