Ch3 TuringMachine

CHNG 3
L thuyt tnh ton
M HNH M
MY TURING
(Theory of Computation)
(Turing Machines Model)
PGS.TS. Phan Huy Kh

Khnh
phkhanh@dut.udn.vn,
phkhanh@dut.udn.vn, khanhph29@gmail.com
Chng 3
Alan Mathison Turing (23/06/1912 07/06/1954)

07/06/1954)
nh
nh to
ton h
hc,logic h
hc v mt m h
hc ng
ngi Anh
th
thng
c xem l
l cha
ca ng
ngnh Khoa h
hc M
My t
tnh
2/189
The Chomsky
Chomskys Language Hierarchy
Handicapped machines
DFA
Languages accepted by
Turing Machines
2-way DFA has equivalent power
Context-Sensitive Languages
anbncn ?
ww ?
Tape is read-only
Tape length is a constant
PDA
2-way PDA has a little more power (can accept { ww | w* }

Tape is read-only, but stack is writable
Stack has only LIFO (last-in, first-out) access
Tape length is constant, but stack is not bounded
Regular Languages
a * a *b*
DFA
(Push-Down Deterministic Finite Automaton) limitations
Tape head moves only one direction
Context-Free Languages
a n b n wwR
PDA
(Deterministic Finite Automaton) limitations
Tape head moves only one direction
What
about:
Writable, 2-way tape?

Random-access stack?
3/189
4/189
Introduction to Turing Machines (TM)

TM)
Computability
Computible
(there is an algorithm)
TM:
a formal model of a (digital) computer

running a particular program
Equivalence of DFAs
The
String membership in a regular language
formalisms equivalent :
Random Access Machines
Uncomputible
Partial recursive functions (Kleene, 1952)
Equivalence of CFGs
-calculus (Church, 1936)
Emptiness of the complement of a CFL

(emptiness of CFL is computible)
Post systems (Post, 1943)

etc...
Whether a C program ever terminates
these were preceded by the work of Kurt Godel, 1931,

which in his incompleteness theorem showed that there was no
way for a computer to answer all mathematical problems
All
How to prove it?
5/189
L thuyt tnh ton -- Ch.3 - My Turing
TM is the invention of Alan Turing, 1936
Other
Determining valid executions of a PDA
6/189
M t
t tng qu
qut m
my Turing
Input string
Blank symbol
Tape
...
A Turing Machine
b a c
...
Read-Write head
Mt bng v h
hn c
c hai ph
pha tr
tri v
v ph
phi
Mt u c-ghi c hay ghi ln bng,

bng, theo hai chi
chiu
Mt b nh
nh ph
ph hu h
hn x
xc
nh tr
trng th
thi
Mt cu v
vo
c
t trn bng
Mt chng tr
trnh l
lm ho
hot
ng m
my Turing
7/189
8/189
M t
t chi ti
tit c
cc th
thnh ph
phn TM
Bng
M t
t chi ti
tit c
cc th
thnh ph
phn (2)
l
lm vi
vic :
Mt
c chia th
thnh c
cc lin ti
tip
b
b nh
nh ph
ph hu h
hn :
Ti
Tip c
cn
c n
ni dung t
ti m
mi th
thi i
im
Mi ch
cha m
mt k t
t no
a S
S c
c ch
cha k t
t tr
trng (blank symbol),, ho
hoc #
C th
th thay
i
c n
ni dung
Ph
t gi
gia t nh
nht hai k t
t #
Phn cu c
c ch (usable part) wS*
Ng
Ngi ta vi
vit #w#
Mi ph
phn t
t ca b
b nh
nh ph
ph xc
nh m
mt tr
trng th
thi
B nh
ngha
nh ph
ph
c li
lit k l
lc
nh ngh
cghi :
Chng
c hay ghi (l
(lm thay
i) n
ni dung t
tng c
ca bng
tr
trnh :
Tp h
hp h
hu h
hn c
cc quy t
tc
Di chuy
chuyn m
mt v
v tr
tr qua ph
phi, ho
hoc tr
tri, tng ng v
vi m
mt
Lm thay
i n
ni dung trn bng v
v ni dung b
b nh
nh ph
ph
Ti m
ch duy nh
nht m
mt
c
c, ho
hoc
c ghi
mi th
thi i
im, ch
9/189
10/189
A Turing Machine
Input tape
The Tape
Input String w *
...
...
...
No boundaries -- infinite length
...
Read-Write head
The head moves Left or Right
Read-Write head
Control Unit / Transition Diagram
The head at each time step:

1. Reads a symbol
2. Writes a symbol
3. Moves Left or Right
11/189
12/189
Example Moves Left :
Example Moves Right :
Time 0
...
a b
...
...
...
...
Time 1
...
b k
c
1. Reads
Time 1
b k
...
...
Time 2
1. Reads
2. Writes
2. Writes
3. Moves Right
3. Moves Left
13/189
14/189
The Input String

