Maximizing the Number of Pieces

Planes, Hyperplanes, and Beyond

Jeremy L. Martin
Department of Mathematics
University of Kansas

KU Mini College
June 6, 2012

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces

Summary

Suppose that you have a cake and are allowed to make ten
straight-line slices. What is the greatest number of pieces you can
produce? What if the slices have to be symmetric — or if the cake
is four-dimensional? How can we possibly see what it looks like to
slice space into pieces using lines, planes, or hyperplanes? Many of
these questions have beautiful answers that can be revealed using
unexpected, yet essentially simple mathematical techniques. Better
yet, the seemingly abstract study of hyperplane arrangements has
many surprising practical applications, ranging from optimization
problems, to the theory of networks, to how a group of cars can
find parking spots.

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces

Themes

I Mathematics is about studying natural patterns using logic.

I Mathematics requires a high standard of proof — we can use

evidence to make conjectures, but not to draw conclusion.

I In order to understand concepts we can’t directly visualize (like

four- and higher-dimensional space), we can use analogies.

I For example, understanding the two- and three-dimensional

versions of a problem can help us understand the
four-dimensional version.

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

The Cake-Cutting Problem

What is the greatest number of pieces that a cake can be cut into
with a given number of cuts?

I The cuts must be straight lines and must go all the way
through the cake.
I The sizes and shapes of the pieces don’t matter.
I For the moment, we’ll focus on 2-dimensional cakes (think of
them as pancakes).
Planes, Hyperplanes, and Beyond
 Dimension 2
Maximizing the Number of Pieces
Dimension 3

Solutions with 2, 3 or 4 Cuts

Let’s write P2 (N) for the maximum number of pieces obtainable
using N cuts. (The 2 is a reminder of the dimension.)

2 cuts: 3 cuts: 4 cuts:

P(2) = 4 P(3) = 7 P(4) = 11

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

More Data

There is no way to obtain more than 2N pieces with n cuts (since

P2 (N) can at most double with each cut)

In fact, P2 (N) is much smaller than 2N .

Cuts N Pieces P2 (N) 2N

1 2 2
2 4 4
3 7 8
4 11 16
5 16 32
... ... ...
20 211 1048576

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

The Pattern

N P2 (N)
0 1
1 2 = 1+1
2 4 = 2+2 = 1+1+2
3 7 = 4+3 = 1+1+2+3
4 11 = 7+4 = 1+1+2+3+4
5 16 = 11 + 5 = 1+1+2+3+4+5
... ... ... ...
20 211 = 1 + 1 + 2 + 3 + · · · + 19 + 20

I How do we prove that the pattern works for every N?

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

Interlude: Counting Stairs

N+1

N N

2 x (1 + 2 + .... + N) N x (N+1)

1 + 2 + .... + N = N(N+1) / 2

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

Maximizing the Number of Pieces

How can we ensure obtaining as many pieces as possible?
I First cut the pancake into P2 (N − 1) pieces using N − 1 cuts.
I Now make the Nth cut, hitting as many pieces as possible.

N−1 cuts

Nth cut

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

Maximizing the Number of Pieces

Key observation: The number of pieces subdivided by the Nth

cut equals one more than the number of previous cuts it meets.

N−1 cuts x5
x4

Nth cut

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

Maximizing the Number of Pieces

If we make sure that

I every pair of cuts meets in some point, and
I no more than two cuts meet at any point,
then the N th cut will meet each of the previous N − 1 cuts, and
therefore will make N new pieces.

Since the original pancake had one piece, we have proved that

N(N + 1)
P2 (N) = 1 + (1 + 2 + · · · + N) = 1 + .
2

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

From 2D to 3D

Let’s write P3 (N) for the maximum number of pieces obtainable

from a 3-dimensional cake with N cuts.

P3 (1) = 2 P3 (2) = 4 P3 (3) = 8

Compare 2D: P(1) = 2, P(2) = 4, P(3) = 7.

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

From 2D to 3D

At first glance it seems that P3 (4) = 14, but in the left-hand

diagram, the four planes all meet at a point, so jiggling one plane
a little increases the number of pieces to 15.

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

From 2D to 3D

N 0 1 2 3 4 5 6 7 8
P2 (N) 1 2 4 7 11 16 22 29 37
P3 (N) 1 2 4 8 15 26 42 64 93

The pattern is

P3 (N) = P3 (N − 1) + P2 (N − 1).

(This observation, together with the formula P2 (N) = N(N+1)2 +1

and a little algebra, can be used to prove the formula
3
P3 (N) = N +5N+6
6 . But what is really important is the pattern!)

