The moment problem on the Wiener space

Frederik S Herzberg

DOI: https://doi.org/10.48550/arXiv.math/0604211

2006-12-05

Abstract:Consider an $L^1$-continuous functional $\ell$ on the vector space of polynomials of Brownian motion at given times, suppose $\ell $ commutes with the quadratic variation in a natural sense, and consider a finite set of polynomials of Brownian motion at rational times, $f_1(\vec b),...,f_m(\vec b)$, mapping the Wiener space to $\mathbb{R}$. In the spirit of Schmüdgen's solution to the finite-dimensional moment problem, we give sufficient conditions under which $\ell$ can be written in the form $\int \cdot d\mu$ for some finite measure $\mu$ on the Wiener space such that $\mu$-almost surely, all the random variables $f_1(\vec b),...,f_m(\vec b)$ are nonnegative.

Probability,Functional Analysis

What problem does this paper attempt to address?

Based on the provided text content, the problems that this paper attempts to solve mainly focus on certain theoretical issues in stochastic analysis and measure theory. Specifically, the paper is concerned with the properties of linear functionals under specific conditions and their relationships with stochastic processes. The following is a more detailed interpretation:

### Problems the paper attempts to solve

1. **Definition and properties of linear functionals**:
- The author defines a linear functional \(\ell\) that acts on a specific space \(\langle\Pi_Q\rangle\) and satisfies some specific conditions. For example, \(\ell(1) = 1\).
- The linear functional \(\ell\) is defined as follows:
\[
\ell: \langle\Pi_Q\rangle \to \mathbb{R}
\]
where \(\langle\Pi_Q\rangle\) is a space generated by certain polynomials.

2. **Representation of stochastic processes**:
- The paper studies how to represent these linear functionals through Brownian motion and other stochastic processes. In particular, it discusses how to construct these linear functionals using the paths of Brownian motion.
- The Brownian motion \(b_t\) and its path space \(\Omega = C_0([0,1], \mathbb{R}) \cap \{f: f(0) = 0\}\) are defined, and the mapping \(b_t: \Omega \to \mathbb{R}\) from the path space to real numbers is considered.

3. **Integral representation**:
- The paper explores how to represent the linear functional \(\ell\) in some integral form. For example, for a certain function \(p\) and measure \(\mu\), we have:
\[
\ell(p(\vec{x})) = E_\gamma[p(c \cdot \vec{\tilde{b}})]
\]
where \(\vec{\tilde{b}}\) is the Brownian motion path after some transformation.

4. **Approximation and convergence**:
- The paper also studies how to approximate continuous stochastic processes by discretization methods and discusses the convergence of these approximation methods. For example, for a certain discretized path \(a \in [-K, K)^d \cap \frac{1}{H!} \mathbb{Z}^d\), we have:
\[
\left| \int_{|x| \leq K} p(x) d\mu - \sum_{a \in [-K, K)^d \cap \frac{1}{H!} \mathbb{Z}^d} p(a) \mu[a, a + (1/H!)^d] \right| < \epsilon
\]

5. **Application of non - standard analysis**:
- The paper introduces the concept of non - standard analysis and uses hyperreal numbers to deal with infinitesimal and infinite quantities. This helps to simplify some complex limit processes and approximation problems.

### Summary

The main purpose of this paper is to explore the properties, representations, and approximation methods of specific linear functionals by defining and studying them, combined with stochastic processes (such as Brownian motion) and non - standard analysis methods. These problems have important theoretical significance in probability theory, stochastic analysis, and measure theory, and may provide new tools and methods in practical applications.