Top Banner
Between 2- and 3- colorability Rutgers University
33

Between 2- and 3-colorability Rutgers University.

Dec 18, 2015

Download

Documents

Frederica Lewis
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Between 2- and 3-colorability Rutgers University.

Between 2- and 3-colorability

Rutgers University

Page 2: Between 2- and 3-colorability Rutgers University.

The problem

GX Bipartite

Graph

O Independent Set

Page 3: Between 2- and 3-colorability Rutgers University.

The problem

GX Bipartite

Graph• tree

O Independent Set

Page 4: Between 2- and 3-colorability Rutgers University.

The problem

GX Bipartite

Graph• tree• forest

O Independent Set

Page 5: Between 2- and 3-colorability Rutgers University.

The problem

GX Bipartite

Graph• tree• forest• of bounded degree

O Independent Set

Page 6: Between 2- and 3-colorability Rutgers University.

The problem

GX Bipartite

Graph• tree• forest• of bounded degree• complete bipartite

O Independent Set

Page 7: Between 2- and 3-colorability Rutgers University.

Examples

• Trees NP-completeA. Brandstädt, V. B. Le, T. Szymczak, The complexity of some problems related to graph 3-colorability. Discrete Appl. Math. 89 (1998) 59--73.

Page 8: Between 2- and 3-colorability Rutgers University.

Examples

• Trees NP-completeA. Brandstädt, V. B. Le, T. Szymczak, The complexity of some problems related to graph 3-colorability. Discrete Appl. Math. 89 (1998) 59--73.

• Forest NP-complete

Page 9: Between 2- and 3-colorability Rutgers University.

Examples

• Trees NP-completeA. Brandstädt, V. B. Le, T. Szymczak, The complexity of some problems related to graph 3-colorability. Discrete Appl. Math. 89 (1998) 59--73.

• Forest NP-complete

• Graphs of bounded vertex degree NP-completeJ. Kratochvíl, I. Schiermeyer, On the computational complexity of (O, P)-partition problems. Discuss. Math. Graph Theory 17 (1997) 253--258.

Page 10: Between 2- and 3-colorability Rutgers University.

Examples

• Trees NP-completeA. Brandstädt, V. B. Le, T. Szymczak, The complexity of some problems related to graph 3-colorability. Discrete Appl. Math. 89 (1998) 59--73.

• Forest NP-complete

• Graphs of bounded vertex degree NP-completeJ. Kratochvíl, I. Schiermeyer, On the computational complexity of (O, P)-partition problems. Discuss. Math. Graph Theory 17 (1997) 253--258.

• Complete bipartite PolynomialA. Brandstädt, P.L. Hammer, V.B. Le, V. Lozin, Bisplit graphs. Discrete Math. 299 (2005) 11--32.

Page 11: Between 2- and 3-colorability Rutgers University.

Question

Is there any boundary separating difficult instances of the (O,P)-partition problem from polynomially solvable ones?

Page 12: Between 2- and 3-colorability Rutgers University.

Question

Is there any boundary separating difficult instances of the (O,P)-partition problem from polynomially solvable ones?Yes ?

Page 13: Between 2- and 3-colorability Rutgers University.

Hereditary classes of graphs

Definition.

A class of graphs P is hereditary if XP implies X-vP for any vertex vV(X)

Page 14: Between 2- and 3-colorability Rutgers University.

Hereditary classes of graphs

Definition.

A class of graphs P is hereditary if XP implies X-vP for any vertex vV(X)

Examples: perfect graphs (bipartite, interval, permutation graphs), planar graphs, line graphs, graphs of bounded vertex degree.

Page 15: Between 2- and 3-colorability Rutgers University.

Speed of hereditary properties

E.R. Scheinerman, J. Zito, On the size of hereditary classes of graphs. J. Combin. Theory Ser. B 61 (1994) 16--39.

Alekseev, V. E. On lower layers of a lattice of hereditary classes of graphs. (Russian) Diskretn. Anal. Issled. Oper. Ser. 1 4 (1997) 3--12.

J. Balogh, B. Bllobás, D. Weinreich, The speed of hereditary properties of graphs. J. Combin. Theory Ser. B 79 (2000) 131--156.

Page 16: Between 2- and 3-colorability Rutgers University.

Lower Layers

• constant

• polynomial

• exponential

• factorial

Page 17: Between 2- and 3-colorability Rutgers University.

Lower Layers

• constant

• polynomial

• exponential

• factorial

permutation graphs line graphs

graphs of bounded vertex degree graphs of bounded tree-width

planar graphs

Page 18: Between 2- and 3-colorability Rutgers University.

