A local Theory of Domains and its (Noncommutative) Symbolic Counterpart

TL;DR

This paper develops a local theory of domains that preserves key algebraic and analytical properties, connecting polylogarithmic calculus with harmonic sums and exploring the properties of generalized Eulerian gamma functions.

Contribution

It introduces a sketched local theory of domains that maintains quasi-shuffle identities, Taylor expansions, and Hadamard products, linking polylogarithmic calculus to harmonic sums.

Findings

Existence of generalized Eulerian gamma functions and their inverses are entire.

Distribution of zeros of these functions is characterized.

Regularization of divergent zeta values using this framework.

Abstract

It is widely accepted nowadays that polyzetas are connected by polynomial relations. One way to obtain relations among polyzetas is to consider their generating series and the relations among these generating series. This leads to the indexation of the generating series of polylogarithms, recently described in \cite{GHM22,BHN,CM}. But, in order to understand the bridge between the extension of this "polylogarithmic calculus" and the world of harmonic sums, a local theory of domains has to be done, preserving quasi-shuffle identities, Taylor expansions and Hadamard products. In this contribution, we present a sketched version of this theory. As an example of generating series, one can consider the eulerian gamma function, \begin{eqnarray*} \Gamma(1+z)=\exp\biggl(-\gamma z+\sum_{n\ge2} \zeta(n)\dfrac{(-z)^n}{n}\biggr) {eqnarray*} and this may suggest to regularize the divergent zeta value $\zeta(1)$, for the quasi-shuffle structure, as to be Euler's $\gamma$ constant. In the same vein, in \cite{BHN}, we introduce a family of eulerian functions, \begin{eqnarray*} \Gamma_{y_k}(1+z)=\exp\biggl(\sum_{n\ge1}\zeta(kn)\dfrac{(-z^k)^n}{n}\biggr), &\mbox{for}&k\ge2,y_k\in Y=\{y_n\}_{n\ge1}. {eqnarray*} This being done, in this work, via their analytical aspects, we establish, on one side, their existence and the fact that their inverses are entire. On the other side, using the same symmetrization technique, we give their distributions of zeroes.