Services & Resources / Wolfram Forums
-----
 /
MathGroup Archive
1996
*January
*February
*March
*April
*May
*June
*July
*August
*September
*October
*November
*December
*Archive Index
*Ask about this page
*Print this page
*Give us feedback
*Sign up for the Wolfram Insider

MathGroup Archive 1996

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

Search the Archive

Re: Chaotic Equation NDSolve....

  • To: mathgroup at smc.vnet.net
  • Subject: [mg3153] Re: Chaotic Equation NDSolve....
  • From: withoff (David Withoff)
  • Date: Fri, 9 Feb 1996 03:00:42 -0500
  • Organization: Wolfram Research, Inc.
  • Sender: owner-wri-mathgroup at wolfram.com

In article <4f4hts$cpu at dragonfly.wolfram.com>  
aspuru at servidor.dgsca.unam.mx writes:
> Greetings from Mexico City,
> 
> We are a pair of students of chemistry doing some exercises on chaotic  
systems.
> So we are not too keen on math as a mathematician would be.
> 
> We programmed this system and get this error:
> "Differential Equation does not evaluate as a number at t=1" Tried with 
> different values of t with no success:
> 
> The system we programmed is as follows:
> 
> NDSolve[ {q1'[t] == p1[t],
>           q2'[t] == p2[t],
>           p1'[t] == ((1-f) q1[t])/((1 - q2[t]) ^ 2  + (q1[t]) ^ 2 ) ^  
0.5,
>           p2'[t] == (1-f) - q2[t] - ((1-f) (1-q2[t]))/((1-q2[t]) ^ 2 +  
(q1[t]) ^ 2) ^ 0.5 ,
> 	  q1 [1] == q2[1] == p1[1] == 1, p2[1] == 2},
>          {p1, q1, p2, q2}, {t, 1, 20}, MaxSteps->3000 ]
> 
> And we dont know what is wrong on it... I hope any one of you can  
guide/help
> us in our quest of solving this system.
> 
> Thanks a lot in advance.
> 
> Alan Aspuru & Octavio Paniagua...
> 
> Chemistry Faculty,
> Universidad Nacional Autonoma de Mexico

You will get this message if f (or any other expressions in your
input) doesn't have numerical values.

A trick that I often use when confronted with a large set of
differential equations and the "Differential Equation does not
evaluate as a number" message is to insert random numerical
values for the things that I know to be numbers, and look for
anything in the result that isn't numerical.

In[1]:= equations = {q1'[t] == p1[t],
        q2'[t] == p2[t],
        p1'[t] == ((1-f) q1[t])/((1 - q2[t]) ^ 2  + (q1[t]) ^ 2 ) ^0.5,
        p2'[t] == (1-f) - q2[t] - ((1-f) (1-q2[t]))/((1-q2[t]) ^ 2 +
                               (q1[t]) ^ 2) ^ 0.5}

Out[1]= {q1'[t] == p1[t], q2'[t] == p2[t], 
 
                     (1 - f) q1[t]
>    p1'[t] == --------------------------, 
                     2              2 0.5
               (q1[t]  + (1 - q2[t]) )
 
                          (1 - f) (1 - q2[t])
>    p2'[t] == 1 - f - -------------------------- - q2[t]}
                             2              2 0.5
                       (q1[t]  + (1 - q2[t]) )

In[2]:= equations /. {p1[t] -> Random[], p2[t] -> Random[],
                      q1[t] -> Random[], q2[t] -> Random[],
                      p1'[t] -> Random[], p2'[t] -> Random[],
                      q1'[t] -> Random[], q2'[t] -> Random[]}

Out[2]= {False, False, 0.404791 == 0.00145649 (1 - f), 
 
>    0.862479 == 0.710355 - 0.999999 (1 - f) - f}

This exposes the non-numerical symbols quite readily.

Dave Withoff
Research and Development
Wolfram Research


==== [MESSAGE SEPARATOR] ====


  • Prev by Date: MathLink-problem
  • Next by Date: Re: [Q] Programming & "Not Floating Point Number" Error
  • Previous by thread: Re: Chaotic Equation NDSolve....
  • Next by thread: Newsgroup and Mailing List Down Last Night