On the Analysis of a Singular Stochastic Volterra Differential Equation driven by a Wiener Noise

Basic Information

• Paper ID: 2501.01373
• Title: On the Analysis of a Singular Stochastic Volterra Differential Equation driven by a Wiener Noise
• Authors: Emmanuel Coffie (University of Liverpool), Olivier Menoukeu-Pamen (University of Liverpool), Frank Proske (University of Oslo)
• Classification: math.PR (Probability Theory)
• Publication Date: January 2, 2025
• Paper Link: https://arxiv.org/abs/2501.01373

Abstract

This paper constructs a unique strong solution for a class of stochastic Volterra differential equations (SVDE) driven by singular drift vector fields and Wiener noise. Furthermore, the Sobolev differentiability of the strong solution with respect to initial values is investigated.

Research Background and Motivation

Problem Background

Stochastic Volterra differential equations (SVDE), originally studied by Berger and Mizel, have been successfully applied to model the dynamics of various phenomena, including:

• Population growth
• Epidemic propagation
• (Rough) stock price volatility in finance

These equations are particularly effective in representing dynamical systems exhibiting non-Markovian behavior.

Core Problem

This paper studies stochastic Volterra differential equations of the following form:

$dX^x_t = x + \int_0^t b(t,s,X^x_s)ds + B_t, \quad 0 \leq t \leq T, \quad x \in \mathbb{R}^d$

where $b: [0,T] \times [0,T] \times \mathbb{R}^d \to \mathbb{R}^d$ is a Borel measurable function and $B_t$ is a Brownian motion.

Research Challenges

The primary challenge lies in the fact that the drift vector field $b$ is spatially singular, meaning:

• It does not satisfy Lipschitz continuity
• It may have discontinuities with respect to the variable $x$ (and $s$)

Research Motivation

In the existing literature:

• The case of non-singular kernels has been studied (e.g., 1, 20)
• The case of singular kernels with spatially Lipschitz continuous vector fields has been investigated (e.g., 10, 11)
• However, the case of singular drift vector fields lacks in-depth investigation

Core Contributions

1. Construction of Unique Strong Solution: For singular drift vector fields with specific structure, the existence and uniqueness of strong solutions to SVDE is established
2. Malliavin Differentiability: The strong solution is shown to possess Malliavin differentiability
3. Sobolev Differentiability: Local Sobolev differentiability of the strong solution with respect to initial values is established
4. Novel Analytical Methods: Based on Malliavin calculus techniques, the methods for handling singular stochastic differential equations are extended

Detailed Methodology

Problem Formulation

The study of the existence, uniqueness, and differentiability properties of strong solutions to SVDE (1) when the drift vector field $b$ belongs to a specific function class $\mathcal{H}$.

Vector Field Structure

The drift vector field $b$ admits the following representation:

$b(t,s,x) = \sum_{m \geq 0} (t-s)^m g_m(s,x), \quad 0 \leq s \leq t \leq T, \quad x \in \mathbb{R}^d$

where $g_m \in L^{\infty}([0,T] \times \mathbb{R}^d; \mathbb{R}^d)$ and satisfies:

$\sum_{m \geq 0} T^m \|g_m\|_{\infty} < \infty$

Core Theoretical Results

Theorem 1 (Main Result): Assume $b \in \mathcal{H}$. Then for sufficiently small time interval $T$, SVDE (1) admits a unique strong solution $X^x_t$ for all initial values $x$. Furthermore, $X^x_t$ is Malliavin differentiable, and the associated stochastic flow $X^{\cdot}_t$ belongs to the space $L^2(\Omega; W^{1,2}_{loc}(\mathbb{R}^d))$.

Technical Methods

1. Malliavin Calculus Techniques

• Employ the Malliavin derivative $D_u$ to analyze the regularity of solutions
• Establish estimates for Malliavin derivatives

2. Picard Iteration

Construct solutions through Picard iteration: $D_u X^x_t = \sum_{n \geq 0} \left\{\int_u^t \int_{t_{n-1}}^u \cdots \int_{t_2}^u \prod_{j=2}^n Db(t_j, t_{j-1}, X^x_{t_{j-1}}) dt_1 \cdots dt_n\right\}$

3. Girsanov Transformation

Employ the Girsanov theorem to establish measure transformations, converting the problem into equations driven by standard Brownian motions.

4. Weak to Strong Solution Conversion

Through approximating sequences and weak convergence techniques, convert the existence of weak solutions into the existence of strong solutions.

