Reversing the order of integration in a double integral

Click For Summary
The discussion revolves around reversing the order of integration in a double integral involving the function y^2 - 4x over a triangular region. The initial integration yields a result of 2/5 when performed with horizontal strips, but confusion arises when attempting to use vertical strips, particularly with limits involving square roots. Participants highlight that the antisymmetry of the integrand -4x leads to a cancellation in the right half of the region, contributing to the total integral being 2/5. The challenge lies in correctly setting the limits for vertical integration, with some participants suggesting that sign errors may be causing discrepancies in the results. Ultimately, the consensus is that while horizontal strips yield the correct answer, vertical strips introduce complications that can lead to incorrect results if not handled carefully.
dyn
Messages
774
Reaction score
63
Homework Statement
Calculate the double integral of y^2 - 4x over the following region -y^2 < x < y^2 , 0 < y < 1
Relevant Equations
I did the integral by dividing the region into strips parallel to the x-axis and got the answer 2/5 which i know is correct but i am trying to reverse the order of integration but i can't get the same answer
Performing the x-integration first the limit are x=y2 and x= -y2 and then the y limits are 0 to 1. This gives the final answer 2/5

But i am getting confused when trying to reverse the order of integration. My attempt is that i have to divide the region in 2 equal halfs and then double my answer at the end. I used y limits from √x to 1 and x limits from 0 to 1 but i am ending up with the answer zero ! Where am i going wrong with my limits here ?
Thanks
 
Physics news on Phys.org
The integration is
\int_S (y^2-4x) \ dS
Area S is two-side-curved triangle with vertex (0,0),(1,-1),(1,1)

dyn said:
My attempt is that i have to divide the region in 2 equal halfs and then double my answer at the end.

That' good for integrand y^2 but not for -4x because of antisymmetry it cancels by right plus and left minus of y-axis
\int_S (-4x) \ dS=0
You should omit it from integration calculation in right half.
 
The answer of the total integral is 2/5 which i got my dividing the region into horizontal strips. If i divide the region into 2 equal halfs along the line x=0 and perform the 2 integrals using horizontal strips i get IR = 0 and IL = 2/5 which when added together give my total answer of 2/5.

If i now use vertical strips for both halfs , i can get IR = 0 as it should but i cannot get 2/5 for IL. I am struggling with the limits because they involve square roots of negative numbers. Is this an example of where reversing the order of integration is impossible to perform or extremely difficult ? Or am i just missing something when trying to perform the integral on the LHS using vertical strips ?
 
I=\int (y^2-4x) dS = \int y^2\ dS -4 \int x\ dS
By vertical strips for halfs
I=\int y^2\ dS_R+ \int y^2\ dS_L -4 \int x\ dS_R-4 \int x\ dS_L
=2\int y^2\ dS_R
because
\int y^2\ dS_R= \int y^2\ dS_L
\int x\ dS_R=- \int x\ dS_L
You said right half integration of y^2-4x is zero so I assume
I=1/5 + 1/5 - 1/5 + 1/5
Confirm this by integration of each term if you please.
 
  • Like
Likes etotheipi
To perform the integral over the RH region using vertical strips i integrate (y2-4x) wrt y from √x to 1 ; i then integrate 1/3 - 2x +(5/3)x3/2 wrt x from 0 to 1. This gives a final answer of 0 which is the same answer as expected if i integrated using horizontal strips.
The problem comes when i try to integrate y2=4x over the LH side using vertical strips
 
dyn said:
The problem comes when i try to integrate y2=4x over the LH side using vertical strips
\int^0_{-1} dx \int_{\sqrt{-x}}^1 dy (y^2-4x)=\int^0_{-1} dx\ [y^3/3-4xy]_{y=\sqrt{-x}}^{y=1}
RH side
\int^1_0 dx \int_{\sqrt{x}}^1 dy (y^2-4x)=\int^1_0 dx\ [y^3/3-4xy]_{y=\sqrt{x}}^{y=1}
 
i apologise ; i was writing this late at night. The function to be integrated is y2- 2x
It's late again so i will look at this again when i wake
 
Do not worry. It does not matter on the result of integration.
 