Input string
...
a b a c
States & Transitions
Write
Read
Blank symbol
...
q1
head
Move Left
a b, L
q2
Move Right
Head starts at the leftmost position

of the input string
Remark: the input string is never empty
q1
a b, R
q2
15/189
16/189
Example Moves Right :

Time 1
...
a b aa c
current state
...
q1
...
Time 1
...
a b a c
current state
q1
Time 2
a b bb c
a b, R
Example Moves Left :
...
...
q2
q1
q2
17/189
...
...
q1
Time 2
a b bb c
a b, L
q2
q2
18/189
Example:
Determinism (n/ti
n/t
n/ti
n/tt
nh)
Time 1
...
a b a c
current state
Time 2
...
Mi k t c xong ch tng ng vi mt dch chuyn duy nht
a b, R
...
q2
b d,L q
3
q1
c e,L
q2
g, R
Not Allowed
Allowed
q1
a b b c g
q1
Turing Machines are deterministic :
...
q2
a b, R
q2
a d, L
q3
q1
q1
No epsilon transitions allowed

19/189
20/189
Partial Transition Function
Halting
Example:
...
a b a c
...
The machine halts if there are
no possible transitions to follow
q1
a b, R
q2
Allowed:
q1
b d, L
No transition
for input symbol c
q3
21/189
22/189
Example:
...
Final States
a b a c
...
q1
a b, R
q2
b d, L
q3
q1
No possible transition
HALT!!!
q2
Allowed
q1
q2
Not Allowed
Final states have no outgoing transitions

In a final state the machine halts
23/189
q1
24/189
Acceptance
Turing Machine Example (1)
Accept Input
If machine halts
in a final state
A Turing machine that accepts the language:
(This Turing machine mimic a DFA)
Reject Input
If machine halts
in a non-final state
or
If machine enters
an infinite loop
aa* = a+
a a, R
q0
, L
q1
25/189

Time 0
26/189

Time 2
a a a
a a a
q0
Time 1
a a a
a a, R
q0
q0
, L
a a a
Time 3
q0
a,
a, R
aa
q1
q0
q0
, L
q1
27/189
28/189
Rejection Example (1)

Time 4
Time 0
a a a
a b a
q0
q1
Halt & Accept
a a, R
q0
, L
q0
q1
29/189
a a, R
, L
q1
30/189
Rejection Example (2)
Infinite Loop Example (1)

A Turing machine
for language aa * +b( a + b) *
a b a
Time 1
q0
No possible Transition
a a, R
q0
Halt & Reject
, L
b b, L
a a, R
q0
q1
, L
q1
31/189
32/189
a b a
Time 0
a b a
Time 1
q0
q0
b b, L
a a, R
q0
b b, L
a a, R
, L
q0
q1
, L
q1
33/189
34/189

Time 2
a b a
Time 2
b b, L
a a, R
q0
q0
q0
, L
Time 3
q0
Time 4
a
q0
q1
Time 5
35/189
q0
36/189
Infinite Loop Result
Another TM Example
T = <, Q, P>
Mt cch khc m t ngn gn TM
= {1, #}, hoc = {1} Nu khng mun ch k t trng ca bng
Q = {q1, q2, q3}
P = { q1, 1 1, R, q2,
c mt s chn s 1, v li q1
q2, 1 1, R, q1,
1|1, R
q1, # #, L, q3,
q3, 1 1, L, q3,
2
1
1|1, R
q3, 1 1, R, q3,
q3, 1 #, R, q1 }
1|#, R
#|#, L
Because of the infinite loop:

The final state cannot be reached
The machine never halts

The input is not accepted
q3
1|1, R
Bi
Bitp
tp::M
Mt
thot
hotng
ng
on
onnhn
nhncu
cu111111
111111??
Tha nhn cu c s chn s 1
37/189
TM for the language anbn (1)
1|1, L
38/189
a a b b
Time 0
q0
q4
y y, R
q3
, L
y y, R
q0
y y, R
a a, R
a x, R
q1
y y, L
a a, L
b y, L
x x, R
q2
X, Y
q3
, L
y y, R
q0
q3
, L
y y, R
q0
b y, L
x x, R
q2
40/189
x a b b
Time 2
q1
y y, R
q1
y y, L
a a, L
x a b b
q4
y y, R
a a, R
a x, R
39/189

Time 1
q4
y y, R
q1
y y, R
a a, R
a x, R
q1
y y, L
a a, L
b y, L
x x, R
q2
41/189
y y, R
q3
q4
, L
y y, R
q0
y y, R
a a, R
a x, R
q1
y y, L
a a, L
b y, L
x x, R
q2
42/189
x a y b
Time 3
x a y b
Time 4
q2
q3
y y, R
a a, R
q4
y y, R
, L
y y, R
q0
q2
a x, R
q1
y y, L
a a, L
b y, L
x x, R
q2
q3
, L
y y, R
q0
q3
, L
y y, R
q0
y y, R
a a, R
q1
y y, L
a a, L
b y, L
q2
q4
y y, R
q3
, L
y y, R
q0
y y, R
q0
q1
q2
46/189
x x y y
Time 8
q2
y y, R
a a, R
a x, R
b y, L
x x, R
q1
q3
q1
y y, L
a a, L
x x y b
, L
y y, R
a a, R
a x, R
45/189
y y, R
44/189
q1
a x, R
q4
q2
x x y b
Time 6
x x, R
Time 7
b y, L
x a y b
q4
q1
y y, L
a a, L
x x, R
q0
y y, R
y y, R
a a, R
a x, R
43/189

