Вы находитесь на странице: 1из 4

Primitive Juggling Sequences

Adam Bender

Graph Theory

3/2/2017

Dr. Donovan
Primitive Juggling Sequences

Fan Chung and Ron Graham Wrote a paper for The Mathematical Association
of America monthly publication #115 about juggling. Fan Chung is a mathematician
at the university of California at San Diego who specializes in Spectral Graph Theory
and Random Graphs. Ron Graham holds the chair of computer and information
science at the same university and specializes in Ramsey Theory and theoretical
computer science. Juggling has always been a personal passion of mine and I even
spent a few years with the circus and the Minnesota Renaissance Festival working
as a juggler and fire breather, so when I heard there was a strong connection to
mathematics and graph theory I had to read that paper. The paper itself was 9
pages long and worked through several different types of mathematics which
required the uninitiated reader to pause and look up what was being referenced on
many occasions.

The primary focus on the paper are juggling sequences. A juggling sequence
is denoted by T=( t1, t2, t3,... tn) where all ti's are nonnegative integers. The values
for ti must also satisfy the following condition:

i+ ti (mod n) must be distant for 1 i n.

In all cases, n will denote the number of balls the juggler is using. and the
various ti will denote how high a ball is being thrown, or how long it will be before
the ball lands again. A valid juggling sequence is one which can be repeated ad
nauseam without having 2 balls land at the same time. Since a ball is thrown at
time i, to a height of ti it will return to the jugglers hand at a time i+t i; which is why
the values for i +ti must be distinct mod n. The authors then list a number of valid
juggling sequences such as T0 = (5, 3, 0, 4, 2, 5, 2, 6, 1, 2) and show how it can be
decomposed into 3 shorter sequences (5, 3, 0, 4) (2, 5, 2) (6, 1, 2). One interesting
thing to note is that the sum of the sequence mod 3 is always 0. An illustration is
then produced of the valid juggling sequence (4,1,5,2), and its associated "state"
(1101) which is a binary vector that represents when the balls fall into the jugglers
hands after they stop repeating the pattern.
These, states are the key to forming the bridge to graph theory. When the
juggler is in any such a state he only has one choice to make, what to do with the
next ball, he can throw it to a height of 1,2,3...n, for the sake of simplicity we
assume the juggler will not throw the ball higher than a 5. For example if a juggler
had 1 ball, and was in state (10000) he could throw it to a height of 3 and then be in
state (00100) since the ball will land 3 time units later. A digraph is then
constructed, whose cycles are valid juggling sequences.

For example, starting with the state (11100) and throwing "3" gives us
(11100) and the juggler can do that exactly as many times as he desires which
makes (3) a valid juggling sequence (and in fact the first one all jugglers learn).
This leads us into the heart of the paper, which is enumerating ground state
juggling sequences.

Some definitions:
let a ground state for b balls be: (b) = (1,1,....1) for a total number of b 1's.
A juggling sequence is "primitive" if it cannot be written as the product of two
shorter sequences.
we now let b be the number of balls and n be the period of the sequence.
J(b,n) denotes the number of juggling sequences with ground state (b).

The proof of theorem 1 was a direct result from matrix theory. A matrix is
constructed whose entries are either zero or one; and the permanent of that matrix
is used to show the above result. Once J(b,p) is established we can use that result to
derive P(b,n), which is the number of primitive ground states.
This result is from finding a generating function and a recurrence relation to
J(b,n). The paper concludes with a number of answered questions, among them is
one that I personally find very intriguing: What are the analogous results for
multiplex juggling patterns? In this case, many
balls (i.e., more than one) are allowed to be thrown and caught at each time instant.
This question is of interest to me for two reasons, that is my favorite way to juggle,
and I believe the graph of all possible states of multiplex juggling would have a very
unique structure.

Вам также может понравиться