Minimal Factorial Classes of graphs

Bipartite graphs

3 subclasses

Complements of bipartite graphs

3 subclasses

Split graphs, i.e., graphs partitionable into an independent set and a clique

3 subclasses

Page 19: Between 2- and 3-colorability Rutgers University.

Three minimal factorial classes

of bipartite graphs

P1 The class of graphs of vertex degree at most 1

Page 20: Between 2- and 3-colorability Rutgers University.

Three minimal factorial classes

of bipartite graphs

P1 The class of graphs of vertex degree at most 1

P2 Bipartite complements to graphs in P1

Page 21: Between 2- and 3-colorability Rutgers University.

Three minimal factorial classes

of bipartite graphs

P1 The class of graphs of vertex degree at most 1

P2 Bipartite complements to graphs in P1

P3 2K2-free bipartite graphs (chain or difference graphs)

Page 22: Between 2- and 3-colorability Rutgers University.

(O,P)-partition problem

Let P be a hereditary class of bipartite graphsProblem. Determine whether a graph G admits a partition into an independent set and a graph in the class P

Page 23: Between 2- and 3-colorability Rutgers University.

(O,P)-partition problem

Let P be a hereditary class of bipartite graphsProblem. Determine whether a graph G admits a partition into an independent set and a graph in the class P

Conjecture

If P contains one of the three minimal factorial classes of bipartite graphs, then the (O,P)-partition problem is NP-complete. Otherwise it is solvable in polynomial time.

Page 24: Between 2- and 3-colorability Rutgers University.

Polynomial-time results

Theorem. If P contains none of the three minimal factorial classes of bipartite graphs, then the (O,P)-partition problem can be solved in polynomial time.

Page 25: Between 2- and 3-colorability Rutgers University.

Polynomial-time results

Theorem. If P contains none of the three minimal factorial classes of bipartite graphs, then the (O,P)-partition problem can be solved in polynomial time.

If P contains none of the three minimal factorial classes of bipartite graphs, then P belongs to one of the lower layers

• exponential

• polynomial

• constant

Page 26: Between 2- and 3-colorability Rutgers University.

Exponential classes of bipartite graphs

Theorem. For each exponential class of bipartite graphs P, there is a constant k such that for any graph G in P there is a partition of V(G) into at most k independent sets such that every pair of sets induces either a complete bipartite or an empty (edgeless) graph.

Page 27: Between 2- and 3-colorability Rutgers University.

Exponential classes of bipartite graphs

Theorem. For each exponential class of bipartite graphs P, there is a constant k such that for any graph G in P there is a partition of V(G) into at most k independent sets such that every pair of sets induces either a complete bipartite or an empty (edgeless) graph.

(O,P)-partition 2-sat

Page 28: Between 2- and 3-colorability Rutgers University.

NP-complete results

J. Kratochvíl, I. Schiermeyer, On the computational complexity of (O,P)-partition problems. Discuss. Math. Graph Theory 17 (1997) 253--258.

Theorem. If P is a monotone class of graphs different from the class of empty (edgeless) graphs, then the (O,P)-partition problem is NP-complete.

Page 29: Between 2- and 3-colorability Rutgers University.

NP-complete results

J. Kratochvíl, I. Schiermeyer, On the computational complexity of (O,P)-partition problems. Discuss. Math. Graph Theory 17 (1997) 253--258.

Theorem. If P is a monotone class of graphs different from the class of empty (edgeless) graphs, then the (O,P)-partition problem is NP-complete.

Corollary. The (O,P)-partition problem is NP-complete if P is the class of graphs of vertex degree at most 1.

Page 30: Between 2- and 3-colorability Rutgers University.

One more result

Yannakakis, M. Node-deletion problems on bipartite graphs. SIAM J. Comput. 10 (1981), no. 2, 310--327.

Page 31: Between 2- and 3-colorability Rutgers University.

One more result

Yannakakis, M. Node-deletion problems on bipartite graphs. SIAM J. Comput. 10 (1981), no. 2, 310--327.

Let P be a hereditary class of bipartite graphsProblem*(P). Given a bipartite graph G, find a maximum induced subgraph of G belonging to P.

Page 32: Between 2- and 3-colorability Rutgers University.

One more result

Yannakakis, M. Node-deletion problems on bipartite graphs. SIAM J. Comput. 10 (1981), no. 2, 310--327.

Let P be a hereditary class of bipartite graphsProblem*(P). Given a bipartite graph G, find a maximum induced subgraph of G belonging to P.

Theorem. If P is a non-trivial hereditary class of bipartite graphs containing one of the three minimal factorial classes of bipartite graphs, then Problem*(P) is NP-hard. Otherwise, it is solvable in polynomial time.

Page 33: Between 2- and 3-colorability Rutgers University.

Thank you