MHB -17.2.05 by undetermined coefficients.

karush
Gold Member
MHB
Messages
3,240
Reaction score
5
$\textrm{Solve the given equation by the method of undetermined coefficients.}$
\begin{align*}\displaystyle
y'' +6y'&=-4xe^{-6x}\\
y_p&=Ax^2e^{-6x}+Bxe^{-6x}
\end{align*}ok just wanted get this posted before I leave campus
so assume finding zeros is next
 
Physics news on Phys.org
Yes, in order to get the general solution to the given ODE, you will need the homogeneous solution, and this involves finding the roots (zeroes) of the characteristic equation. :D
 
$\tiny{17.2.05}$
\begin{align*}\displaystyle
y'' +6y'&=-4xe^{-6x}\\
\end{align*}
$\textrm{So the auxiliary equation is}$
\begin{align*}\displaystyle
r^2+6r&=0\\ r(r+6)&=0\\ r&=0\\ r&=-6
\end{align*}
$\textrm{then}$
\begin{align*}\displaystyle
y_h&=c_1e^{0x}+c_2e^{-6x}=c_1e+c_2e^{-6x}\\
y_p&=Ax^2e^{-6x}+Bxe^{-6x}
\end{align*}
is $y_p$ correct?
 
Last edited:
karush said:
$\tiny{17.2.05}$
\begin{align*}\displaystyle
y'' +6y'&=-4xe^{-6x}\\
\end{align*}
$\textrm{So the auxiliary equation is}$
\begin{align*}\displaystyle
r^2+6r&=0\\ r(r+6)&=0\\ r&=0\\ r&=-6
\end{align*}
$\textrm{then}$
\begin{align*}\displaystyle
y_h&=c_1e^{0x}+c_2e^{-6x}=c_1e+c_2e^{-6x}\\
y_p&=Ax^2e^{-6x}+Bxe^{-6x}
\end{align*}
is $y_p$ correct?

You correctly found the characteristic roots, and so that means the homogeneous solution is:

$$y_h(x)=c_1e^{0x}+c_2e^{-6x}=c_1+c_2e^{-6x}$$

The mistake you made was in using $e^0=e$ when we actually have $e^0=1$ :D

Now, if we were not given the form for the particular solution, we would use:

$$y_p(x)=x^s\left(Ax+B\right)e^{-6x}$$

Since one of the terms of the homogeneous solution is $c_2e^{-6x}$, we need to let $s=1$ so that no term in the particular solution is in the homogeneous solution, and so we have:

$$y_p(x)=x\left(Ax+B\right)e^{-6x}$$

And this is equivalent to what you stated. Now you need to determine $A$ and $B$ using the method of undetermined coefficients.
 
\begin{align*}\displaystyle
y_p&=Ax^2e^{-6x}+Bxe^{-6x}\\
G(x)&=-4xe^{-6x}
\end{align*}
So
\begin{align*}\displaystyle
y_p(x)&=Ax^2+Bx+C\\
y_p^{'}& =2Ax+B\\
y_p^{''}& =2A
\end{align*}
substituting into the differential equation
$$(2A)+(2Ax+B)-2(Ax^2+Bx+C)=-4xe^{-6x}$$
got this from an example but ??
 
We have:

$$y_p(x)=x\left(Ax+B\right)e^{-6x}$$

Using the general product rule for differentiation, we may state:

$$y_p'(x)=\left(Ax+B\right)e^{-6x}+Axe^{-6x}-6x\left(Ax+B\right)e^{-6x}=e^{-6x}\left(Ax+B+Ax-6x(Ax+B)\right)=\left(-6Ax^2+2(A-3B)x+B\right)e^{-6x}$$

And hence:

$$y_p''(x)=\left(-12Ax+2(A-3B)\right)e^{-6x}-6\left(-6Ax^2+2(A-3B)x+B\right)e^{-6x}=2\left(18Ax^2-6(2A-3B)x+A-6B\right)e^{-6x}$$

Now, you want to substitute for $y_p''$ and $y_p'$ into the ODE, and then continue with the process of using the method of undetermined coefficients. :D
 
karush said:
this ?

