Approximation of the Lerch zeta-function (Q1881773)

For \(\sigma > 1\), with real parameters \(\lambda\) and \(\alpha\), \(0 < \alpha \leq 1\), the Lerch zeta--function is defined by \[ L(\lambda, \alpha, s) = \sum_{m=0}^\infty {{e^{2\pi i \lambda m}} \over {(m+\alpha)^s}}, \] and can be continued analytically. Improving on an approximation in the monograph by the author and A. Laurinčikas on the Lerch zeta--function, the author proves: For \( -{1\over 2} \leq \lambda \leq {1\over 2}\), \(\lambda \not= 0, \; \sigma \geq 0\), \(| t| \leq \pi \cdot | \lambda| \cdot x\) the following estimate holds: \[ \left| L(\lambda, \alpha, s) - \sum_{0 \leq m \leq x} {{e^{2\pi i \lambda m}} \over {(m+\alpha)^s}} \right| \leq \left( {4 \over {\pi | \lambda| }} + 10.5 \right) \cdot {1 \over {x^\sigma}}. \] This approximation implies upper estimates for \(| L(\lambda, \alpha,s )| \) in \(\sigma \geq 0\). The approximate functional equation for the Lerch zeta--function leads to the estimate (for \(\sigma \geq 0, \; t \geq 1\)) \[ L(\lambda, \alpha, s) - {1\over {\alpha^{{1\over 2} + it}}} - \left({{2\pi}\over t} \right)^{\sigma - {1\over 2} + it} \cdot e^{it + {{\pi i}\over 4} - 2\pi i \lambda \alpha} \cdot \left({1\over {\lambda^{1-s}}} - {{\exp( -\pi t + \pi i \sigma + 2\pi i \alpha)} \over{(1-\{\lambda\})^{1-s}}} \right) \] \qquad \(\displaystyle \ll {1\over {1-\sigma}} \; t^{(1-\sigma)/2}, \text{ if } 0<\sigma \leq 1 - {1\over {\log t}},\) resp. \qquad \(\displaystyle \ll \log t, \text{ if }\sigma \geq 1 - {1\over {\log t}}, \) uniformly in \(\lambda\) and \(\alpha\).