Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

So the equations is $y''-5y'+6y=g(t)$

I found the characteristic equation to be $r^{2}-5r+6=0$ which factors to $(r-3)(r-2)$

So the fundamental set of solutions to the homogeneous equation is:

$y_{1}=e^{3t}$ and $y_{2}=e^{2t}$

Than I know I had to find $y_{p}$ so I could find the general solution $y(t)=c_{1}y_{1}+c_{2}y_{2}+y_{p}(t)$

\begin{equation*} \begin{pmatrix} e^{3t} & e^{2t}\\ 3e^{3t} & 2e^{2t} \end{pmatrix} \begin{pmatrix} u_{1}'\\ u_{2}' \end{pmatrix} = \begin{pmatrix} 0\\ g(t) \end{pmatrix}. \end{equation*}

I found the inverse of the first matrix and solved for $u_{1}'$ and $u_{2}'$ getting



But after this I am a little stuck. I know the general solution will have integrals in it since you don't really know what g(t) is but the book gives the answer with another variable s and I don't know where it comes from. This is the solution the book gives:

$y=c_{1}e^{2t}+c_{2}e^{3t}+\int{[e^{3(t-s)}-e^{2(t-s)}]g(s) ds}$

where I just have $y_{1}$ and $y_{2}$ switched.

share|cite|improve this question
The variable $s$ is probably the integration variable, no? – 1015 Feb 25 '13 at 22:18
like having u' be the same as du/ds and treating it as a separable diff eq? – USC Feb 25 '13 at 22:19
If you want. But, simply, integrate $u'(s)=e^{-3s}g(s)$ between $0$ and $t$, to get $u(t)-u(0)=\int_0^te^{-3s}g(s)ds$ by the fundamental theorem of calculus. – 1015 Feb 25 '13 at 22:21
Why don't you post the solution? – Mhenni Benghorbal Feb 25 '13 at 22:24

According to the method of variation of parameters, particular solution of 2$^{nd}$ order ODE $$ y''(t)+py'(t)+qy = f(t) $$ is $$ y_p = -y_1(t) \int \frac {y_2(s) f(s)}{W(s)} ds+y_2(t) \int \frac {y_1(s)f(s)}{W(s)}ds $$ where $y_1(t)$ and $y_2(t)$ are fundamental solutions of homogeneous problem, and $W$ is a Wronskian of $y_1$ and $y_2$: $$ W(t) = \left | \begin{array}{cc} y_1 & y_2 \\ y_1' & y_2' \end{array}\right | = y_1y_2'-y_2y_1' $$ In your case $y_1 = e^{2t}$ and $y_2 = e^{3t}$, so $$ W(t) = 3e^{2t}e^{3t} - 2e^{3t}e^{2t} = e^{5t} $$ and therefore $$ y_p = -e^{2t}\int \frac{e^{3s}g}{e^{5s}}ds+e^{3t}\int \frac {e^{2s}g}{e^{5s}}ds = \int{e^{3(t-s)}}gds-\int e^{2(t-s)}gds=\int \left [ e^{3(t-s)}-e^{2(t-s)}\right]gds $$

share|cite|improve this answer

The $s$ you saw likely referred to a Laplace transform variable; the LT was used to solve this problem.

The Laplace transform of a function $f(t)$ is

$$\hat{f}(s) = \int_0^{\infty} dt \: f(t) e^{-s t}$$

When $f$ is differentiable over $[0,\infty)$, we have $\hat{f'}(s) = -f(0) + s \hat{f}(s)$, and etc. for higher derivatives.

You may then show that your differential equation takes the form

$$(-y'(0) - sy(0) + s^2 \hat{y}(s)) - 5 (-y(0)+s \hat{y}(s)) + 6 \hat{y}(s) = \hat{g}(s)$$

or, solving for $\hat{y}(s)$:

$$\hat{y}(s) = \frac{y'(0) + (s-5) y(0) + \hat{g}(s)}{s^2-5 s+6}$$

To find $y(t)$, you need the inverse Laplace transform:

$$y(t) = \frac{1}{i 2 \pi} \int_{a-i \infty}^{a+i \infty} ds \frac{y'(0) + (s-5) y(0) + \hat{g}(s)}{s^2-5 s+6} e^{s t}$$

where $a>0$. Not sure you've seen this form of the inverse; perhaps you looked the inverses up in a table, which is OK, too. In any case, I will not detail the way I got the answer, which involves the residue theorem, of which you may or may not have a working knowledge. My answer is

$$y(t) = (y'(0)-2 y(0) + \hat{g}(3)) e^{3 t} - (y'(0) - 3 y(0) + \hat{g}(2)) e^{2 t}$$


$$\hat{g}(3) = \int_0^{\infty} dt' \: g(t') e^{-3 t'}$$ $$\hat{g}(2) = \int_0^{\infty} dt' \: g(t') e^{-2 t'}$$

The solution is now specified completely in terms of the initial conditions, the homogeneous solution, and integrals over the driving function. You should be able to see that this solution is equivalent to what you posted, with $c_1 = - (y'(0) - 3 y(0))$ and $c_2 = (y'(0)-2 y(0))$.


I see that the title states "using variation of parameters," so I may have goofed in assuming that the LT was used. That said, this is a very worthwhile approach to solving these equations, especially when there is a forcing term as you have here.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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