Trajectory-Aware Flow Matching for Topology Optimisation

DOI10.48550/arXiv.2607.14652MaRDI QIDQ7284382arXivFDO

Authors Shusheng Xiao, Jinshuai Bai, Hyogu Jeong, Y. T. Gu, Yunfei Xi, Yilin Gui, YuanTong Gu

Publication date 16 July 2026

Published in arXiv

Tldr This paper develops a flow matching-based topology optimisation framework for generating diverse topology candidates under varying physical conditions, where mechanically meaningful intermediate states are incorporated via physics-guided optimisation history into generative flow learning without additional inference-time optimisation; trajectory-aware FMTO achieves improved performance and stability with moderate path guidance.

Full work available at URL https://doi.org/10.48550/arXiv.2607.14652

arXiv classification

cs.LGmath.NA

Cites work

• A critical review of established methods of structural topology optimization
• A Fourier neural operator-based lightweight machine learning framework for topology optimization
• A level set method for structural topology optimization.
• Accurate and real-time structural topology prediction driven by deep learning under moving morphable component-based framework
• Evolutionary structural optimisation (ESO) using a bidirectional algorithm
• Neural networks for topology optimization
• Stochastic Interpolants: A Unifying Framework for Flows and Diffusions
• Topology optimization in aircraft and aerospace structures design

This page was built for publication: Trajectory-Aware Flow Matching for Topology Optimisation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7284382)