Scribd Logo
Upload

Welcome to Scribd!

  • 5 views

    6 Jun Um 1022022

    Uploaded by

    anushkadassi

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    6 Jun Um 1022022

    Uploaded by

    anushkadassi
    5 views2 pages

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    5 views2 pages

    6 Jun Um 1022022

    Uploaded by

    anushkadassi

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 2

    1 Recap

    • Non-rectangular domains. Fubini. Triple integrals.

    2 Non-rectangular domains
    Examples:
    RR p R 1 R x2 p
    • Example 3: Calculate [−1,1]×[0,2]
    |y − x2 |dA. We integrate over y first: −1 ( 0 x2 − ydy+
    R2 p 3/2
    x2
    y − x2 dy). Since x is a constant, we can easily integrate to get 23 x3 + 23 (2 − x2 ) .
    The second term can be evaluated by trigonometric substitution. So we get 34 + π2 .
    Note that first integrating over x would have made life worse.

    3 What should an FTC look like?


    We
    R b 0want to formulate a fundamental theorem of calculus. In 1-D, recall that it is
    a
    f (x)dx = f (b) − f (a). In other words, “the integral of a derivative over a region
    boils down to its boundary”. So in 2D, there are a few questions: What regions must we
    consider? If f (x, y) is a scalar field, what “derivative” must we integrate? and since the
    boundary is a curve, what must the integral boil down to ( presumably a line integral
    on the boundary)? For the first question, we must choose a domain whose boundary
    is a piecewise C 1 regular curve ( to make sense of line integrals). Furthermore, the re-
    gion must not have “holes” because then the boundary can be more than one curve (
    such regions are called simply connected). It turns out that every simple closed curve di-
    vides R2 into two regions the interior region is simply connected ( Jordan curve theorem).

    Theorem: Let P, Q be C 1 scalar fields on aRsimply connected closed set S whose bound-
    ary is a piecewise C 1regular curve. Then C (P dx + Qdy) taken in the anti-clockwise
    R R ∂Q ∂P 
    direction equals S ∂x
    − ∂y dA.
    The proof is quite painful and is frankly, beyond the scope of this course. However, let us
    look at a special case of a rectangle: The boundary is piecewise C 1 and is parametrised as
    (t, c) a ≤ t ≤ b, (b, t) c ≤ t ≤ d, (−t, d) −b ≤ t ≤ −a, and (a, −t) −d ≤ t ≤ −c. Thus the
    Rb R −a Rd R −c
    line integral is a P (t, c)dt − −b P (−t, d)dt + c Q(b, t)dt − −d Q(a, −t)dt. By the usual
    FTC, this equals the other side of Green’s theorem. It is not hard to do the same thing
    for Type-III domains (HW). In the general case, one approximates the boundary by a
    many-sided polygon and breaks the interior of this polygon up into a bunch of rectangles
    and triangles ( all Type-III). Then one applies the above proof to each and adds.

    Examples:
    RR
    • Find the area of the circle x2R+y 2 = 1. The area is dxdy. Choose Q = x, P = −y
    and use Green: 2 × Area = C (xdy − ydx). Parametrise C as x = cos(t), y = sin(t).
    Thus xdy − ydx = dt. Thus Area = π. ( A device called the planimeter works on
    this principle!)

    1
    • Consider ydx−xdy
    R
    x2 +y 2
    over the circle of radius r centred at (0, 0). Parametrise it as
    (r cos(t), r sin(t)), 0 ≤ t ≤ 2π. Then the integral is −2π. However, naively applying
    the Green theorem, we get 0 !! What is going wrong? The point is that the domain
    of P, Q is actually the disc minus the origin, i.e., it has a hole. So Green is not
    applicable! ( This way of deducing the shape of regions by what kind of calculus
    one can do on them is a big thing. It is called “Differential Topology”.)

    Change of variables (a digression from FTCs): In one-variable calculus, recall that if


    u : [a, b] → [c, d] is a 1 − 1 onto C 1 map such that du
    dx
    6= 0 for all x ( except possibly at
    R u−1 (b) Rd
    a, b), and f : [c, d] → R is continuous function, then u−1 (a) f (u(x))u0 (x)dx = c f (u)du.
    R2 R1 R2
    Examples: Take 0 u2 du with u = 2x to get 0 8x2 dx. But with u = 2 − 2x, 0 u2 du =
    R0 R1
    − 1 8(1 − x)2 dx = 2 0 4(1 − x)2 dx. If we want to useR only [0, 2]Rand [0, 1] instead of the
    upper and lower limits, then in the second example, [0,2] u du = [0,1] 4(1 − x)2 | du
    2
    dx
    |dx. We
    want a generalisation for multiple integrals. ( Indeed, if we have cylindrical or spherical
    symmetry, it makes sense to choose other coordinates systems like polar coordinates.)

    You might also like