Automated allocation of mesh points for stiff two-point boundary value problems
Nonlinear boundary value problems for ordinary differential equations (34B15) Multiple scale methods for ordinary differential equations (34E13) Numerical solution of boundary value problems involving ordinary differential equations (65L10) Mesh generation, refinement, and adaptive methods for ordinary differential equations (65L50)
The numerical solution of stiff two-point boundary value problems for a system of \(n\) differential equations (1) \(dy/dt=F(t,y)\), \(a<t<b\); \(B(y(a),y(b))=0\), where \(y^ T=(y_ 1,y_ 2,\dots,y_ n)\) is a given vector, is considered. A general procedure requiring negligible prerequisites to the problem is obtained to automate the allocation of mesh points by constructing a diffeomorphism \(t=\psi(u)\) from \([a,b]\) to \([a,b]\). It is the composition of the diffeomorphisms \(t=p(s)\), whose inverse is a solution of an initial value problem (2) \(ds/dt=Q(t,y)\), \(a<t<b\); \(s(a)=c\), where \(Q>0\) is a continuous ``density function to be constructed, and \(s=q(u)\) being a solution of the boundary value problem (3) \(d^ 2s/du^ 2=0\), \(a<b\); \(s(a)=c\), \(s(b)=d\). A properly constructed \(\psi\) smooths out the sharp corners in the solution while it leaves the domain boundary and the boundary conditions unchanged in the augmented and transformed system (1)-(3).
- A uniform scheme for the singularly perturbed Riccati equation
- Graded-Mesh Difference Schemes for Singularly Perturbed Two-Point Boundary Value Problems
- Numerical Methods for Stiff Two-Point Boundary Value Problems
- Singularly Perturbed Spline Difference Schemes on a Non‐Equidistant Grid
- Variable mesh methods for the numerical solution of two-point singular perturbation problems
This page was built for publication: Automated allocation of mesh points for stiff two-point boundary value problems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1179396)
