A classical problem in complex analysis consists in finding the distance of a functionfe L” on the unit circle U to H”, the space of functions which extend to a bounded holomorphic function in the unit disk D. It is closely related to some other questions, such as the Pick-Nevanlinna problem of minimizing the supremum norm over the set of bounded holomorphic functions in D, subject to a finite or infinite set of interpolation conditions [S-lo] or the problem of seeking the largest circular domain of a positive harmonic function whose first Taylor coefficients are given [2]. The problem has a remarkable variety of applications, especially in systems engineering. Recent heightening of theoretical interest was brought about by results of Adamyan, Arov, and Krejn [l] on an equivalent problem in operator theory. Nowadays a series of related interpolation and approximation problems can be handled by several alternative mathemati- cal approaches in a unified treatment (problems with matrix-valued functions included, cf. [7, 141, introduction, for instance). Much less is known about a far-reaching generalization of the above problem, which was brought into discussion in the recent paper [6] by J. W. Helton and R. E. Howe (Unfortunately, this paper is not available to me at present, therefore I refer to [S]). The authors study the following optimization problem: Given a function F T x CN + Iw, find inf sup F( t, w(t)), WEE tt~ (1.1) where E= (H” n C)” denotes the space of all continuous cCN-valued func- tions on U with holomorphic continuation into D. Assuming the existence
Read full abstract