Introduction to

Turing Machines

Lecture 21 & 22

0 01 1 10 1 1 1 1 1 1 0 1 1 0 1 1 0 0 0

1
 The Turing Machine

10 01 1 10 1 1 1 1 1 1 0 1 1 0 1 1 0 0 0

A TM consists of an infinite length tape, on which

input is provided as a finite sequence of symbols.

A head reads the input tape. The TM starts at start

state s0. On reading an input symbol it optionally
replaces it with another symbol, changes its
internal state and moves one cell to the right or left.
2
 The Turing Machine

A TM is defined as:
TM = <S, T, s0, , H> where,

S is a set of TM states
T is a set of tape symbols
s0 is the start state
HS is a set of halting states
 : S x T →S x T x {L,R}
is the transition function

3
 A Turing Machine
Tape
...... ......

Read-Write head
Control Unit

4
 The Tape

No boundaries -- infinite length

...... ......

Read-Write head

The head moves Left or Right

5
 ...... ......

Read-Write head

The head at each time step:

1. Reads a symbol
2. Writes a symbol
3. Moves Left or Right
6
 Example:
Time 0
...... a b a c ......

Time 1
...... a b d c ......

1. Reads a
2. Writes d
3. Moves Left
7
 Time 1
...... a b s c ......

Time 2
...... a f s c ......

1. Reads b
2. Writes f
3. Moves Right
8
 The Input String

Input string Blank symbol

...... # # a b a c # # # ......

head
•Head starts at the leftmost position of the input string.
•The input string is never empty

9
 States & Transitions

Read Write
Move Left

q1 a → b, L q2

Move Right

q1 a → b, R q2
10
Example:
Time 1
...... a b a c ......

q1 current state

Time 2
......   a b d c    ......

q2

q1 a → d, R q2
11
Example:
Time 1
......   a b a c    ......

q1

Time 2
......   a b b c    ......

q2

q1 a → b, L q2
12
Example:
Time 1
...... # # a b a c # #
......

q1

Time 2
...... # # a b b c g #
......

q2
 → g, L
q1 q2
13
 Determinism
Turing Machines are deterministic

Allowed Not Allowed

a → b, R q2 a → b, R q2

q1 q1
q3 a → d, L q3
b → d, L

No lambda transitions allowed

14
 Partial Transition Function
Example:

...... # # a b a c # # # ......
q1

a → b, R q2 Allowed:

q1 No transition
for input symbol c
b → d, L q3
15
 Halting

The machine halts if there are

no possible transitions to follow

...... # # a b c c # # # ......
q1

a → b, R q2 No possible transition
q1
q3
HALT!!!
b → d, L

16
 Final States

q1 q2 Allowed

q1 q2 Not Allowed

• Final states have no outgoing transitions

• In a final state the machine halts

17
 Acceptance

If machine halts
Accept Input
in a final state

If machine halts
in a non-final state
Reject Input or
If machine enters
an infinite loop
18
 Turing Machine Example

A Turing machine that accepts the language:

a*

a → a, R

 → , L
q0 q1

19
Time 0   a a a  

q0

a → a, R

 → , L
q0 q1

20
Time 1   a a a  

q0

a → a, R

 → , L
q0 q1

21
Time 2   a a a  

q0

a → a, R

 → , L
q0 q1

22
Time 3   a a a  

q0

a → a, R

 → , L
q0 q1

23
Time 4   a a a  

q1

a → a, R Halt & Accept

 → , L
q0 q1

24
Rejection Example

Time 0   a b a  

q0

a → a, R

 → , L
q0 q1
25
Time 1   a b a  

q0
No possible Transition
a → a, R Halt & Reject

 → , L
q0 q1
26
 Infinite Loop Example
Let’s run it for a string
aa * +b(a + b) *

b → b, L
a → a, R

 → , L
q0 q1

27
Time 0   a b a  

q0

b → b, L
a → a, R

 → , L
q0 q1

28
Time 1   a b a  

q0

b → b, L
a → a, R

 → , L
q0 q1

29
Time 2   a b a  

q0

b → b, L
a → a, R

 → , L
q0 q1

30
Time 2   a b a  
q0
Time 3   a b a  

Infinite loop
q0
Time 4   a b a  
q0

Time 5   a b a  
q0
31
Because of the infinite loop:

