Finite difference methods for certain singular two-point boundary value problems (Q1919496): Difference between revisions
From MaRDI portal
Latest revision as of 13:37, 24 May 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Finite difference methods for certain singular two-point boundary value problems |
scientific article |
Statements
Finite difference methods for certain singular two-point boundary value problems (English)
0 references
22 January 1997
0 references
The authors consider the boundary value problem (BVP) \[ Lu \equiv \sum^2_{k=0} (-1)^k {d^k \over dx^k} \left[ P_k(x)\;{d^k u\over dx^k}\right] = F(x,u),\quad x \in (0,1),\tag{1} \] \[ \sum^2_{k=0} A_{lk} \left(d^ku \over dx^k\right)_{x\to 0}+\sum^3_{k=0} B_{lk} \left( {{d^k u} \over {dx^k}} \right)_{x=1} = g_l, \quad l = 1,\dots, n, \] where one or more of the coefficients \(p_k(x)\) may be infinite at \(x =0\) and the singularities are of the first kind. At first, the BVP (1) is linearized by Newton's method. For each linear BVP the authors apply Gustafsson's method of series expansion at a small neighborhood of the point \(x = 0\) and a difference method in the rest of the interval. The consistency and stability of the method are proved and an error estimate is presented. In the conclusion numerical experiments and some comments are given.
0 references
finite difference methods
0 references
singular two-point boundary value problems
0 references
Newton's method
0 references
method of series expansion
0 references
consistency
0 references
stability
0 references
error estimate
0 references
numerical experiments
0 references
0 references