Abstract:Xu, Yan and Zhao showed that in even weight, the multiple $T$ value $T(2, 1, \ldots, 1, \overline{1})$ is a polynomial in $\log(2)$, $\pi$, Riemann zeta values, and Dirichlet beta values. Based on low-weight examples, they conjectured that $\log(2)$ does not appear in the evaluation. We show that their conjecture is correct, and in fact follows largely from various earlier results of theirs. More precisely, we derive explicit formulae for $T(2, 1, \ldots, 1, \overline{1})$ in even weight and $S(2, 1, \ldots, 1, \overline{1})$ in odd weight via generating series calculations. We also resolve another conjecture of theirs on the evaluations of $T(\overline{2}, 1, \ldots, 1, \overline{1})$, $S(\overline{2}, 1, \ldots, 1, 1)$, and $S(\overline{2}, 1, \ldots, 1, \overline{1})$ in even weight, by way of calculations involving Goncharov's theory of iterated integrals and multiple polylogarithms.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the evaluation of specific multiple T - values and S - values. Specifically, Xu, Yan and Zhao showed in previous research that under even weights, the multiple T - value \(T(2,1,\ldots,1,1)\) can be expressed as a polynomial form of the logarithm (\(\log(2)\)), the circumference ratio (\(\pi\)), the Riemann zeta function value and the Dirichlet beta function value. Based on examples of low weights, they conjectured that \(\log(2)\) would not appear in these evaluations.
The author Steven Charlton proved that their conjecture was correct and derived explicit formulas to express \(T(2,1,\ldots,1,1)\) under even weights and \(S(2,1,\ldots,1,1)\) under odd weights through the generating series calculation. In addition, he also solved another of their conjectures regarding \(T(2,1,\ldots,1,1)\), \(S(2,1,\ldots,1,1)\) and \(S(2,1,\ldots,1,1(-))\) by using Goncharov's theory of iterated integrals and the polylogarithm function for calculation.
### Main contributions
1. **Verify the conjecture**: Prove that \(\log(2)\) indeed does not appear in the evaluation of \(T(2,\{1\}^{2m - 1},1)\).
2. **Derive formulas**: Give explicit formulas for \(T(2,\{1\}^{2m - 1},1)\) and \(S(2,\{1\}^{2m - 2},1)\).
3. **Expand results**: Solve the evaluation problems of more special values, including \(S(2,\{1\}^{2m})\) and \(S(2,\{1\}^{2m - 1},1)\), and show that these values can be represented by polynomials of \(\log(2)\), \(\pi\), \(\zeta(odd)\), \(\beta(even)\) and some alternating double zeta values.
### Markdown representation of formulas
#### For \(T(2,\{1\}^{2m - 1},1)\)
\[T(2,\{1\}^{2m - 1},1)=- 2(2m + 1)\beta(2m + 2)+\sum_{a + b=m + 1,a,b\geq1}4(1 - 2^{-2a})\zeta(2a)\beta(2b)\]
#### For \(S(2,\{1\}^{2m - 2},1)\)
\[S(2,\{1\}^{2m - 2},1)=-m\left(2^{-2m + 1}-4\right)\zeta(2m + 1)-2(1 - 2^{-2m})\log(2)\zeta(2m)-\pi\beta(2m)+\sum_{a + b=m,a,b\geq1}2(1 - 2^{-2a})\left(4 - 2^{b}-3\cdot2^{-2b + 2}\right)\zeta(2a)\zeta(2b + 1)\]
These formulas not only answer the questions raised by Xu, Yan and Zhao, but also further refine the descriptions they gave in [19, Corollary 4.8].