Where Can We Draw The Line?

On the Hardness of
Satisfiability Problems

Complexity 1
©D.Moshkovits
 Introduction
• Objectives:
– To show variants of SAT and check if
they are NP-hard
• Overview:
– Known results
– 2SAT
– Max2SAT

Complexity 2
©D.Moshkovits
 What Do We Know?
• Checking if a propositional calculus
formula is satisfiable (SAT) is NP-
hard.

Example: propositional calculus formula

(xz(wx))(xy)y

Complexity 3
©D.Moshkovits
 What Do We Know?
• We concentrated on a special case:
CNF formulas.

structure of CNF formulas

(...........)…(...........)

Complexity 4
©D.Moshkovits
 What Do We Know?
maximal P
number of NP-hard
literals per
clause

4
We’ll
3 explore this!

2
1

Complexity 5
©D.Moshkovits
 2SAT
• Instance: A 2-CNF formula 
• Problem: To decide if  is satisfiable

Example: a 2CNF formula

(xy)(yz)(xz)(zy)

Complexity 6
©D.Moshkovits
 PAP 184-185

2SAT is in P
Theorem: 2SAT is polynomial-time
decidable.
Proof: We’ll show how to solve this
problem efficiently using path
searches in graphs…

Complexity 7
©D.Moshkovits
 Searching in Graphs
Theorem: Given a graph G=(V,E) and
two vertices s,tV, finding if there is
a path from s to t in G is polynomial-
time decidable.
Proof: Use some search algorithm
(DFS/BFS). 

Complexity 8
©D.Moshkovits
 Graph Construction
• Vertex for each variable and a
negation of a variable
• Edge (,) iff there exists a clause
equivalent to ()

Complexity 9
©D.Moshkovits
 Graph Construction: Example
(xy)(yz)(xz)(zy)

x
y
x

y
z
z

Complexity 10
©D.Moshkovits
 Observation
Claim: If the graph contains a path
from  to , it also contains a path
from  to .
Proof: If there’s an edge (,), then
there’s also an edge (,).

Complexity 11
©D.Moshkovits
 Correctness
Claim:
a 2-CNF formula  is unsatisfiable iff
there exists a variable x, such that:
1. there is a path from x to x in the
graph
2. there is a path from x to x in the
graph

Complexity 12
©D.Moshkovits
 Correctness (1)
• Suppose there are paths x..x and
x..x for some variable x, but there’s
also a satisfying assignment .
• If (x)=T (similarly for (x)=F):

() is false!

x  . . . x

T T F F
Complexity 13
©D.Moshkovits
 Correctness (2)
• Suppose there are no such paths.
• Construct an assignment as follows:
 2. assign T to
x
1. pick an x all reachable
unassigned y vertices
literal , with x
no path from  3. assign F to
to , and  their negations
y
assign it T  y
z 4. Repeat until all
z
z vertices are assigned
Complexity 14
©D.Moshkovits
 Correctness (2)
Claim: The algorithm is well defined.
Proof: If there were a path from x to
both y and y,
then there would have been a path
from x to y and from y to x.

Complexity 15
©D.Moshkovits
 Correctness
A formula is unsatisfiable iff there are
no paths of the form x..x and x..x.


Complexity 16
©D.Moshkovits
 2SAT is in P
We get the following efficient
algorithm for 2SAT:
– For each variable x find if there is a
path from x to x and vice-versa.
– Reject if any of these tests succeeded.
– Accept otherwise
 2SATP. 

Complexity 17
©D.Moshkovits
 Max2SAT
• Instance: A 2-CNF formula  and a
goal K.
• Problem: To decide if there is an
assignment satisfying at least K of ’s
clauses.
Example: a 2CNF formula

(xy)
(yz)
(xz)
(zy)
Complexity 18
©D.Moshkovits
 PAP 186-187

Max2SAT is in NPC
Theorem: Max2SAT is NP-Complete.
Proof: Max2SAT is clearly in NP.
We’ll show 3SATpMax2SAT.

(......)…(......) p (....)…(....)

Complexity 19
©D.Moshkovits
 Gadgets 

Proof: By checking.
Claim: Let
(x,y,z,w) = (x)(y)(z)(w)
(xy)(yz)(zx)
(xw)(yw)(zw).
• Every satisfying assignment for (xyz)
can be extended into an assignment that
satisfies exactly 7 of the clauses.
• Other assignments can satisfy at most 6 of
the clauses.
Complexity 20
©D.Moshkovits
 The Construction
• For each 1im, replace the i-th
clause of the 3-CNF formula ()
with a corresponding (,,,wi) to
get a 2-CNF formula.
• Fix K=7m.
Make sure this
construction is
poly-time

Complexity 21
©D.Moshkovits
 Correctness
• Every satisfying assignment for the
3-CNF formula can be extended into
an assignment that satisfies 7m
clauses.
• If 7m clauses of the 2-CNF formula
are satisfied, each  has 7 satisfied
clauses, so the original formula is
satisfied.
Complexity 22
©D.Moshkovits
 Corollary
 3SATpMax2SAT and Max2SATNP
 Max2SAT is NP-Complete. 

Complexity 23
©D.Moshkovits
 Summary 
• We’ve seen that checking if a given CNF
formula is satisfiable is:
– Polynomial-time decidable, if every clause
contains up to 2 literals.
– NP-hard, if each clause may contain more than
2 literals.
• We’ve also seen Max2SAT is NP-hard.

Complexity 24
©D.Moshkovits
 Conclusions 
• A special case of a NP-hard problem may
be polynomial time decidable.
• The optimization version of a polynomial-
time decidable problem may be NP-hard.

Complexity 25
©D.Moshkovits