Time 5
q4
y y, R
y y, L
a a, L
b y, L
x x, R
q2
47/189
y y, R
q3
q4
, L
y y, R
q0
y y, R
a a, R
a x, R
q1
y y, L
a a, L
b y, L
x x, R
q2
48/189
x x y y
Time 9
x x y y
Time 10
q2
q4
y y, R
q3
, L
y y, R
q0
q0
y y, R
a a, R
a x, R
q1
y y, L
a a, L
b y, L
x x, R
q2
q3
, L
y y, R
q0
q1
q3
, L
y y, R
q0
q1
x x y y
Time 12
y y, L
a a, L
b y, L
x x, R
q2
51/189
x x y y
Halt & Accept
q4
q4
y y, R
a a, R
y y, R
q3
, L
y y, R
q0
a x, R
q1
y y, R
q3
q4
, L
y y, R
q0
a x, R
q1
y y, L
a a, L
x x, R
b y, L
x x, R
If we modify the TM
for the Context-Free Language
b y, L
y y, L
a a, L
y y, R
a a, R
q2
52/189
Observation:

Time 13
50/189
q3
y y, R
a a, R
a x, R
q2
x x y y
q4
b y, L
x x, R
q3
y y, R
y y, L
a a, L
y y, R
a a, R
a x, R
49/189

Time 11
q4
y y, R
{a b }
we can easily construct a TM

n n n
for the Context-Sensitive Language { a b c }
q2
53/189
54/189
Algorithm for Accepting anbncn, n1
TM Accepting anbncn, n1
a/a, R ; y/y, R
a/x, R
b/b, R ; z/z, R
b/y, R
q0
q1
z/z, L ; b/b, L
y/y, L ; a/a, L
c/z, L
q2
q3
x/x, R
y/y, R
z/z, R
#/#, L
q4
y/y, R
q5
Idea:
Idea: Finding
Finding matching
matching
a,b,c
a,b,c and
and replacing
replacing by
by x,y,z
x,y,z
q6
accepting matched x,y,z
z/z, R
Algorithm
input abc:
q0abc xq1bc xyq2c xq3yz q3xyz xq0yz xyq4z xyzq5B xyzBq6B
input a2b2c2:
q0 a2b2c2 xaq1b2c2 xayq2bc2 xaybq2c2 xayq3bzc xaq3ybzc
xq3aybzc q3xaybzc xq0aybzc xxq1ybzc xxyyq2zc xxyyq3zz
xq3xyyzz xxq0yyzz xxyyq4zz xxyyzzq5B xxyyzzBq6B
55/189
Begin
While Cha g
gp x
x Do
Gp a, ghi x
x ; Do qua R
R While Cha g
gp b
b Od
Gp b, ghi y
y ; Do qua R
R While Cha g
gp c
c Od
Gp c, ghi z
z ; Do qua L
L While Cha g
gp x
x Od
Qua R
R
Od
Do Qua R
R While Gp y
y Od
b/b, R, z/z, R z/z, L, b/b, L
a/a, R, y/y, R
y/y, L, a/a, L
Do Qua R
R While Gp z
z Od
a/x, R
b/y, R
c/z, L
If
Gp #
#
q0
q1
q2
q3
Then Qua L, Yes
Yes
x/x, R
No
Else
No
y/y, R
Fi
z/z, R
End
q4
#/#, L
q5
y/y, R
q6
56/189
z/z, R
Transition Function
A Move to right:
Formal Definitions
for
Turing Machines
q1
a b, R
q2
(q1, a ) = (q2 , b, R)
A Move to left:
q1
c d, L
q2
(q1, c) = (q2 , d , L)
58/189
Turing Machine:
States
Input
alphabet
Configuration
Tape
alphabet
c a b a
q1
M = (Q, , , , q0 , , F )
Instantaneous description:
Transition
function
Initial
state
Set of
Final states
ca q1 ba
blank
59/189
60/189
10
Transition
Time i
Transition process
Time i+1
x a y b
x a y b
q2
q0
A Move:
Time 1
x a y b
Time 2
x a y b
q2
q0
Time 3
x x y b
q2 xayb x q0 ayb
Time 4
x x y b
q1
q1
q2 xayb
x q0
ayb
xx q1
yb
xxy q1 b
61/189
62/189
Transition process
Notation
q2 xayb x q0 ayb xx q1 yb xxy q1 b
Input string
Equivalent notation:
q2 xayb
q0 w
Initial configuration:
a a b b
xxy q1 b
q0
63/189
TM computes palindromes
The Accepted Language
L( M)
| q0
nh du con a
mt tri nht
x1 q f x2 }
Initial state
q3
#|#, R
#|#, R
Vt qua tri
q4
Gp # qua phi
b|#, L
q5
qf F
Odd palindrome
a|#, L
b|#, R
Final state
#|#, L
q6
a|a, L
b|b, L
q7
#|#, R
Odd palindrome
a|a, R
b|b, R
L(M) is the Language Accepted by TM M
#|#, R
Even palindrome
65/189
# | #, L
q2
a|#, R
q1
nh du con a
mt phi nht
a|a, R
b|b, R Gp # li li
Vt qua phi
For any Turing Machine (TM) M:
={w
64/189
Ex1: w=abbbba?
Ex2: Draw M: L(M)=w#w for = {a, b, c} ?
66/189
11
Definitions
Standard Turing Machine

