Factorization of Certain Piecewise Constant Matrix Functions and Its Applications (Q3978595)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 12539
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Factorization of Certain Piecewise Constant Matrix Functions and Its Applications |
scientific article; zbMATH DE number 12539 |
Statements
Factorization of Certain Piecewise Constant Matrix Functions and Its Applications (English)
0 references
25 June 1992
0 references
\(2\times 2\)-matrix function
0 references
factorization
0 references
generalized Hardy-classes
0 references
partial indices
0 references
singular integral equations on the unit circle
0 references
0.8222370147705078
0 references
0.8206928372383118
0 references
The paper is concerned with a piecewise constant \(2\times 2\)-matrix function \[ G(t)=C_ 0\chi_ 0(t)+C_ 1\chi_ 1(t)+C_ 2\chi_ 2(t), \] where \(\chi_ 1\), \(\chi_ 2\) are the characteristic functions of intervals \((t_ 0,t_ 1)\) and \((t_ 1,t_ 2)\), \(-\infty<t_ 0<t_ 1<t_ 2\leq+\infty\), and \(\chi_ 0\) is the characteristic function of \(\mathbb{R}\setminus[t_ 0,t_ 2]\) (resp. \(\mathbb{R}\setminus [t_ 0,+\infty)\) if \(t_ 2=+\infty\)). Necessary and sufficient conditions are given such that \(G\) has a factorization \[ G(t)=G_ +(t)\Lambda(t)G_ - (t)^{-1}, \] with \({1\over {t\pm i}}G_ \pm(t)\), resp. \({1\over {t\pm i}} G_ \pm(t)^{-1}\), in the generalized Hardy-classes \(H_{p,\rho}^ \pm\) resp. \(H_{q,1/\rho}^ \pm\), in the upper (lower) half-plane \(\Pi^ \pm\), where \(1<p<\infty\), \(p^{-1}+q^{-1}=1\), and the weight \(\rho\) is given by \(\rho(t)=(1+t^ 2)^{\nu/2}| t-t_ 0|^{\nu_ 0}| t-t_ 1|^{\nu_ 1}\) (if \(t_ 2=+\infty\)) or \(\rho(t)=(1+t^ 2)^{-(\nu_ 0+\nu_ 1+\nu_ 2)/2}| t-t_ 0|^{\nu_ 0} | t-t_ 1|^{\nu_ 1}| t-t_ 2|^{\nu_ 2}\) (if \(t_ 2\in\mathbb{R}\)), while \(-p^{- 1}<\nu_ k<1-p^{-1}\), \(-p^{-1}<\nu+\nu_ 0+\nu_ 1<1-p^{-1}\), whereas \[ \Lambda(t)=\text{diag}\left(\left({{t-i} \over {t+i}}\right)^{\kappa_ 1}, \left({{t-i} \over {t+i}}\right)^{\kappa_ 2}\right). \] The invertibility of the matrices \(C_ 0\), \(C_ 1\) and \(C_ 2\) is a necessary condition, and admitting this one finds conditions equivalent to the existence of a factorization in terms of the arguments of the eigenvalues of \(C_ 0^{-1}C_ 1\), \(C_ 1^{-1}C_ 2\) and \(C_ 0^{-1}C_ 2\). The partial indices \(\kappa_ 1\), \(\kappa_ 2\) of the factorization can be expressed in terms of these arguments as well. From known stability conditions one then derives necessary and sufficient conditions for the stability of \(\kappa_ 1\), \(\kappa_ 2\) under small perturbations of \(G(t)\). The result can be used to obtain conditions for the invertibility of the operator \(R_ G\) induced by \(G\) in the space \(L_{2,\rho}^ 2\), e.g. in the case where each convex combination of \(C_ 0\), \(C_ 1\) and \(C_ 2\) is invertible. The above results are applied to singular integral equations on the unit circle, where the necessary and sufficient conditions are obtained for the applicability of the projection- (\(\equiv\) finite-section-) method to the operators \({1\over 2}a(I+S^ 0)+{1\over 2} b(I-S^ 0)\) (with \(a\), \(b\) piecewise continuous \(2\times 2\)-matrix functions) and \({1\over 4}((I+S^ 0)a(I+S^ 0))+{1\over 4}((I+S^ 0)b(I-S^ 0)J)+ {1\over 2}(I-S^ 0)\) (with \(a\), \(b\) scalar functions and \((Jf)(\zeta)={1\over\zeta}f(\overline \zeta))\), and these results are applied to the generalized Riemann-Hilbert problem of solving \(\varphi^ +(t)+a(t)\varphi^ -(t)+b(t)\overline{\varphi^ -(t)}=c(t)\) on the unit circle \(\mathbb{T}=\partial\mathbb{D}\) with \(\varphi^ \pm\in H_ 2^ \pm(\mathbb{T})\).
0 references