$y''-y'+e^{2x}y=0$ can you give me a hint which I can start with

Should I use the fact that if $r_1$ and $r_2$ are complex numbers (which happens if $b^2 - 4ac < 0$), then the general solution is

$$y = c_1 \Bbb e ^{\alpha x} \cos \beta x + c_2 \Bbb e ^{\alpha x} \sin \beta x$$


$$r_1 = \bar r_2 = \alpha + \Bbb i \beta = \frac {-b} {2a} + \Bbb i \frac {\sqrt {4ac - b^2}} {2a} ,$$

or make $t=e^x$?

  • $\begingroup$ The method in the gray box is only applicable when the coefficients of the y function and its derivatives are constants. So, that method is not applicable here. $\endgroup$ – Sinister Cutlass Oct 6 '15 at 5:43

There are often multiple ways to get to the final solution of a differential equation. I used $t=ie^x$ and got to $y = c_1 \cos(t) + c_2 \sin t$, yielding the solution $y = c_1 \cosh(e^x) + c_2 \sinh(e^x)$. I haven't checked it, but the substitution $t = e^x$ might work similarly. As a tip, if you show something you have attempted you could likely receive more help, as I (and likely others) don't feel like TeXing a proof for solving that Dif-Eq. Try it your way and let us know where you get stuck.

First few steps I did, using $t = ie^x$ and $\frac{dt}{dx} = t$. First is the original equation with substitution $$\frac{d^2 y}{dx^2} - \frac{dy}{dx} = -t^2y$$ Apply chain rule for both derivatives $$\frac{d^2 y}{dx^2} = \frac{d}{dx} \bigg(\frac{dy}{dt}\frac{dt}{dx}\bigg) = \bigg(\frac{dt}{dx}\bigg)^2\frac{d^2 y}{dt^2} + \frac{d^2t}{dx^2}\frac{dy}{dt} = t^2\frac{d^2 y}{dt^2} + t \frac{dy}{dt}$$ $$\frac{dy}{dx} = \frac{dt}{dx}\frac{dy}{dt} = t \frac{dy}{dt}$$ Substituting back in, we get $$\Rightarrow t^2\frac{d^2 y}{dt^2} + t \frac{dy}{dt} - t \frac{dy}{dt} = -t^2y$$ $$\Rightarrow t^2\frac{d^2 y}{dt^2} = -t^2y$$ $$\Rightarrow \frac{d^2 y}{dt^2} = -y$$

Can you get it from here?

  • $\begingroup$ thanks. can you write what is $\frac{d^2y}{d^2x}$ if $e^x=t$ i mean $\dfrac {d} {dx}\left( \dfrac {tdy} {dt}\right) $ $\endgroup$ – emmett Oct 6 '15 at 5:52
  • $\begingroup$ @emmett I'm not sure what you are trying to say... I'll write out the first couple steps I used with the substitution $t=ie^x$ and hopefully that will clear some things up for you as you do your substitution. $\endgroup$ – Brevan Ellefsen Oct 6 '15 at 5:55
  • $\begingroup$ @emmett I have expanded my answer to show how to expand when $t=ie^x$. Does this help you answer your question? It is very similar to what I show, but I leave the work up to you ;) $\endgroup$ – Brevan Ellefsen Oct 6 '15 at 6:08
  • $\begingroup$ can you please explain this $\dfrac {d} {dx}\left( \dfrac {dy} {dt}\right) =\dfrac {dt} {dx}\left( \dfrac {d^{2}y} {dt^{2}}\right) $ $\endgroup$ – emmett Oct 6 '15 at 6:13
  • $\begingroup$ I've got to go now, but $$\frac{dt}{dx}\frac{d^2y}{dt^2} = \frac{dt}{dx} \frac{d}{dt} \frac{dy}{dt}= \frac{d}{dt} \bigg(\frac{dt}{dx}\bigg)\frac{dy}{dt} = \frac{d}{dx}\frac{dy}{dt}$$. I haven't checked this, but I think its correct. I've just gotten used to these identities. $\endgroup$ – Brevan Ellefsen Oct 6 '15 at 6:25

Ansatz : Guess $y = e^{Ax}cos(e^{Bx})$. Then

$y = e^{Ax}cos(e^{Bx})$

$y' = Ae^{Ax}cos(e^{Bx}) - Be^{Ax+Bx}sin(e^{Bx})$.

$y'' =A^2 e^{A x} cos(e^{B x})-B^2 e^{A x+2 B x} cos(e^{B x})-B (2 A+B) e^{A x+B x} sin(e^{B x})$

From the cosine terms, we see that $e^{A+2} + (A+A^2)e^A - B^2e^{A+2B} = 0$.

From the sine terms, we get $-B-2AB - B^2 = 0$.

A solution that will work for the first equation (got from matching the powers) is $A = -1, B=1$. This also satisfies the second equation.

So, a solution is $y = e^{-x}cos(e^x)$. Use reduction of order to get the second linearly independent solution. (If you're tired, you can check it's just $e^{-x}sin(e^x)$.


Let me try with a change of variable, setting $$y=z(t(x))$$ Differentiating, we get $$y'=t'(x)\, z'(t(x))$$ $$y''=t''(x)\, z'(t(x))+t'(x)^2 \,z''(t(x))$$ After simplifications, the differential equation becomes $$t'(x)^2\, z''(t(x))+z'(t(x)) \left(t''(x)-t'(x)\right)+e^{2 x} \,z(t(x))=0$$ It is clear that the equation simplifies if we set $t''(x)=t'(x)$ and/or $t'(x)^2=e^{2x}$ that is to say $t(x)=e^x$. Using it, the equation write $$z''(t)+z(t)=0$$

I am sure that you can take from here.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.