A functional relation for the Tornheim double zeta function
A functional relation for the Tornheim double zeta function
复制标题
DOI:
10.4064/aa125-3-3
复制
发表时间:
2006
期刊:
影响因子:
0.7
通讯作者:
Takashi Nakamura
中科院分区:
文献类型:
--
作者:
Takashi Nakamura
The case of t = 0, that is T (s, 0, u), is called the Euler–Zagier double zeta function [10]. The values T (a, b, c) for a, b, c ∈ N were first investigated by Tornheim [7] in 1950 and later Mordell [5] in 1958. Tornheim [7, Theorem 7] showed that T (a, b, c) can be expressed as a polynomial in {ζ(j) | 2 ≤ j ≤ a+ b+ c} with rational coefficients when a + b + c is odd, and that the same is true for T (2r, 2r, 2r) and T (2r− 1, 2r, 2r+1) [7, Theorem 8], but he did not give the coefficients. Mordell [5, Theorem III] proved that T (2r, 2r, 2r) = krπ 6r for some rational number kr. In 1985 Subbarao and Sitaramachandrarao [6, Theorem 4.1] explicitly determined T (2p, 2q, 2r)+T (2q, 2r, 2p)+T (2r, 2p, 2q) (p, q, r ∈ N). Then, by taking p = q = r, they gave an explicit formula for T (2r, 2r, 2r) (r ∈ N) [6, Remark 3.1]. In 1996 Huard, Williams and Zhang [3, Theorems 1–3] determined T (r, 0, N−r) (r ∈ N, N ∈ 2N+1, 1 ≤ r ≤ N−2), T (p, q,N − p − q) (p, q ∈ N ∪ {0}, N ∈ 2N + 1, 1 ≤ p + q ≤ N − 1, 0 ≤ p, q ≤ N − 2) and T (r, r, r) (r ∈ N). In 2002 Tsumura [8, Theorem 1]