The machine we described is the standard:
The
languages accepted by TM
are called Recursively Enumerable (lit k quy)
Recursion is related to recursion in programs
Enumerable means there is some TM that can enumerate the strings in the
language
language is Recursive if it is accepted by some TM

that halts on all inputs
Deterministic
Infinite tape in both directions
We'll
usually assume that

TM halt if they enter a final state
Tape is the input/output file (online)
67/189
68/189
Notation
f ( w)
A function
Domain:
Computing Functions
with
Turing Machines
w D
has:
Result Region:
f (w)
f ( w) R
f can be TotaL or Partial Function

70/189
Data representation
Integer Domain
A function may have many parameters
The number of parameters is called arity
Example:
Addition function
f ( x, y ) = x + y
Thp phn
Binary:
101
Nh phn
Unary:
11111
n phn
We prefer unary representation:
with 2-arity
easier to manipulate with TMs, first

71/189
Decimal:
72/189
12
In other words:
Definition:
f
A function
is computable if
there is a Turing Machine M such that:
Initial configuration
For all
q f final state
w D
q f f ( w)
f (w)
q0 initial state
q0 w
Final configuration
f
A function
is computable if
there is a Turing Machine M such that:
Initial
Configuration
Final
Configuration
For all
Domain
w D
Domain
73/189
74/189
Example
The function
Input representation
f ( x, y ) = x + y is computable
x, y
are integers
Start
1 1
1 0 1 L 1
initial state q0
Turing Machine:
Input string:
x0 y
unary
Output string:
xy 0
unary
The 0 is the delimiter that

separates the two numbers
75/189
76/189
Computing Function
y
x
Start
1 1
Computing Function
The 0 helps when we use

the result for other operations
1 0 1 L 1
q0 initial state
x+ y
x+ y
Finish
1 1
1 1 0
q f final state
Finish
1 1
1 1 0
q f final state
77/189
78/189
13
Execution Example (1)
Turing machine for function
f ( x, y ) = x + y
1 1, R
1 1, R
x = 11 (2)
y = 11 (2)
1 1, L
Time 0
q0 0 1, R q1 , L q2 1 0, L q3
, R
q4
x+ y
1 1 0 1 1
Draw TM for inc(x), dec(x), double(x), x*y?
q4
79/189
80/189
1 1 0 1 1
Time 1
1 1 0 1 1
q0
q0
1 1, R
1 1 1 1 0
q0
Exercice:
Time 0
Final Result
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
1 1, R
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
, R
q4
, R
q4
81/189

Time 2
1 1 0 1 1
Time 3
1 1 1 1 1
q0
1 1, R
82/189
q1
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
1 1, R
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
, R
q4
83/189
, R
q4
84/189
14

Time 4
1 1 1 1 1
Time 5
1 1 1 1 1
q1
q1
1 1, R
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
1 1, R
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
, R
q4
, R
q4
85/189

Time 6
1 1 1 1 1
Time 7
1 1 1 1 0
q2
1 1, R
1 1, R
q3
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
1 1, R
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
, R
q4
, R
q4
87/189

Time 8
88/189
1 1 1 1 0
Time 9
1 1 1 1 0
q3
1 1, R
86/189
q3
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
1 1, R
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
, R
q4
89/189
, R
q4
90/189
15

Time 10
1 1 1 1 0
Time 11
q3
1 1, R
1 1 1 1 0
q3
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
1 1, R
1 1, R
1 1, L
q0 0 1, R q1 , L q2 1 0, L q3
, R
q4
91/189

Time 12
, R
q4
Another Example: f(x) = 2x

The function
1 1 1 1 0
f ( x) = 2 x
q4
1 1, R
1 1, R
1 1, L
HALT & accept
Another Example: f(x) = 2x
Output string:
1 1
q4
unary
xx
unary
94/189
Replace every 1 with $
TM Pseudocode for f(x) = 2x
(2)
A
A Tip!
Tip!
Repeat:
Find rightmost $, replace it with 1
2x
1 1
is integer
93/189
q0 initial state
Finish
is computable
, R
x
Start
(1)
Turing Machine:
Input string:
q0 0 1, R q1 , L q2 1 0, L q3
92/189
1 1 1
Go to right end, insert 1

