Kotani’s Ant Problem

Projects for my (and maybe your) classes

Adam E. Parker
aparker@wittenberg.edu

Wittenberg University

Kotani’s Ant Problem – p.1/16

The Problem
Q: An ant is at corner A of a 1 × 1 × 2 box. It crawls
along the surface, along a geodesic, the shortest possible
path, to a point B. Where is B located to make the path
as long as possible?

B EE F EE
EE EE
EE EE
EE EE
EE EE
E E
A EE C G
EE
EE
EE
EE
E
D H

Kotani’s Ant Problem – p.2/16

History of Problem
Originally propsed by Yoshiyuki Kotani. (year???)

Kotani’s Ant Problem – p.3/16

History of Problem
Originally propsed by Yoshiyuki Kotani. (year???)
Martin Gardener, “The Ant on a 1 × 1 × 2”,
math HORIZONS, February 1996, 8-9.

Kotani’s Ant Problem – p.3/16

History of Problem
Originally propsed by Yoshiyuki Kotani. (year???)
Martin Gardener, “The Ant on a 1 × 1 × 2”,
math HORIZONS, February 1996, 8-9.
Dick Hess proposed the problem in Fall 1998 issue
of Pi Mu Epsilon Journal.

Kotani’s Ant Problem – p.3/16

History of Problem
Originally propsed by Yoshiyuki Kotani. (year???)
Martin Gardener, “The Ant on a 1 × 1 × 2”,
math HORIZONS, February 1996, 8-9.
Dick Hess proposed the problem in Fall 1998 issue
of Pi Mu Epsilon Journal.
Later solved by Meagan Tripp (Alma) and Rex Hu
(Brooklyn).

Kotani’s Ant Problem – p.3/16

History of Problem
Originally propsed by Yoshiyuki Kotani. (year???)
Martin Gardener, “The Ant on a 1 × 1 × 2”,
math HORIZONS, February 1996, 8-9.
Dick Hess proposed the problem in Fall 1998 issue
of Pi Mu Epsilon Journal.
Later solved by Meagan Tripp (Alma) and Rex Hu
(Brooklyn).
Dick Hess’ Paper “Geodesics on the 1 × a × b
Rectangular Block.”

Kotani’s Ant Problem – p.3/16

History of Problem
Originally propsed by Yoshiyuki Kotani. (year???)
Martin Gardener, “The Ant on a 1 × 1 × 2”,
math HORIZONS, February 1996, 8-9.
Dick Hess proposed the problem in Fall 1998 issue
of Pi Mu Epsilon Journal.
Later solved by Meagan Tripp (Alma) and Rex Hu
(Brooklyn).
Dick Hess’ Paper “Geodesics on the 1 × a × b
Rectangular Block.”
David Wolfe and Tom Rodgers,
Puzzlers Tribute: A Feast for the mind. 1999.
Kotani’s Ant Problem – p.3/16
History of Problem
Originally propsed by Yoshiyuki Kotani. (year???)
Martin Gardener, “The Ant on a 1 × 1 × 2”,
math HORIZONS, February 1996, 8-9.
Dick Hess proposed the problem in Fall 1998 issue
of Pi Mu Epsilon Journal.
Later solved by Meagan Tripp (Alma) and Rex Hu
(Brooklyn).
Dick Hess’ Paper “Geodesics on the 1 × a × b
Rectangular Block.”
David Wolfe and Tom Rodgers,
Puzzlers Tribute: A Feast for the mind. 1999.
Martin Gardener, A Gardeners’ Walk, 2001. Kotani’s Ant Problem – p.3/16
The Problem
Q: An ant is at corner A of a 1 × 1 × 2 box. It crawls
along the surface, along a geodesic, the shortest
possible path, to a point B. Where is B located to
make the path as long as possible?

B EE F EE
EE EE
EE EE
EE EE
EE EE
E E
A EE C G
EE
EE
EE
EE
E
D H

Kotani’s Ant Problem – p.4/16

The Problem
Q: An ant is at corner A of a 1 × 1 × 2 box. It crawls
along the surface, along a geodesic, the shortest
possible path, to a point B. Where is B located to
make the path as long as possible?

