Ask your own question, for FREE!
Mathematics 8 Online
OpenStudy (idealist10):

Find all solutions of xy'=2-x+(2x-2)y-xy^2.

OpenStudy (anonymous):

Use the homogeneous sub, \(y=vx\).

OpenStudy (idealist10):

But do I divide by x first?

OpenStudy (anonymous):

Yes, it should give you a linear equation in \(v\).

OpenStudy (anonymous):

Oh wait it might not be linear...

OpenStudy (idealist10):

This is what I did. xy'=2-x+2xy-2y-xy^2 xy'=(xy^2-2xy+x)+2(1-y) xy'=x(y-1)^2-2(y-1) y'=(y-1)^2-2(y-1)/x So from there, do I plug in y=vx?

OpenStudy (anonymous):

Just a sec. The homogeneous sub might not work here...

OpenStudy (idealist10):

I use the substitution u=y-1 and u'=y' and u'=u^2-2u/x, v=1/u, v'=-1/u^2*u' and v'=-1/u^2(u^2-2u/x)=-1+2/ux=2v/x-1 and w=2v/x-1 and w'=-2/x^2*v' but it doesn't work.

OpenStudy (idealist10):

The answer in the book is y=1-1/(x(1-cx)).

OpenStudy (anonymous):

I think I might have it, but I need to change computers. This one is being very slow.

OpenStudy (idealist10):

@SithsAndGiggles

OpenStudy (idealist10):

Sorry, I can't see anything. Have you posted your reply?

OpenStudy (anonymous):

Hey, sorry for the wait. I was checking to see if this is a particular type of ODE, and it's called a Ricatti equation, which has the form \[y'=f_0(x)+f_1(x)y+f_2(x)y^2\] i.e. a nonlinear equation that's quadratic in the dependent variable \(y(x)\). We want to be able to reduce this into a linear 2nd order equation. Divide by \(x\): \[y'=\frac{2-x}{x}+\frac{2x-2}{x}y-y^2\] Substitute \(v=yf_2(x)=-y\), so \(v'=-y'\). \[\begin{align*}-v'&=\frac{2-x}{x}-\frac{2x-2}{x}v-v^2\\\\ v'&=\frac{x-2}{x}+\frac{2x-2}{x}v+v^2\end{align*}\] Next you substitute \(v=-\dfrac{t'}{t}\), which gives \(v'=\dfrac{(t')^2-t~t''}{t^2}\). \[\begin{align*} v'&=\frac{x-2}{x}+\frac{2x-2}{x}v+v^2\\\\ \frac{(t')^2-t~t''}{t^2}&=\frac{x-2}{x}+\frac{2x-2}{x}\frac{t'}{t}+\left(\frac{t'}{t}\right)^2\\\\ -\frac{t''}{t}&=\frac{x-2}{x}+\frac{2x-2}{x}\frac{t'}{t}\\\\ -t''&=\frac{x-2}{x}t+\frac{2x-2}{x}t'\\\\ t''+\frac{2x-2}{x}t'+\frac{x-2}{x}t&=0\end{align*}\] To give the equation constant coefficients, we'll substitute \(t=\dfrac{p}{x}\), so \(t'=\dfrac{xp'-p}{x^2}\) and \(t''=\dfrac{x^2p''-2xp'+2p}{x^3}\). \[\begin{align*}\frac{1}{x}p''-\frac{2}{x^2}p'+\frac{2}{x^3}p-\frac{2x-2}{x^2}p'+\frac{2x-2}{x^3}p+\frac{x-2}{x^2}p&=0\\\\ \frac{1}{x}p''-\frac{2}{x}p'+\frac{1}{x}p&=0\\\\ p''-2p'+p&=0\end{align*}\]

OpenStudy (idealist10):

That's fine. We all know that technology has always been an issue at times. Can you please take a snapshot for what you just typed because my computer currently has 0% for processing math.

OpenStudy (anonymous):

OpenStudy (idealist10):

So p^2-2p+1=0 (p-1)(p-1)=0 p=1

OpenStudy (anonymous):

No, that's not an equation that's quadratic in \(p\); there are derivatives involved. It's a homogeneous linear equation. Find the characteristic equation and its roots.

OpenStudy (idealist10):

But if v=-t'/t, shouldn't \[v'=\frac{ t(-t'')+t' }{ t^2 }\]

OpenStudy (anonymous):

\[\frac{d}{dx}\left[-\frac{t'}{t}\right]=-\left(\frac{t\dfrac{d}{dx}[t']-t'\dfrac{d}{dx}[t]}{t^2}\right)=-\frac{t~t''-(t')^2}{t^2}=\frac{(t')^2-t~t''}{t^2}\]

OpenStudy (anonymous):

Screenshot?

OpenStudy (idealist10):

I see. So after finding the characteristic equation and the roots, you have to go back to all the substitutions to find the equation of y, right?

OpenStudy (anonymous):

Yes, might take a while...

OpenStudy (idealist10):

But I found that\[-\frac{ t" }{ t }=\frac{ x-2 }{ x }-\frac{ 2x-2 }{ x }*\frac{ t' }{ t }\]

OpenStudy (idealist10):

What you have is \[-\frac{ t" }{ t }=\frac{ x-2 }{ x }-\frac{ 2x-2 }{ x }*\frac{ t' }{ t }\]

OpenStudy (anonymous):

Yes you're right. It looks like I made that error but corrected it in the end... Didn't even notice it. The equation in \(p\) and its derivatives is correct though.

OpenStudy (idealist10):

But I wonder how y=ce^x+cxe^x is the solution. The root is 1...

OpenStudy (anonymous):

I'll be back in a few minutes. The hard part's over.

OpenStudy (idealist10):

Okay.

OpenStudy (anonymous):

Okay so\[p''-2p'+p=0\] gives the characteristic equation, \[r^2-2r+1=0\] with root \(r=1\), multiplicity 2, hence the general solution \[p=C_1e^x+C_2xe^x\] Back-subbing to \(t\) gives \[\begin{align*}xt&=C_1e^x+C_2xe^x\\\\ t&=C_1\frac{e^x}{x}+C_2e^x \end{align*}\] Back-subbing to \(v\): \[\begin{align*} t&=C_1\frac{e^x}{x}+C_2e^x\\\\ \color{gray}{t'}&\color{gray}{=C_1\frac{x-1}{x^2}e^x+C_2e^x}\\\\ v&=-\frac{C_1\dfrac{x-1}{x^2}e^x+C_2e^x}{C_1\dfrac{e^x}{x}+C_2e^x} \end{align*}\] and finally back-subbing to \(y\): \[\begin{align*} y&=\frac{C_1\dfrac{x-1}{x^2}e^x+C_2e^x}{C_1\dfrac{e^x}{x}+C_2e^x} \end{align*}\] And you should be able to simplify that a bit.

OpenStudy (idealist10):

THANK YOU SO MUCH! YOU'RE AWESOME!

OpenStudy (anonymous):

You're welcome!

Can't find your answer? Make a FREE account and ask your own questions, OR help others and earn volunteer hours!

Join our real-time social learning platform and learn together with your friends!
Can't find your answer? Make a FREE account and ask your own questions, OR help others and earn volunteer hours!

Join our real-time social learning platform and learn together with your friends!