Until no more $ remain
q f final state
95/189
96/189
16
Example TM for f(x) = 2x
TM compute succ(n)
Finish
Start
1 1 1 1
1 1
q0
q3
1 $, R
1 1, L
1 1, R
q 0 , L q1 $ 1, R
1.
1.Replace
Replaceevery
every11with
with$$
2.
2.Repeat:
Repeat:
Find
Findrightmost
rightmost$,
$,
replace
replaceititwith
with11
Go
to
Go toright
rightend,
end,insert
insert11
3.
3.Until
Untilno
nomore
more$$remain
remain
, R
q3
T = <, Q, P> ; = { 0, 1, # } ; Q = {q1, q2, q3}

P = { q1, 1 1, R, q1,
q2, 0 1, L, q3,
q1, 0 0, R, q1,
q2, 1 0 L, q2,
q1, # #, L, q2,
q2, # 1, L, q3 }
1|0, L
1|1, R
q2
0|1, L
q1
1, L
q3
#|1, L
97/189
0|0, R
#|1, R
Find a 1|1, R
X|X, L
X|X, R
q1
a|X, L
q0
0|1, R
#|#, L
98/189
Another Example
The function is
computable
1|0, L
if
x> y
if
x y
Turing Machine for
q3
x0 y
Input:
0|0, R
X|#, L
f ( x, y ) =
Increment 1
1|1, R
q2
Input string:
#aaaaaa#
Output string? #110###
Exercice:
Exercice:Draw
DrawTM
TM for
for Inverse?
Inverse?
Output:
or
99/189
100/189
TM move RWH to the right at nth = 2
Turing Machine Pseudocode:

Repeat
Assume that in the Input Tape, we have Input String w1$w2$$wk
Match a 1 from
Until all of
q2
0|0, R
An example: Unary to Binary Conversion
Erase X
#|#, L
or
x with a 1 from y
y
is matched
If a 1 from x is not matched

erase tape, write 1
else
erase tape, write 0
( x > y)
( x y)
101/189
wi , $
Draw a TM in order to move RWH to the right at w3:

P = { q1, x x, R, q1, vi x
q1, $ $, R, q2, vi $
q2, x x, R, q2, vi x
Tng
Di RWH
RWH v
v v
v tr
tr n,
n,
Tng qut
qut :: Di
q2, $ $, R, q3, vi y
vi
vi nn t
t trc
trc cu
cu vo
vo
q3, x x, L, q4, vi x }
n$w
$w22$$w
$$wkk
n$w11$w
x|x, R
x | x, R
x\
x\
q1
y | y, L
$ | $, R
$
q2
q3
x | x, L
x
q4
102/189
17
CopyG : G* (G*)2 or w w#w
TM move RWH one case to the right

# 0
##
0#
P = { q1, 0
q1, 1
q(0), 0
1#
q(0), 1
q(1), 0
q(1), 1
q(0), #
q(1), #
1 | 0, R Restore 0,
#
q(#), 0
to the right q(#), 1
}
0 | #, R
Or
# 1
##
0 | 0, R
Remember 0
0 | #, R
q1
q(0)
0 | 1, R
#, R, q(0),
#, R, q(1),
0, R, q(0),
0, R, q(1),
1, R, q(0),
1, R, q(1),
0, R, q(#),
1, R, q(#),
#, R, q(0),
#, R, q(1)
0 | 0, R
1 | a, R
Remember 1
b | b, R
a | 1, R
b | b, L
b | 0, R
b | b, L
0 | b, R
# | #, L
q(1)
# | 1, R
q3
q1
# | 1, R
1 | #, R
b | b, R
# | 0, L
q4
q5
q6
Restore 1
1 | 1, L
0 | 0, R
1 | 1, R
103/189
1 | 1, R
1 | 1, L
# | 1, L
q2
q#
1 | 0, R
0 | 0, L
1 | 1, R
w w#wRR?
0 | 0, L
104/189
Block Diagram
Combining Turing Machines

input
Turing
Machine
output
106/189
Example:
f ( x, y ) =
x + y if x > y
0
x, y
x, y
Comparer
if x y
Turings Thesis
Adder
x+ y
Eraser
x> y
x y
107/189
18
Computer Science Law:
Turings thesis*:
Any computation carried out
by mechanical means
can be performed by a Turing Machine
(1930)
A computation is mechanical
if and only if
it can be performed by a Turing Machine
There is no known model of computation

more powerful than Turing Machines
Thesis:
Thesis: AA Accepted
Accepted Statement
Statement but
but not
not Demontration
Demontration
109/189
Definition of Algorithm:
110/189
Algorithms are Turing Machines
An algorithm for function f (w)

is a
Turing Machine which computes f (w)
When we say:
There exists an algorithm
We mean:
There exists a Turing Machine
that executes the algorithm
111/189
112/189
The Standard Model

