Abstract:Utilising classical results on the structure of Hopf algebras, we develop a novel approach for the construction of cubature formulae on Wiener space based on unshuffle expansions. We demonstrate the effectiveness of this approach by constructing the first explicit degree-7 cubature formula on $d$-dimensional Wiener space with drift, in the sense of Lyons and Victoir. The support of our degree-7 formula is significantly smaller than that of currently implemented or proposed constructions.
What problem does this paper attempt to address?
### What problem does this paper attempt to solve?
This paper aims to solve the problem of constructing efficient, high - order Cubature formulas on the Wiener space. Specifically, the author proposes a new method based on unshuffle expansions for constructing Cubature formulas on the Wiener space. This method utilizes the Hopf algebra structure, especially the Eulerian idempotents of the tensor algebra, to simplify and sparsify the moment problem, thus making the construction of high - order Cubature formulas more feasible.
#### Background and Motivation
1. **Basic Concepts of Cubature**:
- Cubature is a classical numerical integration method that precisely matches the integrals of polynomial test functions in a finite - dimensional space by replacing the target measure with a discrete measure.
- In the Wiener space, the Cubature method is extended to the path space, using the stochastic Taylor expansion instead of the traditional Taylor expansion, and using the Stratonovich iterated integral of the Brownian motion instead of the polynomial.
2. **Limitations of Existing Methods**:
- Existing Cubature methods face significant challenges when dealing with high - dimensional noise, especially when the order is high, and the algebraic and computational complexity increases sharply.
- For example, Lyons and Victoir [LV04] proposed third - order and fifth - order Cubature formulas, but these methods are difficult to directly generalize to higher dimensions or higher orders.
3. **Advantages of the New Method**:
- The method based on unshuffle expansions proposed by the author can significantly simplify the moment problem, making it possible to explicitly construct seventh - order Cubature formulas in any dimension.
- This method not only improves efficiency but also reduces the size of the support set, thus being more efficient in practical applications.
#### Specific Problems
- **High - order Cubature Formulas**: How to explicitly construct efficient high - order (such as seventh - order) Cubature formulas on the Wiener space?
- **Multi - dimensional Noise**: How to generalize existing low - dimensional Cubature formulas to higher - dimensional noise scenarios?
- **Computational Complexity**: How to reduce the algebraic and computational complexity in the construction of high - order Cubature formulas?
### Solutions
The author proposes a new method based on unshuffle expansions. The specific steps are as follows:
1. **Expansion of the Expected Signature**: Use the Eulerian idempotents to expand the expected signature of the Brownian motion (with drift).
2. **Simplification of the Moment Problem**: Represent the unknown \( L(\mathbb{R}^{1 + d}) \)-valued random variable \( L \) as a linear combination of the elements appearing in the above expansion, and obtain the constraint conditions of the joint moments through expansion and matching conditions.
3. **Construction of Discrete Random Variables**: Introduce low - order independent auxiliary Gaussian Cubature, construct discrete random variables to satisfy the moment conditions, and thus obtain the Cubature formula on the Wiener space.
This method not only solves the complexity problems of existing methods in high - dimensional and high - order cases but also provides a smaller and more effective support set, making high - order Cubature formulas more feasible in practical applications.
### Summary
The main contribution of this paper is to provide a new method based on unshuffle expansions for explicitly constructing efficient high - order Cubature formulas on the Wiener space. This method not only simplifies the moment problem and reduces the computational complexity but also enables high - order Cubature formulas to be explicitly constructed in any dimension, thus providing new tools for high - precision numerical integration and weak solutions of stochastic differential equations.