Differential Equations Driven by Variable Order H\"older Noise, and the Regularizing Effect of Delay

TL;DR

This paper extends rough path theory to variable H"older exponent processes, enabling the analysis of delay differential equations driven by such signals with low regularity and introducing a canonical method to construct iterated integrals.

Contribution

It develops a framework for variable order rough paths, including a construction algorithm for iterated integrals and applications to delay differential equations.

Findings

Variable order signals can have H"older regularity close to 0.

Delay differential equations driven by variable order signals are well-posed.

A canonical algorithm for iterated integrals of variable order paths is proposed.

Abstract

In this article we extend the framework of rough paths to processes of variable H\"older exponent or variable order paths. We show how a class of multiple discrete delay differential equations driven by signals of variable order are especially well suited to be studied in a path wise sense. In fact, under some assumptions on the H\"older regularity functions of the driving signal, the local H\"older regularity of the variable order signal may be close to 0, without more than C_{b}^{2} diffusion coefficients (in contrast to higher regularity requirements known from constant order rough path theory). Furthermore, we give a canonical algorithm to construct the iterated integral of variable order on the domain \left[0,T\right]^{2}, by constructing the iterated integrals on some well chosen strip around the diagonal of \left[0,T\right]^{2}, and then extending it to the whole domain using Chens relation. At last, we give a recollection of the results of rough paths found in [FriHai], but tailored for variable order paths.