34.

NP Completeness

Click
Hsu,toLih-Hsing
add Text
 Computer Theory Lab.

P and NP
Non-formal description
 P: solvable polynomial time

 NP:

 nondeterministic polynomial time

 Verifiable in polynomial time by
deterministic Turing machine.

Chapter 34 P.2
 Computer Theory Lab.

NP-complete
 NP-Complete: No polynomial-time
algorithm has yet been discovered for an
NP-complete problem, nor has anyone yet
been able to prove that no polynomial-time
algorithm can exist for any one of them.

Chapter 34 P.3
 Computer Theory Lab.

Polynomial time algorithms: on inputs of size n, their

worst-case running time is O(n k ) .

It is natural to wonder whether all problems can be solved

in polynomial time. The answer is no. For example, the
Halting Problem.

Given a description of a program and a finite input, decide whether the program
Finishes running or will run forever.

Chapter 34 P.4
 Computer Theory Lab.
Generally, we think of problems that are solvable by
polynomial-time algorithms are being tractable, and problems
that requires superpolynomial time are being intractable.

Polynomial Intractables

NP-Complete
Problem

Chapter 34 P.5
 Computer Theory Lab.

The subject of this chapter, however, is an interesting class of

problems, called the “ NP-complete” problems, whose status
is unknown. No polynomial-time algorithm has yet been
discovered for an NP-computer problem, nor has anyone yet
been able to prove that no polynomial-time algorithm can
exist for any one of them. This so-called P ≠ NP question has
been one of the deepest, most perplexing open research
problems in theoretical computer science since it was first
posed in 1971.

Chapter 34 P.6
 Computer Theory Lab.

NP-complete problem: status are unknown.

If any single NP-complete problem can be solved in

polynomial time, then every NP-complete problem has a
polynomial time algorithm.

To become a good algorithm designer, you must

understand the rudiments of the theory of
NP-completeness.

Chapter 34 P.7
 Computer Theory Lab.

The difference between these problems

 Shortest vs. longest simple paths:
 Euler tour vs. hamiltonian cycle:
 2-CNF satisfiability vs. 3 CNF satisfiablility
 NP-completeness and the classes P and NP
 Overview of showing problems to be NP-
complete
 Decision problems vs. optimization problems

Chapter 34 P.8
 Computer Theory Lab.

Reductions
Suppose that there is a different decision problem,
say B, that we already know how to solve in
polynomial time. Finally, suppose that we have a
procedure that transforms any instance  of A into
some instance  of B with the following
characteristics:
1.The transformation takes polynomial time.
2.The answer are the same. That is, the answer for  is
“yes” if and only if the answer for  is also “yes.”

Chapter 34 P.9
 Computer Theory Lab.

yes yes
 Polynomial-time  Polynomial-time
reduction algorithm Algorithm to decide B
no no
Polynomial-time algorithm to decide A

We can call such a procedure a polynomial-time

reduction algorithm and, it provides us a way to
solve problem A in polynomial time:
1.Given an instance  of problem A, use a polynomial-
time reduction algorithm to transform it to an instance 
of problem B.
2.Run the polynomial-time decision algorithm for B on
the instance .
3.Use the answer for  as the answer for .

Chapter 34 P.10
 Computer Theory Lab.

A First NP-complete problem

 Because the technique of reduction
relies on having a problem already
known to be NP-complete in order to
prove a different problem NP-complete,
we need a “first” NPC problem.
 Circuit-satisfiability problem

Chapter 34 P.11
 Computer Theory Lab.
34.1 Polynomial time
Polynomial time solvable problem are regarded as tractable.
 Even if the current best algorithm for a problem has a
running time of (n100), it is likely that an algorithm with
a much better running time will soon be discovered.
 Problems for many reasonable models of computation,
can be solved in one model can be solved in polynomial
in another.
 Polynomial-time solvable problems has a nice closure
property.
f , g are polynomial
 f ( g) is also polynomial
Chapter 34 P.12
 Computer Theory Lab.

