Quasinilpotency of the operators in picard-lindelöf iteration (Q4029138)

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 126456
Language Label Description Also known as
default for all languages
No label defined
    English
    Quasinilpotency of the operators in picard-lindelöf iteration
    scientific article; zbMATH DE number 126456

      Statements

      Quasinilpotency of the operators in picard-lindelöf iteration (English)
      0 references
      0 references
      0 references
      7 March 1993
      0 references
      Picard-Lindelöf iteration
      0 references
      quasinilpotent operators
      0 references
      order of convergence
      0 references
      superlinear convergence
      0 references
      continuous time iteration
      0 references
      resolvent operator
      0 references
      Suppose that the matrix \(A\) in the initial value problem \(\dot x+Ax=f\), \(x(0)=x_ 0\), \(f\in L_{1,\text{loc}}\), is decomposed as \(A=M-N\). Then the corresponding iteration, \(\dot x^ k+Mx^ k=Nx^{k-1}+f\), \(x^ k(0)=x_ 0\), can be written in the form \(x^ k=Kx^{k-1}+g\), where the iteration operator \(K\) is given by \[ K u(t):=\int_ 0^ t \exp(-(t- s)M)Nu(s)ds. \] The authors use this model to investigate the (superlinear) convergence of the (continuous time) iteration on finite time intervals by considering the resolvent operator associated with \(\lambda u-Ku=g\) as an operator-valued entire function. They show that the convergence properties of the above iteration when applied to a large and sparse system are essentially determined by the graph properties of the decomposition and are partly insensitive in terms of the values of the elements of \(M\) and \(N\).
      0 references

      Identifiers

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