Infinite Tape
aababbcac a
Variations
of the Turing Machine
Read-Write Head
(Left or Right)
Control Unit
Deterministic
114/189
19
Variations of the Standard Model
Variations of the Standard Model

The variations form different
Turing Machine Classes
Turing machines with:

1. Stay-Option
2. Semi-Infinite Tape
3. Off-Line
4. Multitape
5. Multidimensional
6. Nondeterministic
We want to prove:
(or Equivalence)
Each Class has the same
power
with the Standard Model

What is the Same Power of two classes?
115/189
116/189
Notion
Demonstration
Same Power of two classes means:
Simulation: a technique to prove same power
Both classes of Turing machines accept

the same languages
M1 of first class
there is a machine M 2 of second class
Simulate the machine of one class

with a machine of the other class
For any machine

such that:
L( M1 ) = L( M 2 )
And vice-versa
First Class
Original Machine
M1
Second Class
Simulation Machine
M2
M1
117/189
Correspondence of the
configurations
Correspondence of the
configurations
Configurations in the Original Machine

correspond to configurations
in the Simulation Machine
Original Machine:
Simulation Machine:
d 0 d1
d 0 d1 L
118/189
dn
d n
119/189
Final Configuration
Original Machine:
df
Simulation Machine:
d f
The Simulation Machine

and the Original Machine
accept the same language
120/189
20
Example:
Turing Machines with StayStay-Option

The head can stay in the same position
aababbcac a
Time 1
aababbcac a
q1
Time 2
Left, Right, Stay
b ab abb c ac a
L,R,S: moves
q2
121/189
Theorem:
q1 a b, S
q2
Proof of Part 1 :
Part 1: Stay-Option Machines
are at least as powerful as
Standard machines
Stay-Option Machines
have the same power with
Standard Turing machines
Proof: a Standard machine is also

a Stay-Option machine
(that never uses the S move)
123/189
Proof of Part 2 :
Stay-Option Machine
q1
a standard machine can simulate

a Stay-Option machine
125/189
124/189
Simulation for Left moves
Part 2: Standard Machines

are at least as powerful as
Stay-Option machines
Proof:
122/189
a b, L
q2
Simulation in Standard Machine
q1 a b, L
q2
Similar for Right moves
126/189
21
Example
Simulation for Stay

Stay-Option Machine
q1
a b, S
Stay-Option Machine:
q1 a b, S q2
q2
q1
q2
x x, R
For every symbol
baba
q2
Simulation in Standard Machine:

1
q3
aaba
x
127/189
Turing Machines with Multiple Track Tape
baba
baba
q3
q2
q1
a b a b
b a c d
track 1
track 2
q1
track 1
track 2
a c a b
b d c d
one symbol
track 1
track 2
q2
q1
(b, a ) (c, d ), L
q2
129/189
Turing Machines with SemiSemi-Infinite Tape
# a b a c
128/189
Turing Machines with Double Track Tape
Standard Machine--Multiple Track Tape
a b a b
b a c d
aaba
q1
Simulation in Standard Machine
a b, L
130/189
Standard SemiSemi-Infinite Tape

simulate
Semi-infinite tape machines:
...
Trivial
131/189
132/189
22
SemiSemi-Infinite Tape Standard (1)
Semi-infinite tape machines simulate

Standard Turing machines:
Standard machine
.........
.........
Standard machine
.........
.........
a b c d e
reference point
Semi-infinite tape machine with two tracks
Semi-infinite tape machine

.........
Right part # d e
Left part

# c b a
.........
133/189
134/189
Standard machine
Standard machine
q2
q1
Right part
q2L
q1L
q2R
q1R
Right part
q1R
Left part
q1L
Standard machine
a b c d e
( x, a ) ( x, g ), L
q2R
q2L
136/189
.........
Time 2
.........
Standard machine
g b c d e
q1
.........
q2

Right part # d e
Left part
(a, x) ( g , x), R
For all symbols
135/189
Time 1
.........
q2

Left part
a g, R
q1

c
a
b
#
.........
q1L
137/189
Right part # d e
Left part

g
c
b
#
q2L
.........
138/189
23
At the border:

Right part
Left part
q1R
q1L
(# , # ) (# , # ), R
(# , # ) (# , # ), R
Time 1
Right part # d e
# c b g
Left part
q1L
.........
q1L
q1R
Time 2
Right part # d e
# c b g
Left part
q1R
139/189
Theorem:
.........
140/189
The OffOff-Line Machine (M

(My v
vi bng t
tch r
ri)
Input File
a b c
Semi-infinite tape machines

read-only
Control Unit
read-write
Tape
g d e
141/189
142/189
1. Copy input file to tape

Off-line machines simulate
Standard Turing Machines:
Standard machine
a b c
Off-line machine:
Off-line machine
1. Copy input file to tape
Input File
2. Continue computation as in
Standard Turing machine
a b c
143/189
Tape
a b c
144/189
24
2. Do computations as in Turing machine

