Numerical solution of singular perturbation problems (Q1346991)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Numerical solution of singular perturbation problems |
scientific article |
Statements
Numerical solution of singular perturbation problems (English)
0 references
30 March 1995
0 references
This paper deals with the solution of the boundary value problem \(\varepsilon y'' + s(t)y' + c(t)y = f(t)\), \(t \in [t_{\min}, t_{\max}]\), \(y(t_{\min}) = a\), \(y(t_{\max}) = b\), where \(\varepsilon\) is positive (it could be very small), \(c(t)\) and \(f(t)\) are continuous functions on the integration interval \([t_{\min}, t_{\max}]\), and \(s(t)\) is differentiable. The presented method is based on second order difference schemes where the convergence is not very fast. It applies a mesh selection strategy derived by using sufficient conditions which ensure the well-conditioning of the corresponding tridiagonal matrices. It has the positive feature of choosing the initial mesh very efficiently and this allows it to perform better than higher order methods, especially for small values of \(\varepsilon\) or for not very high tolerance. Numerical tests are reported.
0 references
singular perturbation
0 references
numerical examples
0 references
second order difference schemes
0 references
convergence
0 references
mesh selection strategy
0 references
0 references
0 references
0 references