B EE F EE
EE EE
EE EE
EE EE
EE EE
E E
A EE C G
EE
EE
EE
EE
E
D H
A: We (and your students) would all guess point G.
Kotani’s Ant Problem – p.4/16
What’s the best path?
The first guess always seems to be "in" the box.

B @@ F @@
@@ @@
@@ @@
@@ @@
@@ @@
@@ @@
@ @
A @@/o /o /o /o /o /o C
/o /o /o /o /o /o /o /o // G
@@
@@
@@2 1
@@
@@
@
1
D H
We explain that this isn’t a burrowing ant.

Kotani’s Ant Problem – p.5/16

What’s the best path?
The second guess is typically not the best path.

B @@ F @@
@@ @@
@@ @@
@@ @@
@@ @@
@@ @@
@ @
A @@@@ CKS +3 G
@@@@
@@@@ 2
@@@@ 1
@@@@
@@@@
$ 1
D H
We can easily see that this path has length 4.

Kotani’s Ant Problem – p.5/16

What’s the best path?
The third guess is a little better.

B @@ F @@
@@ @@
@@ @@
@@ @@
@@ @@
@@ @@
@ @
A @@@@ C : B G
@@@@ ~
~~~
@@@@ 2 ~
~~
~
@@@@ 1 ~
~~~
@@@@ ~
~~~
@@@@ ~~~
~
$ ~~~~
1
D H
√
We can easily see that this path has length 2 + 2 =
3.414 . . ..
Kotani’s Ant Problem – p.5/16
What’s the best path?
The fourth guess tends to be better still.

B @@ F @@
@@ @@
@@ @@
@@ @@
4 @@WW @@
q < @
q @ WWWWWW W @@
q @ WWWWW @
qqq W '/
A @@ C G
@@
@@
@@2 1
@@
@@
@
1
D H
But how do we find this length?

Kotani’s Ant Problem – p.5/16

The Best Path
Let’s "open up" the solid, and ask what’s the shortest
path between A and G?

Kotani’s Ant Problem – p.6/16

The Best Path

F G

F B C G

E A D H

E H

F G

Kotani’s Ant Problem – p.6/16

The Best Path

F x 8@ G
xxxx
x
xxxx
xx
x
x
xxxx
xx
F B x
x
x
xxx
x
x C G
xxxxxx
x
xxx
x
xxxxx
E A D H
√
It’s easy to see that this path has length 8, and it is the
shortest path from A to G.

Kotani’s Ant Problem – p.6/16

A farther point!
It turns out that there is a farther point away
(otherwise this is a particularly worthless talk).

Kotani’s Ant Problem – p.7/16

A farther point!
It turns out that there is a farther point away
(otherwise this is a particularly worthless talk).
I’ll tell you that the furthest point lies on the line
connecting G and C.

Kotani’s Ant Problem – p.7/16

A farther point!
It turns out that there is a farther point away
(otherwise this is a particularly worthless talk).
I’ll tell you that the furthest point lies on the line
connecting G and C.
So we choose a point and have it depend upon x

C }
G
}}
P
~~~ 
~~~ x
~~ x 
D H

Kotani’s Ant Problem – p.7/16

A farther point!
It turns out that there is a farther point away
(otherwise this is a particularly worthless talk).
I’ll tell you that the furthest point lies on the line
connecting G and C.
So we choose a point and have it depend upon x

C }
G
}}
P
~~~ 
~~~ x
~~ x 
D H
What is the shortest distance from A to P ?
Kotani’s Ant Problem – p.7/16
First Path

B C z
G
{
P
{{x{  x
A D H

Kotani’s Ant Problem – p.8/16

First Path

B C z
G
h hh P
ha
h hhhhh
{{  x
hh {{
hhhhh x
A D H

Kotani’s Ant Problem – p.8/16

First Path

B C z
G
h hh P
ha
h hhhhh
{{  x
hh {{
hhhhh x
A D H

a2 = (2 + x)2 + x2

Kotani’s Ant Problem – p.8/16

First Path

B C z
G
h hh P
ha
h hhhhh
{{  x
hh {{
hhhhh x
A D H
p
a2 = (2 + x)2 + x2 a = (2 + x)2 + x2

Kotani’s Ant Problem – p.8/16

First Path

B C z
G
h hh P
ha
h hhhhh
{{  x
hh {{
hhhhh x
A D H
p
a2 = (2 + x)2 + x2 a = (2 + x)2 + x2
3

2
.

3
2

8
.
2

6
.
2

4
.
2

2
.

Kotani’s Ant Problem – p.8/16

.

.
First Path

B C z
G
h hh P
ha
h hhhhh
{{  x
hh {{
hhhhh x
A D H
p
a2 = (2 + x)2 + x2 a = (2 + x)2 + x2
3

2
.

3
2

8
.
2

6
.
2

4
.
2

2
.

Kotani’s Ant Problem – p.8/16

.

.
First Path

B C z
G
h hh P
ha
h hhhhh
{{  x
hh {{
hhhhh x
A D H
p
a2 = (2 + x)2 + x2 a = (2 + x)2 + x2
3

2
.

3
2

8
.
2

6
.
2

4
.
2

2
.

Kotani’s Ant Problem – p.8/16

.

.
Second Path

C G H
P BBB

B C x D

A D

Kotani’s Ant Problem – p.9/16

Second Path

C G H
w P BB
www  B
B b wwwC
ww x D
www
ww
A D

