Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon.

Pages: 1-

Diff. Eq.

Name: Anonymous 2007-09-02 10:03 ID:D6QdaoP5

I'm at a loss about how to solve the following differential equation:

x'' = x^0.25

That is, double derivative of x equals x to the 0.25th power.

It's very frustrating because I have a feeling it's very trivial. Can anybody help, please? Thanks.

Name: Anonymous 2007-09-02 11:24 ID:ZY7vOegH

0.25x?

Name: Anonymous 2007-09-02 12:45 ID:D6QdaoP5

Differential equations work that way. x is a function, over a variable (y, t, or whatever). The solution will be x(y,c), that is, a function of the variable and an arbitrary constant. Well, two arbitrary constants, since it's a second derivative. I think.

Name: Anonymous 2007-09-02 12:46 ID:D6QdaoP5

Meant to say "don't work that way".

Name: Anonymous 2007-09-02 13:16 ID:c8mw+SRX

(16*x^(9/4)/45)

Name: Anonymous 2007-09-02 13:22 ID:c8mw+SRX

int(x^0.25)=0.8x^1.25

int(0.8x^1.25)=(16*x^(9/4))/45

Name: Anonymous 2007-09-02 13:35 ID:D6QdaoP5

No, no, no.

To check if a answer "f" is valid, you must:

1. Calculate "f" to the 0.25th power
2. Calculate the second derivative of "f"
3. The results from 1 and 2 must be equal

Name: Anonymous 2007-09-02 16:29 ID:yr5AgAKM

constant of integration, anyone?

x'' = x^(1/4)

x' = (4/5)x^(5/4) + c

x = (4/5)(4/9)x^(9/4) + cx + d

Name: Anonymous 2007-09-02 16:44 ID:vxqCpsDY

cant you jus revers solve this thing?

x'' = x ^ 0.25
x'' = (x^1/8)^2
x' = (1/3)(x^1/8)^3
x = (1/4)(1/3)(x^1/8)^3

Name: Anonymous 2007-09-02 16:56 ID:vxqCpsDY

oops

x = (1/4)(1/3)(x^1/8)^4

Name: Anonymous 2007-09-02 16:57 ID:yr5AgAKM

>>9

uh, you clearly don't know how to integrate.

if you think that the integral of (x^(1/8))^2 is (1/3)(x^(1/8))^3, then try differentiating the latter and see if you can get the former. and if you do, then you don't know how to differentiate, either.

Name: Anonymous 2007-09-02 17:00 ID:vxqCpsDY

(1/3)(x^1/8)^3
power rule
3 * (1/3)(x^1/8)^(3-1)
(x^1/8)^2

is this not derivative stuff?

Name: Anonymous 2007-09-02 17:02 ID:yr5AgAKM

>>12

no. because you drop the power, _DIFFERENTIATE THE BRACKET_ and drop the power by one.

(1/3)(x^(1/8))^3 goes to 3(1/3) [ (1/8) x^(-7/8) ]^2

Name: Anonymous 2007-09-02 17:05 ID:Th1MQ4pn

Everyone is mistaking the dependent variable for the independent variable.

Name: Anonymous 2007-09-02 17:06 ID:yr5AgAKM

>>14

no, everyone is forgetting the constant of inte-fucking-gration and something called the chain rule

Name: Anonymous 2007-09-02 17:14 ID:vxqCpsDY

you cant chain rule (1/3)(x^1/8)^3 because its only x^1/8 inside the bracket

Name: Anonymous 2007-09-02 17:16 ID:D6QdaoP5

x is a function, not a variable.

Perhaps it'll be more clear if it's written like this:

d²f/dx² = f^0.25

Function: f
Variable: x

Name: Anonymous 2007-09-02 17:16 ID:yr5AgAKM

>>16

no. let u = x^(1/8), differentiate through by x and you'll get a du/dx

(1/3)(u)^3 --diff--> (du/dx)^3

u = x^(1/8)
du/dx = (1/8) x^(-7/8)

substitute it in.

Name: Anonymous 2007-09-02 17:18 ID:yr5AgAKM

>>18 that index is meant to be a 2

Name: Anonymous 2007-09-02 17:27 ID:Th1MQ4pn

>>15

No, everyone is mistaking the dependent variable for the independent variable.

Name: Anonymous 2007-09-02 17:37 ID:D6QdaoP5

Alright, I'll provide a solution for some cases I know how to solve so you can at least see what this is about.

   x''+x=0
General solution: c1*sin(t)+c2*cos(t), with c1,c2 arbitrary constants

   x''+x'=0
General solution: c1+c2*exp(-t), with c1,c2 arbitrary constants, exp() exponential function

Name: Anonymous 2007-09-02 17:56 ID:yr5AgAKM

>>21

i assume that by t you mean x?

