schur class

In complex analysis, the Schur class is the set of holomorphic functions f(z) defined on the open unit disk \mathbb{D} = \{ z\in \mathbb{C} : |z|<1\} and satisfying |f(z)| \leq 1 that solve the Schur problem: Given complex numbers c_0,c_1,\dotsc,c_n, find a function

:f(z) = \sum_{j=0}^{n} c_j z^j + \sum_{j=n+1}^{\infty}f_j z^j

which is analytic and bounded by {{math|1}} on the unit disk.{{citation|last=Schur|first=J. |title=Über die Potenzreihen, die im Innern des Einheitkreises beschränkten sind. I, II |journal = Journal für die reine und angewandte Mathematik |series=Operator Theory: Advances and Applications |volume = 147|year = 1918 |pages = 205–232, I. Schur Methods in Operator Theory and Signal Processing in: Operator Theory: Advances and Applications, vol. 18, Birkhäuser, Basel, 1986 (English translation) |isbn=978-3-0348-5484-9| doi=10.1007/978-3-0348-5483-2}} The method of solving this problem as well as similar problems (e.g. solving Toeplitz systems and Nevanlinna-Pick interpolation) is known as the Schur algorithm (also called Coefficient stripping or Layer stripping). One of the algorithm's most important properties is that it generates {{math|n + 1}} orthogonal polynomials which can be used as orthonormal basis functions to expand any {{mvar|n}}th-order polynomial.{{cite book | last1=Chung | first1=Jin-Gyun | last2=Parhi | first2=Keshab K. | series=The Kluwer International Series in Engineering and Computer Science | title=Pipelined Lattice and Wave Digital Recursive Filters | publisher=Springer US | publication-place=Boston, MA | year=1996 | isbn=978-1-4612-8560-1 | issn=0893-3405 | doi=10.1007/978-1-4613-1307-6 | page=79}} It is closely related to the Levinson algorithm though Schur algorithm is numerically more stable and better suited to parallel processing.{{cite book | last=Hayes | first=Monson H.|title= Statistical digital signal processing and modeling | url=https://www.worldcat.org/oclc/34243409 | page=242| year = 1996 | publisher = John Wiley & Son | isbn =978-0-471-59431-4| oclc=34243409}}

Schur function

Consider the Carathéodory function of a unique probability measure d\mu on the unit circle \mathbb{T} =\{z\in\mathbb{C} :|z|=1\} given by

: F(z) = \int \frac{e^{i\theta} + z}{e^{i\theta} - z} d\mu(\theta)

where \int d\mu(\theta) = 1 implies F(0)=1.{{Citation | last1=Simon | first1=Barry | author1-link=Barry Simon | title=Orthogonal polynomials on the unit circle. Part 1. Classical theory | url=https://books.google.com/books?id=d94r7kOSnKcC | publisher=American Mathematical Society | location=Providence, R.I. | series=American Mathematical Society Colloquium Publications | isbn=978-0-8218-3446-6 | mr=2105088 | year=2005 | volume=54}} Then the association

: F(z) = \frac{1+zf(z)}{1-zf(z)}

sets up a one-to-one correspondence between Carathéodory functions and Schur functions f(z) given by the inverse formula:

: f(z) = z^{-1}\left( \frac{F(z)-1}{F(z)+1} \right)

Schur algorithm

Schur's algorithm is an iterative construction based on Möbius transformations that maps one Schur function to another.{{Cite book |last=Conway |first=John B. |title=Functions of One Complex Variable I (Graduate Texts in Mathematics 11) |title-link=Graduate Texts in Mathematics |publisher=Springer-Verlag |year=1978 |isbn=978-0-387-90328-6|page=127}} The algorithm defines an infinite sequence of Schur functions f\equiv f_0,f_1,\dotsc,f_n,\dotsc and Schur parameters \gamma_0,\gamma_1,\dotsc,\gamma_n,\dotsc (also called Verblunsky coefficient or reflection coefficient) via the recursion:{{Citation | last =Simon| first =Barry | author-link=Barry Simon | title=Szegő's theorem and its descendants: spectral theory for L² perturbations of orthogonal polynomials| url=https://books.google.com/books?id=e9R5gcz_x0YC | date=2010|publisher=Princeton University Press|isbn=978-0-691-14704-8}}

:f_{j+1}=\frac{1}{z}\frac{f_j(z)-\gamma_j}{1-\overline{\gamma_j}f_j(z)}, \quad f_j(0)\equiv \gamma_j \in \mathbb{D},

which stops if f_j(z)\equiv e^{i\theta} = \gamma_j \in \mathbb{T} . One can invert the transformation as

: f(z)\equiv f_0 (z) = \frac{\gamma_0 + zf_1(z)}{1 + \overline{\gamma_0} z f_1(z) }

or, equivalently, as continued fraction expansion of the Schur function

: f_0(z)=\gamma_0+\frac{1-|\gamma_0|^2}{\overline {\gamma_0}+\frac{1}{z \gamma_1+\frac{z(1-|\gamma_1|^2)}{\overline {\gamma_1}+\frac{1}{z\gamma_2+\cdots}}}}

by repeatedly using the fact that

: f_j(z)=\gamma_j+\frac{1-|\gamma_j|^2}{\overline {\gamma_j}+\frac{1}{zf_{j+1}(z)}}.

See also

References