Kotani’s Ant Problem – p.9/16

Second Path

C G H
w P BB
www  B
B b wwwC 
ww x D
www 
ww 
A D1−x

Kotani’s Ant Problem – p.9/16

Second Path

C G H
w P BB
www  B
B b wwwC 
ww x D
www 
ww 
A D1−x

b2 = (3 − x)2 + (1 + x)2

Kotani’s Ant Problem – p.9/16

Second Path

C G H
w P BB
www  B
B b wwwC 
ww x D
www 
ww 
A D1−x
p
b2 = (3 − x)2 + (1 + x)2 b= (3 − x)2 + (1 + x)2

Kotani’s Ant Problem – p.9/16

Second Path

C G H
w P BB
www  B
B b wwwC 
ww x D
www 
ww 
A D1−x
p
b2 = (3 − x)2 + (1 + x)2 b= (3 − x)2 + (1 + x)2
3

2
.

3
2

8
.
2

6
.
2

4
.
2

2
.

Kotani’s Ant Problem – p.9/16

2

1
Maximum distance from A
Let’s examine the graph closer
3

2
.

3
2

8
.
2

6
.
2

4
.
2

2
.

8
4

1
.

Kotani’s Ant Problem – p.10/16

Maximum distance from A
We could also solve algebraically.

Kotani’s Ant Problem – p.10/16

Maximum distance from A
We could also solve algebraically.
p p
(2 + x)2 + x2 = (3 − x)2 + (1 + x)2

Kotani’s Ant Problem – p.10/16

Maximum distance from A
We could also solve algebraically.
p p
(2 + x)2 + x2 = (3 − x)2 + (1 + x)2

2 2 2 2
4 + 4x + x + x = 9 − 6x + x + 1 + 2x + x

Kotani’s Ant Problem – p.10/16

Maximum distance from A
We could also solve algebraically.
p p
(2 + x)2 + x2 = (3 − x)2 + (1 + x)2

2 2 2 2
4 + 4x + x + x = 9 − 6x + x + 1 + 2x + x

4 + 4x = 10 − 4x

Kotani’s Ant Problem – p.10/16

Maximum distance from A
We could also solve algebraically.
p p
(2 + x)2 + x2 = (3 − x)2 + (1 + x)2

2 2 2 2
4 + 4x + x + x = 9 − 6x + x + 1 + 2x + x

4 + 4x = 10 − 4x

3
x=
4

Kotani’s Ant Problem – p.10/16

Maximum distance from A
We could also solve algebraically.
p p
(2 + x)2 + x2 = (3 − x)2 + (1 + x)2

2 2 2 2
4 + 4x + x + x = 9 − 6x + x + 1 + 2x + x

4 + 4x = 10 − 4x

3
x=
4

Kotani’s Ant Problem – p.10/16

Maximum distance from A
When x = 43 , the distance from A to P is thus
p √
(2.75)2 + (.75)2 = 8.125 = 2.850 . . .

Kotani’s Ant Problem – p.11/16

Maximum distance from A
When x = 43 , the distance from A to P is thus
p √
(2.75)2 + (.75)2 = 8.125 = 2.850 . . .
Recall that the distance from A to G was
√
8 = 2.828 . . .

Kotani’s Ant Problem – p.11/16

Maximum distance from A
When x = 43 , the distance from A to P is thus
p √
(2.75)2 + (.75)2 = 8.125 = 2.850 . . .
Recall that the distance from A to G was
√
8 = 2.828 . . .

We’ve really found a farther point!

Kotani’s Ant Problem – p.11/16

Logistics
A precalculus project ( though I’d probably use it in
a geometry course if I taught one).

Kotani’s Ant Problem – p.12/16

Logistics
A precalculus project ( though I’d probably use it in
a geometry course if I taught one).
Work in groups.

Kotani’s Ant Problem – p.12/16

Logistics
A precalculus project ( though I’d probably use it in
a geometry course if I taught one).
Work in groups.
Have a worksheet with guiding questions.

Kotani’s Ant Problem – p.12/16

Logistics
A precalculus project ( though I’d probably use it in
a geometry course if I taught one).
Work in groups.
Have a worksheet with guiding questions.
Give a wooden block.

Kotani’s Ant Problem – p.12/16

Logistics
A precalculus project ( though I’d probably use it in
a geometry course if I taught one).
Work in groups.
Have a worksheet with guiding questions.
Give a wooden block.
After covering basic modeling with functions.

Kotani’s Ant Problem – p.12/16

Logistics
A precalculus project ( though I’d probably use it in
a geometry course if I taught one).
Work in groups.
Have a worksheet with guiding questions.
Give a wooden block.
After covering basic modeling with functions.
After reviewing Pythgorean Theorem, distance
formula, etc.

Kotani’s Ant Problem – p.12/16

Logistics
A precalculus project ( though I’d probably use it in
a geometry course if I taught one).
Work in groups.