Abstract Problems: An abstract problem Q is a binary

relation on a set of problem instances and a set S of
problem solutions.

Decision problems: those having yes/no solution.

Optimization problems: recast by imposing a bound on the

value to be optimized.

An encoding of a set S of abstract objects is a mapping e

from S to the set of binary string, for example:
Chapter 34 {0,1,2,3,…}={0,1,10,11,…} P.13
 Computer Theory Lab.

We call a problem whose instance sets is the set of binary

strings a concrete problem.

We say that an algorithm solves a concrete problem in time

O( T ( n )) if when it is provided a problem instance i if
length n=|i|, the algorithm can produce the solution in at
most O( T ( n )) time.

Chapter 34 P.14
 Computer Theory Lab.

A concrete problem is polynomial-time solvable if there

exists an algorithm to solve it in time O( n k ) for some
constant k.

The complexity class P is the set of concrete decision

problems that are solvable in polynomial time.

Chapter 34 P.15
 Computer Theory Lab.

Abstract problem  concrete problem

e: I 
{ 0 ,1 } *
encoding

Problem input k complexity O(k)

unary k 11...1 ( k )
binary n   lg k  ( k )  ( 2 n )

Chapter 34 P.16
 Computer Theory Lab.

We say that a function f : { 0 ,1 }*  { 0 ,1 } * is

polynomial-time computable if there exists a
polynomial-time algorithm A that given any x { 0 ,1 } * ,
produces as output f ( x ) .

Chapter 34 P.17
 Computer Theory Lab.

For any set I of problem instances, we say that two

encodings e1 and e2 are polynomial related if there exist
two polynomial-time computable functions f12 and f 21
such that for any i  I , we have f12 ( e1( i ))  e2 ( i ) and
f21( e2 ( i ))  e1( i ) .

Chapter 34 P.18
 Computer Theory Lab.

Lemma 34.1. Let Q be an abstract decision problem on

an instance set I, let e1 and e2 be polynomially related
encodings on I. Then e1( Q )  P if and only if e2 ( Q )  P .

Using reasonable encoding to neglect the distinction

between abstract and concrete problems.

Chapter 34 P.19
 Computer Theory Lab.

A fo rm a l-la n g u a g e fra m e w o rk
A n a lp h a b e t  is a fin ite s e t o f s y m b o ls .

A la n g u a g e L o v e r  is a n y s e t o f s trin g s m a d e u p o f

s y m b o ls fro m  .
e m p ty s tr in g :  .

e m p ty la n g u a g e :  .

*


L e t L1 , L 2 b e tw o la n g u a g e s . W e c a n d e fin e

Chapter 34 P.20
 Computer Theory Lab.

L 1  L 2 (union)
L 1  L 2 (intersect ion)
L (complemen t)
L 1 L 2  { x 1 x 2 | x 1  L 1 and x 2  L 2 }
(concatena tion)
The closure (Kleen star) of L :
L *  { }  L  L  L  ...
2 3

Chapter 34 P.21
 Computer Theory Lab.

The set of instances of any decision problem Q is the set of

 * , where   { 0 ,1 } . Since Q is entirely characterized by
those problem instances that produces a 1 (yes) answer. We
can view Q as the language L over  * , where
L  { x   * | Q( x )  1 } .

Chapter 34 P.22
 Computer Theory Lab.

Algorithm A accepts a string x { 0 ,1 } * if the given input

x, the algorithm output A(x)=1.

The language accepts by an algorithm A is the set

L  { x   * | A( x )  1 } .

The algorithm A rejects a string x if A(x)=0.

Chapter 34 P.23
 Computer Theory Lab.

Even if language L is accepted by an algorithm A, the

algorithm will not necessarily reject a string x  L
provided as input to it. For example, the algorithm may
loop forever.

A language L is decided by an algorithm A if every binary

string is either accepted or rejected by the algorithm.

Chapter 34 P.24
 Computer Theory Lab.

