Webpages of the old version pde2path 1.0, outdated!
We strongly recommend to switch to pde2path 2.0
________________________________________________________________________
Tipps’n tricks
- Check the m2html documentation of pde2path functions for an overview.
- Check the Matlab pdetoolbox help/users guide, and the alphabetic
function list.
- To load a previously saved structure p use p=loadp(pre,fname,name), or
simply double click on a pXXX file. Then manipulate it (in case of ”load
by click” perhaps after renaming or copying to a new variable name) and
call, e.g., cont.
- Manipulate memory of a stored p-file by editing p.branch. For instance
reset memory: p.branch = []; or append to branches (check that the point
ordering fits): p.branch = [p.branch;p1.branch];
- If pde2path indicates a bifurcation (due to a sign change of det(A)), but
no eigenvalue of ∂uG crosses the origin, consider increasing p.neig. Check
whether this has an effect using the utility function sgndetA(p,neig). Also
check p.eigref = ′sr′, which computes the eigenvalues with smallest real
part, but is much slower than the default p.eigref = 0.
- For bifurcations from an explicitly known (parameter-dependent)
equilibrium u0(λ), consider shifting u to v = u−u0(λ) for faster and more
robust numerical computations.
- to be continued ...
octave
For users who have access to the Matlab pdetoolbox but prefer octave over
Matlab we provide a compatibility mode with (currently) very restricted
functionality. For this, in octave call path(path-to-pdetoolbox,path); and after
setting the path to p2plib, call, e.g., path(’pde2pathhome/octcomp’,path); as the
functions in octcomp must mask the functions in p2plib. In fact, in octcomp we only
reset p2interpol.m and plotsol.m since in octave only mesh-plotting (plotstyle 1)
works reliably.
With these changes the demos bratu and ac work (where it is useful to turn off
octave pagination with more), while the other demos require some more adaptions.
In particular there currently is no numjac in octave. However, quick adaptions should
be no problem for expert users. Sometimes it is helpful to compare Matlab’s
performance with octave.
FAQ
(Hopefully) coming soon ..