Sunteți pe pagina 1din 10

Maths Mini Exploration

Topic - Graham’s Number

Anuj Kejriwal

Grade 11

Introduction

When asked to select a topic to investigate, I was bewildered. Many of the topics
required skills we hadn't even scraped the surface of as yet. However, one of these
topics was particularly interesting: Graham’s Number. It is a number so big, that if
one was to try and comprehend it in his/her mind, the information would be so
complex that his/her mind would in fact collapse into a blackhole. This is in fact not
exaggeration. The data content of the number is gargantuan and transcends the
amount of entropy that could be processed or even stored in a space the size of
our brain 1. It is impossible to visualize a number this big.

Take for example, a knowingly huge number, googol. The number is equivalent to
10100 - 1 with a 100 zeroes after it, or -
10,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,00
0,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000.
Now, a googolplex, is 10googol -
1010,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,
000,000,000,000,000,000,000.

These numbers are unimaginably big. To write out the actual number googolplex,
considering one writes 2 digits a second, it would take 1.51×1092 years, which is
1.1×1082 times the actual age of the universe 2.

Graham’s number is so large, that it dwarfs googol and googolplex.


The reason however, why this number is so recognizable is that it is part of a
solution to a maths problem.

1http://ibmathsresources.com/2013/04/15/grahams-number-literally-big-enough-to-collapse-
your-head-into-a-black-hole/
2 http://www.googolplexwrittenout.com
The question

Graham was trying to solve a problem in the Ramsey theory. This theory is the facet
of mathematics that investigates under which conditions, what kind of order must
occur3. Such kind of investigation occurs in many aspects of science as well. For
example, Physics. When studying the collision theory, physicists put many
conditions on the two objects colliding in order to derive certain rules and figures.

The question that Graham researched was -

Connect each pair of geometric vertices of an n-dimensional hypercube to obtain a


complete graph on 2n vertices. Color each of the edges of this graph either red or
blue. What is the smallest value of n for which every such coloring contains at least
one single-colored complete subgraph on four coplanar vertices?4

To understand this question, we use the diagram -

Figure 1 Figure 2

Figure 1 is a bicolored 3-Dimensional drawing of a cube representing a shape that


would satisfy the conditions of the question. Figure 2 shows the part of Figure 1
that does this: one single-colored 4-vertex coplanar complete subgraph.

3 http://www.numdam.org/item?id=CM_1935__2__463_0
4 http://isu.indstate.edu/ge/GEOMETRY/cubes.html
However, if the bottom edge of figure 2 was colored blue instead, it would no longer
satisfy the conditions.

And so, what the question really wants, is that no matter how one
color’s the said shape, it should always have a complete
subgraph of a single-colored 4-vertex coplanar.

To investigate such a question by trial and error proves greatly difficult because of
the complexity of shapes with increasing dimensions -

4-Dimensional 5-Dimensional 6-Dimensional

Thus a range is produced. A range in which the answer lies.

Graham and Rothschild gave this solution a name N*.


And the range - 6 ≤ N* ≤ N.
This number range was later specified to 13≤ N* ≤ N by Jerome Barkley in 2008.
The tool to further investigate this topic is the hyper operation sequence.
This is a sequence of mathematical operations that increases in complexity at every
stage.

Operation Level 0 - Counting - 1,2,3,4,5…

Operation Level 1 - Addition - 3 + 6 = 9 = shortcut to reach certain numbers

Operation Level 2 - Multiplication - 3 x 6 = 3+3+3+3+3+3 = 18 = Shorter cut to


reach numbers

Operation Level 3 - Exponentiation - This is where 3 x 3 x 3 x 3 x 3 x 3 = 36.


Now if the base and the exponent were made massive - Googolplexgoogolplex, it
reaches a point of a number one may think of the biggest there is.
However, at this operation level of 3, one can increase the complexity by increasing
the exponential power = # 36
6

To really exploit the power of exponents, we use Knuth’s up-arrow notation.


so 36 becomes 3 ↑ 6, so -
3 ↑ 6 = 729
2 ↑ 5 = 32
5 ↑ 5 = 3,125

Operation Level 4 - Tetration (↑↑)


At this level, power towers are observed.
For example - # 2 2
2

Power towers initiate with the topmost number and then operate downwards.
So, 2 2 = 24 = 16
2

33
Now, if one looks at the example of # 33
33 ( 33)
# 33 = # 33 = # 33 = # 3(3 = # 37,625,597,484,987
27 27 )

And this final number is a 3.6 trillion-digit number.


This is with a power tower of only 4.
33
Using Knuth’s notation, # 33 = 3↑(3↑(3↑3))
These four 3s can be joined together with a double arrow -
3↑↑4
This signifies that 3, is exponentiated with a power tower of 4 copies of 3.
So,
44

4↑↑5 = # 4 44

5↑↑2 = # 5 5

aa
Tetration = a↑↑b = a↑(a↑(…↑a)) =# a a } b copies of a

Now we move up another level…

Operation Level 5 - Pentation (↑↑↑)

Pentation is similar to tetration -

3↑↑↑3 = a sequence of 3↑↑3, 3 terms long -


3↑↑↑3 = 3↑↑(3↑↑3)
↑↑ = a single number b terms high, 3↑↑3 = # 33 = 7,625,597,484,987
3

3↑↑7,625,597,484,987 = a tower of 3s, 7,625,597,484,987 terms high


To put that into perspective, if one were to write out these threes, considering they
write out each 3 1cm long, the tower would rise 76 million kilometers high. That is
100 times to the moon and back.

a↑↑↑b = a↑↑(a↑↑(…↑↑a)) } - b copies of a


The first tower at the bottom has a height of x so - # a x . When multiplied out, the
next tower has a height of # a x , so =# a (a .
x)

Operation Level 6 - Hexation (↑↑↑↑)


Continuing with the sequence, we get -
3↑↑↑↑3 = a sequence of 3↑↑↑3, 3 terms long -
3↑↑↑↑3 = 3↑↑↑(3↑↑↑3)
and that is equivalent to - 3↑↑↑(3↑↑(3↑↑3)) - that is 3↑↑3, 3↑↑(3↑↑3) long.
Now that the toolkit is understood, we can talk about Graham’s number.

