1. Not finding help here? Sign up for a free 30min tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Maximising area for constant perimeter

  1. Mar 11, 2007 #1
    1. The problem statement, all variables and given/known data
    A curve y = y(x) meets the x-axis at x = a, x = -a and has fixed length pi*a between these points. Show that the curve which encloses the maximum area between itself and the x-axis is the semi circle x^2 + y^2 = a^2 for y => 0


    2. Relevant equations
    INT[-a,a] ds = INT[-a,a] (1 + (dy/dx)^2)^(1/2) = pi*a
    y(a) = y(-a) = 0
     
    Last edited: Mar 11, 2007
  2. jcsd
  3. Mar 11, 2007 #2

    Dick

    User Avatar
    Science Advisor
    Homework Helper

    Apply the Euler-Lagrange equations to the unconstrained functional:
    (1 + (dy/dx)^2)^(1/2)+lambda*y.
    The lambda is a Lagrange multiplier representing your constraint and you solve for it in the end.
     
  4. Mar 11, 2007 #3
    (1 + (dy/dx)^2)^(1/2) + lambda*y = F + lambda*G

    lambda = C
    dy/dx = y'

    No explicit dependence on x so:
    (1 + (y')^2)^(1/2) + C*y - (y')^2 (1 + (y')^2)^(-1/2) = A (constant)
    (1 + (y')^2) + C*y*(1 + (y')^2)^(1/2) - (y')^2 = A*(1 + (y')^2)^(1/2)
    1 + C*y*(1 + (y')^2)^(1/2) = A*(1 + (y')^2)^(1/2)
    1 = (A - C*y)*(1 + (y')^2)^(1/2)
    1/(A - C*y) = (1 + (y')^2)^(1/2)
    1/(A - C*y)^2 = (1 + (y')^2)
    (y')^2 = (1/(A - C*y)^2) - 1
    y' = [(1/(A - C*y)^2) - 1]^(1/2)

    Substitution: 1/(A - Cy) = coshz
    y = A/C - sechz/C
    dy = -(tanhz.sechz)/C

    Making substitutions and seperating variables:
    -1/C INT[sech^2 zdz] = x + B
    tanh^2 z = C^2*(x + B)^2
    sech^2 z = 1 - C^2*(x + B)^2
    coshz = [1 - C^2*(x + B)^2]^(-1/2)
    A - Cy = [1 - C^2*(x + B)^2]^(1/2)
    Cy = A - [1 - C^2*(x + B)^2]^(1/2)
    y = A/C - 1/C*[1 - C^2*(x + B)^2]^(1/2)
    y = A/C - [(1/C^2) - (x + B)^2]^(1/2)

    Is this okay so far? I then tried:
    y(a) = y(-a) = 0, which gives B = 0
    Similarly I can obtain A/C = (1/C^2 - a^2)^(1/2)

    But then I run out of ideas, is this right so far?
     
  5. Mar 11, 2007 #4
    Are you sure it shouldn't be

    lambda*(1 + (dy/dx)^2)^(1/2) + y ?
     
  6. Mar 12, 2007 #5

    Dick

    User Avatar
    Science Advisor
    Homework Helper

    In this case it doesn't matter much which you consider to be the constraint. Just changes lambda->1/lambda. Your solution looks to be a bit more complicated than it needs to be. But when you get down to here:

    y = A/C - [(1/C^2) - (x + B)^2]^(1/2)

    you are basically done. Rearrange and rename constants and it is (y+A)^2+(x+B)^2=R^2, a general equation for a circle. Now apply the contraints to determine A,B and R.
     
  7. Mar 12, 2007 #6
    I tried another way and got the equation of a circle!
    But I'm still stuck!

    I have (x + B)^2 + (y + D)^2 = C^2
    y(a) = y(-a) = 0
    (a + B)^2 + (y + D)^2 = C^2
    (-a + B)^2 + (y + D)^2 = C^2
    so: (a + B)^2 = (-a + B)^2 and B = 0

    x^2 + (y + D)^2 = C^2
    Using the constraints:
    y(a) = y(-a) = 0
    INT[-a,a] (1 + (dy/dx)^2)^(1/2) dx = pi*a

    How can I show that D = 0 and C = a, as the question requires?
     
  8. Mar 12, 2007 #7

    Dick

    User Avatar
    Science Advisor
    Homework Helper

    From this point, think more geometrically. You KNOW it's circle. What circle arc connects the two points, has a center on the y-axis and a length of a*pi? Going back through the arc length formalism is needlessly painful.
     
  9. Mar 12, 2007 #8
    Ah I can just write it like that. It makes sense but I was worried perhaps the lecturer was looking for something more. Thanks very much!
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?



Similar Discussions: Maximising area for constant perimeter
  1. Perimeter vs. Area (Replies: 3)

Loading...