The Effect of Quadrature on the Convergence of Policy Iteration for Hamilton-Jacobi-Bellman Equations
math.NA, cs.NA
Submitted: 2026-06-23
Updated: 2026-06-23
Journal ref: Journal of Scientific Computing, volume 109, article number 79 (2026)
DOI: 10.1007/s10915-026-03480-9
Code: https://github.com/tomhall-git/hjb-policy-iteration-quadrature
License: http://creativecommons.org/licenses/by/4.0/
The gist: Modern finite element libraries allow users to express partial differential equations directly in variational form, with the added convenience of automatic quadrature selection.
Terminology
Abstract
Modern finite element libraries allow users to express partial differential equations directly in variational form, with the added convenience of automatic quadrature selection. In the context of Hamilton-Jacobi-Bellman (HJB) equations, automatic quadrature selection can result in nonmatching quadratures between different terms that may lead to loss of convergence of the policy iteration, which is otherwise expected from theory to converge superlinearly. The simple remedy of enforcing matching quadrature recovers the expected superlinear convergence.
Related papers
- Do physics-informed neural networks (PINNs) need to be deep? Shallow PINNs using the Levenberg-Marquardt algorithm
- A Neural-preconditioned Poisson Solver for Mixed Dirichlet and Neumann Boundary Conditions
- Advanced Linear Algebra with Applications - Part I (Numerical linear algebra for PDEs, machine learning, and data assimilation)
- Operator learning meets inverse problems: A probabilistic perspective
- Second-order consistency for learning chaotic dynamics via randomized Jacobian matching
- Windowed thinning and query complexity for the bouncy particle and Zigzag samplers