Reduction of Order: Professor didn't have time to go over it, but still assigned it for hw. Anyone mind working me through a problem?
\[x^{2}y'' - xy' + 2y = 0; y _{1}= xsin(lnx)\]
And sorry if I reply slow, working a bit. Any help is appreciated still.
reduction of order tends to imply solving a u' equaivalent setup
Yeah, finding a way to makeit linear in u using a sub, correct?
yep
I just dont know the steps in doing that, lol. My book is pathetic :/
Pauls tends to be very readable to me http://tutorial.math.lamar.edu/Classes/DE/ReductionofOrder.aspx
assume y2 = u y1
Right, so y2 = uxsin(lnx)
and y2', y2'' from that gets us?
Oh joy. y' = u'xsin(lnx) + usin(lnx) + ucos(lnx) y'' = u''xsin(lnx) + u'sin(lnx) + u'cos(lnx) + u'sin(lnx) + ucos(lnx)/x + u'cos(lnx) - usin(lnx)/x I know Ill have to simplify that, assuming I did it right. But Yeah, plug it in... x^2(u''xsin(lnx)+2u'sin(lnx) + 2u'cos(lnx) + (u/x)[cos(lnx) - sin(lnx)) - x(u'xsin(lnx) + usin(lnx) + ucos(lnx)) + 2(uxsin(lnx) Yeah, thats awful.
:) \[u''x^3\sin(\ln x)+2u'x^2[\sin(\ln x)+\cos(\ln x)]-ux[\sin(\ln x)-\cos(\ln x)]\\ ~~~~~~~~~~~~~~~~~~~~~~-u'x^2[\sin(\ln x)]~~~~~~~~~~~~~~~~~-ux[\sin(\ln x+\cos(\ln x))]\\ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +2ux[\sin(\ln x)] \]
Ill take your word for it. Dunno if you simplifie it more or if I just made a mistake in all that mess. Bit distracted, so forgive me xD
\[u''x^3\sin(\ln x)+u'x^2\sin(\ln x)+2u'x^2\cos(\ln x)]+ux[\cos(\ln x)]\\ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-ux[\cos(\ln x))]\\\] \[u''x^3\sin(\ln x)+u'x^2[\sin(\ln x)+2\cos(\ln x)=0 \]
Nice :P
since the term in u drops out, we can "reduce" this by an order: let w=u'
looks like i forgot some brackets and what not in the latex code :) but i think its still readable
looks like i forgot some brackets and what not in the latex code :) but i think its still readable
\[w'x^3\sin(\ln x)+w~x^2[\sin(\ln x)+2\cos(\ln x)]=0\] \[w'\sin(\ln x)+\frac1xw[\sin(\ln x)+2\cos(\ln x)]=0\] \[w'+\frac1x[1+2cos(\ln x)]~w=0\]
cotangent in that last one, not cos
The term in u drops out? Obviously not talking about derivatives. So is that just when the u term with no derivatives drops? And you said because of that, does that mean its supposed to happen or did we get lucky with something?
its spose to happen in order to have the u'' and u' parts all alone
since the name of a variable is not important, we can state this in terms of say: u'' = w'; and u' = w to "reduce" the order
So integrating factor is xsin^2(x)
maybe ... with any luck :)
cot, not cos
Yeah, Im pretty sure I did it with cot, lol.
\[w'+\frac1x[1+2\cot(\ln x)]~w=0\]
So, assuming Im right, I would have just : \[\int\limits_{}^{}(wxsin^{2}x)'= 0 \] \[wxsin^{2}x = C\] Im just guessing that its supposed to do that, but im not 100% positive considering its 0 over there.
oops, meant the x in sin to be lnx, lol
\[\int~\frac1x+\frac2x\cot(\ln x)~dx\] i spose since the lnx, that makes the right part simpler to deal with \[\ln x+2\ln(\sin(\ln x))\]
Okay, so I got it Im pretty sure, just forgot to keep the angle of sin as lnx.
\[e^{\ln x+2\ln(\sin(\ln x))}w=C\] etc ... looks like a hoot
you did fine
since w=u', we can determine u; since we know y2 = u y1, we can then determine y2
Okay, cool. so....that means \[\mu'(x) = (wsin^{2}(lnx) + C)\]?
