Zeta-like multiple zeta values in positive characteristic (Q2137881)

Let \(A=\mathbb{F}_q[\theta]\) be the polynomial ring over \(\mathbb{F}_q\) and let \(K=\mathbb{F}_q(\theta)\) be the fraction field of \(A\). Then, the multiple zeta values \(\zeta_A(s)\) over \(A\), introduced by D. Thakur as analogs of multiple zeta values of Euler, were defined for any \(r\)-tuple \(s=(s_1,\dots,s_r)\) of positive integers by \[ \zeta_A(s)=\sum\prod_{i=1}^r\frac{1}{a_i^{s_i}}, \] where the sum is over \(r\)-tuples \((a_1,\dots,a_r)\) of monic polynomials in \(A\) with \(\deg a_1>\cdots>\deg a_r\). A multiple zeta value \(\zeta_A(s_1,\dots,s_r)\) is called zeta-like if the ratio \[ \frac{\zeta_A(s_1,\dots,s_r)}{\zeta_A(s_1+\cdots+s_r)}\in K. \] In the paper under review, the authors provide a complete list of zeta-like multiple zeta values of weight at most \(q^2\) and depth \(2\). As consequence, this result solves a conjecture proposed by \textit{J. A. Lara Rodríguez} and \textit{D. Thakur} [Indian J. Pure Appl. Math. 45, No. 5, 787--801 (2014; Zbl 1365.11104)]. Furthermore, they determine all zeta-like multiple zeta values of weight at most \(q^2\) and arbitrary depth. More precisely, they prove that all zeta-like multiple zeta values of weight at most \(q^2\) are \(\zeta_A(i,j(q-1))\) with \(1\le i\le q\) and \(i\le j\le \left[\frac{q^2-i}{q-1}\right]\), and \(\zeta_A(1,q-1,q^2-q)\). Therefore, there are no zeta-like multiple zeta values of weight at most \(q^2\) and depth at least \(4\).