A language L is accepted in polynom ial tim e by an

algorithm A if for any length n string x  L , the algorithm
k
accepts x in tim e O ( n ) for som e constant k.

A language L is decided in polynom ial tim e by an

algorithm A if for any length n string x  { 0 ,1 } * , the
O ( n k)
algorithm decides x in tim e for som e constant k.
Chapter 34 P.25
 Computer Theory Lab.

Example:
PATH PROBLEM:
PATH={<G,u,v,k> | G=(V,E) is an undirected graph,
u, v V , k  0 is an integer, and there is a path from u to v
whose length is at most k}.

Chapter 34 P.26
 Computer Theory Lab.

 Can be accepted in polynomial time.

 Can be decided in polynomial time.

HALTING PROBLEM:
There exists an accepting algorithm, but no decision
algorithm exists.

Chapter 34 P.27
 Computer Theory Lab.

We can informally define a complexity class as a set of

languages, membership in which is determined by a
complexity measure, such as running time, on an algorithm
that determines whether a string x belongs to language L.

Chapter 34 P.28
 Computer Theory Lab.

We define the complexity class P as: P  { L  { 0 ,1 } *|

there exists an algorithm A that decides L in polynomial
time}.

Theorem 34.2. P  { L | L is accepted by a polynomial

algorithm}.

Chapter 34 P.29
 Computer Theory Lab.

HAMILTONIAN CYCLE PROBLEM:

HAM_CYCLE={<G> | G is a hamiltonian graph}

verification: polynomial
decision problem: ?
Chapter 34 P.30
 Computer Theory Lab.

34.2 Polynomial-time verification

PATH PROBLEM:
PATH={<G,u,v,k> | G=(V,E) is an undirected graph,
u, v V , k  0 is an integer, and there is a path from u to v
whose length is at most k}.

verification: linear time.

Decision problem: polynomial

Chapter 34 P.31
 Computer Theory Lab.

naive algorithm:
input size: If we use the reasonable encoding of a graph as
its adjacency matrix, the number m of vertices is ( n ),
where n = |<G>| is the length of the encoding of G. There
are m! possible permutations of the vertices. Therefore the
running time is ( m !)  ( n !)  ( 2 n ) . This is not
a polynomial algorithm.

Chapter 34 P.32
 Computer Theory Lab.

Verification algorithms:
A verification algorithm is a two-argument algorithm A,
where one argument is an ordinary input string x and the
other is a binary string y called a certificate. A
two-argument algorithm A verifies an input x if there exists
a certificate y such that A(x,y)=1. The language verified by
a verification algorithm A is
L  { x  { 0 ,1 } *| y  { 0 ,1 } * s.t . A( x , y )  1 } .

Chapter 34 P.33
 Computer Theory Lab.

The complexity class NP

The complexity class NP is the class of languages that can
be verified by a polynomial-time algorithm. More precise, a
language L belongs to NP if and only if there exists a
two-input polynomial-time algorithm A and a constant c
such that
L  { x { 0 ,1 } * | there exists a certificate y with
| y|  O(| x|c ) such that A(x,y)=1}.
 NP   (HAM_CYCLE  NP.)
 P  NP .
Chapter 34 P.34
 Computer Theory Lab.

Problem:
1. P  NP?
2. Complexity class co-NP
co  NP  { L | L  NP } .
NP  co  NP?
3. Obviously P  NP  co  NP .
P  NP  co  NP ?

Chapter 34 P.35
 Computer Theory Lab.

NP=co-NP
P=NP=co-NP
P

P
NP co-NP

P  NP  co  NP

P
NP co-NP

P  NP  co  NP
Chapter 34 P.36
 Computer Theory Lab.

34.3 NP-completeness and reducibility

NP-completeness problem: if any one NP-complete
problem can be solved in polynomial time, then every
problem in NP has a polynomial-time solution, that is
NP=P.

Chapter 34 P.37
 Computer Theory Lab.

Reducibility:
ax  b  0
ax 2  bx  c  0

Chapter 34 P.38
 Computer Theory Lab.

