Webpages of the old version pde2path 1.0,
outdated!
We strongly recommend to switch to
pde2path 2.* Overview.pde2path uses (pseudo) arclength continuation
to compute branches of stationary solutions and their bifurcations
for PDE systems of the
$\huga{\label{gform}
G(u,\lam):=-\nabla\cdot(c\otimes\nabla u)+a u-b\otimes\nabla u-f=0,
}$
where $u=u(x)\in\R^N$, $x\in\Omega\subset\R^2$ some
bounded domain, $\lam\in\R$ is a parameter,
$c\in\R^{N\times N\times 2\times 2}$,
$b\in\R^{N\times N\times 2}$,
$a\in\R^{N\times N}$ and $f\in\R^N$
can depend on $x,u,\nabla u$, and, of course,
parameters. The software is based on the Matlab pdetoolbox.
The standard assumption is that $c,a,f,b$ depend on
$u,\nabla u,\ldots$ locally, e.g., $f(x,u)=f(x,u(x))$; however,
the dependence of $c,a,f,b$ on arguments can in fact be quite general,
for instance involving global coupling.
The current version supports "generalized Neumann" boundary conditions (BC)
of the form
\begin{align}\label{e:gnbc}
{\bf n}\cdot (c \otimes\nabla u) + q u = g,
\end{align}
where ${\bf n}$ is the outer normal and again $q\in \R^{N\times N}$
and $g\in \R^N$ may depend on $x$, $u$, $\nabla u$ and
parameters. These BC include zero flux BC for $q,g=0$,
while large prefactors in $q$, $g$ can be used to generate "stiff spring"
approximations of Dirichlet BC that we found to work
reasonably well.
There are a number of predefined functions
to specify domains $\Om$ and boundary conditions, or these can be
exported from Matlab's pdetoolbox GUI, thus making it easy to deal
with (almost) arbitrary geometry and boundary conditions.
The software can also be used to time-integrate parabolic
problems of the form
\begin{equation}\label{pprob}
\pa_t u=-G(u,\lam),
\end{equation}
with $G$ as in \eqref{gform}. This is mainly intended to easily find initial
conditions for continuation. Finally
any number of eigenvalues
of the Jacobian $G_u(u,\lam)$ can be computed, thus allowing stability
inspection for stationary solutions of \eqref{pprob}.
Go to equation to see
the equation and BCs in more detail, jump
directly to Demos or
Download, or see
Basics and
First steps
for a little mathematical background and a short overview on the
usage of pde2path.