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

Pancakes, Cakes and Beyond

How about four-dimensional pancakes?

(Never mind whether they actually exist!)
More generally, if you have a d-dimensional cake and you can
make N cuts, how many pieces can you make? (Call this Pd (N).)
We already know the answer for d = 2 and d = 3. If d = 1, it is
pretty clear that N cuts give N + 1 pieces.

N
0 1 2 3 4 5 6 7 8
P1 (N) 1 2 3 4 5 6 7 8 9
P2 (N) 1 2 4 7 11 16 22 29 37
P3 (N) 1 2 4 8 15 26 42 64 93

Planes, Hyperplanes, and Beyond

 Dimension 2
Maximizing the Number of Pieces
Dimension 3

Pancakes, Cakes and Beyond

I Each number is the sum of the numbers immediately “west”

(←) and “northwest” (-).
I Formula: Pd N) = Pd (N − 1) + Pd−1 (N − 1).

N
0 1 2 3 4 5 6 7 8
P1 (N) 1 2 3 4 5 6 7 8 9
P2 (N) 1 2 4 7 11 16 22 29 37
P3 (N) 1 2 4 8 15 26 42 64 93
P4 (N) 1 2 4 8 16 31 57 99 163
P5 (N) 1 2 4 8 16 32 63 120 219
... ...

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Symmetric Cake-Cutting
What are the possible ways to cut a perfectly round cake so that
all pieces are congruent (i.e., geometrically the same)?

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Symmetric Cake-Cutting

The symmetry shows up in the equations.

x = 0, y = 0, z = 0 x = y , x = z, y = z

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Symmetric Cake-Cutting in Higher Dimensions

Question: If we can cut up a 3-dimensional sphere into congruent

pieces, using planes defined by the equations

x = 0, y = 0, z = 0 or x = y , x = z, y = z

then what happens if we cut up a 4-dimensional sphere into pieces

using the “hyperplanes”

w = 0, x = 0, w = x, w = y, w = z,
or ?
y = 0, z =0 x = y, x = z, y =z

(Note: We can replace “sphere” with “space”.)

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Symmetric Cake-Cutting in Higher Dimensions

Some tools for visualizing 4-dimensional space:

I Work by analogy: understanding low-dimensional space can
help us understand higher dimensions
I Project into lower dimension to make visualization easier
I Reexpress high-dimensional problems mathematically

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

The Braid Arrangement

The arrangement of planes x = y , x = z, y = z is called the
3-dimensional braid arrangement (Braid3 for short).

x=z

x=y

y=z
Projecting from 3D to 2D makes the diagrams simpler, and retains
both the number and symmetry of the regions.
Planes, Hyperplanes, and Beyond
 Maximizing the Number of Pieces
Symmetric Cake-Cutting

A Brief Note on Algebra

I The equations of the planes are x = y , x = z, y = z.

I All three planes meet in the line L defined by x = y = z.
I Passing from 3D to 2D can be viewed either as “squashing”
the arrangement along L, or as “slicing” perpendicularly to L.
I The two sides of each plane are defined by inequalities. E.g.,
the two sides of y = x are x < y and y < x.

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Regions Between The Planes of Braid3

Each region of Braid3 is described by three inequalities:

I either x < y or y < x, and
I either x < z or z < x, and
I either y < z or z < y .

Therefore, which region you are in corresponds to the order of the

three coordinates x, y , z. There are six possibilities:

x <y <z y <x <z z <x <y

x <z <y y <z <x z <y <x

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Regions of Braid3

111111111
0000000000000000000
1111111111
x=z
000000000
111111111
000000000
1111111110000000000
1111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
111111111
0000000001111111111
1111111110000000000
0000000001111111111
1111111110000000000
0000000001111111111
z<x<y
111111111
000000000
111111111
0000000000
0000000000
1111111111 x<z<y
0000000001111111111
1111111110000000000
0000000000
1111111111
000000000
111111111
0000000000
11111111110000000000
1111111111
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
00000000000
11111111111
0000000000
1111111111
00000000000
11111111111
0000000000
1111111111
00000000000
11111111111
0000000000
1111111111
00000000000
11111111111
y=z
000000000
111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
00000000000
11111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
00000000000
11111111111
0000000000
1111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
00000000000
11111111111
000000000
111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
00000000000
11111111111
z<y<x 0000000000
1111111111
0000000000
1111111111
0000000000
11111111110000000000
1111111111
00000000000
11111111111
0000000000
1111111111 x<y<z
0000000000
1111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111 00000000000
11111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
x=y
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
y<z<x
0000000000
1111111111
0000000000
1111111111
y<x<z
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Why “Braid”?
Crossing a border corresponds to reversing one inequality.

