Functional it{\^o} versus banach space stochastic calculus and strict solutions of semilinear path-dependent equations

TL;DR

This paper compares functional Itô calculus with Banach space stochastic calculus, reformulates the former for better clarity, explores their relationship, and investigates solutions to path-dependent PDEs related to stochastic differential equations.

Contribution

It introduces a reformulation of functional Itô calculus that separates time and past, and analyzes the relation with Banach space approaches, along with studying solutions to related path-dependent PDEs.

Findings

Reformulation of functional Itô calculus enhances understanding.

Clarification of the relation between functional and Banach space approaches.

Existence and uniqueness results for smooth solutions to path-dependent PDEs.

Abstract

Functional It\^o calculus was introduced in order to expand a functional $F(t, X\_{\cdot+t}, X\_t)$ depending on time $t$, past and present values of the process $X$. Another possibility to expand $F(t, X\_{\cdot+t}, X\_t)$ consists in considering the path $X\_{\cdot+t}=\{X\_{x+t},\,x\in[-T,0]\}$ as an element of the Banach space of continuous functions on $C([-T,0])$ and to use Banach space stochastic calculus. The aim of this paper is threefold. 1) To reformulate functional It\^o calculus, separating time and past, making use of the regularization procedures which matches more naturally the notion of horizontal derivative which is one of the tools of that calculus. 2) To exploit this reformulation in order to discuss the (not obvious) relation between the functional and the Banach space approaches. 3) To study existence and uniqueness of smooth solutions to path-dependent partial differential equations which naturally arise in the study of functional It\^o calculus. More precisely, we study a path-dependent equation of Kolmogorov type which is related to the window process of the solution to an It\^o stochastic differential equation with path-dependent coefficients. We also study a semilinear version of that equation.