Standard Turing machines simulate
Off-line machines:
Standard machine
a b c
q1
Off-line machine
Tape
Input File
a b c
a b c
Use a Standard machine

with four track tape to keep track
of the Off-line input file
and tape contents
q1
145/189
Reference point
Off-line Machine
Tape
Input File
e f g
a b c d
Four track tape -- Standard Machine
# a b
# 0 0
e f
0 1
c d
1 0
g
0
146/189
Input File
head position
Tape
head position
147/189
# a b
# 0 0
e f
0 1
Input File
head position
Tape
head position
c d
1 0
g
0
Repeat for each state transition:

Return to reference point
Find current input file symbol
Find current tape symbol
Make transition
148/189
Theorem:
Multitape TMs (1)

Control unit
Off-line Machines
Stansard Machines
Tape 1
a b c
Tape 2
e f g
Input
149/189
150/189
25
Multitape Standard
Multitape TMs (2)

Tape 1
Time 1
Tape 2
Multitape machines simulate

Standard Machines:
e f g
q1
a b c
q1
Time 2
a g c
q2
q1
Use just one tape
e d g
q2
(b, f ) ( g , d ), L, R
q2
151/189
152/189
Standard Multitape
Multitape Machine
Standard machines simulate

Multitape machines:
Tape 1
Tape 2
e f g h
a b c
Standard machine:
Standard machine with four track tape
Use a multi-track tape

A tape of the Multiple tape machine
corresponds to a pair of tracks
a
0
e
0
b
1
f
0
c
0
g h
1 0
Tape 1
head position
Tape 2
head position
153/189
Reference point
#
#
#
#
a
0
e
0
b
1
f
0
c
0
g h
1 0
154/189
Theorem:
Tape 1
head position
Tape 2
head position
Repeat for each state transition:

Return to reference point
Find current symbol in Tape 1
Find current symbol in Tape 2
Make transition
155/189
Multi-tape machines
Standard Turing Machines
156/189
26
L = {a nb n }
Same power doesnt imply same speed:

Language
Standard machine:
L = {a nb n }
Go back and forth
Acceptance Time
Standard machine
Two-tape machine
n2
times
Two-tape machine:
b n to tape 2
n
Leave a on tape 1
( n steps)
Compare tape 1 and tape 2
( n steps)
Copy
( n steps)
157/189
158/189
MultiDimensional Turing Machines

Two-dimensional tape
y
Multidimensional machines simulate
Standard machines:
c a
MOVES: L,R,U,D
U: up
D: down
x
Use one dimension
HEAD
Position: +2, -1
159/189
160/189
Two-dimensional machine
y
Standard machines simulate
Multidimensional machines:
c
a
Standard machine:
Use a two track tape
Standard Machine
Store symbols in track 1

Store coordinates in track 2
161/189
x
q1
c
symbols
a
b
1
1
2
1
#
#
#
#
coordinates
1
q1
162/189
27
Standard machine:
Theorem:
Repeat for each transition
MultiDimensional Machines
have the same power
with Standard Turing Machines
Update current symbol

Compute coordinates of next position
Go to new position
163/189
164/189
NonDeterministic Turing Machines
Nondeterministic choices
a b, L
q2
a b, L
a c, R
q3
Choice 1
b b c
Non Deterministic Choice

165/189
q2
Time 0
a b c
q1
q1
a c, R
q2
q3
Time 1
q1
Choice 2
c b c
q3
166/189
Notion
Input string w is accepted if
this a possible computation
q0 w
Initial configuration
NonDeterministic Machines simulate

Standard (deterministic) Machines:
x qf y
Final Configuration
Every deterministic machine

is also a nondeterministic machine
Final state
167/189
168/189
28
Non-Deterministic Choices
Computation 1
Deterministic machines simulate

NonDeterministic machines:
Computation 2
q1
q1
Keeps track of all possible computations
q4
q5
q4
q6
169/189
q3
q2
q3
q2
Deterministic machine:
q5
q6
q7
q7
170/189
Simulation
a b, L
Deterministic machine:
q2
NonDeterministic machine
Time 0
q1
Keeps track of all possible computations
a b c
a c, R
q3
q1
Deterministic machine
Stores computations in a
two-dimensional tape
#
#
#
#
# # #
a b c
q1
# # #
# #
#
#
#
Computation 1
171/189
172/189
NonDeterministic machine
a b, L
Time 1
q2
b b c
Choice 1
c b c
q3
Choice 2
q2
q1
a c, R
#
#
#
#
q3
Deterministic machine
# # # # # #
b b c #
Computation 1
q2
#
c b c #
Computation 2
q3
#
173/189
Repeat
Execute a step in each computation:
If there are two or more choices
in current computation:
1. Replicate configuration
2. Change the state in the replica
174/189
29
Remark:
The simulation
in the Deterministic machine
takes time exponential time compared
to the NonDeterministic machine
Theorem: NonDeterministic Machines

Deterministic machines
175/189
176/189
A limitation of Turing Machines:

