AI-Driven Synthesis for High-Tech System Design: Automating Innovation
06:00 · June 29, 2026 · arXiv cs.AI RSS
This article addresses the combinatorial complexity inherent in modern high-tech system design by presenting automation-in-design (AiD) as a transformative paradigm. We propose computational design synthesis (CDS), a framework utilising deep learning and generative AI to automate the creation of novel systems. Two case studies (e-drive system design and spatial dimensioning problem) serve as proof-points for this approach. The AI-driven methods used in the case studies represent a fundamental shift in engineering, advancing from simulation-based optimisation towards autonomous design with minimal human supervision.
Summary
The article presents automation-in-design (AiD) as a response to the combinatorial explosion that arises when engineers must select components, define interconnections, and tune continuous parameters for complex dynamical systems. It frames computational design synthesis (CDS) as a structured framework that applies deep learning and generative models to generate valid topologies, map them to performance variables, and optimise outcomes with limited manual intervention. The approach treats topology selection, dimensioning, and control design as interdependent rather than sequential tasks, using graph-based representations and knowledge-extraction methods to constrain the search space.
Two case studies illustrate the framework. In e-drive system design, reinforcement learning combined with nonlinear programming iterates over feasible topologies while optimising continuous parameters. A second study addresses spatial dimensioning and packaging, where predictive models reduce reliance on repeated CAD iterations by directly mapping constraints to component placement and routing solutions. Both examples demonstrate a progression from simulation-driven optimisation loops toward generative schemes that can propose high-performing architectures almost instantaneously once trained on prior optimisation data.
The authors, affiliated with the Engineering Systems Design lab at Eindhoven University of Technology, position CDS as a practical step toward co-design of physical and control systems. By codifying design rules through expert input, physics analysis, or automated extraction from documents, the method supports both iterative refinement and fully predictive generation of system architectures. The work aligns with Dutch strengths in high-tech mechatronics and powertrain engineering, where managing tightly coupled discrete and continuous design variables remains a persistent bottleneck.
Why it matters
Directly targets AI-driven engineering innovation relevant to Dutch high-tech sectors (e.g., automotive, robotics); offers actionable CDS framework for researchers and advanced practitioners in NL/EU context.
