Abstract:The numerical analysis of stochastic time fractional evolution equations presents considerable challenges due to the limited regularity of the model caused by the nonlocal operator and the presence of noise.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is how to increase the convergence order of the time - discretization method for stochastic fractional evolution equations. Specifically, the article proposes new high - order numerical methods for linear stochastic fractional evolution equations driven by integral - additive noise.
### Main Problem Description
1. **Limitations of Existing Methods**:
- Due to the non - local operator of the model and the presence of noise, limited regularity is caused.
- Existing time - stepping methods exhibit a significantly lower convergence order, usually \(O(\tau^{\min\{ \alpha + \gamma - \frac{1}{2}, 1 \}})\).
2. **Research Objectives**:
- Propose a new high - order time - discretization scheme to overcome the problem of low convergence order.
- By introducing smoothing techniques, develop a novel high - order BDF convolution quadrature method (ID m - BDFk method) capable of handling integral - additive noise.
### Solution
The author proposes a method that combines m - fold integral - differential calculus and smoothing techniques, specifically including:
- **Smoothing Technique**: Regularize the noise by performing m - fold integral - differential calculus on the additive noise.
- **Time - Discretization**: Use the k - step BDF convolution quadrature method to discretize the evolution equation.
- **New Method**: Called the ID m - BDFk method, where ID stands for integral - differential, m represents the integration order, and BDFk represents the k - step backward difference formula.
### Key Results
For different parameter ranges, this method can achieve a higher convergence order:
- For \(1 < \alpha + \gamma \leq \frac{5}{2}\), the convergence order of the ID2 - BDF2 method is \(O(\tau^{\alpha + \gamma - \frac{1}{2}})\).
- For \(\frac{5}{2} < \alpha + \gamma < 3\), the convergence order of the ID2 - BDF2 method is \(O(\tau^2)\).
- When using the ID3 - BDF3 method, for any \(\alpha \in (1, 2)\) and \(\gamma \in (0, 1)\), the convergence order can be further increased to \(O(\tau^{\alpha + \gamma - \frac{1}{2}})\).
### Mathematical Expressions
The key mathematical expressions involved in the article include:
- Riemann - Liouville fractional - order derivatives and integrals:
\[
\partial_t^\alpha u(t)=\frac{1}{\Gamma(2 - \alpha)}\frac{d^2}{dt^2}\int_0^t(t - s)^{1 - \alpha}u(s)ds
\]
\[
\partial_t^{-\gamma}u(t)=\frac{1}{\Gamma(\gamma)}\int_0^t(t - s)^{\gamma - 1}u(s)ds
\]
- Expression of additive noise:
\[
dW(t)=\sum_{j = 1}^\infty\sigma_j(t)\dot{\beta}_j(t)\phi_j
\]
- Key formula in the ID m - BDFk method:
\[
\partial_t^{-\gamma}\frac{dW(t)}{dt}=\partial_t^{m - \gamma}\left(\frac{t^{m - 1}}{\Gamma(m)}*\frac{dW(t)}{dt}\right)
\]
These formulas ensure the regularization treatment of noise and provide a theoretical basis for high - order time - discretization.