报告题目：IGANets: Isogeometric Analysis Networks

报 告 人：Matthias Möller 副教授（Delft University of Technology）

报告时间：2026年10月19日（星期一）14:00—15:00

报告地点：数学科学学院115（大报告厅）      

校内联系人：朱春钢 教授         联系方式：84708351-8618

报告摘要：Fast predictions across varying geometries and loading conditions are valuable in engineering design, but repeated high-fidelity simulations can be costly. This talk introduces IGANets, a spline-based, physics-informed learning framework that connects CAD representations with isogeometric analysis. IGANets map problem parameters to spline coefficients of the solution and incorporate the governing equations through an isogeometric collocation loss. Numerical examples for the Poisson equation and linear elasticity illustrate the approach. The talk will also discuss its extension to nonlinear elasticity. A family of structural problems with varying geometries and boundary conditions further demonstrates its ability to predict solutions for previously unseen problem instances within the training range.

报告人简介：Matthias Möller is an Associate Professor of Numerical Analysis in the Department of Applied Mathematics at Delft University of Technology and the Scientific Director of the Delft Institute for Computational Science and Engineering. His research focuses on computational methods for numerical simulation and design optimization, including scientific machine learning, isogeometric analysis, and quantum-accelerated scientific computing.