Probabilistic renormalization and analytic continuation (Q2167497)

In the paper under review, the authors introduce a theory of probabilistic renormalization for series. The renormalized values are encoded by the expectation of a certain random variable on the set of natural numbers. Then, they identify a large class of weakly renormalizable series of Dirichlet type, whose analysis depends on the properties of a difference operator called Bernoulli operator. It is also shown that the probabilistic renormalization is compatible with analytic continuation. The main results of the paper are the following: \begin{itemize} \item If \(s\neq 0\), the series corresponding to \(\mathbf{a}=(a_n)_{n=0}^\infty\) is weakly renormalizable and \[ \mathbb{E}[X_{\mathbf{a}}]=-\frac{1}{s}\left(\mathbf{B}f(s,t_0)-m^s\mathbf{B}f^m\left(s,\frac{t_0-1}{m}+1\right)\right). \] \item The value of \(\mathbb{E}[X_{\mathbf{a}}]\) can be expressed in terms of the analytic continuation of the Dirichlet series associated with \(f(s,t)\): \[ \mathbb{E}[X_{\mathbf{a}}]=\mathbf{D}^f(1-s,t_0)-\mathbf{D}^{\mathbf{S}_mf}\left(1-s,\frac{t_0+m-1}{m}\right). \] \item Let \(s\neq 1\). The series associated to the sequence \(\mathbf{a}\) defined by \(a_0=0\), \({a_n=1/n^s}\), \(n\geq 1\), is strongly renormalizable and \[ \mathbb{E}[X_{\mathbf{a}}]=(1-m^s)\zeta(s), \] where \(\zeta(s)\) is the Riemann zeta function. \item The complex vector space consisting of continuous \(s\)-Kubert functions is two-dimensional and consists of real analytic functions. Actually, it is spanned by \(-s\zeta(1-s,t)\) and \(\ell(s,t)\). \end{itemize}