Sharp bounds for Neuman-Sándor's mean in terms of the root-mean-square (Q484567)

The Neuman-Sándor mean \(M(a,b)\) and the root-mean-square \(S(a,b)\) are defined respectively by \[ M(a,b)=\frac{a-b}{\sinh ^{-1}\left( \frac{a-b}{a+b}\right)}\text{ and }S(a,b)=\sqrt{\frac{a^{2}+b^{2}}{2}}. \] The result of this paper is given by the following theorem: the double inequality \[ S(\alpha a+(1-\alpha)b,\alpha b+(1-\alpha)a)<M(a,b)<S(\beta a+(1-\beta)b,\beta b+(1-\beta)a) \] holds for all \(a,b>0\) with \(a\neq b\) if and only if \[ \frac{1}{2}\leq \alpha \leq \frac{1}{2}\left\{ 1+\sqrt{\frac{1}{\left[ \ln\left( 1+\sqrt{2}\right) \right] ^{2}}-1}\right\} =0.76\ldots \] and \[ 1\geq \beta \geq \frac{3+\sqrt{3}}{6}=0.78\ldots \text{.} \]