Modified BDF2 Schemes for Subdiffusion Models with a Singular Source Term

Minghua Chen, Jingkai Shi, Zhi ZhouInternational Journal of Numerical Analysis and Modeling 2025 · DOI

TL;DR

This paper introduces a novel numerical method, IDk-BDF2, to accurately solve subdiffusion equations with "weakly singular" source terms. These types of terms often cause standard numerical methods to lose accuracy, but IDk-BDF2 maintains second-order accuracy even for these challenging cases, opening doors for more reliable simulations of anomalous diffusion phenomena.

What Problem Does This Paper Address?

Anomalous diffusion, where particle movement doesn't follow standard Brownian motion, is often described by subdiffusion equations. These equations frequently involve source terms that are "weakly singular," meaning they become very large or ill-behaved near time zero. Many existing numerical methods, even sophisticated ones like high-order BDF (Backward Differentiation Formulas), struggle with these singular source terms, leading to a loss of their expected accuracy. This paper aims to bridge this gap by developing a new time-stepping scheme that can reliably handle these singular source terms and preserve second-order accuracy.

Methods

The core of the paper is the development of the "IDk-BDF2" (Integral-Derivative k-step BDF2) method. The key innovation lies in how they handle the singular source term. Instead of directly discretizing it, they first "regularize" it by applying an "integral-derivative" operation multiple times (an $m$-fold integral-derivative). This process smooths out the singularity. Then, they discretize the regularized equation using a modified BDF2 convolution quadrature. This combination allows them to achieve second-order accuracy even when the original source term is nonsmooth or incompatible with the initial conditions. The analysis relies heavily on operational calculus and complex analysis techniques, particularly Laplace transforms and contour integration, to derive and prove the convergence properties of the scheme.

Datasets

Not discussed in the paper. The paper focuses on theoretical analysis and numerical verification using simulated data derived from the equations themselves.

Key Insights and Ideas

  • Regularization is Key: The paper highlights that directly discretizing singular source terms is problematic. The novel approach of regularizing the source term before discretization is crucial for maintaining accuracy.
  • BDF2's Potential: While standard BDF2 methods struggle with singularities, a modified BDF2 convolution quadrature, when combined with the regularization technique, proves to be a powerful tool for achieving high-order accuracy.
  • Beyond Smoothness: The work demonstrates that achieving high-order accuracy is possible even when the source term is not smooth in time, a common scenario in many physical applications of subdiffusion.
  • Theoretical Foundation: The rigorous mathematical analysis using Laplace transforms and contour integration provides a strong theoretical backing for the proposed method's accuracy.

Key Results

The paper proves that the proposed IDk-BDF2 time-stepping scheme achieves second-order accuracy for solving subdiffusion equations with weakly singular source terms. This accuracy is maintained even when the source term is nonsmooth in time and not compatible with the initial data. Numerical experiments presented in the paper support these theoretical findings, showing that the IDk-BDF2 method outperforms standard BDF2 and corrected BDF2 methods in terms of accuracy for these challenging cases.

Limitations and Future Work

  • Focus on Time Discretization: The paper primarily focuses on the time-stepping scheme. While it mentions that spatial discretization is well-understood, the detailed analysis of the combined spatio-temporal discretization is not the main focus.
  • Specific Regularization: The paper uses an $m$-fold integral-derivative for regularization. The optimal choice of $m$ might depend on the specific singularity, and further exploration could be beneficial.
  • Extension to Higher Orders: The current work focuses on a second-order scheme. Extending the IDk-BDF approach to higher-order schemes would be a natural next step.
  • Nonlinear Models: The authors suggest adapting the idea to nonlinear fractional models in future studies.
  • Stochastic Subdiffusion: The abstract mentions applicability to stochastic subdiffusion equations, which could be an area for further investigation.

Summary updated

Abstract

The aim of this paper is to study the time stepping scheme for approximately solving the subdiffusion equation with a weakly singular source term. In this case, many popular time stepping schemes, including the correction of high-order BDF methods, may lose their high-order accuracy. To fill in this gap, in this paper, we develop a novel time stepping scheme, where the source term is regularized by using an $m$-fold integral-derivative and the equation is discretized by using a modified BDF2 convolution quadrature. We prove that the proposed time stepping scheme is second-order, even if the source term is nonsmooth in time and incompatible with the initial data. Numerical results are presented to support the theoretical results. The proposed approach is applicable for stochastic subdiffusion equation.
BibTeX
@article{chen2025modified,
  author = {Minghua Chen and Jingkai Shi and Zhi Zhou},
  title = {{Modified BDF2 Schemes for Subdiffusion Models with a Singular Source Term}},
  journal = {International Journal of Numerical Analysis and Modeling},
  year = {2025},
  doi = {10.4208/ijnam2025-1039},
}

Loading related content...

Figure