A language L1 is polynomial-time reducible to a language

L2 , written L1  P L2 if there exists a polynomial-time
computable function f : { 0 ,1 }*  { 0 ,1 } * such that for all
x  L1 if and only if f ( x )  L2 . We call the function f
the reduction function, and a polynomial algorithm F that
computes f is called a reduction algorithm.

Chapter 34 P.39
 Computer Theory Lab.

f
{ 0 ,1 }* { 0 ,1 }*

L1 L2

Chapter 34 P.40
 Computer Theory Lab.

Lemma 34.3. If L1 , L2 { 0 ,1 } * are languages

such that L1  P L2 , then L2  P implies L1  P .

yes, f ( x)  L2 yes, f ( x)  L1

x f (x)
F A2
no, f ( x)  L2
no, f ( x)  L1
A1

Chapter 34 P.41
 Computer Theory Lab.

NP-Completeness
A language L { 0 ,1 } * is NP-complete if
1. L  NP , and
2. L'  P L for every L'  NP

Chapter 34 P.42
 Computer Theory Lab.

 If a language L satisfies property 2, but not necessarily

property 1, we say that L is NP-hard.

 We also define NPC to be the class of NP-complete

language.

Chapter 34 P.43
 Computer Theory Lab.

Theorem 34.4. If any NP-complete problem is

polynomial-time solvable, then NP=P. If any problem is
not polynomial-time solvable, then all NP-complete
problem are not polynomial-time solvable.

Proof. By Lemma 34.3.

NPC
NP

Chapter 34 P.44
 Circuit satisfiability Computer Theory Lab.

x1 1
1
1
1
x2 1
0 1
1 1
1
0
1
1 1 Satisfiable
x3
0 1
1

x1

x2

Not Satisfiable!
x3
Chapter 34 P.45
 Computer Theory Lab.

Circuit-satisfiability problem: Given a boolean

combinational circuits composed of AND, OR, or NOT
gates, is it satisfiable?

CIRCUIT_SAT={<C> | C is a satisfiable boolean

combinational circuit}.

Chapter 34 P.46
 Computer Theory Lab.

Lemma 34.5. The circuit-satisfiability problem belongs

to the class NP.

Lemma 34.6. The circuit-satisfiability problem is NP-hard.

Proof. L  P CIRCUIT _ SAT L  NP.

Theorem 34.7. The circuit-satisfiability problem is

NP-Complete.

Chapter 34 P.47
 Computer Theory Lab.

Input bits

C0 A Program Counter Aux Machine State x y working storage

C1 A Program Counter Aux Machine State x y working storage

C2 A Program Counter Aux Machine State x y working storage

C T(n) A Program Counter Aux Machine State x y working storage

0/1 output
Chapter 34 P.48
 Computer Theory Lab.

34.4 NP-Completeness Proof

Lemma 34.8. If L is a language such that L'  P L for

some L'  NPC , then L is NP-hard. Moreover, if L  NP
then L  NPC .

Chapter 34 P.49
 Computer Theory Lab.

Method for proving a language L is NPC:

1. Prove L  NP .
2. Select a known NPC language L'
3. Describe an algorithm that computes a function f
mapping every instance of L' to an instance of L.
4. Prove that the function f satisfies x  L' if and only if
f ( x )  L for all x { 0 ,1 } * .
5. Prove that the algorithm computing f runs in polynomial
Chapter 34 time. P.50
 Computer Theory Lab.

Formula satisfiability:
An instance of SAT is a boolean formula  composed of
1. boolean variables: x1 , x2 ,...
2. boolean connectives: any boolean function with one or
two input and one output
3. parentheses

Chapter 34 P.51
 Computer Theory Lab.

SAT={< > |  is a satisfiability formula}

  (( x1  x2 )  (( x1  x3 )  x4 ))  x2
x1  0 , x2  0 , x3  1, x4  1
  (( 0  0 )  (( 0  1 )  1 ))  0
Example:
 (1  (1  1 ))  1
 (1  0 )  1
1

Chapter 34 P.52
 Computer Theory Lab.