Or no, replace w with u', right
solve for w
yes, replace u' with the results for w
\[\mu' (x)\sin^{2}(lnx) = C\] \[\mu'(x) = C*\csc^{2}(lnx) \]
Integrate again, right?
\[w=C\frac{\csc^2(\ln x)}{x}\] \[u'=C\frac{\csc^2(\ln x)}{x}\]
integrate again, yes
Oops, x in there x_x
u = -C*cot(lnx) + C2
looking good; c1 and c2 might clean it up more
Right. \[y _{2}=(-C _{1}\cot(lnx) + C _{2})xsin(lnx)\]
We need a C1 and C2, though, right?
since c1 is arbitrary, then -c1 is redundant
yes, y'' tells us that there are 2 arbitrary constants
Right, lol. I just didnt want to mess up, so I put it in still xD
Okay, so the book just decided to pick 2 numbers for C1 and C2. Does that mean they can be any 2 numbers?
notice that y1 is incorporated into y2 at the moment, so the full solution would amount to y2
yes; and if we had some initial conditiions we could then specify values of c1 and c2 as well
Yeah, I wasnt sure because usually textbooks would put something liek "answers may vary" in the backbut all of these have specific answers, so i wasnt sure.
\[y = c_1 x \sin(\ln x)+c_2 x \cos(\ln x)\]
So I guess it doesnt matter what C1 and C2 I pick. Okay, cool. SO another thing I wasnt sure what to do with. Basically you do all the substituting and then replace u'' with w', but all I had left was \[w'e^{2x} = 0\]So how would I get my integrating factor out of that without a regular w term?
theres no need to play with that .... our goal was in determining \(u\): \[u'=w\] \[u=\int w\]
Oh O.o Okay, cool, wouldnt have know. Alright, thanks a lot for taking thte time ^_^
\[xy'' + y' = 0; y _{1}=lnx\] \[y_{2} = \mu(x)lnx\] \[y_{2}' = \mu'(x)lnx + \frac{ \mu(x) }{ x }\] \[y''_{2}=\mu''(x)lnx + \frac{ \mu'(x) }{ x }+\frac{ x \mu'(x)-\mu(x) }{ x^{2} }\] \[x(\mu''(x)lnx + \frac{\mu'(x)}{x}+\frac{x \mu'(x)- \mu(x)}{x^{2}})+\mu'(x)lnx+ \frac{\mu(x)}{x}=0\] \[x \mu''(x)lnx + (2+lnx) \mu'(x)\] \[\mu''(x) = w'(x) = \frac{ dw }{ dx }\] \[\mu'(x) = w(x)\] \[(xlnx)\frac{dw}{dx}+(2+lnx)w = 0\] \[\frac{ dw }{ w }+\frac{ (2+lnx)dx }{ xlnx }=0\] \[\int\limits_{}^{}\frac{dw}{w}=-\int\limits_{}^{}\frac{ 2+lnx }{ xlnx }dx\] u = 2+lnx du = (1/x)dx dx = xdu \[\ln|w| = \int\limits_{}^{}\frac{ u }{ x(u-2) }*xdu\] \[\ln|w| = -\int\limits_{}^{} 1du+2\int\limits_{}^{}\frac{ 1 }{ u-2 }du\] \[\ln|w| =-u-2\ln|u-2|+C\] \[\ln|w| = -2-lnx-2\ln|lnx| + C\] Ill stop there. In the end, it says y2 is 1 in my book :/ So I want to make sure I did this lead-up stuff correct and then yeah, somehow figure out how the book gets 1, lol.
\[xy'' + y' = 0;~y _{1}=lnx\] \[y=u~lnx\] \[y'=u'~lnx+ux^{-1}\] \[y''=u''~lnx+2u'x^{-1}-ux^{-2}\] --------------------- \[u''x~lnx+2u'-ux^{-1}+u'~lnx+ux^{-1}=0\] \[u''x~lnx+(2+lnx)u'=0\] ---------------------- \[w'x~lnx+(2+lnx)w=0\] \[w'+(\frac{2}{x~lnx}+\frac1x)w=0\] \[\frac 1w dw=-(\frac{2}{x~lnx}+\frac1x)dx \] \[ln~w=-2ln(ln~x)-ln~x+C \] \[w=C~\frac{1}{x~ln^{2}(x)} \] im up to here
\[u = C_1\frac{1}{ln(x)}+C_2\]
\[y=u lnx\] \[y=( C_1\frac{1}{ln(x)}+C_2) lnx\] \[y= C_1\frac{ln(x)}{ln(x)}+C_2 lnx\]
Join our real-time social learning platform and learn together with your friends!