I'm attempting to solve for $y$ as a function of $x$ if $y''''-12y'''+36y''=256e^{-2x}$, with initial conditions: $y(0)=5, y'(0)=-1, y''(0)=40, y'''(0)=-8$. I've tried to solve this multiple times and each time the solution I obtain is the same, although it's apparently incorrect. Here's my work so far:

$y(x)=y_h(x)+y_p(x)$. So first we solve for the homogeneous solution of $y''''-12y'''+36y''=0$. Let $y=e^{mx}$, and then we have $m^4-12m^3+36m^2=0 \implies m=0,0,6,6$. So our solution is of the form $y_h(x)=c_1e^{0x}+c_2xe^{0x}+c_3e^{6x}+c_4xe^{6x}$. Now we use initial conditions to solve for the constants:

$y(0)=5 \implies c_1+c_3=5$

$y_h'=c_2+6c_3e^{6x}+c_4e^{6x}+6c_4xe^{6x} \implies y'(0)=c_2+6c_3+c_4=-1$

$y_h''=36c_3e^{6x}+6c_4e^{6x}+6c_4e^{6x}+36c_4xe^{6x} \implies y''(0)=36c_3+12c_4=40$

$y_h'''=216c_3e^{6x}+36c_4e^{6x}+36c_4e^{6x}+36c_4e^{6x}=216c_4xe^{6x}\implies y'''(0)=216c_3+108c_4=-8$

Now we have a system of four equations, four unknowns, and solving , we find that $c_1=\frac{43}{27}, c_2=\frac{-131}{9}, c_3=\frac{92}{27}, c_4=\frac{-62}{9}$ (I have verified this with WolframAlpha) and so $y_h(x)=\frac{43}{27}-\frac{131}{9}x+\frac{92}{27}e^{6x}-\frac{62}{9}xe^{6x}$

Now to solve for the particular solution $y_p(x)$ we use the method of undetermined coefficients: $y_p$ is of the form $ae^{-2x}$. $y_p''=4ae^{-2x}, y_p'''=-8ae^{-2x}, y_p''''=16ae^{-2x}$. Plugging these into the D.E.:

$16ae^{-2x}-12\cdot-8ae^{-2x}+36\cdot4ae^{-2x}=256ae^{-2x}=256e^{-2x} \implies a=1$

Thus $y(x)=\frac{43}{27}-\frac{131}{9}x+\frac{92}{27}e^{6x}-\frac{62}{9}xe^{6x} + e^{-2x}$

However, according to the program we use for assignments, this incorrect. If anyone could find my error it would be greatly appreciated.

  • 1
    $\begingroup$ Your (first) mistake is in the form of $y_h$. Since $0$ is a repeated root, you need an $xe^{0x}$ term. But a more important mistake is that you shouldn't solve for the coefficients of $y_h$ until you have combined it with $y_p$, as it is the sum that must satisfy those initial conditions. $\endgroup$ – Gerry Myerson Mar 1 '15 at 8:06
  • 1
    $\begingroup$ @GerryMyerson Forgot that in the beginning, but I did use that in my solution, if you look at the final solution. I had a realization: is my error in finding the coefficients only for $y_h$ and not after I'd found $y_p$? $\endgroup$ – jofl Mar 1 '15 at 8:09
  • $\begingroup$ This is the reason. If you use your last results, the conditions are not satisfied. $\endgroup$ – Claude Leibovici Mar 1 '15 at 8:20
  • $\begingroup$ Ah, silly mistake, but great to learn. Thank you. $\endgroup$ – jofl Mar 1 '15 at 8:21

Answer: $$y=-6 x\, {{ e}^{6 x}}+3 {{ e}^{6 x}}+{{ e}^{-2 x}}-11 x+1$$

Solved with CAS Maxima.

  • $\begingroup$ I promise you, I will download and try it :-) But please admit, it is really not a very HQ answer. I hate to click "Recommend Deletion", particularly for posts with a clearly visible goodwill. I suggest this: 1) write a lot HQ answers 2) meanwhile, if you find such, work-intensive questions, solve them with CAS Maxima and put the result in a comment. | You can also put its link into your profile, if you are active, you will have a lot of visitors. $\endgroup$ – peterh Jun 4 '18 at 21:40
  • $\begingroup$ @peterh load(odes); eq:'diff(y,x,4)-12*'diff(y,x,3)+36*'diff(y,x,2)=256*exp(-2*x); expand(odeL_ic(eq,y,x,[0,5,-1,40,-8])); $\endgroup$ – Aleksas Domarkas Jun 4 '18 at 21:59
  • $\begingroup$ The problem with it is that it is understable only for the people knowing that CAS Maxima. Also I don't know it (yet). $\endgroup$ – peterh Jun 4 '18 at 22:02
  • $\begingroup$ Maxima CAS is free computer algebra system. maxima.sourceforge.net $\endgroup$ – Aleksas Domarkas Jun 4 '18 at 22:04
  • $\begingroup$ I know, and probably all the reviewers of the site already know it :-) $\endgroup$ – peterh Jun 4 '18 at 22:06

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.