An algorithm for corona solutions on H^(D)

From MaRDI portal
Publication:2476173





Let \(D\) be the unit disk in \(\mathbb{C}\). The corona theorem states that for any \(\{f_j\}_{j=1}^m\subset H^\infty(D)\) satisfying \(0<\varepsilon^2\leq\sum_{j=1}^m| f_j(z)| ^2\leq1\), \(z\in D\), there exists \(\{u_j\}_{j=1}^m\subset H^\infty(D)\) such that \(\sum_{j=1}^mf_j(z)u_j(z)\equiv 1\) in \(D\) and \(\sup_{z\in D}\sum_{j=1}^n| u_j(z)| ^2\leq M(\varepsilon)\). The best known value for \(M(\varepsilon)\) was found by Uchiyama: for \(0<\varepsilon<1/e\) we can take \(M(\varepsilon)=M_0(\varepsilon):=(8\varepsilon^{-2}\log\varepsilon^{-2})^2\). A constructive method to solve the problem was suggested by \textit{Peter W. Jones} [J. Funct. Anal. 39, 162--181 (1980; Zbl 0471.30023)] but it is too complicated to see how it could be realized explicitly. The author gives a new constructive method for the case when \(f_j\) are polynomials of maximal degree \(n\) (for \(M(\varepsilon)=M_0(\varepsilon)\)). It is subdivided into four steps. The first one is to find an analytic polynomial \(F\) with prescribed square of modulus on \(\partial D\) (Fejer-Riesz factorization). Three other steps have pure algebraic nature and use matrix operations such as Cholesky decomposition and finding of inverse matrix. The elements of the considered matrices depend on the coefficients of \(f_j\) and \(F\). The solution to the corona problems is obtained as a set \(\{u_j\}_{j=1}^n\subset H^\infty(D)\) of rational functions of maximal order \(n\). The suggested method is illustrated by an example for \(m=2\) and \(n=1\) (\(f_1(z)=1-z\), \(f_2(z)=z\)).











This page was built for publication: An algorithm for corona solutions on \(H^{\infty}(D)\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2476173)