$y'' +6y'=-4xe^{-6x}$
$(-6Ax^2+2(A-3B)x+B)e^{-6x}+6[2(18Ax^2-6(2A-3B)x+A-6B)e^{-6x}]=-4xe^{-6x}$

Yes. :D
 
$$(-6Ax^2+2(A-3B)x+B)e^{-6x}+6[2(18Ax^2-6(2A-3B)x+A-6B)e^{-6x}]=-4xe^{-6x}\\
\textrm{expanded}$$
$$210Ae^{-6x}x^2-142Ae^{-6x}x+12Ae^{-6x}+210Be^{-6x}x-71Be^{-6x}
=-4e^{-6x}x$$
$$210Ax^2-142Ax+12A+210Bx-71B=-4x$$
 
Last edited:
I mistakenly told you "Yes" when I should have pointed out that you substituted incorrectly (sorry about that :o)...you should have:

$$2\left(18Ax^2-6(2A-3B)x+A-6B\right)e^{-6x}+6\left(\left(-6Ax^2+2(A-3B)x+B\right)e^{-6x}\right)=-4xe^{-6x}$$

Divided through by $4^{-6x}$ since it cannot be zero:

$$2\left(18Ax^2-6(2A-3B)x+A-6B\right)+6\left(\left(-6Ax^2+2(A-3B)x+B\right)\right)=-4x$$

Arrange both sides in standard form:

$$-12Ax+2A-6B=-4x+0$$

Okay, now what do you get when you equate coefficients?
 
  • #10
karush said:
$\displaystyle-12Ax+2A-6B=-4x+0$\\
$\displaystyle x=0$\\
$\displaystyle 2A-6B=0$\\
$\displaystyle x=1$\\
$\displaystyle -12A+2A-6B=0$\\
$\displaystyle -10A-6B=4$
this doesn't look to good!

What you want to do is equate corresponding coefficients, like so:

$$-12A=-4$$

$$2A-6B=0$$

Now you can uniquely determine $A$ and $B$...:D
 
  • #11
When Karush set x= 0, he got the "constant terms", 2A- 6B= 0. When he set x= 1 he got the sum of the constant term and the coefficient of x, -10A- 6B= 4, effectively the same thing. I don't know why he says "this doesn't look too good".

Subtracting the second equation from the first, 12A= -4 so A= -1/3. Then 2A- 6B= -2/3- 6B= 0 so 6B= -2/3 and B= -1/9[FONT=MathJax_Main]−[FONT=MathJax_Main]10[FONT=MathJax_Math]A[FONT=MathJax_Main]−[FONT=MathJax_Main]6[FONT=MathJax_Math]B[FONT=MathJax_Main]=[FONT=MathJax_Main]4[FONT=MathJax_Main]−[FONT=MathJax_Main]10[FONT=MathJax_Math]A[FONT=MathJax_Main]−[FONT=MathJax_Main]6[FONT=MathJax_Math]B[FONT=MathJax_Main]=[FONT=MathJax_Main]4
 
  • #12
updated
$A=\frac{1}{3}$
$B=\frac{1}{9}$
so
$$\displaystyle
y_p=\frac{1}{3}x^2e^{-6x}+\frac{1}{9}x e^{-6x}$$
 
Last edited:
  • #13
karush said:
I got

$A=-\frac{1}{3}$
$B=-\frac{1}{9}$

i was expecting integers

I got:

$$(A,B)=\left(\frac{1}{3},\frac{1}{9}\right)$$

And so, this would give me:

$$y_p(x)=x\left(\frac{1}{3}x+\frac{1}{9}\right)e^{-6x}=\frac{x}{9}(3x+1)e^{-6x}$$