Have a worksheet with guiding questions.
Give a wooden block.
After covering basic modeling with functions.
After reviewing Pythgorean Theorem, distance
formula, etc.
After finding intersections using the TI-XX

Kotani’s Ant Problem – p.12/16

Logistics
A precalculus project ( though I’d probably use it in
a geometry course if I taught one).
Work in groups.
Have a worksheet with guiding questions.
Give a wooden block.
After covering basic modeling with functions.
After reviewing Pythgorean Theorem, distance
formula, etc.
After finding intersections using the TI-XX
Take < one class.

Kotani’s Ant Problem – p.12/16

Logistics
A precalculus project ( though I’d probably use it in
a geometry course if I taught one).
Work in groups.
Have a worksheet with guiding questions.
Give a wooden block.
After covering basic modeling with functions.
After reviewing Pythgorean Theorem, distance
formula, etc.
After finding intersections using the TI-XX
Take < one class.
I ask for them to come up with other questions.
Kotani’s Ant Problem – p.12/16
Why I like this project
It’s a surprising result.

Kotani’s Ant Problem – p.13/16

Why I like this project
It’s a surprising result.
It’s a relatively new result.

Kotani’s Ant Problem – p.13/16

Why I like this project
It’s a surprising result.
It’s a relatively new result.
It involves modeling something that isn’t a falling
objects, bank account, or radioactive waste.

Kotani’s Ant Problem – p.13/16

Why I like this project
It’s a surprising result.
It’s a relatively new result.
It involves modeling something that isn’t a falling
objects, bank account, or radioactive waste.
It shows the graphical / algebraic/ numerical
approaches to problem sovling.

Kotani’s Ant Problem – p.13/16

Why I like this project
It’s a surprising result.
It’s a relatively new result.
It involves modeling something that isn’t a falling
objects, bank account, or radioactive waste.
It shows the graphical / algebraic/ numerical
approaches to problem sovling.
It has a calculus feel, with no calculus.

Kotani’s Ant Problem – p.13/16

Why I like this project
It’s a surprising result.
It’s a relatively new result.
It involves modeling something that isn’t a falling
objects, bank account, or radioactive waste.
It shows the graphical / algebraic/ numerical
approaches to problem sovling.
It has a calculus feel, with no calculus.
It’s easy to make easier. For example, rather than
finding the point P , you can tell the students where
it is. This avoids the modeling.

Kotani’s Ant Problem – p.13/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3.

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n
What happens if, instead of n > 1 we have n < 1?

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n
What happens if, instead of n > 1 we have n < 1?
What happens if, instead of a rectangular solid, we
have a pyramid?

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n
What happens if, instead of n > 1 we have n < 1?
What happens if, instead of a rectangular solid, we
have a pyramid? With a triangular base?

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n
What happens if, instead of n > 1 we have n < 1?
What happens if, instead of a rectangular solid, we
have a pyramid? With a triangular base? With a
square base?

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n
What happens if, instead of n > 1 we have n < 1?
What happens if, instead of a rectangular solid, we
have a pyramid? With a triangular base? With a
square base? With a . . .. base?

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n
What happens if, instead of n > 1 we have n < 1?
What happens if, instead of a rectangular solid, we
have a pyramid? With a triangular base? With a
square base? With a . . .. base?
