DESIGN AND ANALYSIS OF

ALGORITHMS

P, NP, and NP-Complete

snistforum.com
 Tractability
Some problems are intractable:
as they grow large, we are unable to solve
them in reasonable time
What constitutes reasonable time?
Standard working definition: polynomial time
On an input of size n the worst-case running
time is O(nk) for some constant k
O(n2), O(n3), O(1), O(n lg n), O(2n), O(nn), O(n!)
Polynomial time: O(n2), O(n3), O(1), O(n lg n)
Not in polynomial time: O(2n), O(nn), O(n!)

snistforum.com
Polynomial-Time Algorithms
Are some problems solvable in polynomial
time?
Of course: many algorithms we’ve studied
provide polynomial-time solutions to some
problems
Are all problems solvable in polynomial time?
No: Turing’s “Halting Problem” is not solvable
by any computer, no matter how much time is
given
Most problems that do not yield polynomial-
time algorithms are either optimization or
decision problems.
snistforum.com
Optimization/Decision
Problems
Optimization Problems
An optimization problem is one which asks,
“What is the optimal solution to problem X?”
Examples:
 0-1 Knapsack
 Fractional Knapsack
 Minimum Spanning Tree
Decision Problems
An decision problem is one with yes/no
answer
Examples:
 Does a graph G have a MST of weight  W?

snistforum.com
Optimization/Decision
Problems
 An optimization problem tries to find an optimal
solution
 A decision problem tries to answer a yes/no question
 Many problems will have decision and optimization
versions
 Eg: Traveling salesman problem
 optimization: find hamiltonian cycle of minimum weight
 decision: is there a hamiltonian cycle of weight  k
 Some problems are decidable, but intractable:
as they grow large, we are unable to solve them in
reasonable time
 Is there a polynomial-time algorithm that solves the problem?

snistforum.com
 The Class P
P: the class of decision problems that have
polynomial-time deterministic algorithms.
 That is, they are solvable in O(p(n)), where p(n) is a
polynomial on n
 A deterministic algorithm is (essentially) one that always
computes the correct answer

Why polynomial?
 if not, very inefficient
 nice closure properties
 the sum and composition of two polynomials are always polynomials
too

snistforum.com
Sample Problems in P
Fractional Knapsack
MST
Sorting
Others?

snistforum.com
 The class NP
NP: the class of decision problems that are solvable
in polynomial time on a nondeterministic machine
(or with a nondeterministic algorithm)
 (A determinstic computer is what we know)
 A nondeterministic computer is one that can
“guess” the right answer or solution
 Think of a nondeterministic computer as a parallel
machine that can freely spawn an infinite number of
processes
 Thus NP can also be thought of as the class of
problems
 whose solutions can be verified in polynomial time
 Note that NP stands for “Nondeterministic
Polynomial-time”
snistforum.com
Sample Problems in NP
Fractional Knapsack
MST
Sorting
Others?
Hamiltonian Cycle (Traveling Salesman)
Graph Coloring
Satisfiability (SAT)
 the problem of deciding whether a given
Boolean formula is satisfiable

snistforum.com
The Satisfiability (SAT)
Problem
Satisfiability (SAT):
Given a Boolean expression on n variables,
can we assign values such that the
expression is TRUE?
Ex: ((x1 x2)  ((x1  x3)  x4)) x2
Seems simple enough, but no known
deterministic polynomial time algorithm
exists
Easy to verify in polynomial time!

snistforum.com
Example: CNF satisfiability
 This problem is in NP. Nondeterministic
algorithm:
 Guess truth assignment
 Check assignment to see if it satisfies CNF formula

 Example:
(A¬B  ¬C )  (¬A  B)  (¬ B  D  F )  (F  ¬ D)
 Truth assignments:
ABCDEF
 0 1 1 0 1 0
 1 0 0 0 0 1
 1 1 0 0 0 1
 ... (how many more?)

Checking phase: Θ(n)

snistforum.com
Review: P And NP Summary
P = set of problems that can be solved in
polynomial time
Examples: Fractional Knapsack, …
NP = set of problems for which a solution can
be verified in polynomial time
Examples: Fractional Knapsack,…, Hamiltonian
Cycle, CNF SAT, 3-CNF SAT
Clearly P  NP
Open question: Does P = NP?
Most suspect not
An August 2010 claim of proof that P ≠ NP, by
Vinay Deolalikar, researcher at HP Labs, Palo Alto,
has flaws
snistforum.com
NP-complete problems
 A decision problem D is NP-complete iff
1. D  NP
2. every problem in NP is polynomial-time
reducible to D

 Cook’s theorem (1971): CNF-sat is NP-

complete

snistforum.com
 Reduction
A problem R can be reduced to another problem
Q if any instance of R can be rephrased to an
instance of Q, the solution to which provides a
solution to the instance of R
This rephrasing is called a transformation
Intuitively: If R reduces in polynomial time to Q,
R is “no harder to solve” than Q
Example: lcm(m, n) = m * n / gcd(m, n),
lcm(m,n) problem is reduced to gcd(m, n)
problem

snistforum.com
 NP-Hard and NP-Complete
If R is polynomial-time reducible to Q, we
denote this R p Q
Definition of NP-Hard and NP-Complete:
If all problems R  NP are polynomial-time
reducible to Q, then Q is NP-Hard
We say Q is NP-Complete if Q is NP-Hard
and Q  NP
If R p Q and R is NP-Hard, Q is also NP-Hard
(why?)

snistforum.com
 Practical Implication
Given a problem that is known to be NP-
Complete
Try to solve it by designing a polynomial-time
algorithm?
 Prove P=NP

Alleviate the intractability of such problems

 To make some large instances of the problem solvable
(like solving some instances of Knapsack problem in
polynomial time)
 To find good approximations (Chap 12)

snistforum.com
Appendix

The following slides are from a document

by Dr. Robins, University of Virginia

snistforum.com
snistforum.com
snistforum.com