Conical zeta values and their double subdivision relations (Q2445295)

In the paper under review, the authors introduce the concept of a conical zeta value, which is an infinite sum \[ \zeta(C;s_1,\ldots,s_k)=\sum\limits_{(n_1,\ldots,n_k)\in C\cap \mathbb{Z}_{\ge 0}^k}\frac{1}{n_1^{s_1}\cdots n_k^{s_k}}, \] where \(C\) is an open (polyhedral) cone in \(\mathbb{R}_{\ge 0}^k\) and \(s_1,\ldots,s_k\) are non-negative integers. If \(C\) is the open Chen cone \(\{x_1>\cdots>x_k>0\}\), then the conical zeta value \(\zeta(C;s_1,\ldots,s_k)\) is just the multiple zeta value \[ \zeta(s_1,\ldots,s_k)=\sum\limits_{n_1>\cdots>n_k>0}\frac{1}{n_1^{s_1}\cdots n_k^{s_k}}. \] The main purpose of this paper is to generalize the quasi-shuffle and shuffle relations of multiple zeta values to the family of conical zeta values. Let \(\{C_i\}_i\) be an open subdivision of the cone \(C\), then we get the open subdivision relation \[ \zeta(C;s_1,\ldots,s_k)=\sum\limits_{i}\zeta(C_i;s_1,\ldots,s_k), \] which is a generalization of quasi-shuffle relation of multiple zeta values. On the other hand, the authors call \([v_1]^{s_1}\cdots [v_k]^{s_k}\) a decorated smooth cone for \(\{v_1,\ldots,v_k\}\) a part of a basis of \(\mathbb{Z}^n\), and define an algebraic subdivision of this decorated smooth cone as \[ \sum\limits_{i}\frac{1}{(s_1-1)!\cdots(s_k-1)!}\delta_{v_1^{\ast}}^{s_1-1}\cdots \delta_{v_k^{\ast}}^{s_k-1}\left([v_{i1}]\cdots [v_{ik}]\right), \] where \(\{\langle v_{i1},\ldots,v_{ik}\rangle^c\}_i\) is a smooth subdivision of the closed cone \(\langle v_1,\ldots,v_k\rangle^c\), \(\{v_i^{\ast}\}_i\) is the dual basis to \(\{v_i\}_i\) with respect to the usual inner product of \(\mathbb{R}^n\), and \[ \delta_{e_i^{\ast}}\left([v_1]^{s_1}\cdots [v_k]^{s_k}\right)=\sum\limits_{j}s_j\left(e_i^{\ast},v_j\right)[v_1]^{s_1}\cdots[v_j]^{s_j+1}\cdots[v_k]^{s_k}, \] with \(e_1,\ldots,e_n\) the standard basis of \(\mathbb{R}^n\). If we denote this algebraic subdivision of \(D=[v_1]^{s_1}\cdots [v_k]^{s_k}\) as \(\sum\limits_{i}a_iD_i\), then the authors show the closed subdivision relation \[ \zeta^c(D)=\sum\limits_{i}a_i\zeta^c(D_i), \] and prove that it is a generalization of shuffle relation of multiple zeta values. Here for \(v_i=\sum_{j}a_{ij}e_j\), the linearly constrained zeta value \(\zeta^c\left([v_1]^{s_1}\cdots[v_k]^{s_k}\right)\) is defined as the infinite sum \[ \sum\limits_{m_1=1}^\infty\cdots\sum\limits_{m_n=1}^\infty\frac{1}{(a_{11}m_1+\cdots+a_{1n}m_n)^{s_1} \cdots(a_{k1}m_1+\cdots+a_{kn}m_n)^{s_k}}. \] Note that if \(C\) is the transpose cone of \(\langle v_1,\ldots v_k\rangle^c\), then the authors show that \[ \zeta(C;s_1,\ldots,s_k)=\zeta^c\left([v_1]^{s_1}\cdots[v_k]^{s_k}\right). \] As a consequence, the authors get the double subdivision relation among conical zeta values as a generalization of double shuffle relation of multiple zeta values. The authors also introduce the extended double subdivision relation conjecture for conical zeta values. Another result of the paper is that the authors prove that the conical zeta values and Shintani zeta values span the same linear space over \(\mathbb{Q}\). Here a Shintani zeta value \(\zeta(M;s_1,\ldots,s_k)\) is defined as the infinite sum \[ \sum\limits_{m_1=1}^\infty\cdots\sum\limits_{m_n=1}^{\infty}\frac{1}{(c_{11}m_1+\cdots+c_{1n}m_n)^{s_1}\cdots (c_{k1}m_1+\cdots+c_{kn}m_n)^{s_k}} \] with the matrix \(M=(c_{ij})\in M_{k\times n}(\mathbb{Z}_{\geqslant 0})\).