x=z x
y
z
1111111111
0000000000
0000000000
1111111111 y=z
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
0000000000
1111111111
x<y<z
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
000000000
111111111 x=y

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Why “Braid”?
Crossing a border corresponds to reversing one inequality.

x=z x
y
z
y=z
1111111111
0000000000
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
000000000
111111111
0000000000
1111111111
0000000000
1111111111
x=y
0000000000
1111111111
y<x<z
0000000000
1111111111
0000000000
1111111111
Planes, Hyperplanes, and Beyond
 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Why “Braid”?
Crossing a border corresponds to reversing one inequality.

x=z x
y
z
y=z
111111111
000000000
0000000001
111111111 0
0000000001
111111111 0
0
1
000000000
111111111
000000000
111111111 0
1
0000000001
111111111 0
0
1
000000000
111111111 x=y
y<z<x 101010
000000000
111111111
000000000
111111111
000000000
111111111
0000000001
111111111 0
0
1
Planes, Hyperplanes, and Beyond
 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Why “Braid”?
Crossing a border corresponds to reversing one inequality.

x=z x
y
z
11111111111y=z
00000000000
00000000000
11111111111
00000000000
11111111111
00000000000
11111111111
00000000000
11111111111
00000000000
11111111111
x<y<z
00000000000
11111111111
00000000000
11111111111
00000000000
11111111111
00000000000
11111111111
00000000000
11111111111
x=y

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

The 2-Dimensional Braid Arrangement

Before we tackle Braid4, let’s make sure we understand Braid2.

1111111111111111111
0000000000000000000
0000000000000000000 x=y
1111111111111111111
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
x<y
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
x<y y<x
0000000000000000000
1111111111111111111 1111111111111111
0000000000000000
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
x=y
y<x
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
0000000000000000000
1111111111111111111
Braid2 Projection into 1D

Note that there are 2 regions.

Planes, Hyperplanes, and Beyond
 Maximizing the Number of Pieces
Symmetric Cake-Cutting

The 4-Dimensional Braid Arrangement

The arrangement Braid4 consists of the hyperplanes defined by

the equations

w = x, w = y, w = z, x = y, x = z, y =z

in four-dimensional space.

Key observation: We can project Braid2 from 2D to 1D, and

Braid3 from 3D to 2D,
so, by analogy, we should be able to project Braid4 from 4D
to 3D!

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Here’s what Braid4 looks like!

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Suppose we put a dot in each region and connect adjacent dots. . .

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

. . . and then remove the hyperplanes, leaving only the dots.

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Regions of Braid4

The regions of Braid4 correspond to the orderings of the four

coordinates w , x, y , z:

wxyz wxzy wyxz wyzx wzxy wzyx

xwyz xwzy xywz xyzw xzwy xzyw
ywxz ywzx yxwz yxzw yzwx yzxw
zwxy zwyx zxwy zxyw zywx zyxw

I There are 4 possibilities for the first letter;

I 3 possibilities for the second, once the first is determined;
I 2 possibilities for the third, once the first two are determined;
I only 1 possibility for the last letter.

Total: 4 × 3 × 2 × 1 = 24 orderings = 24 regions.

Planes, Hyperplanes, and Beyond
 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Regions of Braid4

I We have just seen that Braid4 has 24 regions.

I The regions correspond to permutations of w , x, y , z.
I If two regions are adjacent, the corresponding permutations
differ by a single flip:

x z wy ←→ x w zy

(In particular, each region has exactly 3 neighboring regions.)

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces
Symmetric Cake-Cutting

Beyond the Fourth Dimension

The n-dimensional braid arrangement consists of the hyperplanes
defined by the equations

x1 = x2 ,
x1 = x3 , x2 = x3 ,
...
x1 = xn , x2 = xn , . . . , xn−1 = xn

I There are n(n − 1)/2 hyperplanes (by the staircase formula!)

I The regions correspond to the possible orderings of the
coordinates x1 , . . . , xn .
I The number of regions is n × (n − 1) × · · · × 3 × 2 × 1 (also
known as n factorial; notation: n!).
I Each region has n − 1 neighboring regions.
Planes, Hyperplanes, and Beyond
 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking Cars

I A group of cars enter a parking lot, one by one.

I # of parking spaces = # of cars (say n).

I The parking spaces are arranged along a one-way road.

I Each car has a preferred parking space that it drives to first. If
that spot is not available, it continues to the first empty space.