N.B. you can also do$$\begin{align*}
\int_S (y^2 - 4x) dy dx &= 2\int_{S_R} y^2 dy dx\\

&= 2\int_0^1 \int_0^{y^2} y^2 dx dy\\

&= 2\int_0^1 y^4 dy\\

&= \frac{2}{5}

\end{align*}
$$which also agrees with @anuttarasammyak's solution.
 
  • #10
anuttarasammyak said:
\int^0_{-1} dx \int_{\sqrt{-x}}^1 dy (y^2-4x)=\int^0_{-1} dx\ [y^3/3-4xy]_{y=\sqrt{-x}}^{y=1}
I am still not getting the correct answer for integral using vertical strips over the LHS. The function to be integrated is actually y2 -2x
Using the limits you have given i get
1/3 - 2x -(5/3)(-x)5/2 to be integrated from -1 to 0 giving me a final answer of 2 which is still wrong
 
  • #11
Changing the coefficient to 2 and introducing -x=t
I_L=\int_0^1 dt [y^3/3 +2ty]_{y=\sqrt t}^{y=1}=I_1+I_2
corresponding to
I_R=\int_0^1 dx [y^3/3 -2xy]_{y=\sqrt x}^{y=1}=I_1-I_2
where
I_1=\int_0^1 dt [y^3/3]_{y=\sqrt t}^{y=1}=1/5
I_2=\int_0^1 dt [2ty]_{y=\sqrt t}^{y=1}=1/5
though we do not have to calculate ##I_2## to get the answer.
 
Last edited:
  • Like
Likes Charles Link
  • #12
The answer should be 2/5 and the function to be integrated is y2 -2x.

Integrating this wrt y gives y3/3 - 2xy

Inserting the limits of y from√(-x) to1 gives the following

1/3 - 2x - [ (-x)3/2/3 - 2x(-x)1/2] = 1/3 - 2x - (7/3)(-x)3/2

Integrating this wrt x gives x/3 - x2 - (14/15)(-x)5/2

Inserting the limits of x from -1 to 0 give the final answer

(1/3) + 1 + (14/15) = 34/15

I have spent hours on this problem in total and i can't get the answer 2/5 by using vertical strips ! It is easy using horizontal strips but i wanted to check that i could do it using vertical strips but i can't !
 
  • #13
I think the substitution mentioned in post 11 is the best way to proceed. Otherwise I think the ## -x ## to the fractional power is causing problems that aren't readily resolved.
Edit: I think I see the problem though: In the integration, in going from ##3/2 ## to ##5/2 ## in the exponent, you introduce another minus sign, which alternatively, you can apply the chain rule and check your integration by taking derivatives, and I think you are off by a minus sign on that term.
 
Last edited:
  • Like
Likes dyn and anuttarasammyak
  • #14
dyn said:
1/3 - 2x - [ (-x)3/2/3 - 2x(-x)1/2] = 1/3 - 2x - (7/3)(-x)3/2
Integrating this wrt x gives x/3 - x2 - (14/15)(-x)5/2
How did you integrate the RHS last term,
-\frac{7}{3}\int_{-1}^0 (-x)^{3/2} dx?
 
  • Like
Likes dyn and Charles Link
  • #15
See also my edit in post 13.
 
  • #16
dyn said:
Integrating this wrt x gives x/3 - x2 - (14/15)(-x)5/2

Inserting the limits of x from -1 to 0 give the final answer

(1/3) + 1 + (14/15) = 34/15
Sign errors when you plugged in -1.
 
  • #17
Observe: ##dy/dx=(-7/3)(-x)^{3/2} ##.
##y=+(14/15)(-x)^{5/2} ##. Take the derivative with the chain rule, and you will see I have the sign right. The OP has it incorrect.
 
Last edited:
  • Like
Likes dyn
  • #18
Thank you everyone so much. It was driving me nuts ! Yes , it was that negative sign when integrating the (-x). I am so relieved. Thank you all again !
 
  • Like
Likes Charles Link

Similar threads

  • · Replies 9 ·
Replies
9
Views
2K
Replies
3
Views
2K
  • · Replies 19 ·
Replies
19
Views
2K
  • · Replies 4 ·
Replies
4
Views
1K
Replies
2
Views
2K
Replies
24
Views
3K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 4 ·
Replies
4
Views
2K
Replies
10
Views
2K