More NP-completeness

Sipser 7.5 (pages 283-294)

CS 311Mount Holyoke College 2

NP’s hardest problems• Definition 7.34:

A language B is NP-complete if1. B NP∈2. A≤pB, for all A NP∈

NP

A1

SATA3

A2

CLIQUE

CS 311Mount Holyoke College 3

Hamiltonian paths• HAMPATH = {<G,s,t> | Hamiltonian path from s to t}∃• Theorem 7.46: HAMPATH is NP-complete.

s t

CS 311Mount Holyoke College 4

Hamiltonian paths• HAMPATH = {<G,s,t> | Hamiltonian path from s to t}∃• Theorem 7.46: HAMPATH is NP-complete.

s t

CS 311Mount Holyoke College 5

Remember… HAMPATH NP∈• N = "On input <G,s,t>:

1. Guess an orderings, p1, p2,..., pn, of the nodes of G

2. Check whether s = p1 and t = pn

3. For each i=1 to n-1, check whether (pi, pi+1) is an edge of G. If any are not, reject. Otherwise, accept.”

CS 311Mount Holyoke College 6

3SAT≤pHAMPATH

NP

A1

3SATA3

A2

HAMPATH

CS 311Mount Holyoke College 7

Proof outline• Given a boolean formula φ,

we convert it to a directed graph G such that φ has a valid truth assignment iff G has a Hamiltonian graph

CS 311Mount Holyoke College 8

3SAT’s main features• Choice:

Each variable has a choice between two truth values. • Consistency:

Different occurrences of the same variable have the same value.

• Constraints: Variable occurrences are organized into clauses that provide constraints that must be satisfied.

*We model each of these three features by a different a "gadget" in the graph G.

CS 311Mount Holyoke College 9

The choice gadget• Modeling variable xi

CS 311Mount Holyoke College 10

Zig-zagging and zag-zigging

Zig-zag (TRUE) Zag-zig (FALSE)

CS 311Mount Holyoke College 11

The consistency gadget

CS 311Mount Holyoke College 12

Clauses• Modeling clause cj

cj

CS 311Mount Holyoke College 13

The global structure

CS 311Mount Holyoke College 14

The constraint gadget• Modeling when clause cj contains xi

CS 311Mount Holyoke College 15

The constraint gadget• Modeling when clause cj contains xi

CS 311Mount Holyoke College 16

A situation that cannot occur

CS 311Mount Holyoke College 17

TSP is NP-complete• TSP: Given n cities,

1, 2, ..., n, together with a nonnegative distance dij between any two cities, find the shortest tour.

CS 311Mount Holyoke College 18

HAMPATH ≤p TSP

NP

A1

HAMPATHA3

A2

TSP

CS 311Mount Holyoke College 19

SUBSET-SUM is NP-complete• SUBSET-SUM=

{<S,t> | S = {x1,…,xk} and, for some {y1,…,yl} S, ⊆

Σ yi=t} • Why is SUBSET-SUM in NP?

CS 311Mount Holyoke College 20

3SAT ≤p SUBSET-SUM

CS 311Mount Holyoke College 21

And…if that’s not enough• There are more than 3000 known

NP-complete problems!

http://en.wikipedia.org/wiki/List_of_NP-complete_problems

CS 311Mount Holyoke College 22

Other types of complexity• Space complexity• Circuit complexity• Descriptive complexity• Randomized complexity• Quantum complexity• …