The first deriving term is called # g1 and it is equivalent to 3↑↑↑↑3 and 3↑↑↑↑3 =
3↑↑↑(3↑↑(3↑↑3))
But first -
3=3
# 33 = 27
# 33 = 7,625,597,484,987
3

33
# 33 = a 3.6 trillion digit number
3
33
# 33 = 3 raised to a 3.6 trillion-digit exponent

And so, with 3↑↑↑(3↑↑(3↑↑3)) we have a sequence of 3↑↑3, 3↑↑(3↑↑3) terms


long.
The explanation may sound repetitive, but to move on, one must understand the
gargantuan nature of this number. This number, is in fact, only # g1 .

To get from # g1 to # g2 , this is the rule -

You take the number when # g1 is completely


multiplied out, and you put those many arrows in
# g2 .

Now that is a number unimaginably large.


Graham’s number lies at # g64 -
Calculating the rightmost digits of Graham’s Number -
To do this, we use modular exponentiation.

When looking at power towers, the right most digits ‘d’ of the multiplied form
remain same provided that the power tower has a height of more than ‘d’+2.

For example, 3↑x has the values, 1,3,9,27 etc.


However, the rightmost digit of these numbers are - 1,3,9,7.
3↑3↑x has the values 3,27 whose rightmost digits - 3,7
3↑3↑3↑x has the values 27, 7,625,597,484,987 whose rightmost digits - 7,7
In all of these power towers, the rightmost digit, when x is taken as 3, is 7. That
means, that any power tower of more than 3 digits, with only 3s in the tower, will
completely multiply to form a number which has a rightmost digit, 7. Hence, 7 is the
last number of Graham’s number.

Using the same rule, we can form a table to obtain the last three digits-

Number of 3↑x 3↑3↑x 3↑3↑3↑x 3↑3↑3↑3↑x 3↑3↑3↑3↑3↑x 3↑3↑3↑3↑3↑3


digits (d) ↑x

1 (1,3,9,7) (3,7) (7) (7) (7) (7)

(01,03,…87,… (03,27,…83,…
2 (27,87) (87) (87) (87)
67) 87)

(001,003,…, (003,027,… (027,987,227,


3 (987,387) (387) (387)
387,…,667) 387,…,587) 387)

A more concise rule can be formed -

This rule requires modular exponentiation -

If # 33 = 27, then to obtain the number 7, we divide 27 by 10 and the remainder is 7.


If # 33 = 7,625,597,484,987, we again divide this number by 10 and obtain 7 as the
3

remainder.
However, if were to divide this number by # 10 2 , the remainder would be 87. This
shows that when looking at number of digits d, a power tower of height greater
than d+2 gives the same remainder, when divided by #10 d .
Modular exponentiation takes the form # c = b e mod m
c = remainder
b = base
e = exponent
# b e = dividend
m = divisor

The rule formed then is -


The remainder/rightmost digit = x
Number of rightmost digits wanted = d

x = # 3x mod # 10 d .
# 33 = 27
27# ÷ 10 = 2, remainder 7
# 33 = 7,625,597,484,987
3

7,625,597,484,987 # ÷ 10 = 762,559,748,498 remainder 7


7,625,597,484,987 # ÷ 100 = 76,255,974,849 remainder 87

The value of has to be an exponent multiple of 3, since only then will it be part of
Graham’s number. Using this equation, the values for the 12 rightmost digits of
Graham’s number can be found - …262464195387.
An algorithm derived by this equation was able to give the last 500 digits of
Graham’s Number.

So going back to the question,

Connect each pair of geometric vertices of an n-dimensional hypercube to obtain a


complete graph on 2n vertices. Color each of the edges of this graph either red or
blue. What is the smallest value of n for which every such coloring contains at least
one single-colored complete subgraph on four coplanar vertices?

The question was derived from a similar question which asked the smallest value of
n for which only 3 coplanar vertices were completely sub graphed…
The answer to that was # 6 ≤ x .

Thus Graham derived that in order for his conditions to coexist, the number of
dimensions would have to be between 6 and a very very great number. So great,
that it can't be completely expressed in any notation currently existing. Another
mathematician, Jerome Barkley, specified this to 13. This answer, 13≤ N* ≤ g64 is in
fact an immensely specific answer, because there is no practical method to
investigate this beyond the 13th, 14th or the 20th dimension.

Conclusion

The reason Graham’s number is recognized is because it isn't just a random large
number, but because it is a part of the proof to a mathematical question.
Today, there are other notations such as the Steinhaus–Moser notation 5 that allow
us to produce numbers such as the Skewes’ Number 6. These are also part of
mathematical proofs to various questions.

The challenge when dealing with such numbers is the fact that there is no way of
digitally representing such numbers. Thus it is impossible to visualize, calculate with
or operate on any of such numbers.

5 http://mathworld.wolfram.com/Steinhaus-MoserNotation.html
6 http://mathworld.wolfram.com/SkewesNumber.html

S-ar putea să vă placă și