Maximising area for constant perimeter

Click For Summary
SUMMARY

The discussion centers on maximizing the area enclosed by a curve with a fixed perimeter of πa, specifically demonstrating that the optimal shape is a semicircle defined by the equation x² + y² = a² for y ≥ 0. Participants apply the Euler-Lagrange equations to derive the necessary conditions for this optimization problem, leading to the conclusion that the curve must be a circle segment. The conversation also touches on the implications of Lagrange multipliers and geometric interpretations of the problem.

PREREQUISITES
  • Understanding of Euler-Lagrange equations
  • Familiarity with Lagrange multipliers
  • Knowledge of arc length calculations in calculus
  • Basic geometry of circles and semicircles
NEXT STEPS
  • Study the application of Euler-Lagrange equations in optimization problems
  • Explore the concept of Lagrange multipliers in constrained optimization
  • Learn about arc length and its derivation in calculus
  • Investigate geometric properties of circles and their equations
USEFUL FOR

Students and educators in mathematics, particularly those focusing on calculus, optimization, and geometric analysis. This discussion is beneficial for anyone looking to deepen their understanding of optimization techniques and their applications in real-world problems.

SunGod87
Messages
30
Reaction score
0

Homework Statement


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

Homework Equations


INT[-a,a] ds = INT[-a,a] (1 + (dy/dx)^2)^(1/2) = pi*a
y(a) = y(-a) = 0
 
Last edited:
Physics news on Phys.org
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.
 
(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?
 
Are you sure it shouldn't be

lambda*(1 + (dy/dx)^2)^(1/2) + y ?
 
SunGod87 said:
Are you sure it shouldn't be

lambda*(1 + (dy/dx)^2)^(1/2) + y ?

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.
 
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?
 
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.
 
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!
 

Similar threads

Replies
3
Views
2K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 14 ·
Replies
14
Views
2K
Replies
3
Views
2K
Replies
11
Views
3K
Replies
20
Views
2K
  • · Replies 13 ·
Replies
13
Views
1K
  • · Replies 7 ·
Replies
7
Views
2K
Replies
6
Views
2K
Replies
12
Views
2K