Approximate convexity and concavity of generalized Grötzsch ring function (Q2505268): Difference between revisions

The authors study generalized Grötzsch ring function defined as \[ \mu_a(r)\equiv{\pi\over 2{\mathrm {sin}}(\pi a)} {F(a,1-a;1;1-r^2)\over F(a,1-a;1;r^2)} \] for \(a \in(0, 1/2], r \in(0,1)\) where \(F(a,1-a;1;r^2)\) is the Gaussian hypergeometric function introduced by \textit{G. D. Anderson, S.-L. Qiu, M. Vamanamurthy}, and \textit{M. Vuorinen} [Pac. J. Math. 192, 1-37 (2000; Zbl 0951.33012)] This function is neither convex nor concave, but the authors obtain the following result. For \(a \in(0, 1/2],\) the function \(f(x)= \mu_a(1-e^{-x})\) is strictly decreasing and convex from \((0, \infty)\) onto \((0, \infty)\). In particular, for all all \(s,t,p \in(0,1),\) \[ \mu_a( 1-(1-s)^p (1-t)^{1-p}) \leq p \mu_a(s) + (1-p)\mu_a( t) \leq \mu_a(s^p t^{1-p} ) \] with equality in each instance if and only if \(s=t \,.\)