Lecture 2: Q&A: Gaussian Integral

Lecture 2: Q&A
Gaussian integral
1 How do you derive
x2 e
dx =
For that story see Section 7.2 in the notes (page 40ff). 2 What if = 1 in the integral
x2 e
2
dx?
Then the exponent becomes ex , so the exponential, and its integral, blows up as |x| . 3 For
x2 e
dx, why cant you choose [x] = L2 with [] = L4 ?
You can! You then get the same conclusion: that the integral is proportional to 1/2 . Indeed you can choose any dimensions for x, so long as x is not dimensionless. You can even choose x to be dimensionless; if you do, however, the method of dimensions will not help you guess the integral. 4 Why use in
x2 e
dx?
Without , the exponent is x2 , and it has to be dimensionless. So x would be dimensionless; the method of dimensions then cannot help you guess the integral. 5 I understood
x2 e
dx 1/2 but got lost at the end.
The last step is guring out the constant of proportionality. I didnt derive it in lecture, and instead just used the result (explained in Section 7.2 of the notes) that
ex dx =
. as the result, the full integral has
Putting in = 1 turns to be
x2 e
dx into
x2 e
dx, so to get .
ex dx =
How can I know right away that f () = L3 rather than k/ ? Its ne to say that f () = L3 . But how do you make a quantity with dimensions L3 ? All you can use in your construction is . Since [] = L2 , the L3 must come from A3/2 , where A is a (dimensionless) constant. Thus L3 is 1/2 times a dimensionless constant.
2008-01-15 02:38:17 / rev:14d5071118f3
Cite as: Sanjoy Mahajan, course materials for 18.098 / 6.099 Street-Fighting Mathematics, IAP 2008.
MIT OpenCourseWare (http://ocw.mit.edu/), Massachusetts Institute of Technology. Downloaded on [DD Month YYYY].
18.098/6.099. Street-ghting mathematics (IAP 2008)
Sanjoy Mahajan
Dimensions
7 It is very strange that the integral of x, where [x] = L, is also a length. As another example, here is the relation between distance and velocity: s= v dt.
The integral is the area under the v curve. The vertical axis, which is velocity, has dimensions of LT1 . The horizontal axis, which is time, has dimensions of T. So the area has dimensions of LT1 T, which is a length. And it should be a length, since the integral gives a distance. 8 Why does xx dx become impossible to integrate when x has dimensions?
2
The same problem happens when doing a sum instead of an integral, and is easier to see the problem in that case. Consider this sum: xx . The exponent x2 is dimensionless. Lets say that, to do the sum, you evaluate the terms when x2 = 1, 2, 3, . . .. There is no problem if x is dimensionless. However, now let x have dimensions of length. Then the rst term, which is x1 , has dimensions of L1 . The second term, which is x2 , has dimensions of L2 . And so on. These terms, which have to be added together, do not have the same dimensions, which is the problem. The same problem happens with the integral, which is just a continuous sum.
2
Pyramid
9 For the truncated pyramids volume, why not use h 0 as another easy case? Thats a good suggestion, which I incorporated into the derivation on Friday. Thanks! 10 Would h be an easy case for the pyramid? Thats also a good suggestion, which I incorporated into the derivation on Friday. Thanks! 11 Im confused about the tests a 0, b 0. The test a 0 is the special case of a normal pyramid (not truncated), with a square base and a point above it. Whatever formula we invent for the truncated pyramid must work for the normal pyramid, which is the special case a 0 of the truncated pyramid. The test b 0 is the same case with the pyramid ipped over (so the point points downward). 12
1 How was V = 3 h(a2 + b2 ) derived?
By guessing a formula that would work when a 0 or when b 0. For a 0, a valid formula is hb2 /3. For b 0, a valid formula is ha2 /3. The combination V = 1 h(a2 + b2 ) works in both cases. 3
2008-01-15 02:38:17 / rev:14d5071118f3
18.098/6.099. Street-ghting mathematics (IAP 2008)
Sanjoy Mahajan
13
Where does the one-third come from in the pyramid volume? From the very special case a = 0, b = 2, h = 1. You take six of those pyramids and form a cube of side 2 and volume 8. So you know the volume of each pyramid (it is 4/3). The only way to get the volume correct is to use V = hb2 /3. (See Section 8.5.)
14
How do you use the a = 0, b = 0, a = b cases to nd the volume of the truncated pyramid? I hope that the summary on Friday or Section 8.3 of the notes clears that up. But ask again if it is not clear.
15
Why is V h? The volume should be proportional to the height because (and this is a rough argument) it usually is. Imagine a cylinder of radius r and height h. If you stack two of those cylinders to make a longer one, with height 2h, then youve doubled the volume. So its volume is proportional to h. And that relation is true for the pyramid as well.
16
Why, when a = b, do you get a cube rather than a sheet? Its a cube if a = h as well. In general it is a rectangular prism: It has a square base and square top, each with side b (or a), and it has height h.
General
17 How do you gure out what equation to start with when guring out equations from dimensional analysis and easy cases (such as for the pyramid)? What is the best way to start the easy cases method? Start with the easiest, most familiar cases. For example, a = 0 is good when analyzing the trun cated pyramid because it produces a familiar shape (the ordinary pyramid). 18 How do you guess the constant term when there is no easy case (for example, for
x2 e
dx)?
You usually cannot guess it without an easy case. But there is almost always an easy case! And if the easy case is not easy enough to solve analytically, you can sometimes nd its value from experiment. An example is the drag coefcient of the cones. The argument in lecture will show that the drag coefcient is constant for the easy case of high-speed ow. But our mathematical knowledge is not advanced enough, and may never be advanced enough, to deduce the constant from the NavierStokes equations. So instead we measure the constant, for example in a wind tunnel. 19 If the constant is not known, should it just be carried through as a constant such as C? What if there is more than one constant? Thats a good idea to use C. Its what we do with the drag coefcient cd . In most cases we cannot calculate it, so we just use the symbol, and compute it from experiment. 20 We are spending too much time on dimensions! You are probably right, which is why I went ahead with easy cases without answering the above questions in class.
2008-01-15 02:38:17 / rev:14d5071118f3

Lecture 2: Q&A: Gaussian Integral

Uploaded by

Copyright:

Available Formats

Lecture 2: Q&A: Gaussian Integral

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture 2: Q&A: Gaussian Integral

Uploaded by

Copyright:

Available Formats

Lecture 2: Q&A

dx, why cant you choose [x] = L2 with [] = L4 ?

dx 1/2 but got lost at the end.

. as the result, the full integral has

2008-01-15 02:38:17 / rev:14d5071118f3

18.098/6.099. Street-ghting mathematics (IAP 2008)

2008-01-15 02:38:17 / rev:14d5071118f3

18.098/6.099. Street-ghting mathematics (IAP 2008)

2008-01-15 02:38:17 / rev:14d5071118f3

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.