List Archives >  Maple User Group List Archive >  Archive by date >  This Month By Date >  This Month By Topic

[MUG] Stopping integration due to what?

Search email archive for  

[MUG] Stopping integration due to what?
Author: Angelique Lamour    Posted: 27/11/2000 07:40:05 GMT
>> From: Dear MUG, In running my model (7 ODEs) using DEplot, I sometimes get the error message "Stopping integration due to, undefined", depending on the parameter values used. In some cases, decreasing the stepsize gives indeed the desired plot. COULD the size of the stepsize BE the reason for the error message? In other cases I have to decrease the stepsize so much that my computer crashes (no simulation succeeded for stepsize=0.0001 and t=0..500). My computer is a Pentium 794MHz, 128 Mb RAM. Is it possible to write the points that are calculated by DEplot to a file, such that after each time step the value for each variable is written to a file? If the computer gets stuck in a particular run, then at least the points until the crash are available, which may provide information on the cause of the crash. Thanks for any advice, best wishes, Angelique Lamour (PhD student) Wageningen University, The Netherlands

[MUG] Re: Stopping integration due to what?
Author: Allan Wittkopf    Posted: 27/11/2000 23:00:51 GMT
>> From: Allan Wittkopf By default, DEplot uses a fixed step method to integrate the input system. There could be a number of reasons for the failure. 1) The problem could be too difficult for a fixed step method (there are regions that need a very small step size). 2) The problem could be stiff One of the options for DEplot is 'method=<your method of choice>'. The first thing I would suggest is using method=rkf45 - a variable step integrator. If this does not work, then I suggest trying method=lsode - a stiff integrator. In any event, the message being returned to you is not helpful, and requires a fix. Would it be possible for you to post the actual system that is giving you the problem so I could try it out? Thanks, Allan Wittkopf | >> From: | | In running my model (7 ODEs) using DEplot, I sometimes get the error message | "Stopping integration due to, undefined", depending on the parameter values | used. In some cases, decreasing the stepsize gives indeed the desired plot. | COULD the size of the stepsize BE the reason for the error message? | | In other cases I have to decrease the stepsize so much that my computer | crashes (no simulation succeeded for stepsize=0.0001 and t=0..500). My | computer is a Pentium 794MHz, 128 Mb RAM. Is it possible to write the points | that are calculated by DEplot to a file, such that after each time step the | value for each variable is written to a file? If the computer gets stuck in a | particular run, then at least the points until the crash are available, which | may provide information on the cause of the crash.

[MUG] Re: Stopping integration due to what?
Author: Edgardo S Cheb-Terrab    Posted: 04/12/2000 23:56:11 GMT
>> From: "Edgardo S. Cheb-Terrab" ----- Original Message ----- >> From: Dear MUG, In running my model (7 ODEs) using DEplot, I sometimes get the error message "Stopping integration due to, undefined", depending on the parameter values used. In some cases, decreasing the stepsize gives indeed the desired plot. COULD the size of the stepsize BE the reason for the error message? In other cases I have to decrease the stepsize so much that my computer crashes (no simulation succeeded for stepsize=0.0001 and t=0..500). My computer is a Pentium 794MHz, 128 Mb RAM. Is it possible to write the points that are calculated by DEplot to a file, such that after each time step the value for each variable is written to a file? If the computer gets stuck in a particular run, then at least the points until the crash are available, which may provide information on the cause of the crash. Thanks for any advice, best wishes, Angelique Lamour (PhD student) Wageningen University, The Netherlands -------------------------------------------------------------------------- Hi, I'm not sure about what ODE systems you have in mind, but, if you have Maple 6 (not otherwise), you could always use the new DNA libraries for numerically solving ODEs (see http://lie.uwaterloo.ca/dna.htm). Apart from numerically solving ODEs 30 times faster than in standard Maple 6, these DNA libraries are basically free of the problems you are mentioning. In practice, you just install the library and that automatically updates all of dsolve (numeric solutions), odeplot and DEplot to take advantage of them. For odeplot, the plots are then also computed 30 faster, use to be free of these "undefined" or "large number of function evaluations" problem, and are much smooth than without using DNA. More info is in the web-site mentioned above. Hope this is of use. Edgardo ___________________________________________________________________________ Edgardo S. Cheb-Terrab http://lie.uwaterloo.ca/ecterrab Centre for Experimental and Constructive Mathematics SFU, Canada Theoretical Physics Department UERJ,Brazil ___________________________________________________________________________ -- "Whoever has the most fun wins" --

Previous by date: [MUG] Re: laplace transform and dsolve (or Heaviside), Maple Group
Next by date: [MUG] Integration bug, i think, Halvor Mehlum
Previous thread: [MUG] How do I get Maple output into a report?,  Larry Fasnacht
Next thread: [MUG] Integration bug, i think, Halvor Mehlum