Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Stochastic Calculus as Operator Factorization

Ramiro Fontes

TL;DR

The paper proposes that stochastic calculus is fundamentally an operator factorization, revealing a unifying identity $(\mathrm{Id}-\mathbb{E})F=\delta(\varPi DF)$ on an isonormal Gaussian space and extending it to general energy spaces via an operator--covariant derivative $D_X$ defined by Riesz representation. This leads to a unified Clark--Ocone representation $F=\mathbb{E}[F]+\delta_X(\varPi_X D_XF)$ that encompasses Malliavin, Volterra–Malliavin, and functional Itô derivatives, and clarifies the operator geometry behind Itô/change-of-variables formulas. The framework naturally handles mixed Gaussian drivers and diffusion–jump models by working on a direct-sum energy space, yielding a correct evolution identity and a nonlocal generator in the mixed setting. Overall, the work provides a canonical, energy-space–driven view of stochastic calculus with broad implications for analysis of Gaussian and hybrid drivers, and for extensions to fractional, jump, and mild SPDEs.

Abstract

We present a unified operator-theoretic formulation of stochastic calculus based on two principles: fluctuations factor through differentiation, predictable projection, and integration, and the appropriate stochastic derivative is the Hilbert adjoint of the stochastic integral on the energy space of the driving process. On an isonormal Gaussian space we recover the identity (Id - E)F = delta Pi D F, where D is the Malliavin derivative, Pi is predictable projection, and delta is the divergence operator. Motivated by this factorization, we define for a square-integrable process X admitting a closed stochastic integral an operator-covariant derivative on L2(Omega) via Riesz representation. This yields a canonical Clark-Ocone representation that unifies Malliavin, Volterra-Malliavin, and functional Ito derivatives and clarifies the operator geometry underlying stochastic calculus.

Stochastic Calculus as Operator Factorization

TL;DR

The paper proposes that stochastic calculus is fundamentally an operator factorization, revealing a unifying identity on an isonormal Gaussian space and extending it to general energy spaces via an operator--covariant derivative defined by Riesz representation. This leads to a unified Clark--Ocone representation that encompasses Malliavin, Volterra–Malliavin, and functional Itô derivatives, and clarifies the operator geometry behind Itô/change-of-variables formulas. The framework naturally handles mixed Gaussian drivers and diffusion–jump models by working on a direct-sum energy space, yielding a correct evolution identity and a nonlocal generator in the mixed setting. Overall, the work provides a canonical, energy-space–driven view of stochastic calculus with broad implications for analysis of Gaussian and hybrid drivers, and for extensions to fractional, jump, and mild SPDEs.

Abstract

We present a unified operator-theoretic formulation of stochastic calculus based on two principles: fluctuations factor through differentiation, predictable projection, and integration, and the appropriate stochastic derivative is the Hilbert adjoint of the stochastic integral on the energy space of the driving process. On an isonormal Gaussian space we recover the identity (Id - E)F = delta Pi D F, where D is the Malliavin derivative, Pi is predictable projection, and delta is the divergence operator. Motivated by this factorization, we define for a square-integrable process X admitting a closed stochastic integral an operator-covariant derivative on L2(Omega) via Riesz representation. This yields a canonical Clark-Ocone representation that unifies Malliavin, Volterra-Malliavin, and functional Ito derivatives and clarifies the operator geometry underlying stochastic calculus.
Paper Structure (24 sections, 8 theorems, 34 equations)

This paper contains 24 sections, 8 theorems, 34 equations.

Table of Contents

  1. Introduction
  2. Contributions.
  3. Organization.
  4. Gaussian operator factorization
  5. Preliminaries
  6. Main identity
  7. An operator--covariant derivative on $L^2(\Omega)$
  8. Energy space and stochastic integral
  9. Definition by Riesz representation
  10. Unified Clark--Ocone representation
  11. From representation to Itô and functional Itô
  12. Markov functionals
  13. Path-dependent functionals
  14. Canonical examples
  15. Analytical consequences
  16. ...and 9 more sections

Key Result

Theorem 2.1

For every $F\in \mathcal{D}^{1,2}$,

Theorems & Definitions (26)

  • Theorem 2.1: Gaussian operator identity
  • proof
  • Remark 2.2
  • Definition 3.1: Operator--covariant derivative
  • Proposition 3.2: Riesz representation
  • proof
  • Proposition 3.3: Linearity, closedness, adjointness
  • proof
  • Theorem 3.4: Unified Clark--Ocone representation
  • proof
  • ...and 16 more