Find a particular solution for the differential equation by the method of undetermined coefficients. $$2y'' - 16y' + 32y = -e^{4x}$$ Also, find the general solution of this equation.

The steps I took to solve this problem,

Find the auxiliary equation which is $2m^2-16m+32=0$ for which the roots are $m_1=4$ and $m_2=4$ so $m=4$ of multiplicity 2.

Solve for a general equation of $y_h(x) = C_1e^{4x}x + C_2e^{4x}$

When I try to find a particular solution by taking the derivates of the right hand side, I get \begin{align} y_p &= Ae^{4x}\\ y_p' &= 4Ae^{4x}\\ y_p'' &= 16Ae^{4x} \end{align} Substituting these values into the left hand side results in $0 = -e^{4x}$ which is not possible. Can someone identify what I am missing?


Your problem is $e^{4x}$ is part of your homogeneous solution, which explains why you get $0$ when you try $Ae^{4x}$ as a particular solution.

I would try $Ax^2e^{4x}$ as a particular solution.

Just to show it works, let me show you another way to solve the problem. Let

$$z=y'-4y,z'=y''-4y'$$ $$y''-8y'+16y=-\frac12e^{4x}$$ $$(y'-4y)'-4(y'-4y)=z'-4z=-\frac12e^{4x}$$ $$e^{-4x}z'-4e^{-4x}z=(e^{-4x}z)'=-\frac12$$ $$e^{-4x}z=-\frac12x+k_1,z=-\frac12xe^{4x}+k_1e^{4x}$$ $$y'-4y=-\frac12xe^{4x}+k_1e^{4x}$$ $$e^{-4x}y'-4e^{-4x}y=(e^{-4x}y)'=-\frac12x+k_1$$ $$e^{-4x}y=-\frac14x^2+k_1x+k_2,y=-\frac14x^2e^{4x}+k_1xe^{4x}+k_2e^{4x}$$

So our particular solution turns out to be $-\frac14x^2e^{4x}$

  • $\begingroup$ Why would we do this? $\endgroup$ – Mike Beta Nov 20 '12 at 6:09
  • $\begingroup$ We start with a different starting point because a double root does not give us the two Linearly Independent solutions we are looking for. $\endgroup$ – Aseem Dua Nov 20 '12 at 6:12
  • $\begingroup$ @MikeBeta Often in these kinds of problems, the solution is to multiply by $x$ when one particular solution doesn't work. $xe^{4x}$ is also part of the homogeneous solution, so multiply by $x$ again. If you want to try solving this problem a different way, try making the substitution $z=y'-4y$. It might help you understand. $\endgroup$ – Mike Nov 20 '12 at 6:16

In case of a double root, the method fails, we need to start over with a different assumption of the solution. This time though, we approach the problem in a more informed manner.

hint: when we solve a simple homogeneous system, when we have a double root, to obtain a second LI solution we simply multiply root by x.

  • $\begingroup$ This relates to the multiplicity of the root? Does the same happen for when the root is m = 4, 4, 4? $\endgroup$ – Mike Beta Nov 20 '12 at 6:10
  • $\begingroup$ You start with an "anticipated" solution based on you qualitative understanding of the system. If it doesn't satisfy your requirements, you make an understanding and proceed to a next best possible attempt. In this case multiplicity of roots gives an understanding that the system will behave differently than "anticipated". If you are working with a higher order differential eq than 2, according to this logic, i feel yes you may have to change the initial attempt in case of a triple root. $\endgroup$ – Aseem Dua Nov 20 '12 at 6:19
  • $\begingroup$ Another thing is that, you have to proceed with an iterative process till you find a suitable starting solution. $\endgroup$ – Aseem Dua Nov 20 '12 at 6:26

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.