Analysis of Newton's method to compute travelling waves in discrete media (Q2574148)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Analysis of Newton's method to compute travelling waves in discrete media |
scientific article |
Statements
Analysis of Newton's method to compute travelling waves in discrete media (English)
0 references
18 November 2005
0 references
Consider the scalar differential difference equation \[ -\gamma\phi''(\xi)- c\phi'(\xi)= F(\phi(\xi), \phi(\xi+ r_1),\dots, \phi(\xi+ r_N),\rho).\tag{\(*\)} \] Here, \(\gamma\geq 0\) is a fixed parameter, \(\rho\) is a determining parameter, the parameter \(c\) is the unknown wavespeed, the numbers \(r_i\) may have either sign. The study of the existence of a travelling wave solution \(u_{i,j}(t)= \phi(ik_1+ jk_2- ct)\) of a lattice differential equation on the lattice \(Z^2\), \(\dot u_{i,j}= h(u_{i,j},\lambda)\) leads to a forward-backward differential difference equation which represents a special case of \((*)\) with \(\gamma= 0\). The authors give a detailed analysis of a numerical method to solve \((*)\). The presented algorithm consists of a combination of a Newton-type method with parameter continuation techniques. They prove that the algorithm provides sequences of functions converging to a solution of \((*)\). This requires a detailed analysis of operators whose fixed point provides a solution of \((*)\). The authors prove existence and uniqueness of connecting solutions to \((*)\) and show that these solutions depend \(C^1\)-smoothly on the determining parameter \(\rho\). Additionally, they investigate the limiting behaviour of such solutions as \(\gamma\) tends to zero. Numerical examples demonstrate the robustness of the algorithm and illustrate phenomena like propagation failure, which are encountered in studying lattice differential equations.
0 references
Computation of travelling waves
0 references
functional differential equations
0 references
Newton's method
0 references
bistable lattice differential equations
0 references
numerical computation
0 references
Ising model
0 references
discrete media
0 references
myelinated nerve fibers
0 references