
Dom V. answered 10/18/15
Tutor
5.0
(119)
Cornell Engineering grad specializing in advanced math subjects
Using x=e(rt) as a trail solution is only valid when dealing with constant coefficients. Remember, x(t) is the function you are solving for. If x appears as a coefficient in the characteristic equation, that means you'll end up with answers that look like X(t)=Ae({function of X}), which doesn't make sense.
The coefficients involve x(t) itself; the ODE is nonlinear. This would be classified as a homogeneous second-order nonlinear autonomous ODE. Using power series methods is as close to impossible as you can get. This needs to be solved numerically (computationally) by recasting it as a system of differential equations.
(x2)(x'') + (x-x2)(x') +(-k)(x)=0
Introduce these variables:
- F(t)=x(t)
- G(t)=x'(t)=F'
- H(t)=x''(t)=G'=F''
The original ODE becomes F2H+(F-F2)G-kF=0, which we can solve for H(F,G). At this point, because H=G', we no longer require H explicitly. We can express F' and G' as functions of F and G. Hence we have a system of two ODEs, which can be written as
- G'=[kF-(F-F2)G]/[F2]
- F'=G
I'm not sure what type of numerical solvers you're accustomed to using, but this is a fairly straightforward system for Matlab or something similar. As the original ODE was written, it is very similar to what's called the Van der Pol equation. The VdP equation is used pretty extensively as an introductory example to nonlinear ODEs. Researching that may give you a few tips on proceeding.
=========================================================================
As was mentioned before, the ODE system in this case is nonlinear. A linear system could have been written as something like
- F'=aF+bG
- G'=cF+dG
where a,b,c,d are constants. Problems like this are amenable to linear algebra, and we could have found solutions as linear combinations of eigenvectors. Not at all relevant to the problem at hand, but just an FYI if this pops up in the future!
Hope that helped!
Koshila D.
10/13/15