Multiple polylogarithms, a regularisation process and an admissible open domain of convergence

TL;DR

This paper explores the analytic properties of multiple polylogarithms, expanding their domain of convergence through translation formulas and introducing a regularisation process to study their behavior at integer points.

Contribution

It introduces a new regularisation method for multiple polylogarithms and extends their domain of convergence using novel translation formulas.

Findings

Larger open domain of convergence for multiple polylogarithms.

A new regularisation process extending the series to integer points.

Derivation of a generalized Euler-Boole summation formula.

Abstract

In this article, we study the analytic properties of the multiple polylogarithms in the $s$-aspect. Although the domain of absolute convergence of the series defining the multiple polylogarithms is well-known, the study towards a larger open domain of (conditional) convergence has been limited, particularly when the depth is $\ge 2$. Here, we exhibit a larger open domain of (conditional) convergence for this series by writing certain translation formulas satisfied by them. The series moreover defines a holomorphic function in this open set. We then introduce a regularisation process for the multiple polylogarithms, extending an earlier work of the second author. This regularisation process requires a generalisation of the Euler-Boole summation formula that we derive in the appendix of this article. The regularisation process leads to a larger open domain, where the series (conditionally) converges at integer points. The holomorphicity at such points is a more delicate question and this regularisation process is to be used to study the local behaviour of the multiple polylogarithms around such points.