Now, suppose I wish to check my answer, I would then compute:

$$y_p'=\frac{1}{9}(1-18x^2)e^{-6x}$$

$$y_p''=\frac{2}{3}(18x^2-6x-1)e^{-6x}$$

Now plug into the LHS of the original ODE and simplify to make sure we get the RHS of the original ODE:

$$\frac{2}{3}(18x^2-6x-1)e^{-6x}+6\left(\frac{1}{9}(1-18x^2)e^{-6x}\right)=\frac{2}{3}e^{-6x}\left(18x^2-6x-1+1-18x^2\right)=\frac{2}{3}e^{-6x}\left(-6x\right)=-4xe^{-6x}\quad\checkmark$$

So, we find this particular solution works. Check your algebra to find where you went wrong with the signs. :D
 
  • #14
much mahalo
these are ? max problems

note
I had to repeatedly log in?
 
  • #15
Here's an alternative approach to working the problem...we are given:

$$y''+6y'=-4xe^{-6x}$$

Let:

$$u=y'\implies u'=y''$$

And we have:

$$u'+6u=-4xe^{-6x}$$

Now, if we multiply through by an integrating factor of:

$$\mu(x)=e^{6x}$$

We obtain:

$$e^{6x}u'+6e^{6x}u=-4x$$

The LHS may now be rewritten as the differentiation of a product:

$$\frac{d}{dx}\left(e^{6x}u\right)=-4x$$

Integrate w.r.t $x$:

$$e^{6x}u=-2x^2+c_1$$

And so we have:

$$u=-2x^2e^{-6x}+c_1e^{-6x}$$

Back-substitute for $u$:

$$y'=-2x^2e^{-6x}+c_1e^{-6x}$$

Integrate w.r.t $x$:

$$y(x)=\int -2x^2e^{-6x}+c_1e^{-6x}\,dx=-2\int x^2e^{-6x}\,dx-\frac{1}{6}c_1e^{-6x}$$

Now, if we note that the factor in the second term of $$-\frac{1}{6}c_1$$ is just an arbitrary constant, we may write:

$$y(x)=-2\int x^2e^{-6x}\,dx+c_2e^{-6x}$$

Now, on the remaining integral, let's use IBP, where:

$$u=x^2\implies du=2x\,dx$$

$$dv=e^{-6x}\,dx\implies v=-\frac{1}{6}e^{-6x}$$

And so we have:

$$y(x)=-2\left(-\frac{1}{6}x^2e^{-6x}+\frac{1}{3}\int xe^{-6x}\,dx\right)+c_2e^{-6x}=\frac{1}{3}x^2e^{-6x}-\frac{2}{3}\int xe^{-6x}\,dx+c_2e^{-6x}$$

For the remaining integral, let's use IBP again, where:

$$u=x\implies du=dx$$

$$dv=e^{-6x}\,dx\implies v=-\frac{1}{6}e^{-6x}$$

And so we have:

$$y(x)=\frac{1}{3}x^2e^{-6x}-\frac{2}{3}\left(-\frac{1}{6}xe^{-6x}+\frac{1}{6}\int e^{-6x}\,dx\right)+c_2e^{-6x}=\frac{1}{3}x^2e^{-6x}+\frac{1}{9}xe^{-6x}+\frac{1}{18^2}e^{-6x}+c_1+c_2e^{-6x}$$

$$y(x)=c_1+\left(c_2+\frac{1}{18^2}\right)e^{-6x}+\frac{x}{9}\left(3x+1\right)e^{-6x}$$

Noting that $$\left(c_2+\frac{1}{18^2}\right)$$ is still just an arbitrary constant, we may write:

$$y(x)=c_1+c_2e^{-6x}+\frac{x}{9}\left(3x+1\right)e^{-6x}$$

And this is equivalent to the solution obtain through the method of undetermined coefficients (but with a lot more work). :D
 
  • #16
MarkFL said:
Divided through by $4^{-6x}$ since it cannot be zero:

$$2\left(18Ax^2-6(2A-3B)x+A-6B\right)+6\left(\left(-6Ax^2+2(A-3B)x+B\right)\right)=-4x$$

did you mean $e^{-6x}$
 
  • #17
karush said:
did you mean $e^{-6x}$

Yes, 'twas a typo. :D
 
Back
Top