Search All of the Math Forum:

Views expressed in these public forums are not endorsed by NCTM or The Math Forum.

Topic: nonlinear differential equation
Replies: 8   Last Post: Mar 1, 2005 2:14 AM

 Messages: [ Previous | Next ]
 wes Posts: 24 Registered: 12/13/04
Re: Re: nonlinear differential equation
Posted: Feb 28, 2005 3:55 AM

In article <cvrqae\$p3s\$1@smc.vnet.net>, drbob@bigfoot.com says...
> Yikes!!! Good luck inverting the functions involved.
>
> Off[Solve::verif, Solve::tdep]
> deqn = Derivative[2][s][t] -
> a*s[t]^2 - b*s[t] - c == 0;
> ddeqn =
> ((Integrate[#1, t] & ) /@
> Expand[Derivative[1][s][t]*
> #1] & ) /@ deqn
> s /. DSolve[{%}, s, t]
> (-c)*s[t] - (1/2)*b*s[t]^2 -
> (1/3)*a*s[t]^3 +
> (1/2)*Derivative[1][s][t]^
> 2 == 0
>
> {Function[{t}, InverseFunction[
> (I*EllipticF[I*ArcSinh[
> (2*Sqrt[3]*Sqrt[
> c/(3*b + Sqrt[9*b^2 -

The solution is formed from the integral wrt s of the reciprical square
root of (1/3)a s^3 + (1/2)b s^2 + c s + v0^2. The result depends
strongly on the roots of the polynomial in s. If they are real the
result will be an inverse Jacobi elliptic function equal to d t, d a
constant. Inverting the result is usually something like

sin[y[s]] = sn[d t|m]

with y[s] not being too complicated. Mathematica doesn't handle this
type of problem gracefully yet. See "Handbook of Elliptic Integrals For
Engineers And Physicists", Bryd & Friedman, Springer-Verlag 1954.

Date Subject Author
2/23/05 Umby
2/25/05 Jens-Peer Kuska
2/27/05 Bobby R. Treat
2/28/05 Peter Pein
3/1/05 Paul Abbott
2/28/05 wes
3/1/05 Bobby R. Treat