Decomposition and construction of preconditioners for Wiener-Hopf integral operators (Q1815604)

From MaRDI portal





scientific article; zbMATH DE number 946769
Language Label Description Also known as
default for all languages
No label defined
    English
    Decomposition and construction of preconditioners for Wiener-Hopf integral operators
    scientific article; zbMATH DE number 946769

      Statements

      Decomposition and construction of preconditioners for Wiener-Hopf integral operators (English)
      0 references
      0 references
      0 references
      15 December 1996
      0 references
      The authors consider the finite section Wiener-Hopf equation \((\alpha I+ A_\tau) x_\tau = \alpha x_\tau (t)+ \int^t_0 a(t-s)x_\tau (s) ds=g(t)\), \(0\leq t\leq\tau\). They construct the decomposition of the operator \(A_\tau\) in a sum of \(w_v\)-circulant operators \(P_\tau^{ (u,v)} :A_\tau= {1\over u} \sum^{u-1}_{v=0} P_\tau^{(u,v)}\). With the help of the operators \(P_\tau^{(u,v)}\) the preconditioners may be defined as \[ B_\tau^{(u)} = {1\over u} \sum^{u-1}_{v=0} (\alpha I+ P_\tau^{(u,v)})^{-1}, \] the spectra of which are clustered around 1. The main result is contained in Theorem 3.1: Let \(a(x) \in L_1(-\infty, \infty)\) and its Fourier transform \(\widehat a(t)\geq 0\). Then for any given \(\varepsilon>0\), there exists a positive integer \(N\) and a \(\tau^* >0\) such that for all \(\tau> \tau^*\), the spectrum of \((B_\tau^{(u)})^{1/2} (\alpha I+ A_\tau) (B_\tau^{(u)})^{1/2}\) has at most \(N\) eigenvalues outside the interval \((1-\varepsilon, 1 +\varepsilon)\). This theorem shows that the conjugate gradient method, when applied to solving the preconditioned operator equations \[ \bigl(B_\tau^{(u)} (\sigma I+ A_\tau)x_\tau \bigr) (t) = (B_\tau^{(u)} g) (t),\;0\leq t\leq \tau \] converges superlinearly.
      0 references
      superlinear convergence
      0 references
      finite section Wiener-Hopf equation
      0 references
      decomposition
      0 references
      preconditioners
      0 references
      conjugate gradient method
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references