On the hybrid finite difference scheme for a singularly perturbed Riccati equation (Q5963413)
From MaRDI portal
scientific article; zbMATH DE number 6543036
Language | Label | Description | Also known as |
---|---|---|---|
English | On the hybrid finite difference scheme for a singularly perturbed Riccati equation |
scientific article; zbMATH DE number 6543036 |
Statements
On the hybrid finite difference scheme for a singularly perturbed Riccati equation (English)
0 references
19 February 2016
0 references
This paper proposes an hybrid scheme for the numerical solution of singularly perturbed Riccati initial value problem of the form \[ \begin{cases} \varepsilon\,u'(x) + a(x) \left(u^2(x) - g^2(x)\right) = 0, \quad 0 \leq x \leq l,\\ u(0) = A, \end{cases} \] where \(0 < \varepsilon \ll 1\), the coefficients \(a,g \in C^2([0,l])\) such that \(\alpha^* \geq a(x) \geq \alpha >0\), \(g(x) \geq \gamma > 0\), \(u(0) + \gamma > 0\). The solution of the above problem exhibits an initial layer near \(x=0\). Therefore, to obtain \(\varepsilon\)-uniform numerical results, one has to discretize the domain with layer-adapted meshes and to apply the implicit Euler difference scheme, which is of first-order of convergence. Here, the authors discretize the domain with the piecewise-uniform Shishkin mesh and apply an hybrid scheme to approximate the differential equation. Instead of using the implicit Euler difference scheme throughout the domain, here they use the midpoint scheme, where the local mesh width is smaller than \(\varepsilon\) and the implicit Euler scheme elsewhere. By this proper combination, they improve the order of convergence from first-order to second-order (with a logarithmic factor). Truncation errors are derived, the stability analysis is carried out. \(\varepsilon\)-uniform second-order error estimates are derived. Numerical experiments are provided to validate the theoretical estimates.
0 references
singular perturbation
0 references
Riccati equation
0 references
Shishkin mesh
0 references
uniform convergence
0 references
initial value problem
0 references
implicit Euler difference scheme
0 references
stability
0 references
error estimate
0 references
numerical experiment
0 references
0 references
0 references
0 references
0 references
0 references
0 references