From Wikipedia, the free encyclopedia
This article is now available on my blog: https://www.sligocki.com/2009/10/08/green-numbers.html
Milton Green discovered a class of Turing machines which produce extremely large results in the busy beaver game.[ 1]
The machines defined recursively and the number of 1s they leave on the tape is likewise defined recursively. We examine those numbers here:
Let us define the numbers
B
n
{\displaystyle B_{n}}
for n odd.
B
n
(
0
)
=
1
{\displaystyle B_{n}(0)=1}
B
1
(
m
)
=
m
+
1
{\displaystyle B_{1}(m)=m+1}
B
n
(
m
)
=
B
n
−
2
[
B
n
(
m
−
1
)
+
1
]
+
1
{\displaystyle B_{n}(m)=B_{n-2}[B_{n}(m-1)+1]+1}
Then, Green's numbers
B
B
n
{\displaystyle BB_{n}}
are defined as:
B
B
n
=
B
n
−
2
[
B
n
−
2
(
1
)
]
{\displaystyle BB_{n}=B_{n-2}[B_{n-2}(1)]}
for odd n
B
B
n
=
B
n
−
3
[
B
n
−
3
(
3
)
+
1
]
+
1
{\displaystyle BB_{n}=B_{n-3}[B_{n-3}(3)+1]+1}
for even n
B
1
(
m
)
=
m
+
1
{\displaystyle B_{1}(m)=m+1}
B
3
(
m
)
=
3
m
+
1
{\displaystyle B_{3}(m)=3m+1}
B
5
(
m
)
=
7
2
⋅
3
m
−
5
2
{\displaystyle B_{5}(m)={\frac {7}{2}}\cdot 3^{m}-{\frac {5}{2}}}
B
7
(
m
)
>
3
↑↑
m
{\displaystyle B_{7}(m)>3\uparrow \uparrow m}
and
B
7
(
1
)
=
B
5
(
2
)
+
1
=
29
{\displaystyle B_{7}(1)=B_{5}(2)+1=29}
B
9
(
m
)
>
3
↑↑↑
m
{\displaystyle B_{9}(m)>3\uparrow \uparrow \uparrow m}
and
B
9
(
1
)
=
B
7
(
2
)
+
1
=
B
5
(
30
)
+
2
=
720618962331271
{\displaystyle B_{9}(1)=B_{7}(2)+1=B_{5}(30)+2=720618962331271}
B
B
3
=
3
{\displaystyle BB_{3}=3}
B
B
4
=
7
{\displaystyle BB_{4}=7}
B
B
5
=
13
{\displaystyle BB_{5}=13}
B
B
6
=
35
{\displaystyle BB_{6}=35}
B
B
7
=
B
5
(
8
)
=
22961
{\displaystyle BB_{7}=B_{5}(8)=22961}
B
B
8
=
B
5
(
93
)
+
1
=
3
(
7
⋅
3
92
−
1
)
/
2
=
824792557184288824246737061810550733633916929
{\displaystyle BB_{8}=B_{5}(93)+1=3(7\cdot 3^{92}-1)/2=824792557184288824246737061810550733633916929}
B
B
9
=
B
7
(
B
7
(
1
)
)
=
B
7
(
29
)
>
3
↑↑
29
>
10
↑↑
28
{\displaystyle BB_{9}=B_{7}(B_{7}(1))=B_{7}(29)>3\uparrow \uparrow 29>10\uparrow \uparrow 28}
> a googolplex -plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex-plex (25 plexes).
B
B
10
=
B
7
(
B
7
(
3
)
+
1
)
+
1
>
3
↑↑
3
↑↑
3
=
3
↑↑↑
3
{\displaystyle BB_{10}=B_{7}(B_{7}(3)+1)+1>3\uparrow \uparrow 3\uparrow \uparrow 3=3\uparrow \uparrow \uparrow 3}
= the third Ackermann number
B
B
11
=
B
9
(
B
9
(
1
)
)
=
B
9
(
720618962331271
)
>
3
↑↑↑
720618962331271
{\displaystyle BB_{11}=B_{9}(B_{9}(1))=B_{9}(720618962331271)>3\uparrow \uparrow \uparrow 720618962331271}
We will show that
B
n
(
m
)
{\displaystyle B_{n}(m)}
grows like
3
↑
n
/
2
m
{\displaystyle 3\uparrow ^{n/2}m}
and thus that
B
B
n
{\displaystyle BB_{n}}
grows like
3
↑
n
/
2
3
{\displaystyle 3\uparrow ^{n/2}3}
(See Knuth's up-arrow notation and User:sligocki/up-arrow properties ).
