[R-sig-dyn-mod] Hybrid simulation of nonlinear ode and discrete state machine

Eike Petersen eike.petersen at uni-luebeck.de
Mon Dec 12 17:25:37 CET 2016


Maybe I'm mistaken, but I think you misunderstood me: I'm really dealing 
with a nonlinear function of the first derivative of the state on my 
left-hand side. By

y'^2 + y'

I really mean the squared of the first derivative, plus the same 
derivative itself. I am *not* talking about a second derivative here. I 
stand to be corrected, but as far as I know, there are no general 
techniques to transform (fully implicit?) ODEs of this type into a 
system of linear first-order ODEs.

Hence I am currently using the daspk solver for my problem, since it can 
deal with fully implicit ODEs. However, returning to my initial 
question, I'd like to use the output of a finite state machine (FSM) as 
the input function to my problem. This requires the efficient and 
reliable detection of state transitions, for which I wanted to use the 
'root function' solver option. According to the deSolve documentation 
however, this option is currently not implemented for the daspk solver, 
which is why I wrote to this list, asking for my options.

Hopefully this clarifies things a bit.

Kind regards and thanks for your time,

Eike


On 12/09/2016 11:59 PM, Thomas Petzoldt wrote:
> Thanks for clarification. The transformation of a second order 
> derivative is briefly shown here:
>
> https://journal.r-project.org/archive/2010-2/RJournal_2010-2_Soetaert~et~al.pdf#3 
>
>
> and some more may be found in the book of Soetaert, Cash and Mazzia:
>
> http://www.springer.com/cn/book/9783642280696
>
>
> Thomas
>
> _______________________________________________
> R-sig-dynamic-models mailing list
> R-sig-dynamic-models at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-dynamic-models



More information about the R-sig-dynamic-models mailing list