What if instead of having a prizm with square base,
we had a prizm with a triangle base?

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n
What happens if, instead of n > 1 we have n < 1?
What happens if, instead of a rectangular solid, we
have a pyramid? With a triangular base? With a
square base? With a . . .. base?
What if instead of having a prizm with square base,
we had a prizm with a triangle base? With a
pentagon base?

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n
What happens if, instead of n > 1 we have n < 1?
What happens if, instead of a rectangular solid, we
have a pyramid? With a triangular base? With a
square base? With a . . .. base?
What if instead of having a prizm with square base,
we had a prizm with a triangle base? With a
pentagon base?
Last time I checked, there were still only a small
finite number of regular polyhedra.

Kotani’s Ant Problem – p.14/16

Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n
What happens if, instead of n > 1 we have n < 1?
What happens if, instead of a rectangular solid, we
have a pyramid? With a triangular base? With a
square base? With a . . .. base?
What if instead of having a prizm with square base,
we had a prizm with a triangle base? With a
pentagon base?
Last time I checked, there were still only a small
finite number of regular polyhedra. What about
some non-regular polyhedra?
Kotani’s Ant Problem – p.14/16
Ideas for making it harder
What happens if, instead of 1 × 1 × 2, we have a
1 × 1 × 3. 1 × 1 × 3 1 × 1 × n
What happens if, instead of n > 1 we have n < 1?
What happens if, instead of a rectangular solid, we
have a pyramid? With a triangular base? With a
square base? With a . . .. base?
What if instead of having a prizm with square base,
we had a prizm with a triangle base? With a
pentagon base?
Last time I checked, there were still only a small
finite number of regular polyhedra. What about
some non-regular polyhedra?
Higher dimensional analogues? Kotani’s Ant Problem – p.14/16
Ideas for making it harder (con’t)
What happens if don’t require that the ant start at
corner A. In other words, what are the farthest that
two points can be away from each other.

Kotani’s Ant Problem – p.15/16

Ideas for making it harder (con’t)
What happens if don’t require that the ant start at
corner A. In other words, what are the farthest that
two points can be away from each other. Easy to
see that the distance must be at least 3.

Kotani’s Ant Problem – p.15/16

Ideas for making it harder (con’t)
What happens if don’t require that the ant start at
corner A. In other words, what are the farthest that
two points can be away from each other. Easy to
see that the distance must be at least 3. In fact the
longest distance is
q
√
16 − 4 3 = 3.0119

Kotani’s Ant Problem – p.15/16

Ideas for making it harder (con’t)
What happens if don’t require that the ant start at
corner A. In other words, what are the farthest that
two points can be away from each other. Easy to
see that the distance must be at least 3. In fact the
longest distance is
q
√
16 − 4 3 = 3.0119

You get the idea

Kotani’s Ant Problem – p.15/16

Conclusion
I believe that this is a fun way of using the
Pythagorean Theorem to model distances in a
(perhaps) new an unexpected way.

Kotani’s Ant Problem – p.16/16

Conclusion
I believe that this is a fun way of using the
Pythagorean Theorem to model distances in a
(perhaps) new an unexpected way.
References were at the beginning.

Kotani’s Ant Problem – p.16/16

Conclusion
I believe that this is a fun way of using the
Pythagorean Theorem to model distances in a
(perhaps) new an unexpected way.
References were at the beginning.
Thank you for listening.

Kotani’s Ant Problem – p.16/16