Regular perturbations of discrete nonstationary problems in mathematical physics connected with linear restrictions. (Q1848645)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Regular perturbations of discrete nonstationary problems in mathematical physics connected with linear restrictions. |
scientific article |
Statements
Regular perturbations of discrete nonstationary problems in mathematical physics connected with linear restrictions. (English)
0 references
13 November 2002
0 references
The author investigates the correctnesses of a special type of discrete nonstationary problems of mathematical physics containing stricly saddle operators with small parameter \(\alpha \geq 0\). When \(\alpha = 0\) the problems can be considered as well known approximations of classical parabolic and hyperbolic equations considered on a given subspace \(H_0\) of the Hilbert space \(H_1\). It is essentially important that the subset \(H_0\) is the kernel of a normally reversible operator \(A_{1, 2} \ni L(H_1, H_2)\), where \(L(H_1, H_2)\) designates a linear normed space of the linear bounded operators mapping \(H_1\) into the Hilbert space \(H_2\). The author considers only real Hilbert spaces. For stationary problems of the type \[ A_{\alpha}u_{\alpha} \equiv \left[ \begin{matrix} A_{1,1} & A_{1,2}\\ A_{2,1} & -\alpha A_{2,2} \end{matrix} \right] \left[\begin{matrix} u_{1, \alpha}\\ u_{2, \alpha} \end{matrix} \right] = \left[ \begin{matrix} f_1\\ f_2 \end{matrix} \right] \] with strictly saddle operator \(A_{\alpha}=A_{\alpha}^\star(A_{1,1} \asymp I_1, A_{2,2}\geq 0)\) their correctness is proved with estimation \(\| A_{\alpha}^{-1} \|\leq K\) and leading to the regular estimation of perturbation \[ \| u_{\alpha} - u_{\widetilde\alpha}\| \leq \| K_1 | \alpha - \widetilde\alpha |\,\| .\tag \(*\) \] The aim of the article is to get regular estimations of perturbations (similar to (\(*\))) for nonstationary problems in \({\mathbf H}\equiv H_1 \times H_2\) of the following types: \[ D_0M_1u_1(t) + A_{1,1}u_1(t) + A_{1,2}u_2 = f_1(t), \quad A_{2,1}u_1(t)=0, \] where \[ M_1 \in {L}(H_1; H_1), D_0u = \frac{\partial u(t)}{\partial x_0}, x_0 \equiv t\in [0, T] \] and also for the problems in \({\mathbf H}\) of the types: \[ D_0^2M_2u_1(t) + D_0M_1u_1(t) + A_{1,1}u_1(t) + A_{1,2}u_2(t)= f_1(t), \] \[ D_0^2A_{2,1}u_1(t) + \alpha A_{2,2}u_2(t) = f_2(t), \] \(M_i \in {L}(H_1;H_1), M_i = M_i^{\star}, Mi \geq 0, M_2 \asymp I_1.\)
0 references
mathematical physics
0 references
nonstationary problems
0 references
elliptic and hyperbolic equations
0 references
regular estimations of perturbation
0 references
saddle operators
0 references
Hilbert space
0 references