14,557 reputation
23769
bio website
location
age
visits member for 2 years, 11 months
seen Dec 17 at 0:27

Jul
27
comment Efficient Langevin Equation Solver
@Gabriel I suppose it can be guessed by analogy to Kibble-Zurek type arguments (well, they don't really apply since there's not actual transition, but maybe one could estimate the domain size like this--but maybe not).
Jul
27
comment Efficient Langevin Equation Solver
if I start at a high temperature and decrease it at a fixed rate, can I predict the characteristic size of the domains? is it related to the rate?
Jul
27
comment Efficient Langevin Equation Solver
@LeonidShifrin Yes, I think you pointed this out to me once before. I find it more readable though (small brain) so I'll leave it. thanks
Jul
26
comment Efficient Langevin Equation Solver
You mention that "A follow up would be to extend all this to systems of Langevin equations". What would you like that to look like? Clearly you know that you could just use vectors to obtain a solution for a single realization of the noise in the same way you are doing it now. What more would you want? To do it for many realizations with a single NestList?
Jul
26
revised Efficient Langevin Equation Solver
added 1255 characters in body
Jul
26
revised Efficient Langevin Equation Solver
added 80 characters in body; added 110 characters in body
Jul
26
comment Efficient Langevin Equation Solver
@sebhofer a copy of the package :)
Jul
26
comment Efficient Langevin Equation Solver
Yes you're right, I screwed up. Let me fix it later. The only advantage is that it is trivially parallelizable and it scales much easier (at least, for me). Try to solve the discretised KPZ equation (which is notorious for slow convergence) for large systems and thousands of runs and you will see what I mean. For obtaining intuition, though, I agree that the Nest approach is neater. In practice, anyway, for something as simple algorithmically as this, C is much faster and not very hard to code if performance is needed.
Jul
26
comment Efficient Langevin Equation Solver
quizz: what package was used for figs 2 and 3 of this?
Jul
26
answered Efficient Langevin Equation Solver
Jul
26
comment Efficient Langevin Equation Solver
@R.M it can be parallelized if you calculate each realization independently (I ended up doing this when I was working on stochastic processes a few years back)
Jul
24
answered Sharp ListDensityPlot
Jul
23
answered How to generate table without comma and bracket or get rid of those?
Jul
19
comment How can I access documentation from the Workbench?
apparently, we can't. would be nice if someone has a way though, that's the main problem I have with workbench.
Jul
17
comment Programmatically quitting the FrontEnd or running without one?
@sebhofer thanks--I think that depends on what application the file type is associated to. in my case, doing that opens the script in an editor, because that's also what open does (from the command line). it probably works in the default OS X config
Jul
17
comment Programmatically quitting the FrontEnd or running without one?
@Jagra the iCal thing is robust, but it could be that the script fails for some reason. You could try checking (with applescript) that mma has indeed quit, but that won't work if the script stalls. I agree that it would be more robust to use something other than applescript, but this should work most of the time.
Jul
17
comment Programmatically quitting the FrontEnd or running without one?
@R.M of course, but that gives a warning about a process running in the background, preventing mma from quitting. One could script the UI, but sebhofer's idea was to use SystemOpen in the hope of avoiding this. I can't see how, unfortunately.
Jul
17
comment Programmatically quitting the FrontEnd or running without one?
@sebhofer good idea, let me see if I can work out how to run shell scripts (SystemOpen["bash /Users/acl/Desktop/qmma.sh"] just opens the script in Xcode on my machine)
Jul
17
comment Programmatically quitting the FrontEnd or running without one?
@sebhofer the problem with that is that it pops up a window telling you that a process is running...
Jul
17
comment Programmatically quitting the FrontEnd or running without one?
you could just run an applescript with 'tell application "Mathematica" to quit' (as @sebhofer says, it's applescript, just called from a shell). you could for instance create a calendar event to run this periodically, or use command-line methods (which I could explain if you want, but I guess it's easy to google them and in any case that would better be done in chat)