Weak convergence rates for temporal numerical approximations of stochastic wave equations with multiplicative noise

Sonja Cox, Arnulf Jentzen, Felix Lindner

TL;DR

This paper introduces a novel technique to achieve optimal weak convergence rates for numerical approximations of stochastic wave equations with multiplicative noise. This is a significant advancement because existing methods struggle with this type of noise, and the derived rates are twice as fast as previously known strong convergence rates.

What Problem Does This Paper Address?

Stochastic partial differential equations (SPDEs) are crucial for modeling complex systems influenced by random forces, such as the motion of DNA or heat flow. A key challenge in working with SPDEs is accurately approximating their solutions numerically. While strong convergence rates (measuring the average squared error) are well-understood for many SPDEs, achieving optimal weak convergence rates (measuring the error in expectation of a smooth function of the solution) is considerably harder, especially for SPDEs with multiplicative noise. This paper specifically tackles the problem of temporal discretization for hyperbolic SPDEs with multiplicative noise, a problem for which existing analytical tools are not directly applicable.

Methods

The authors develop a new technique to establish sharp weak convergence rates for temporal discretizations of hyperbolic SPDEs with multiplicative noise. The core of their approach involves:

  • Sophisticated Error Splitting: The overall error is decomposed into several terms, each requiring specific analytical treatment.
  • Mild Itô Formula: A key ingredient is the application of a recently introduced "mild Itô formula." Unlike the classical Itô formula, this version allows for the avoidance of unbounded operators, which is crucial for dealing with the roughing effects introduced by multiplicative noise. This formula is used to rewrite error terms as integrals over short time intervals, enabling better control.
  • Focus on Hyperbolic Anderson Model: The paper specifically analyzes the hyperbolic Anderson model, a canonical example of a hyperbolic SPDE with multiplicative noise.

Datasets

Not discussed in the paper. The paper focuses on theoretical analysis and numerical simulations to validate the theoretical findings.

Key Insights and Ideas

  • Twice the Strong Rate: A major finding is that the weak convergence rates obtained are twice the known strong convergence rates for similar problems. This highlights the efficiency of weak convergence analysis when applicable.
  • Novelty for Multiplicative Noise: The paper claims to be the first to provide essentially sharp weak convergence rates for temporal discretizations of hyperbolic SPDEs with multiplicative noise, overcoming a significant hurdle in the field.
  • Mild Itô Formula's Power: The mild Itô formula proves to be a powerful tool, enabling the analysis of multiplicative noise by circumventing issues with unbounded operators that plague classical Itô calculus in this context.
  • Generalizability: The authors suggest that their approach can be extended to other temporal discretizations, spatial discretization methods (like finite elements), and even to Banach space settings, indicating broad applicability.

Key Results

The paper establishes theoretical upper bounds for weak errors in temporal discretizations of hyperbolic SPDEs with multiplicative noise. A key result, Theorem 1.1, demonstrates a weak convergence rate of $1 - \epsilon$ for any $\epsilon > 0$. This rate is shown to be twice the known strong convergence rates. The analytical findings are complemented by numerical simulations in Python, which illustrate the decay of the weak approximation error for the hyperbolic Anderson model using four different test functions, supporting the theoretical predictions.

Limitations and Future Work

The authors acknowledge that their approach is currently focused on a specific class of SPDEs (hyperbolic wave equations with multiplicative noise) and a specific discretization scheme. They suggest future work could involve extending their methodology to:

  • Other temporal discretization schemes.
  • Spatial discretization methods, such as finite element methods.
  • Banach space settings, allowing for more general semilinear drift and diffusion coefficients.
  • More general classes of nonlinear Nemytskii operators.

Summary updated

Abstract

In numerical analysis for stochastic partial differential equations one distinguishes between weak and strong convergence rates. Often the weak convergence rate is twice the strong convergence rate. However, there is no standard way to prove this: to obtain optimal weak convergence rates for stochastic partial differential equations requires specially tailored techniques, especially if the noise is multiplicative. In this work we establish weak convergence rates for temporal discretisations of stochastic wave equations with multiplicative noise, in particular, for the hyperbolic Anderson model. The weak convergence rates we obtain are indeed twice the known strong rates. To the best of our knowledge, our findings are the first in the scientific literature which provide essentially sharp weak convergence rates for temporal discretisations of stochastic wave equations with multiplicative noise. Key ideas of our proof are a sophisticated splitting of the error and applications of the recently introduced mild Ito formula.
BibTeX
@inproceedings{cox2019weak,
  author = {Sonja Cox and Arnulf Jentzen and Felix Lindner},
  title = {{Weak convergence rates for temporal numerical approximations of stochastic wave equations with multiplicative noise}},
  booktitle = {SAM},
  year = {2019},
}

Loading related content...

Figure