Claim.
B
2
k
+
3
(
m
)
≥
3
↑
k
m
{\displaystyle B_{2k+3}(m)\geq 3\uparrow ^{k}m}
for any
k
≥
−
2
{\displaystyle k\geq -2}
and
m
≥
0
{\displaystyle m\geq 0}
Proof by induction.
Base Case
B
2
k
+
3
(
0
)
=
1
≥
1
=
3
↑
k
0
{\displaystyle B_{2k+3}(0)=1\geq 1=3\uparrow ^{k}0}
B
1
(
m
)
=
m
+
1
≥
m
+
1
=
3
↑
−
2
m
{\displaystyle B_{1}(m)=m+1\geq m+1=3\uparrow ^{-2}m}
Inductive Step
Assume that
B
2
k
′
+
3
(
m
′
)
≥
3
↑
k
′
m
′
{\displaystyle B_{2k^{\prime }+3}(m^{\prime })\geq 3\uparrow ^{k^{\prime }}m^{\prime }}
(for all
k
′
=
k
,
m
′
<
m
{\displaystyle k^{\prime }=k,m^{\prime }<m}
or
k
′
<
k
,
m
′
<
3
↑
k
(
m
−
1
)
{\displaystyle k^{\prime }<k,m^{\prime }<3\uparrow ^{k}(m-1)}
)
B
2
k
+
3
(
m
)
=
B
2
k
+
1
[
B
2
k
+
3
(
m
−
1
)
+
1
]
+
1
>
B
2
k
+
1
[
3
↑
k
(
m
−
1
)
]
>
3
↑
k
−
1
[
3
↑
k
(
m
−
1
)
]
=
3
↑
k
m
{\displaystyle B_{2k+3}(m)=B_{2k+1}[B_{2k+3}(m-1)+1]+1>B_{2k+1}[3\uparrow ^{k}(m-1)]>3\uparrow ^{k-1}[3\uparrow ^{k}(m-1)]=3\uparrow ^{k}m\,}
QED
Claim.
B
2
k
+
3
(
m
)
<
1
2
(
3
↑
k
(
m
+
2
)
)
<
(
3
↑
k
(
m
+
2
)
)
{\displaystyle B_{2k+3}(m)<{\frac {1}{2}}(3\uparrow ^{k}(m+2))<(3\uparrow ^{k}(m+2))}
for any
k
≥
1
{\displaystyle k\geq 1}
and
m
≥
0
{\displaystyle m\geq 0}
(In fact the bound can be tightened to m +1 for k ≥ 2)
Proof by induction.
Base Case
B
2
k
+
3
(
0
)
=
1
<
1
2
(
3
↑
k
2
)
{\displaystyle B_{2k+3}(0)=1<{\frac {1}{2}}(3\uparrow ^{k}2)}
B
5
(
m
)
=
9
2
3
m
<
1
2
(
3
↑
(
m
+
2
)
)
{\displaystyle B_{5}(m)={\frac {9}{2}}3^{m}<{\frac {1}{2}}(3\uparrow (m+2))}
B
7
(
m
)
<
1
2
(
3
↑
2
(
m
+
1
)
)
{\displaystyle B_{7}(m)<{\frac {1}{2}}(3\uparrow ^{2}(m+1))}
(left as an exercise to the reader)
Inductive Step
Assume that
B
2
k
′
+
3
(
m
′
)
<
1
2
(
3
↑
k
′
(
m
′
+
2
)
)
{\displaystyle B_{2k^{\prime }+3}(m^{\prime })<{\frac {1}{2}}(3\uparrow ^{k^{\prime }}(m^{\prime }+2))}
(for
k
′
=
k
,
m
′
<
m
{\displaystyle k^{\prime }=k,m^{\prime }<m}
or
k
′
<
k
,
m
′
<
3
↑
k
m
+
1
{\displaystyle k^{\prime }<k,m^{\prime }<3\uparrow ^{k}m+1}
)
B
2
k
+
3
(
m
)
=
B
2
k
+
1
[
B
2
k
+
3
(
m
−
1
)
+
1
]
+
1
<
B
2
k
+
1
[
1
2
(
3
↑
k
(
m
+
1
)
)
+
1
]
+
1
<
1
2
(
3
↑
k
−
1
[
1
2
(
3
↑
k
(
m
+
1
)
)
+
3
]
)
+
1
≤
?
