# In a 3-dimension isoperimetric problem, a ball maximizes the volume

• I
• graphking
In summary, the conversation discusses the difficulty of using functional derivative in R^3, specifically in solving the isoperimetric problem. The use of Lagrange multipliers is mentioned, but it is noted that it can only solve the problem in R^2. The conversation then shifts to finding a proof for the isoperimetric problem in R^3, with the realization that it is a difficult task. The suggestion of using the Brunn-Minkowski theorem is mentioned as a potential solution.

#### graphking

TL;DR Summary
isoperimetric problem: in R^n, fix the surface area, when the volume can be max? the answer is ball, like B(0,1). in R^2 you can find many ways, such as using the variation of fixed end curve, functional derivative. But in R^3 I found it hard to use functional derivative (the equation get from derivative=0 is complicated, I can't further get to B(0,1))
...

graphking said:
But in R^3 I found it hard to use functional derivative (the equation get from derivative=0 is complicated, I can't further get to B(0,1)
I suppose you are referring to Lagrange multipliers ?
If so, what's the problem ?

##\ ##

BvU said:
I suppose you are referring to Lagrange multipliers ?
If so, what's the problem ?

##\ ##
Using the lagrange multiplier is a way can only solve the isoperimetric problem in R^2, I show you the result you get in R^3:
assuming the surface is z(x,y), with fixed boundary in XOY plane, then
##(z_x/(1+z^2_x+z^2_y)^{1/2})_x+(z_y/(1+z^2_x+z^2_y)^{1/2})_y \equiv -1/\lambda
##

Is there a question here? If i understand you correctly, you tried to use the ideas of one of the solutions to the problem in dimension two to solved it in higher dimensions, but you couldn't. So? The problem is not easy. May be this approach doesn't generalize or may be it does and you couldn't do it. Are you asking how it is done?

graphking
martinbn said:
Is there a question here? If i understand you correctly, you tried to use the ideas of one of the solutions to the problem in dimension two to solved it in higher dimensions, but you couldn't. So? The problem is not easy. May be this approach doesn't generalize or may be it does and you couldn't do it. Are you asking how it is done?
please teach me a good way to proof the isoperimetric problem in R^3

Infrared and graphking