Theorem 34.9 Satisfiability of boolean formula is

NP-complete.
Proof.
 SAT  NP
 CIRCUIT _ SAT  P SAT
x1

x5 x8
x2

x6 x 10

x9

x3 x7
x4
Chapter 34 P.53
 Computer Theory Lab.

  x10  ( x4  x3 )  ( x5  ( x1  x2 ))
 ( x6  x4 )  ( x7  ( x1  x2  x4 ))
 ( x8  ( x5  x6 ))  ( x9  ( x6  x7 ))
 ( x10  ( x7  x8  x9 ))

Chapter 34 P.54
 Computer Theory Lab.

3-CNF satisfiability
 literal
 conjunction normal form (CNF)
 3-conjunction normal form (3-CNF)

( x1  x1  x2 )  ( x3  x2  x4 )
 (x1  x3  x4 )

Chapter 34 P.55
 Computer Theory Lab.

Theorem 34.10. Satisfibility boolean formula in 3-CNF

is NP complete.
Proof.
 3  CNF  SAT  NP
 SAT  P 3  CNF  SAT

Chapter 34 P.56
   (( x1  x2 )  (( x1  x3 )  x4 )  x4 Computer Theory Lab.

y1


y2

x4

y3 y4
 

y5

x1 x2

y6

x4

Chapter 34 x1 x3 P.57

 Computer Theory Lab.

  y1  ( y1  ( y2   x4)  ( y2  ( y3  y4)
 ( y3  (x1  x2)  ( y 4   y5)
 ( y5  ( y6  x4)  ( y6  ( x1  x 3)

Chapter 34 P.58
 Computer Theory Lab.

1  y1  ( y2  x2 )
Truth Table 
1  ( y1  y2  x2 )  ( y1  y2  x2 )
 ( y1  y2  x2 )  (y1  y2  x2 )
De Morgan rule 
1  (y1  y2  x2 )  (y1  y2  x2 )
 (y1  y2  x2 )  ( y1  y2  x2 )

Chapter 34 P.59
 Computer Theory Lab.

| Ci|  3 Ci

| Ci | 2

Ci  l1  l2  (l1  l2  p)  (l1  l2  p)

| Ci|  1

Ci  l  ( l  p  q )  ( l  p  q )
( l  p  q )  ( l  p  q )
Chapter 34 P.60
 Computer Theory Lab.

34.5 NP-Complete Problems

Chapter 34 P.61
 Computer Theory Lab.

CIRCUIT-SAT

SAT

3CNF-SAT

Clique HAM-CYCLE

VERTEX COVER TSP

SUBSET-SUM

Chapter 34 P.62
 Computer Theory Lab.

34.5.1 The clique problem

A clique in a undirected graph G = (V,E) is a subset
V '  V of vertices, each pair of which is connected by an
edge in E. The size of a clique is the number of vertices it
contains. The clique problem is the optimization problem
of finding a clique of maximum size in a graph.

Chapter 34 P.63
 Computer Theory Lab.

CLIQUE={<G, k> |G is a graph with clique size k}

2 |V |
naïve algorithm: ( k   )
 k

Chapter 34 P.64
 Computer Theory Lab.

Theorem 34.11. The clique problem is NP-complete.

Proof.
 clique  NP
 3  CNF  SAT  P clique

  ( x1  x2  x3 )  (x1  x2  x3 )

 ( x1  x2  x3 )

Chapter 34 P.65
 Computer Theory Lab.

C1  x1  x2  x3

x1 x2 x3

x1 x1

x2 x2

x3 x3

C2  x1  x2  x3 C3  x1  x2  x3
Chapter 34 P.66
 Computer Theory Lab.

   C1  C2 ... Ck

r s (1) rs
 ( vi , v j )  E  r s
(2) li  l j

 clique size k

Chapter 34 P.67
 Computer Theory Lab.

