Search papers, labs, and topics across Lattice.
This paper introduces Graph-Instructed Neural Networks (GINNs) to efficiently solve parametric Partial Differential Equations (PDEs) with varying boundary conditions, a challenge for traditional Galerkin methods due to the need to reformulate the discrete problem for each new boundary configuration. GINNs learn the mapping between the parametric description of the computational domain and the PDE solution, avoiding the need for re-formulation. The results show that GINN-based models can efficiently represent complex parametric PDEs, offering a scalable alternative to fully connected architectures.
Forget re-meshing: Graph-Instructed Neural Networks learn to solve parametric PDEs with varying boundary conditions directly from the domain description.
This work addresses the accurate and efficient simulation of physical phenomena governed by parametric Partial Differential Equations (PDEs) characterized by varying boundary conditions, where parametric instances modify not only the physics of the problem but also the imposition of boundary constraints on the computational domain. In such scenarios, classical Galerkin projection-based reduced order techniques encounter a fundamental bottleneck. Parametric boundaries typically necessitate a re-formulation of the discrete problem for each new configuration, and often, these approaches are unsuitable for real-time applications. To overcome these limitations, we propose a novel methodology based on Graph-Instructed Neural Networks (GINNs). The GINN framework effectively learns the mapping between the parametric description of the computational domain and the corresponding PDE solution. Our results demonstrate that the proposed GINN-based models, can efficiently represent highly complex parametric PDEs, serving as a robust and scalable asset for several applied-oriented settings when compared with fully connected architectures.
