MathGroup Archive 2002

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: DSolve for a system of 3 equations

  • To: mathgroup at smc.vnet.net
  • Subject: [mg36318] Re: DSolve for a system of 3 equations
  • From: Neil Klepeis <nklepeis at uclink4.berkeley.edu>
  • Date: Sat, 31 Aug 2002 01:26:02 -0400 (EDT)
  • Organization: Lawrence Berkeley National Laboratory
  • References: <akkctj$81o$1@smc.vnet.net> <akn0f8$lkv$1@smc.vnet.net>
  • Reply-to: nklepeis at uclink4.berkeley.edu
  • Sender: owner-wri-mathgroup at wolfram.com

Guillermo,

Thanks for the package.  I am also doing compartmental modeling, for 
indoor air exposure to particulate matter.   Does your package use 
Runge-Kutta or some other numerical algorithm?   I'm thinking of 
something like that, but with a fixed (non-adaptive) step size.

Neil



Guillermo Sanchez wrote:
> Neil,
> 
> They are not a general analitical solution for a system of ODE with 3
> or more variables where all coffs. are parameters. I have a package
> for solving SODE. Perhaps you can find it useful.
> 
> http://www.mathsource.com/Content/WhatsNew/0211-554
> 
> Guillermo Sanchez
> 
> Neil Klepeis <nklepeis at uclink4.berkeley.edu> wrote in message news:<akkctj$81o$1 at smc.vnet.net>...
> 
>>I try to solve a three equation linear ODE system with constant 
>>coefficients and Mathematica 3.0 (Intel/Linux) chokes (see below).
>>
>>When I try the same thing on Mathematica 4.0 (Solaris) I get a very 
>>strange result containing things like "E^x#1" OR "#1^2".
>>
>>Am I doing something wrong, or is there a bug/problem with Mathematica.
>>
>>Regards,
>>Neil
>>
>>
>>
>>-- Solving a Three Equation ODE system (Linear with Constant 
>>Coefficients) with Mathematica 3.0 ----
>>
>>INPUT:
>>
>>DSolve[{y1'[x] + c11*y1[x] + c12*y2[x] +c13*y3[x] +c1==0,
>>     y2'[x]+c21*y1[x]+c22*y2[x]+c23*y3[x]+c2==0,
>>     y3'[x]+c31*y1[x]+c32*y2[x]+c33*y3[x]+c3==0,
>>     y1[0]==y1i, y2[0]==y2i, y3[0]==y3i},
>>     {y1[x], y2[x], y3[x]}, x]
>>
>>OUTPUT:
>>
>>Eigensystem::"eivec":
>>     "Unable to find eigenvector for eigenvalue \!\(Root[\(\(\(\(c13\\ 
>>c22\\ \
>>c31\) - \(c12\\ c23\\ c31\) - \(c13\\ c21\\ c32\) + \(c11\\ c23\\ c32\) + \
>>\(\\[LeftSkeleton] 20 \\[RightSkeleton]\)\) &\), 1\)]\)."
>>Eigensystem::"eivec":
>>     "Unable to find eigenvector for eigenvalue \!\(Root[\(\(\(\(c13\\ 
>>c22\\ \
>>c31\) - \(c12\\ c23\\ c31\) - \(c13\\ c21\\ c32\) + \(c11\\ c23\\ c32\) + \
>>\(\\[LeftSkeleton] 20 \\[RightSkeleton]\)\) &\), 2\)]\)."
>>Eigensystem::"eivec":
>>     "Unable to find eigenvector for eigenvalue \!\(Root[\(\(\(\(c13\\ 
>>c22\\ \
>>c31\) - \(c12\\ c23\\ c31\) - \(c13\\ c21\\ c32\) + \(c11\\ c23\\ c32\) + \
>>\(\\[LeftSkeleton] 20 \\[RightSkeleton]\)\) &\), 3\)]\)."
>>General::"stop":
>>     "Further output of \!\(Eigensystem :: \"eivec\"\) will be suppressed \
>>during this calculation."
> 
> 



-- 
______________________________________________________
Neil E. Klepeis, UC Berkeley, School of Public Health,
and Lawrence Berkeley National Laboratory,
Berkeley, CA USA.  Voice: 831-768-9510


  • Prev by Date: Re: Resizing Graphics
  • Next by Date: Re: NDSolve with integral equation
  • Previous by thread: Re: DSolve for a system of 3 equations
  • Next by thread: Sub- and Superscripts in FrameLabel