34.5.2 The vertex-cover problem

 A vertex cover of an undirected graph G=(V,E) is a
subset V '  V such that if ( u, v )  E then u V ' or
v V ' (or both).
 The vertex cover problem is to find a vertex cover of
minimum size in a given graph.
 VERTEX-COVER={<G, k> | graph G has a vertex cover
of size k}.

Chapter 34 P.68
 Computer Theory Lab.

Theorem 34.12. The vertex-cover problem is NP-complete.

Proof.
 VERTEX  COVER  NP
 CLIQUE  P VERTEX  COVER

Chapter 34 P.69
 Computer Theory Lab.

u v u v

z w z w

y x y x

Chapter 34 P.70
 Computer Theory Lab.

34.5.3The hamiltonian-cycle problem

Theorem 34.13. The hamiltonian cycle problem is

NP-complete.
Proof.
 HAM  CYCLE  NP
 3CNF  SAT  P HAM  CYCLE

 kinds of wedges
Chapter 34 P.71
   ( x1  x2  x3 )  ( x1  x2  x3 )  ( x1  x2  x3 ) Computer Theory Lab.

b1 ,1

x 1'

b1 ,2
A e1
e1
B
x 1''
A
b1 ,3

b1 ,4

b2 ,1 A

A
x 2'

b2 ,2
e2
B A e2

b2 ,3
A x 2''

b2 ,4
A
b3 ,1
A

x 3'
b3 ,2 e3
e3
B
b3 ,3
A x 3''

b3 ,4

Chapter 34 P.72
 Computer Theory Lab.
b1,3 x 3'

b1 ,3

b1,4
b1 ,4
A x 3'

e3
b3 ,3
A x 3''

b3,3
b3 ,4

b3,4 x 3''

Chapter 34 P.73
 Computer Theory Lab.

34.5.4 The traveling-salesman problem

TSP={<G, c, k> | G = (V, E) is a complete graph, c is a

function from V  V into Z, k  Z , and G has a traveling
salesman tour with cost at most k}.

u v
4

3 2
1
1

x w
Chapter 34
5 P.74
 Computer Theory Lab.

Theorem 34.13
 The hamiltonian cycle problem is NP-
complete.

Chapter 34 P.75
 Computer Theory Lab.

Proof.
 First, show that HAM-CYCLE belongs to NP.
 We now prove that VERTEX-COVER p HAM-
CYCLE, which shows that HAM-CYCLE is NP-
complete.
 Given an undirected graph G=(V,E) and an
integer k, we construct an undirected graph
G’=(V’,E’) that has a hamiltonian cycle iff G
has a vertex cover of size k.

Chapter 34 P.76
 Computer Theory Lab.

widget

[u,v,1] [v,u,1] [u,v,1] [v,u,1] [u,v,1] [v,u,1] [u,v,1] [v,u,1]

[u,v,2] [v,u,2]

[u,v,3] [v,u,3]
Wuv Wuv Wuv Wuv
[u,v,4] [v,u,4]

[u,v,5] [v,u,5]

[u,v,6] [v,u,6] [u,v,6] [v,u,6] [u,v,6] [v,u,6] [u,v,6] [v,u,6]

(a) (b) (c) (d)

Chapter 34 P.77
 The reduction of an instance of Computer Theory Lab.

the vertex-cover problem to an

instance of the hamiltonian-cycle
problem.
 (a) An undirected graph G with a vertex of
size 2, consisting if the lightly shaded vertices
w and y.
 (b) the undirected graph G’ produced by the
reduction, with the hamiltonian path
corresponding to the vertex cover shaded.
 The vertex cover {w,y} corresponds to edges
(s1,[w,x,1]) and (s2,[y,x,1]) appearing in the
hamiltonian cycle.
Chapter 34 P.78
 Computer Theory Lab.
w x

(a)

z y

(b)

[w,x,1] [x,w,1] [x,y,1] [y,x,1] [w,y,1] [y,w,1] [w,z,1] [z,w,1]

Wuv Wuv Wuv Wuv

[w,x,6] [x,w,6] [x,y,6] [y,x,6] [w,y,6] [y,w,6] [w,z,6] [z,w,6]

Chapter 34 P.79
 Computer Theory Lab.

