Introduction to the theory of Toeplitz operators with infinite index. Transl. from the Russian by Andrei Iacob (Q1856387)

Let \(H^p(\mathbb{R})\) and \(H^p_-(\mathbb{R})\) be the Hardy subspaces in \(L^p(\mathbb{R})\) of all functions that have an analytic extension into the upper and lower half-plane, respectively, and for \(a\in L^\infty(\mathbb{R})\) let \(T(a)\) be the Toeplitz operator \(f\mapsto P(a f)\) on \(H^p(\mathbb{R})\), where \(Pf(x)=\frac 12 f(x) + \frac 1{2\pi i}\text{(p.v.)}\int_{\mathbb{R}} f(\tau) (\tau-x)^{-1} d\tau\) is the Riesz projection of \(L^p(\mathbb{R})\) onto \(H^p(\mathbb{R})\). If~the function \(a\) admits a factorization of the form \(a=a^- B a^+\), with \(a^+\in H^\infty(\mathbb{R})\) and \(a^-\in H^\infty_- (\mathbb{R})\) bounded away from zero and \(B\) an infinite Blaschke product, then \(T(a)\) is a left-invertible semi-Fredholm operator, whose left inverse is \(T(\overline B/a^+)T(1/a^-)\) and whose range can also be described fairly explicitly. This is the simplest example of the phenomena which are treated systematically, and in far more general contexts, in~the present book. More specifically, the~authors consider the subspaces \(L^p_+(\Gamma,\rho)\) and \(L^p_-(\Gamma,\rho)\) of all functions in \(L^p(\Gamma,\rho)\) that extend analytically to the interior and exterior of~\(\Gamma\), respectively, where \(\Gamma\) is a closed Jordan curve and \(\rho\) a weight on \(\Gamma\) of a certain type. (Thus taking \(\rho=1\) and replacing \(\Gamma\) by~\(\mathbb{R}\) and its interior and exterior by the upper and the lower half-plane, respectively, we~recover the above Hardy spaces \(H^p(\mathbb{R})\) and \(H^p_-(\mathbb{R})\).) One~can then again define Toeplitz operators \(T(a)\), \(a\in L^\infty(\Gamma)\), in the same way as above. If the function \(a\) has a factorization \(a=bh\) or \(a=bh^{-1}\), where \(b\) is ``nice'' in some sense while \(h\in L^\infty_+(\Gamma)\) and \(h^{-1}\in L^\infty(\Gamma)\), then the description of the kernel, the cokernel, and the index of \(T(a)\) can be reduced to the analogous problem for the ``model operator'' \(T(h^{-1})\). Further, under suitable hypothesis the latter turns out to be independent of the curve \(\Gamma\), thus making it possible to restrict attention to \(\Gamma=\mathbb{R}\). The exposition then continues by an extensive discussion of the most frequent types of functions and model operators for which the above theory applies, namely, the almost periodic, semi almost periodic, and \(u\)-periodic functions (that is, the compositions of an almost periodic function with an inner function), and functions with oscillatory discontinuities of ``whirl'' types (i.e., points \(t_0\) where the argument of \(a(t)\) blows up at a logarithmic, power, exponential, etc. rate as \(t\to t_0\)). In this way, the authors are able to cover a remarkable range of discontinuous symbols. Finally, the book concludes by examining the case when the symbol is no longer assumed to be invertible in \(L^\infty\), so that the corresponding Toeplitz operator need not have closed range (hence is not semi-Fredholm); here the main idea is to modify the norm in the source or the target space so as to make the operator normally solvable (the ``renormalization technique'' in the authors' terminology). Though some parts of the exposition are rather technical, the authors have managed to keep the required prerequisites to a minimum, and thus the book can safely be recommended to anyone past the level of a beginning graduate student who is interested in the theory of singular integral equations or Toeplitz operators.