## Geodesics: Donuts are Delicious

For the discussion of math. Duh.

Moderators: gmalivuk, Moderators General, Prelates

spectacu-awesome
Posts: 14
Joined: Tue Mar 06, 2007 11:28 am UTC
Location: the big C down under
Contact:

### Geodesics: Donuts are Delicious

Given two points on the surface of a torus, A(theta_1, phi_1) and B(theta_2, phi_2), what is the distance over the surface of the torus between the two points, and what is that path?

Put another way, an ant on the surface of a donut wants to travel from sprinkle A to sprinkle B. How far must the ant travel, and what is the shortest path?

To get you going, have some appreciation in advance!
Graphing rational functions is a pain in the asymptote
We're number -[e^(pi*i)]

Posts: 780
Joined: Sat Oct 27, 2007 5:51 pm UTC

### Re: Geodesics: Donuts are Delicious

Well, it depends what sort of metric you have on the torus, but I'll assume it's the standard one you get from identifying the sides of a square with sidelength 1.

Try drawing 9 copies of this sqaure in a 3x3 pattern, and put your first point in the middle square. Now you can put 9 copies of your second point, one in each sqaure. Drawing the 9 line segments from your first point to the 9 other points gives you 9 candidates for shortest path. After thinking about it a bit, you should be able to figure out under which conditions you use which path.
Let's have a fervent argument, mostly over semantics, where we all claim the burden of proof is on the other side!

gmalivuk
GNU Terry Pratchett
Posts: 26022
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

### Re: Geodesics: Donuts are Delicious

MartianInvader wrote:Well, it depends what sort of metric you have on the torus, but I'll assume it's the standard one you get from identifying the sides of a square with sidelength 1.

I would assume in general that, unless stated otherwise, you're talking about a torus in the real world, i.e. sitting in Euclidean 3-space. Is the inherited metric in this case the same as you'd get identifying the sides of a square (or variously-proportioned rectangles)?
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

rhino
Posts: 123
Joined: Fri Dec 08, 2006 3:43 pm UTC
Location: Cambridge, UK

### Re: Geodesics: Donuts are Delicious

gmalivuk wrote:I would assume in general that, unless stated otherwise, you're talking about a torus in the real world, i.e. sitting in Euclidean 3-space. Is the inherited metric in this case the same as you'd get identifying the sides of a square (or variously-proportioned rectangles)?

You're right to ask that question- in fact it's impossible to embed a locally Euclidean torus into R3 no matter how you distort it. You can be absolutely sure that a normal-shaped torus does not have a Euclidean metric!

HenryS
Posts: 199
Joined: Mon Nov 27, 2006 9:16 am UTC
Location: Melbourne
Contact:

### Re: Geodesics: Donuts are Delicious

gmalivuk wrote:
MartianInvader wrote:Well, it depends what sort of metric you have on the torus, but I'll assume it's the standard one you get from identifying the sides of a square with sidelength 1.

I would assume in general that, unless stated otherwise, you're talking about a torus in the real world, i.e. sitting in Euclidean 3-space.
This would make the question much, much harder, so I'd assume like MartianInvader...

gmalivuk
GNU Terry Pratchett
Posts: 26022
Joined: Wed Feb 28, 2007 6:02 pm UTC
Location: Here and There
Contact:

### Re: Geodesics: Donuts are Delicious

HenryS wrote:
gmalivuk wrote:
MartianInvader wrote:Well, it depends what sort of metric you have on the torus, but I'll assume it's the standard one you get from identifying the sides of a square with sidelength 1.

I would assume in general that, unless stated otherwise, you're talking about a torus in the real world, i.e. sitting in Euclidean 3-space.
This would make the question much, much harder, so I'd assume like MartianInvader...

It would indeed. But at what level is this question being asked?

