×

Perturbed projection method for linear compact operator equations. (Chinese. English summary) Zbl 0612.47011

The authors discuss the perturbed projection solutions for linear compact operator equations by compactness principles in operator approximation theory introduced by P. M. Anselone and R. Ansorge [see Numer. Funct. Anal. Optimization 1, 589-618 (1979; Zbl 0466.65036)]. They mainly prove the following result:
Let \(K\in L(X)\) be sequentially compact, \((\lambda -K)^{-1}\) be bounded, \(S_ n\to^{ac}0\) and \(\| g_ n\| \to 0\) as \(n\to \infty\). Suppose projection operators \(P_ n:X\to X_ n\) satisfy \(P_ n\to^{p}I\) furthermore. Then the operator equation \(\lambda x=Kx+f\) and its perturbed projection equations \(\lambda x_ n=K_ nx_ n+P_ nf\) have unique solutions \(x_ 0\) and \(x_ n\), respectively, if for n large enough, also \(x_ n\to x_ 0\) as \(n\to \infty\), and the following estimations hold for large enough \(n:\) \[ \| (\lambda -K_ n)^{- 1}\| \leq (1+\| (\lambda -K)^{-1}\| \| K_ n\|)/(| \lambda | -\Delta_ n), \]
\[ \| x-x_ 0\| \leq \| (\lambda -K)^{-1}\| (\| (K-K_ n)K_ n\| \| x_ 0\| +\| (K- K_ n)P_ nf\| +| \lambda | \| P_ nf-f\|)/(| \lambda | -\Delta_ n), \]
\[ \| x_ n-x_ 0\| \leq \| (\lambda -K_ n)^{-1}\| ([\| P_ nK\| +\| \lambda -K_ n\|]\| P_ nx_ 0-x_ 0\| +\| S_ n\| \| P_ nx_ 0\| +\| g_ n\|), \]
\[ \| x_ n-x_ 0\| \leq \| (\lambda -K_ n)^{-1}\| \| (K-K_ n)x_ 0+f-P_ nf\|. \] Where \(\Delta_ n=\| (\lambda -K)^{-1}(K_ n-K)K_ n\|.\)
Here, X is a Banach space, \(X_ n\) are subspaces of X, L(X) is the usual space consisting of all bounded linear operators from X into X, \(x,f\in X,x_ n,g_ n\in X_ n\), \(K_ nx_ n=P_ nKx_ n+S_ nx_ n+g_ n\), \(T_ n\to^{ac}T\) denotes \(x_ n\to x\Rightarrow T_ nx_ n\to Tx\) and \(S\subset X\) bounded \(\Rightarrow\) every sequence \(\{x_ n:x_ n\in T_ n(S)\), \(n\in N'\subset N\) (the positive integer set)\(\}\) has convergent subsequences, \(T_ n\to^{p}T\) denotes \(T_ nx\to Tx\) for every \(x\in X\), I denotes the identical operator on X.
Reviewer: Zhang Guanglu

MSC:

47A50 Equations and inequalities involving linear operators, with vector unknowns
65J10 Numerical solutions to equations with linear operators
47B06 Riesz operators; eigenvalue distributions; approximation numbers, \(s\)-numbers, Kolmogorov numbers, entropy numbers, etc. of operators

Citations:

Zbl 0466.65036