scientific article; zbMATH DE number 1269445

Equations and inequalities involving linear operators, with vector unknowns (47A50) Numerical solutions to equations with linear operators (65J10) Finite difference methods for boundary value problems involving PDEs (65N06) Stability and convergence of numerical methods for boundary value problems involving PDEs (65N12)

Review text

The paper is concerned with the two-layer operator-difference scheme \[ B{y_{n+ 1}- y_n\over \tau}+ Ay_n= 0\quad (n= 0,1,\dots),\quad \tau> 0 \] and the three-layer one \[ B{y_{n+ 1}- y_{n- 1}\over 2\tau}+ \tau^2R {y_{n+ 1}- 2y_n+ y_{n- 1}\over \tau^2}+ Ay_n= 0\quad (n= 1,2,\dots) \] in an Euclidean space \(H\) with the norm \(\|y\|= \sqrt{(y, y)}\), where the element \(y_n= y(t_n)\in H\) is sought for, \(t_n= n\tau\), the initial data \(y_0\) and \(y_1\) are given, and \(A\), \(B\), \(R\) are linear operators in \(H\). A finite difference scheme is said to be stable in the Euclidean space \(H_D\) if \(\|y_{n+ 1}\|_D\leq \|y_n\|_D\) for any initial data, where \(\|y\|_D= \sqrt{(Dy, y)}\) and \(D: H\to H\) is a positive self-adjoint operator. The necessary and sufficient conditions for the above stability in the space \(H_A\) are well known. The authors study the stability of operator difference schemes in other spaces \(H_D\) under the condition that either \(A\) or \(B\) is not self-adjoint and find some stability conditions that cannot be improved by the choice of a norm in \(H\). Moreover, analoguous results related to schemes with variable weight coefficients are shown.

Recommendations

Cited in

(8)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4236306)