(Also, general practice when someone asks what looks like an obvious homework question, especially when they don't admit it as such and haven't even deigned to come to the forum in months, is to overly complicate things. See, for example, the "mass of a feather" thread over in the Science forum.)
Unless stated otherwise, I do not care whether a statement, by itself, constitutes a persuasive political argument. I care whether it's true.
---
If this post has math that doesn't work for you, use TeX the World for Firefox or Chrome

(he/him/his)

dosboot
Posts: 143
Joined: Sun Jul 01, 2007 5:26 am UTC

### Re: Geodesics: Donuts are Delicious

Would the shortest path still have to be one of those 9 segments, even though their arc lengths are not what we expect? That would reduce the problem to calculations.

HenryS
Posts: 199
Joined: Mon Nov 27, 2006 9:16 am UTC
Location: Melbourne
Contact:

### Re: Geodesics: Donuts are Delicious

gmalivuk wrote:(Also, general practice when someone asks what looks like an obvious homework question, especially when they don't admit it as such and haven't even deigned to come to the forum in months, is to overly complicate things. See, for example, the "mass of a feather" thread over in the Science forum.)
Fair enough. Well in that case, depending on the metric, the answer can be anything in (0,\infty).

spectacu-awesome
Posts: 14
Joined: Tue Mar 06, 2007 11:28 am UTC
Location: the big C down under
Contact:

### Re: Geodesics: Donuts are Delicious

gmalivuk wrote:
HenryS wrote:
gmalivuk wrote:
MartianInvader wrote:Well, it depends what sort of metric you have on the torus, but I'll assume it's the standard one you get from identifying the sides of a square with sidelength 1.

I would assume in general that, unless stated otherwise, you're talking about a torus in the real world, i.e. sitting in Euclidean 3-space.
This would make the question much, much harder, so I'd assume like MartianInvader...

It would indeed. But at what level is this question being asked?

(Also, general practice when someone asks what looks like an obvious homework question, especially when they don't admit it as such and haven't even deigned to come to the forum in months, is to overly complicate things. See, for example, the "mass of a feather" thread over in the Science forum.)

yeah, sorry I've been neglecting the forums...

this is a problem of my own devising, and in true me-style, it has turned out RIDICULOUSLY complicated.

finding A distance is easy (parametri-fy (you know what I mean (OH GOD, how many more parentheses can I nest?(You, know, before I lose track of how many I have to use to close (including this, it's 5)))) the torus, then use the distance of a curve formula), finding the SHORTEST distance isn't. Wikipedia tells me to use "standard techniques of calculus and differential equations" to minimise this distance.
Graphing rational functions is a pain in the asymptote

We're number -[e^(pi*i)]

LoopQuantumGravity
Posts: 416
Joined: Mon Oct 01, 2007 4:19 am UTC

### Re: Geodesics: Donuts are Delicious

You should use the "geodesic equation" in general to do this kind of thing. You also need to be familiar with the calculus of variations.

I found this with google, but you can probably easily find others by searching for "geodesic equation."
http://oregonstate.edu/~drayt/MTH437/hw/geodesic.pdf

It shows how to calculate the shortest distance between points on a sphere using the geodesic equation. The gab is the metric tensor, which you can calculate for a torus given the equations that generate it.
http://en.wikipedia.org/wiki/Metric_tensor (their G is the matrix gab from before).

If you really want, you can derive the geodesic equation from wikipedia's distance formula and the calculus of variations.

(I haven't slept and am sick, so I apologize in advance for any crazy errors that may or may not be here.)
I study theoretical physics & strings, and am a recipient of the prestigious Jayne Cobb Hero of Canton award.

And the science gets done and you make a neat gun
For the people who are still alive!

mudge
Posts: 292
Joined: Sun Sep 30, 2007 10:14 pm UTC
Location: Chicago
Contact:

### Re: Geodesics: Donuts are Delicious

if we're resorting to google, why aren't we googling "torus geodesic"?
http://www.cdbaby.com/cd/mudge <-- buy my CD (Now back in stock!)