Maximum Surface AreaDate: 07/03/2003 at 09:13:59 From: John Hickman Subject: Maximum Surface area of shape (ellipse?) within a rectangle I have a rectangle, X long by Y wide. Within that rectangle, I wish to draw a shape that is no more than X across in any direction, but which has the *largest* surface area possible within the confines of my rectangle. Specifically I'm after a solution where X = 20 inches, and Y = 12 inches. I think the shape might be an ellipse, but I don't know how to calculate its parameters. If it is an ellipse, then I know how to draw it if the 2 fixed points and distance of the loci from them can be calculated. If it isn't an ellipse, then I'm stumped. Date: 07/03/2003 at 16:58:23 From: Doctor Douglas Subject: Re: Maximum Surface area of shape (ellipse?) within a rectangle Hi John, Thanks for writing to the Math Forum. Can you clarify the problem a bit? What do you mean by surface area? Do you mean perimeter of the confined shape? Does the confined shape have to be convex? - Doctor Douglas, The Math Forum http://mathforum.org/dr.math/ Date: 07/04/2003 at 06:09:18 From: John Hickman Subject: Re: Maximum Surface area of shape (ellipse?) within a rectangle Thanks for the reply. By maximum surface area, I want the shape to occupy as much space within the confines of the rectangle as possible. As for convex, I didn't specify that originally, but I would prefer the shape to be curved for its entire perimeter. . For instance, I could take a circle, X diameter (where X > Y), and put that over the rectangle, snip off the parts that would fall outside it, and that might give me a solution that would have the largest surface area. But it would have some flat sides. Apologies for my inability to specify this problem well. Thanks, John Date: 07/04/2003 at 16:29:46 From: Doctor Douglas Subject: Re: Maximum Surface area of shape (ellipse?) within a rectangle Hi John, You are looking for a shape that lies within the rectangle R and possibly touching its sides, and has a maximum area A <= R. Of course, the solution to this problem is R itself. But this doesn't sound like what you want. Now, let's add the restriction that the shape is curved (never flat) over its entire perimeter. This will force the shape to touch the sides of the rectangle at isolated tangent points, rather than over straight line segments. Let me also modify the problem a little bit so that the mathematics becomes a little easier - it will be relatively easy to go back to the original problem. Let the rectangle be centered on the origin (0,0) and have length 2A and width 2B. Then the top right point is at (x,y) = (A,B). Let's shift the point (A,B) by scaling in the x and y directions so that this becomes (1,1): f(x,y) = (x/A,y/B) Now the problem is to work in the square S, which is defined by S = {-1 < x < 1, -1 < y < 1}. Note that an ellipse that touches the original rectangle R at the midpoints of the four sides becomes a circle with its center at the origin and radius 1, within the square S, in this transformed picture. Furthermore, let me also just consider the first quadrant, as we can do the same treatment symmetrically in the other three quadrants. You've already found that a circle/ellipse is not a bad shape to start with: x^2 + y^2 = 1. In the first quadrant, this algebraic relation generates a quarter-circle, which of course has area = pi/4. But we can do better: x^3 + y^3 = 1 [or y = (1-x^3)^(1/3)] yields a curve that lies inside the square, but outside the circle and thus with more area than the circle. And x^4 + y^4 = 1 will do even better at surrounding more area, while touching the square only at the points (0,1) and (1,0). In fact, if you choose x^n + y^n = 1 with some very large integer n, you will get a shape that fills out most of the square/rectangle, while still touching the border at the rectangle midpoints. Furthermore, this family of curves is never flat (except for n=1, where we get a diamond shape). For n>1, the curve yields a convex region. By choosing n sufficiently large, you can get an area that approaches the entire area of the square S (and by rescaling, the rectangle R). I hope this helps. Please write back if you have more questions about this. - Doctor Douglas, The Math Forum http://mathforum.org/dr.math/ Date: 07/04/2003 at 18:13:49 From: John Hickman Subject: Re: Maximum Surface area of shape (ellipse?) within a rectangle Thanks for the answer. I've had the aha moment now. I can do the scaling and increase n until it gets the shape I want. Cheers, John |
Search the Dr. Math Library: |
[Privacy Policy] [Terms of Use]
Ask Dr. Math^{TM}
© 1994-2015 The Math Forum
http://mathforum.org/dr.math/