4
$\begingroup$

The isoperimetric problem is the following:

Among all curves of length L in the upper half-plane passing through the points $( - a, 0)$ and $(a, 0)$, find the one which together with the interval $[ - a, a]$ encloses the largest area.

Solution from Formin and Gelfand's book of calculus of variations (page 49):

We are looking for the function $y = y(x)$ for which the integral

$$J[y]= \int_{-a}^{a}ydx $$

takes the the lagest value subject to the conditions

$$y( - a) = y(a) = 0$$ $$ K[y] = \int_{-a}^{a} \sqrt{1 + y'^2}dx = L$$

Thus, we are dealing with an isoperimetric problem. Then, we form the functional:

$$ J[y] + \lambda K[y] = \int_{-a}^{a}(y + \lambda \sqrt{ 1+ y'^2}) dx $$

and write the corresponding Euler equation

$$1+ \lambda \frac{d}{dx}\frac{y'}{\sqrt{1+y'^2}}=0$$

which implies

$$ x+ \lambda \frac{y'}{\sqrt{1+y'^2}}= C_1 $$

integratin, we obtain the equation

$$(x - C_1)^2 + (y - C_2)^2 = \lambda^2$$

The values of $C_l$, $C_2$ and $\lambda$ are then determined from the conditions

$$y( - a) = y(a) = 0 $$ $$ K[y] = L$$

My doubt is if there are other solutions in which the curves obtained can't be represented by a function $y=y(x)$ or $y'$ doesn't exists in some points. For example, if the answer is really a circumference and $L$ is large enough, thus we have the following figure (in the figure $a=2$):

enter image description here

There're points on the curve which $y'$ doesn't exist. Besides, $y$ isn't a function of $x$. Could someone help me? If doesn't exist other solutions, how you prove that?

I thought of considering $y$ and $x$ as a function of another variable $t$, but I don't know how calculate the area as a integral in this case.

Remark: There're other questions in which I came across the same problem. If you can generalize your argument for other problems I would appreciate.

  • 0
    Derivations are done with y=y(x), assuming y is a function of x. But even with L< πa you find as solutions (complete) circles, where with domain of x restricted , I.e. -a $\le x \le a , y is a function of x, and y' is OK ( except for L= πa , at x= -a,a ).For L>πa still think that circles are the solutions, your drawing above, but there is a problem: y is not a fct. of x, and y' does not exist at certain points.2017-04-13
  • 1
    Interestingly I came with almost the same question independently, please see my post https://math.stackexchange.com/questions/2229845/a-particular-case-of-the-isoperimetric-problem2017-04-13

0 Answers 0