Turing Machines are hardwired
they execute
only one program
A Universal Turing Machine
Real Computers are re-programmable

Solution: Universal Turing Machine
(My Turing vn nng, hay ph qut)
Attributes:
Reprogrammable machine
Simulates any other Turing Machine
178/189
Universal Turing Machine
Simulates any other Turing Machine M
Description of transitions of M
Tape 1
Description of M
Universal
Turing
Machine
Input of Universal Turing Machine:
Tape 2
Tape contents of M
Tape 3
Initial tape contents of M
State of M
179/189
Three tapes
180/189
30
Encoding in Tape 1
Encoding in Tape 1
Tape 1
Tape 1
Description of M
Description of M
We describe TM M as a string of symbols:
We encode
Alphabet Encoding
Symbols:
Encoding:
State Encoding
M as a string of symbols
11
111
1111
States:
q1
q2
q3
q4
Encoding:
11
111
1111
181/189
182/189
Encoding in Tape 1
Encoding in Tape 1
Tape 1
Tape 1
Description of M
Description of M
Head Move Encoding
Move:
Encoding:
11
Transition Encoding
Transition:
(q1, a) = (q2 , b, L)
Encoding:
1 0 1 0 11 0 11 0 1
Using 0 for Separator
183/189
184/189
Encoding in Tape 1
Tape 1
Tape
Description of M
Transitions:
(q1, a) = (q2 , b, L)
1 contents of Universal TM M:
encoding of the simulated machine

as a binary string of 0s and 1s
Machine Encoding
TM is described
with a binary string of 0s and 1s
(q2 , b) = (q3 , c, R)
Encoding:
1 0 1 0 11 0 11 0 1 00 11 0 1 10 111 0 111 0 11
Using 00 for Transition Separator
185/189
186/189
31
Question: How powerful is RAM?
Language of Turing Machines
Can
RAMs compute functions/languages that TMs

cannot?
The set of Turing machines forms a language,

therefore :
Can
RAMs compute functions/languages significantly

faster then TMs?
each string of the language

is the binary encoding of a TM
Theorem:
L = { 010100101, 00100100101111, 111010011110010101, }
TM 1,
TM 2,
A TM can simulate a RAM, provided that the elementary RAM

instructions can themselves be simulated by a TM
...
187/189
188/189
Proof
The idea is to use a multitape TM M
to perform simulation:
Tape
1:
keeps program
# and *, i.e.,
instructions vi in binary, numbered, and separated by

#0*v0#1*v1#10*v2##i*vi#
Tape
2: keeps data registers:
Tape
3: keeps program counter
Tape
4: keeps data address register
#0*r0#1*r1#10*r2##i*ri#
189/189
32

Ch3 TuringMachine

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ch3 TuringMachine

Uploaded by

Copyright:

Available Formats

CHNG 3

L thuyt tnh ton

(Turing Machines Model)

PGS.TS. Phan Huy Kh

Alan Mathison Turing (23/06/1912 07/06/1954)

2-way DFA has equivalent power

2-way PDA has a little more power (can accept { ww | w* }

(Push-Down Deterministic Finite Automaton) limitations

Tape head moves only one direction

(Deterministic Finite Automaton) limitations

Tape head moves only one direction

Writable, 2-way tape?

Introduction to Turing Machines (TM)

a formal model of a (digital) computer

String membership in a regular language

Random Access Machines

Partial recursive functions (Kleene, 1952)

-calculus (Church, 1936)

Emptiness of the complement of a CFL

Post systems (Post, 1943)

Whether a C program ever terminates

these were preceded by the work of Kurt Godel, 1931,

How to prove it?

L thuyt tnh ton -- Ch.3 - My Turing

TM is the invention of Alan Turing, 1936

Determining valid executions of a PDA

Mt u c-ghi c hay ghi ln bng,

No boundaries -- infinite length

The head at each time step:

L thuyt tnh ton -- Ch.3 - My Turing

Example Moves Left :

Example Moves Right :

The Input String

States & Transitions

Head starts at the leftmost position

Remark: the input string is never empty

Example Moves Right :

Example Moves Left :

L thuyt tnh ton -- Ch.3 - My Turing

Mi k t c xong ch tng ng vi mt dch chuyn duy nht

Turing Machines are deterministic :

No epsilon transitions allowed

Partial Transition Function

Final states have no outgoing transitions

L thuyt tnh ton -- Ch.3 - My Turing

Turing Machine Example (1)

A Turing machine that accepts the language:

(This Turing machine mimic a DFA)

Turing Machine Example (2)

Turing Machine Example (3)

Rejection Example (1)

Turing Machine Example (4)

L thuyt tnh ton -- Ch.3 - My Turing

Rejection Example (2)

Infinite Loop Example (1)

Halt & Reject

Infinite Loop Example (2)

Infinite Loop Example (3)

Infinite Loop Example (4)

Infinite Loop Example (5)

L thuyt tnh ton -- Ch.3 - My Turing

Infinite Loop Result

Because of the infinite loop: