The second coefficient of bounded univalent functions which cover a fixed disk (Q1084222)

Denote by S(d), \(0<d<1\), the class of functions \[ f(z)=a_ 1z+a_ 2z^ 2+...,\quad a_ 1>0, \] which are univalent in the unit disk \(\Delta\) and satisfy \(d\Delta\subset f(\Delta)\subset \Delta\). Let \(F_ d\) be the ''Koebe function'' for S(d), that is, \(F_ d\) maps \(\Delta\) onto \(\Delta\setminus (-1,-d].\) In an earlier paper with M. Schiffer the authors considered the problem of maximizing \(| a_ 2|\) in S(d). They proved that \(F_ d\) is extremal if \(d\leq d_ 0=3-2\sqrt{2}\), but not for larger d. The present paper is devoted to a study of the case \(d_ 0<d<1\). Let \(S(d,a_ 1)\) denote the subclass with fixed \(a_ 1\), and let \(F_{d,\theta}\) be the conformal mapping of \(\Delta\) onto \(\Delta \setminus \{(-1,-d]\cup \{z=de^{it}\), \(\pi -\theta \leq t\leq \pi +\theta \}\}\) with \(F_{d,\theta}(0)=0\), \(F'_{d,\theta}(0)=1\). Then for each d and allowable \(a_ 1\) there is a unique \(\theta\) for which \(F_{d,\theta}\in S(d,a_ 1)\). The \(L^ 2\) case of an integral means theorem of Netanyahu and Pinchuk implies that \(F_{d,\theta}\) has the largest second coefficient in \(S(d,a_ 1)\). Thus, to maximize \(a_ 2\) in S(d) it suffices to look at the functions \(F_{d,\theta}\). The authors devise a numerical technique for doing this, and assert that it determines \(\max | a_ 2|\) as a function of d with an accuracy of at least eight decimal places, provided \(d_ 0<d<0.98\).