Name: Anonymous 2007-09-02 18:00 ID:D6QdaoP5

>>22

No. x is a function which depends on a variable, t in this case.

To check a solution, you take its expression (such as "c1*sin(t)+c2*cos(t)") and substitute x by that. Then you substitute x' by its first derivative ("c1*cos(t)-c2*sin(t)" in this example). Then x'' by its second derivative ("-c1*sin(t)-c2*cos(t)" in this example).

Name: Anonymous 2007-09-02 18:02 ID:yr5AgAKM

>>23

all of this would've been much clearer if it had been y = f(x), then my mad DE skilz wud b up in yo gril

Name: Anonymous 2007-09-02 19:13 ID:Th1MQ4pn

>>24

The question, in clearer terms, is this;

d^2y/dx^2 = y^(0.25)

Name: Anonymous 2007-09-02 21:14 ID:bObOiDMP

>>25
The equation is a non-linear second order ODE of the form y''(x)=f(y(x)).  This is known as the "autonomous equation"

The solution to this problem is can be found here http://eqworld.ipmnet.ru/en/solutions/ode/ode0301.pdf

Integral[(C1+2*Integral[f(y),y])^(-1/2),y]=C2+/-x

Just replace f(y) with y^(0.25).  So,

Integral[(C1+8/5*y^(5/4))^(-1/2),y]=C2+/-x

So, you're not going to be able to solve explicitly for y I don't think, but you can give it a shot.

Name: 4tran 2007-09-02 21:52 ID:qxemZd8A

>>26
Mathematica dies if C1 =/= 0.
If C1 = 0, then y = [3(C +/- x)/Sqrt(40)]^(8/3)

Name: Anonymous 2007-09-02 22:03 ID:bObOiDMP

Well, it's easy to do it if C1=0.  If C1 != 0, then you can't solve for y explicitly.  On top of that, I'm pretty confident that there's no closed form for Integral[(a+b*y^(5/4))^(-1/2),y], so you're not even going to be able to get that out.

Name: Anonymous 2007-09-02 22:12 ID:bObOiDMP

And Mathematica should be able to solve it directly without inputing the solution form for the second order autonomous ODE.

DSolve[y''[x]-y[x]^(1/4)== 0,y,x] should do it for you.

Name: Anonymous 2007-09-02 22:15 ID:Heaven

DSolve[y''[x]-y[x]^(1/4)== 0,y[x],x] Actually

Name: 4tran 2007-09-03 2:03 ID:BYBb5Kwc

>>30
I got 10 monstrous compounds of Beta and inverse Beta functions.  I have no idea why they're relevent or why they're not outputted when doing the integral.

Name: Anonymous 2007-09-03 4:23 ID:Ipge4331

apparently its not that trivial

Name: RedCream 2007-09-03 23:16 ID:Cx27+Bzr

Just trying some tags:

4
│x²dx = ?
1

Name: RedCream 2007-09-03 23:17 ID:Cx27+Bzr

Hmm.  Let's put that into the CODE tag:

⌠[sup]4[/sup]
│x²dx = ?
⌡[sub]1[/sub]

Name: 4tran 2007-09-03 23:20 ID:BWFHbitp

>>34
What does CODE stand for in this context?

Name: Anonymous 2007-09-03 23:26 ID:Ipge4331

>>34

lol, you forgot to close the tag to enter the CODE

Name: Anonymous 2007-09-04 0:18 ID:Heaven

>>26
y''=f(y).  Say y'=v=dy/dx therefore y'=dv/dx=dv/dy dy/dx=v dv/dy

So, v dv/dy = f(y) => integral{v dv}=integral{f(y) dy}
v^2/2=Integral[f(y) dy] + a => v = +/-sqrt(2 integral{f(y) dy}+a)

v=dy/dx => +/-dx = dy/sqrt(2 integral{f(y) dy}+a)

Finally, integrating that,

b+/-x = integral{dy/sqrt(2 integral{f(y) dy}+a)}

Name: RedCream 2007-09-04 1:06 ID:r3xUwQoe

Aargh, let's try it again:

4
│x²dx = ?
1

Name: RedCream 2007-09-04 1:09 ID:r3xUwQoe

OK, that worked.  What I did was use the BBCode tags to make it all pretty-like.  Replace all curly brackets "{" and "}" in the following with straight brackets "[" and "]" to achieve the same result:

{code}⌠{/code}{sup}4{/sup}{code}
│x²dx = ?
⌡{/code}{sub}1{/sub}

Name: Anonymous 2007-09-04 5:13 ID:3yhV/DgO

>>39
you blatant, blatant faggot.

Don't change these.
Name: Email:
Entire Thread Thread List