Technical Innovations

1. Singularity Handling: Innovatively addresses spatially singular drift vector fields
2. Power Series Representation: Exploits the power series expansion structure of $(t-s)^m$
3. Malliavin Methods: Extends Malliavin calculus techniques to Volterra-type equations
4. Uniform Estimates: Establishes critical uniform estimates to control singularities

Theoretical Analysis

Key Lemmas

Lemma 3: Establishes boundedness of Malliavin derivatives: $\sup_{0 \leq u \leq T} \|D_u X^x_t\|_{L^2(\Omega)} \leq C_{d,T}(\sup_{\mu \geq 0} \|b_{\mu}\|_{\infty})$

and Hölder continuity: $\|D_r X^x_s - D_v X^x_s\|_{L^2(\Omega)} \leq C_{d,T}(\sup_{\mu \geq 0} \|b_{\mu}\|_{\infty}) |r-v|^{\alpha}$

Weak Convergence Results

Lemma 4: Establishes weak convergence of approximating solutions, proving that for bounded continuous functions $\varphi$: $\varphi(X^{x,n}_t) \to E[\varphi(X^x_t)|{\mathcal F}_t] \text{ weakly in } L^2(\Omega; {\mathcal F}_t)$

Weak Solution Existence

Lemma 6: Under appropriate integrability conditions, establishes the existence and uniqueness of weak solutions.

Application Examples

Example 7

Consider the SVDE: $X^x_t = x + \int_0^t \sin(t-s)g(s,X^x_s)ds + B_t, \quad 0 \leq t \leq T$

where $g \in L^{\infty}([0,T] \times \mathbb{R}^d; \mathbb{R}^d)$.

Since $\sin(t-s) = \sum_{n \geq 0} \frac{(-1)^n}{(2n+1)!}(t-s)^{2n+1}$, the drift vector field belongs to class $\mathcal{H}$. Therefore, by Theorem 1, this SVDE admits a unique strong solution.

Extended Applications

Remark 8: The method can also be applied to study SVDE of the form: $X^x_t = x + \int_0^t (t-s)^{\alpha} g(s,X^x_s)ds + B_t$

where $-\frac{1}{2} < \alpha < 0$, by handling the singular kernel through binomial expansion.

Related Work

Historical Development

• Berger & Mizel (5, 6): First studied stochastic Volterra equations
• Non-singular Case: 1, 20 obtained solutions via the Yamada-Watanabe theorem
• Singular Kernels: 10, 11 studied singular kernels with Lipschitz continuous vector fields
• PDE Methods: 19, 21 established pathwise uniqueness based on PDE techniques

Innovation of This Work

Compared to existing work, this paper is the first to systematically address the case of singular drift vector fields, filling an important gap in the field.

Conclusions and Discussion

Main Conclusions

1. Successfully constructs unique strong solutions for singular SVDE
2. Establishes Malliavin differentiability and Sobolev differentiability of solutions
3. Provides a new methodological framework for handling singular stochastic Volterra equations

Limitations

1. Time Interval Restriction: Requires sufficiently small time interval $T$
2. Vector Field Structure: Restricted to specific power series representation forms
3. Convergence Conditions: Requires convergence of the series $\sum_{m \geq 0} T^m \|g_m\|_{\infty} < \infty$

Future Directions

1. Extension to more general singular kernel structures
2. Investigation of long-time behavior
3. Development of numerical methods
4. Applications in finance and biomathematics

In-Depth Evaluation

Strengths

1. Theoretical Innovation: First systematic treatment of Volterra equations with singular drift vector fields
2. Advanced Methods: Skillfully combines Malliavin calculus and Girsanov transformation
3. Complete Results: Not only establishes existence and uniqueness but also differentiability properties
4. Rigorous Techniques: Rigorous proofs with precise estimates

Weaknesses

1. Limited Applicability: Strong restrictions on time interval and vector field structure
2. Practical Applications: Lacks concrete numerical experiments and application cases
3. Comparative Analysis: Insufficient comparison with existing methods

Impact

1. Theoretical Contribution: Provides new tools for the theory of singular stochastic differential equations
2. Application Potential: Important application prospects in rough volatility modeling and related fields
3. Methodological Value: The extension of Malliavin methods has universal significance

Applicable Scenarios

• Rough volatility models in financial mathematics
• Stochastic dynamical systems with memory effects
• Modeling of non-Markovian stochastic processes

References

The paper includes 21 major references covering important works in stochastic Volterra equations, Malliavin calculus, singular stochastic differential equations, and related fields, providing a solid theoretical foundation for the research.