1
2
(
3
↑
k
−
1
[
1
2
(
3
↑
k
(
m
+
1
)
)
+
4
]
)
≤
?
1
2
(
3
↑
k
−
1
(
3
↑
k
(
m
+
1
)
)
)
=
1
2
(
3
↑
k
(
m
+
2
)
)
{\displaystyle {\begin{array}{rll}B_{2k+3}(m)=B_{2k+1}[B_{2k+3}(m-1)+1]+1&<&B_{2k+1}[{\frac {1}{2}}(3\uparrow ^{k}(m+1))+1]+1\\&<&{\frac {1}{2}}(3\uparrow ^{k-1}[{\frac {1}{2}}(3\uparrow ^{k}(m+1))+3])+1\\&\leq ?&{\frac {1}{2}}(3\uparrow ^{k-1}[{\frac {1}{2}}(3\uparrow ^{k}(m+1))+4])\\&\leq ?&{\frac {1}{2}}(3\uparrow ^{k-1}(3\uparrow ^{k}(m+1)))\\&=&{\frac {1}{2}}(3\uparrow ^{k}(m+2))\end{array}}}
≤? statements seem obvious, but grungy to prove (left as an exercise to the reader :)
Claim.
3
↑
k
3
<
B
B
2
k
+
4
,
B
B
2
k
+
5
<
3
↑
k
+
1
3
{\displaystyle 3\uparrow ^{k}3<BB_{2k+4},BB_{2k+5}<3\uparrow ^{k+1}3}
for
k
≥
1
{\displaystyle k\geq 1}
Proof.
B
B
2
k
+
5
=
B
2
k
+
3
(
B
2
k
+
3
(
1
)
)
>
3
↑
k
(
3
↑
k
1
)
=
3
↑
k
3
{\displaystyle BB_{2k+5}=B_{2k+3}(B_{2k+3}(1))>3\uparrow ^{k}(3\uparrow ^{k}1)=3\uparrow ^{k}3}
B
B
2
k
+
4
>
B
2
k
+
1
(
B
2
k
+
1
(
3
)
)
>
3
↑
k
−
1
(
3
↑
k
−
1
3
)
=
3
↑
k
3
{\displaystyle BB_{2k+4}>B_{2k+1}(B_{2k+1}(3))>3\uparrow ^{k-1}(3\uparrow ^{k-1}3)=3\uparrow ^{k}3}
B
B
2
k
+
5
=
B
2
k
+
3
(
B
2
k
+
3
(
1
)
)
<
1
2
(
3
↑
k
(
1
2
(
3
↑
k
(
1
+
2
)
)
+
2
)
)
<
3
↑
k
(
3
↑
k
3
)
=
3
↑
k
+
1
3
{\displaystyle BB_{2k+5}=B_{2k+3}(B_{2k+3}(1))<{\frac {1}{2}}(3\uparrow ^{k}({\frac {1}{2}}(3\uparrow ^{k}(1+2))+2))<3\uparrow ^{k}(3\uparrow ^{k}3)=3\uparrow ^{k+1}3}
B
B
2
k
+
4
=
B
2
k
+
1
[
B
2
k
+
1
(
3
)
+
1
]
+
1
<
1
2
(
3
↑
k
−
1
(
1
2
(
3
↑
k
−
1
(
3
+
2
)
)
+
3
)
)
+
1
<
3
↑
k
−
1
(
3
↑
k
−
1
5
)
<
3
↑
k
4
<
3
↑
k
+
1
3
{\displaystyle BB_{2k+4}=B_{2k+1}[B_{2k+1}(3)+1]+1<{\frac {1}{2}}(3\uparrow ^{k-1}({\frac {1}{2}}(3\uparrow ^{k-1}(3+2))+3))+1<3\uparrow ^{k-1}(3\uparrow ^{k-1}5)<3\uparrow ^{k}4<3\uparrow ^{k+1}3}
QED
^ Milton W. Green. A Lower Bound on Rado's Sigma Function for Binary Turing Machines. In Preceedings of the IEEE Fifth Annual Symposium on Switching Circuits Theory and Logical Design , pages 91–94, 1964.