I A “parking function” is a list of preferences that allows all

cars to park.

I Application: database indexing, hash tables)

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking Functions

1 2 3 4 5 6

1 4 5 1 4 1

1 2 3 4 5 6

1 4 5 1 4

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking Functions

1 2 3 4 5 6

1 4 5 1

1 2 3 4 5 6

1 4 5

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking Functions

1 2 3 4 5 6

1 4

1 2 3 4 5 6

Success!
1

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking Functions

different 1 2 3 4 5 6

1 4 5 4 4 1

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking Functions

1 2 3 4 5 6

1 4 5 4 4

1 2 3 4 5 6

1 4 5 4

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking Functions

1 2 3 4 5 6

1 4 5

1 2 3 4 5 6

1 4

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking Two Cars

There are 4 = 22 possible lists of preferred spots.

3 of them successfully park both cars.

Car #1 preference Car #2 preference Success?

Space 1 Space 1 Yes
Space 1 Space 2 Yes
Space 2 Space 1 Yes
Space 2 Space 2 No

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking Three Cars

There are 27 = 33 possible lists of preferred spots.

16 of them successfully park all three cars.
Parking functions (the ones that work):

111 112 122 113 123 132

121 212 131 213 231
211 221 311 312 321

Non-parking functions (the ones that don’t work):

133 222 223 233 333

313 232 323
331 322 332

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking n Cars

Observation #1: Whether or not all the cars can park depends
on what their preferred spaces are, but not on the order in which
they enter the parking lot.

For example, if there are 6 cars and the preference list includes two
5’s and one 6, not all cars will be able to park.

Also, every parking function must include at least one 1. (What

are some other conditions that must be satisfied?)

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Parking n Cars

Observation #2: 3 cars =⇒ 16 parking functions.

Number of cars (n) Number of parking functions

1 1 = 20
2 3 = 31
3 16 = 42
4 125 = 53
5 1296 = 64

Conjecture: n cars =⇒ (n + 1)n−1 parking functions.

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Connecting Points
Problem: Connect n points with as few links as possible.

7 7
1 1
4 4

2 6 2 6

5 3 5 3
8 8

I It doesn’t matter where the points are or how you draw the
links — just which pairs of points are linked.
I These structures are called trees.
Planes, Hyperplanes, and Beyond
 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

How Many Trees?

1 point: 1 1 tree

2 points: 1 2 1 tree

3 points: 1 2 3 2 1 3 1 3 2 3 trees

1 2 3 4

2 3 4 1 3 4 1 2 4 1 2 3

1 2 3 4 1 4 2 3 2 3 1 4

4 points: 1 2 4 3 1 4 3 2 2 3 4 1 16 trees
1 3 2 4 2 1 3 4 3 2 1 4

1 3 4 2 2 1 4 3 3 2 4 1

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

The Shi Arrangement

I Draw two parallel lines in 2-dimensional space.

I There are three regions (“above”, “between”, “below”).

I Equations of the lines: x = y and x = y + 1.
Planes, Hyperplanes, and Beyond
 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

The Shi Arrangement

The n-dimensional Shi arrangement consists of the n(n − 1)

hyperplanes defined by the equations

x1 = x2 , x1 = x2 + 1,
x1 = x3 , x1 = x3 + 1,
...
xn−1 = xn , xn−1 = xn .

(“Take the braid arrangement, make a copy of it, and push the
copy a little bit.”)

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

The 3D Shi Arrangement

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

The 3D Shi Arrangement

x=y x = y+1

y = z+1
1 2 3

y=z
4 5 6

7 8 9 10

11 12 13
x = z+1

14 15 16
x=z

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Scoring with a Handicap

I A group of marathon runners are ranked 1 through n.

I You score one point for each other runner you beat
head-to-head.
I But, in order to score a point against a lower-ranked runner,
you must beat him/her by at least one minute.
I The possible outcomes correspond to regions of the Shi
arrangement!

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

Slicing n-Dimensional Space

(n + 1)n−1 = number of regions of the Shi arrangement

= number of handicapped-scoring outcomes

= number of trees on n + 1 points

= number of ways to park n cars

Why are all these numbers the same?

The next figure shows the correspondence between

Shi-arrangement regions and parking functions for n = 3.

Planes, Hyperplanes, and Beyond

 Maximizing the Number of Pieces Parking Cars
Parking. . . Building Trees

x=y x = y+1

y = z+1
120 110 210

y=z
020 100 200

021 010 000 201

011 001 101

x = z+1

012 002 102 x=z

Planes, Hyperplanes, and Beyond