|
Post by Rob on Mar 11, 2006 23:19:00 GMT 1
Please! I have this inhomogenous second order differential equation and I keep mucking it up. You mathemagicians are such clever people... a hint would be useful!
d^2y/dx^2 + 4y = 2sin(2x)
Thank you.
|
|
|
Post by Ben on Mar 12, 2006 1:06:55 GMT 1
The solution of the homogeneous solution is: y=Asin(2x)+Bcos(2x) === A*cos(2x+B) ((although be careful to note that they're not the same A and B)) then the particular solution is done by taking y=C*x*sin(2x)+D*x*cos(2x) and plugging it into the ODE and equating LHS and RHS. dy/dx = C(sin(2x)+2x*cos(2x))+D(cos(2x)-2x*sin(2x)) d(dy/dx)/dx = C(cos(2x)+2(cos(2x)-2x*sin(2x))-D(sin(2x)+2(sin(2x)+2x*cos(2x))) trying to solve that horrific algebraic mess isn't fun. However, (Dy)^2=C*cos(2x)+2*C*cos(2x)-4*C*x*sin(2x)-D*sin(2x)-2*D*sin(2x)-4*D*x*cos(2x) (Dy)^2=Cos(2x)[C+2C-4Dx]-sin(2x)[D+2D-4Cx] So then we have to add in the y term, giving us LHS = Cos(2x)[3C-4Dx]-Sin(2x)[3D-4Cx]+Cxsin(2x)+Dxcos(2x) which simplifies a bit to Cos(2x)[3C-3Dx]-Sin(2x)[3D-5Cx] So we have: Cos(2x)[3C-3Dx]-Sin(2x)[3D-5Cx] = 2Sin(2x) Then i forget what to do This is probably all wrong anyway. Arse...
|
|
|
Post by Rob on Mar 12, 2006 1:12:37 GMT 1
Yeah, I got something similar... the problem is, that when equating the coefficients C comes out as a number AND zero and so does D. I hate my fucking maths lecture person. He likes to test us on how we deal with the longest line of simple algebra as possible. The only joy he has brought to me is that his name sounds exactly the same as the noise Mario and Wario make on Mario Party...
GABOOOOORRR!
|
|
|
Post by Ben on Mar 13, 2006 12:36:32 GMT 1
haha, That's not his first name is it? because we have a lecturer called Gabor Megyesi. Is your lecturer bald, wears glasses, russian, and NASAL?
|
|
|
Post by oasis2000 on Mar 13, 2006 13:06:35 GMT 1
|
|
|
Post by Ben on Mar 13, 2006 13:09:28 GMT 1
Oh, i've solved it, using a maths package, but I just wanted to know how to get the answer, i suppose.
|
|
|
Post by Rob on Mar 13, 2006 16:57:40 GMT 1
For those interested (i.e. Ben), I think this is the answer. As ever, it was silly maths mistakes that put me into the wrong to start with...
y'' + 4y = 2sin(2x)
Complementary function:
y = Acos(2x) + Bsin(2x)
Particular integral:
y = Cxcos(2x) + Dxsin(2x)
y'' = -2Csin(2x) - 2Csin(2x) - 4Cxcos(2x) + 2Dcos(2x) + 2Dcos(2x) -4Dxsin(2x)
Now, substituting these back into the original equation and equating the coefficents, we get
C = -0.5 and D = 0
So the particular integral is
y = -0.5xcos(2x)
And the general solution is this plus the complementary function
y = -0.5cos(2x) + Acos(2x) + Bsin(2x)
Gabor makes life so boring. I had to take a 4x4 matrix to the power of three by hand too, and lots more boring, long things...
|
|
|
Post by Ben on Mar 14, 2006 21:34:02 GMT 1
haha. the method i was about to suggest for making life simpler involved just as many calculations, so never mind. But if you decomposed A=PDP T, where P was orthogonal and D was diagonal, then A 3=PD 3P T, which is pretty cool. Hmm, i must have made a mistake as well then. Buggery. Still, never mind, i knew the technique and that's all that I needed!
|
|