•The final state cannot be reached

•The machine never halts

•The input is not accepted

32
 Another Turing Machine Example
n n
Turing machine for the language {a b }

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
33
 Time 0  a a b b  

q0

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
34
 Time 1  x a b b  

q1

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
35
 Time 2  x a b b  

q1

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
36
 Time 3  x a y b  

q2

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
37
 Time 4  x a y b  

q2

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
38
 Time 5  x a y b  

q0

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
39
 Time 6  x x y b  

q1

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
40
 Time 7  x x y b  

q1

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
41
 Time 8  x x y y  

q2

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
42
 Time 9  x x y y  

q2

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
43
 Time 10  x x y y  

q0

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
44
 Time 11  x x y y  

q3

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
45
 Time 12  x x y y  

q3

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
46
 Time 13  x x y y  

q4
Halt & Accept

q4 y → y, R y → y, L
y → y, R a → a, R a → a, L
 → , L
y → y, R a → x, R b → y, L
q3 q0 q1 q2
x → x, R
47
Observation:

If we modify the
machine for the language n n
{a b }

we can easily construct

n n n
a machine for the language {a b c }

48
Construct a TM to erase the input string

• Language recognizers such as DFA’s cannot

perform computational tasks such as
erasing strings.
• So, no DFA can be used for erasing
strings.
• Language generators such as CFG’s cannot
perform computational tasks such as
erasing strings.
• So, no CFG can be used for erasing strings.
• TM’s are more powerful than language
recognizers and language generators. A
TM can be used for erasing strings
49
50
51
Formal Definitions
for
Turing Machines

52
 Transition Function

q1 a → b, R q2

 (q1, a) = (q2 , b, R)

53
 Transition Function

q1 c → d, L q2

 (q1, c) = (q2 , d , L)

54
Turing Machine:

Input Tape
alphabet alphabet
States

M = (Q, , ,  , q0 , , F )

Transition Final
function states
Initial blank
state 55
 Configuration

  c a b a  

q1

Instantaneous description: ca q1 ba

56
 Time 4 Time 5
 x a y b    x a y b  

q2 q0

A Move: q2 xayb  x q0 ayb

57
 Time 4 Time 5
 x a y b    x a y b  

q2 q0

Time 6 Time 7
 x x y b    x x y b  

q1 q1

q2 xayb  x q0 ayb  xx q1 yb  xxy q1 b

58
q2 xayb  x q0 ayb  xx q1 yb  xxy q1 b


Equivalent notation: q2 xayb  xxy q1 b

59
Initial configuration: q0 w

Input string

 a a b b  

q0

60
 The Accepted Language

For any Turing Machine M


L( M ) = {w : q0 w  x1 q f x2 }

Initial state Final state

61
 Standard Turing Machine

The machine we described is the standard:

• Deterministic

• Infinite tape in both directions

•Tape is the input/output file

62
Computing Functions
with
Turing Machines

63
A function f (w) has:

Domain: D Result Region: S

f (w)
w D f ( w)  S

64
A function may have many parameters:

Example: Addition function

f ( x, y ) = x + y

65
 Integer Domain

Decimal: 5

Binary: 101

Unary: 11111

We prefer unary representation:

easier to manipulate with Turing machines

66
Definition:

A function f is computable if
there is a Turing Machine M such that:

Initial configuration Final configuration

 w   f (w) 

q0 initial state q f final state

For all w D Domain

67
In other words:
A function f is computable if
there is a Turing Machine M such that:


q0 w  q f f ( w)

Initial Final
Configuration Configuration

For all w D Domain

68
 Example

The function f ( x, y ) = x + y is computable

x, y are integers

Turing Machine:

Input string: x0 y unary

Output string: xy 0 unary

69
 x y

Start  1 1  1 0 1  1 

q0
initial state

The 0 is the delimiter that

separates the two numbers

70
 x y

Start  1 1  1 0 1  1 

q0 initial state

x+ y

Finish  1 1  1 1 0 

q f final state
71
 The 0 helps when we use
the result for other operations

x+ y

Finish  1 1  1 1 0 

q f final state
72
Turing machine for function f ( x, y ) = x + y

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
73
Execution Example: Time 0
x y
x = 11 (2)
 1 1 0 1 1 
y = 11 (2)
q0

