Learning to Optimize Multigrid PDE Solvers

Abstract

Constructing fast numerical solvers for partial differential equations (PDEs) is crucial for many scientific disciplines. A leading technique for solving large-scale PDEs is using multigrid methods. At the core of a multigrid solver is the prolongation matrix, which relates between different scales of the problem. This matrix is strongly problem-dependent, and its optimal construction is critical to the efficiency of the solver. In practice, however, devising multigrid algorithms for new problems often poses formidable challenges. In this paper we propose a framework for learning multigrid solvers. Our method learns a (single) mapping from a family of parameterized PDEs to prolongation operators. We train a neural network once for the entire class of PDEs, using an efficient and unsupervised loss function. Experiments on a broad class of 2D diffusion problems demonstrate improved convergence rates compared to the widely used Black-Box multigrid scheme, suggesting that our method successfully learned rules for constructing prolongation matrices.

arXiv commentProceedings of the 36th International Conference on Machine Learning (ICML 2019)

Cited by in corpus (17)

• Fourier Neural Operator for Parametric Partial Differential Equations
• Physics-Informed Neural Operator for Learning Partial Differential Equations
• Multiwavelet-based Operator Learning for Differential Equations
• Variational training of neural network approximations of solution maps for physical models
• Data-Driven Multi-Grid Solver for Accelerated Pressure Projection
• Meta-MgNet: Meta Multigrid Networks for Solving Parameterized Partial Differential Equations
• Analysis of Deep Ritz Methods for Laplace Equations with Dirichlet Boundary Conditions
• Learning optimal multigrid smoothers via neural networks
• Optimization-Based Algebraic Multigrid Coarsening Using Reinforcement Learning
• Neural Multigrid Architectures
• Semi-Implicit Neural Solver for Time-dependent Partial Differential Equations
• Tuning Multigrid Methods with Robust Optimization
• Spectral Analysis of the Neural Tangent Kernel for Deep Residual Networks
• Semi matrix-free twogrid shifted Laplacian preconditioner for the Helmholtz equation with near optimal shifts
• Physics-Aware Downsampling with Deep Learning for Scalable Flood Modeling
• Direct optimization of BPX preconditioners
• On Adapting Nesterov's Scheme to Accelerate Iterative Methods for Linear Problems