The representation and approximation of the Drazin inverse of a linear operator in Hilbert space. (Q1406116): Difference between revisions

The authors obtain a unified representation theorem for the Drazin inverse of a bounded linear operator in Hilbert space \[ T^D=\lim_{n\rightarrow \infty} S_n(\tilde{T})T^kT^{2k+1}T^k, \] where \(T\) is an operator with \(\text{Ind}(T)=k\) and \(R(T^k)\) closed, \(\tilde{T}\) is the restriction of \(T^kT^{2k+1}T^{k+1}\) on \(R(T^k)\), and \(\{S_n(x)\}\) is a sequence of functions uniformly convergent to \(1/x\) on the spectrum of \(\tilde{T}\). On the base of this representation five computational methods for the approximation of \(T^D\) are proposed. Corresponding error bounds of the approximations are also presented.