Construction of pre-and post-Lie algebras for stochastic PDEs

TL;DR

This paper constructs and analyzes pre-Lie and post-Lie algebra structures on decorated rooted trees, with applications to stochastic PDEs, providing new algebraic tools and explicit examples for these complex systems.

Contribution

It introduces a new framework for constructing pre-Lie and post-Lie algebras on decorated rooted trees, including conditions for tree-compatibility and explicit examples relevant to stochastic PDEs.

Findings

Characterization of tree-compatible maps for algebra construction

Explicit isomorphism with classical pre-Lie algebra in noise-free case

Construction of tree-compatible maps as sums in noisy cases

Abstract

We give and study a construction of pre-Lie algebra structures on rooted trees whose edges and vertices are decorated, with a grafting product acting, through a map $\phi$, both on the decoration of the created edge and on the vertex that holds the grafting. We show that this construction gives a pre-Lie algebra if, and only if, the map $\phi$ satisfies a commutation relation, called tree-compatibility. We show how to extend this pre-Lie algebra structure to a post-Lie one by a semi-direct extension with another post-Lie algebra. We also define several constructions to obtain tree-compatible maps, and give examples, including a description of all tree-admissible maps when the space of decorations of the vertices is $2$-dimensional and the space of decorations of the edges is finite-dimensional. A particular example of such a construction is used by Bruned, Hairer and Zambotti for the study of stochastic partial differential equations: when no noise is involved, we show that the underlying tree-compatible map is the exponential of a simpler one and deduce an explicit isomorphism with a classical pre-Lie algebra of rooted trees; when a noise is involved, we obtain the underlying tree-compatible map as a direct sum. We also obtain with our formalism the post-Lie algebras described by Bruned and Katsetsiadis.