Three types of edges in E’

1. Edges in widget.
2. {([u,u(i),6],[u,u(i+1),1]) : 1  i 
degree(u)-1}
3. {(sj,[u,u(1),1]) : uV and 1  j  k} 
{(sj,[u,u(degree(u)),6]) : uV and 1  j 
k}

Chapter 34 P.80
 Computer Theory Lab.

The reduction performed in

polynomial time
 |V’| = 12|E| + k
 12|E| + |V|

 |E’| = (14|E|) + (2|E| -|V|) + (2k|V|)

= 16|E| + (2k-1)|V|
 16|E| + (2|V|-1)|V|

Chapter 34 P.81
 Computer Theory Lab.

 The transformation from graph

G to G’ is a reduction.
 That is, G has a vertex cover of
size k iff G’ has a hamiltonian
cycle.

Chapter 34 P.82
 Computer Theory Lab.

Theorem 34.14. The traveling salesman problem is

NP-complete.
Proof.
 TSP  NP
HAM  CYCLE  P TSP

Chapter 34 P.83
 Computer Theory Lab.

34.5.5 The subset-sum problem

S={1,4,16,64,256,1040,1041,1093,1284,1344}
t=3754
S' ={1,16,64,256,1040,1093,1284}

SUBSET-SUM={<S, t> | there exists a subset S'  S such

that t   s
sS'

Chapter 34 P.84
 Computer Theory Lab.

Theorem 34.15
 The subset-sum problem is NP-
complete.

Chapter 34 P.85
 Computer Theory Lab.

Proof.
 First, show that SUBSET-SUM is in NP.
 We now show that 3-CNF-SAT p SUBSET-
SUM.
 Given a 3-CNF formula  over variables x1,
x2,…, xn with clauses C1, C2,…, Ck, each
containing exactly three distinct literals.
 The reduction algorithm constructs an
instance <S,t> of the subset-sum problem
such that  is satisfiable iff there is a subset
of S whose sum is exactly t.
Chapter 34 P.86
 Computer Theory Lab.

Example
 The formula in 3-CNF is  = C1  C2 
C3  C4, where C1 = (x1  ¬x2  ¬x3), C2
= (¬x1  ¬x2  ¬x3), C3 = (¬x1  ¬x2 
x3), and C4 = (x1  x2  x3).
 A satisfying assignment of  is <x1 = 0,
x2 = 0, x3 = 1>.

Chapter 34 P.87
 Computer Theory Lab.

The reduction of 3-CNF-SAT to

SUBSET-SUM C 4 has no ¬x1
x1 x2 x3 C1 C2 C3 C4
v1 = 1 0 0 1 0 0 1
v1’ = 1 0 0 0 1 1 0
v2 = 0 1 0 0 0 0 1 C4 has x2
v2’ = 0 1 0 1 1 1 0
v3 = 0 0 1 0 0 1 1
v3’ = 0 0 1 1 1 0 0
s1 = 0 0 0 1 0 0 0
s1 ’ = 0 0 0 2 0 0 0
s2 = 0 0 0 0 1 0 0
s2 ’ = 0 0 0 0 2 0 0
s3 = 0 0 0 0 0 1 0
s3 ’ = 0 0 0 0 0 2 0
s4 = 0 0 0 0 0 0 1
s4 ’ = 0 0 0 0 0 0 2
t = 1 1 1 4 4 4 4
Chapter 34 P.88
 Computer Theory Lab.

The reduction performed in

polynomial time
 The set S contains 2n+2k values, each
of which has n+k digits, and the time to
produce each digit is polynomial in n+k.
 The target t has n+k digits, and the
reduction produces each in constant
time.

Chapter 34 P.89
 Computer Theory Lab.

 3-CNF formula  is satisfiable

iff there is a subset S’  S
whose sum is t.

Chapter 34 P.90