Final Result
x+ y
 1 1 1 1 0 

q4
74
 Time 0  1 1 0 1 1 

q0

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
75
 Time 1  1 1 0 1 1 

q0

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
76
 Time 2  1 1 0 1 1 

q0

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
77
 Time 3  1 1 1 1 1 

q1

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
78
 Time 4  1 1 1 1 1 

q1

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
79
 Time 5  1 1 1 1 1 

q1

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
80
 Time 6  1 1 1 1 1 

q2

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
81
 Time 7  1 1 1 1 0 

q3

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
82
 Time 8  1 1 1 1 0 

q3

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
83
 Time 9  1 1 1 1 0 

q3

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
84
 Time 10  1 1 1 1 0 

q3

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
85
 Time 11  1 1 1 1 0 

q3

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
q4
86
 Time 12  1 1 1 1 0 

q4

1 → 1, R 1 → 1, R 1 → 1, L

q0 0 → 1, R q1  → , L q 1 → 0, L q3
2

 → , R
HALT & accept q4
87
 Another Example

The function f ( x) = 2 x is computable

x is integer

Turing Machine:

Input string: x unary

Output string: xx unary

88
 x

Start  1 1  1 

q0 initial state

2x

Finish  1 1  1 1 1 

q f final state
89
Turing Machine Pseudocode for f ( x) = 2 x

• Replace every 1 with $

• Repeat:
• Find rightmost $, replace it with 1

• Go to right end, insert 1

Until no more $ remain

90
Turing Machine for f ( x) = 2 x

1 → $, R 1 → 1, L 1 → 1, R

q0  → , L q1 $ → 1, R q2
 → , R  → 1, L
q3
91
 Example
Start Finish
 1 1   1 1 1 1 
q0 q3
1 → $, R 1 → 1, L 1 → 1, R

q0  → , L q1 $ → 1, R q2
 → , R  → 1, L
q3
92
 Simple TM Examples

Turing Machine U+1:

Given a string of 1s on a tape (followed by

an infinite number of 0s), add one more 1 at the
end of the string.

#111100000000…….
➔
#1111100000000……….

93
 Simple TM Examples

TM: U+1
(s0, 1) |-- (s0, 1, R)
(s0, 0) |-- (h, 1, STOP)

#s0111100000….. →
#1s011100000….. →
#11s01100000….. →
#111s0100000….. →
#1111s000000….. →
#11111h0000….. STOP
94
 Another Example

1 if x y
The function f ( x, y ) =
0 if x y
is computable

95
Turing Machine for

1 if x y
f ( x, y ) =
0 if x y

Input: x0 y

Output: 1 or 0
96
Turing Machine Pseudocode:

• Repeat

Match a 1 from x with a 1 from y

Until all of x or y is matched

• If a 1 from x is not matched

erase tape, write 1 ( x  y)
else
erase tape, write 0 ( x  y) 97
Combining Turing Machines

98
 Block Diagram

Turing
input output
Machine

99
 Example: x + y if x  y
f ( x, y ) =
0 if x  y

x, y
Adder x+ y
x, y x y
Comparer

x y Eraser 0

100
 Turing’s Thesis

Any mathematical problem solving that

can be described by a mechanical procedure
(algorithm) can be modeled by a
Turing machine.

All computers today perform only mechanical

problem solving. They are no more expressive
than a Turing machine.

101
 Turing’s Thesis
Turing’s thesis is not a “theorem” there is no
“proof” for the thesis.
The theorem may be refuted by showing at least
one task that is performed by a digital computer
which cannot be performed by a Turing machine.
Many contentions have been made to this end.
However, till date there have not been any
conclusive evidence to refute Turing’s thesis.

102
 Conclusions
TMs are at a level that is much below the assembly
language of any typical microprocessor.
So in the practical world, TMs are more useful in
what they cannot do rather than in what they
can.

103
 Lab2
Write some simpleTuring machine programs

QUESTIONS?

104