Ma/CS 6b Class 5: Graph Connectivity

Similar documents
Ma/CS 6b Class 4: Matchings in General Graphs

MC 302 GRAPH THEORY 10/1/13 Solutions to HW #2 50 points + 6 XC points

Ma/CS 6b Class 11: Kuratowski and Coloring

Ma/CS 6b Class 10: Ramsey Theory

Ma/CS 6b Class 26: Art Galleries and Politicians

Assignment 4 Solutions of graph problems

Adjacent: Two distinct vertices u, v are adjacent if there is an edge with ends u, v. In this case we let uv denote such an edge.

Ma/CS 6a Class 8: Eulerian Cycles

Ma/CS 6b Class 12: Ramsey Theory

Ma/CS 6b Class 2: Matchings

Ma/CS 6b Class 13: Counting Spanning Trees

CPCS Discrete Structures 1

Section 8.2 Graph Terminology. Undirected Graphs. Definition: Two vertices u, v in V are adjacent or neighbors if there is an edge e between u and v.

Math 170- Graph Theory Notes

Ma/CS 6a Class 9: Coloring

Matching Algorithms. Proof. If a bipartite graph has a perfect matching, then it is easy to see that the right hand side is a necessary condition.

DO NOT RE-DISTRIBUTE THIS SOLUTION FILE

Today. Types of graphs. Complete Graphs. Trees. Hypercubes.

Theorem 3.1 (Berge) A matching M in G is maximum if and only if there is no M- augmenting path.

Number Theory and Graph Theory

CPS 102: Discrete Mathematics. Quiz 3 Date: Wednesday November 30, Instructor: Bruce Maggs NAME: Prob # Score. Total 60

(5.2) 151 Math Exercises. Graph Terminology and Special Types of Graphs. Malek Zein AL-Abidin

HW Graph Theory SOLUTIONS (hbovik)

11.4 Bipartite Multigraphs

Ma/CS 6b Class 10: Kuratowski's Theorem

Ma/CS 6b Class 10: Kuratowski's Theorem

CS70 - Lecture 6. Graphs: Coloring; Special Graphs. 1. Review of L5 2. Planar Five Color Theorem 3. Special Graphs:

1 Matchings in Graphs

Graph Theory S 1 I 2 I 1 S 2 I 1 I 2

DEFINITION OF GRAPH GRAPH THEORY GRAPHS ACCORDING TO THEIR VERTICES AND EDGES EXAMPLE GRAPHS ACCORDING TO THEIR VERTICES AND EDGES

Discrete Wiskunde II. Lecture 6: Planar Graphs

CHAPTER 2. Graphs. 1. Introduction to Graphs and Graph Isomorphism

Math 776 Graph Theory Lecture Note 1 Basic concepts

Ma/CS 6b Class 7: Minors

Graphs. Introduction To Graphs: Exercises. Definitions:

Graph Definitions. In a directed graph the edges have directions (ordered pairs). A weighted graph includes a weight function.

CS 311 Discrete Math for Computer Science Dr. William C. Bulko. Graphs

HW Graph Theory SOLUTIONS (hbovik) - Q

K 4 C 5. Figure 4.5: Some well known family of graphs

Matching Theory. Figure 1: Is this graph bipartite?

Characterizing Graphs (3) Characterizing Graphs (1) Characterizing Graphs (2) Characterizing Graphs (4)

2. Lecture notes on non-bipartite matching

Section 3.1: Nonseparable Graphs Cut vertex of a connected graph G: A vertex x G such that G x is not connected. Theorem 3.1, p. 57: Every connected

8.2 Paths and Cycles

Paths, Flowers and Vertex Cover

Definition For vertices u, v V (G), the distance from u to v, denoted d(u, v), in G is the length of a shortest u, v-path. 1

Math236 Discrete Maths with Applications

Approximation slides 1. An optimal polynomial algorithm for the Vertex Cover and matching in Bipartite graphs

Lecture 19 Thursday, March 29. Examples of isomorphic, and non-isomorphic graphs will be given in class.

Math 778S Spectral Graph Theory Handout #2: Basic graph theory

γ(ɛ) (a, b) (a, d) (d, a) (a, b) (c, d) (d, d) (e, e) (e, a) (e, e) (a) Draw a picture of G.

DO NOT RE-DISTRIBUTE THIS SOLUTION FILE

5 Graphs

Discrete Mathematics and Probability Theory Fall 2013 Vazirani Note 7

Math.3336: Discrete Mathematics. Chapter 10 Graph Theory

The Structure of Bull-Free Perfect Graphs

Module 7. Independent sets, coverings. and matchings. Contents

Extremal Graph Theory: Turán s Theorem

Chapter 3: Paths and Cycles

Potential Bisections of Large Degree

Bipartite Roots of Graphs

CS473-Algorithms I. Lecture 13-A. Graphs. Cevdet Aykanat - Bilkent University Computer Engineering Department

MATH20902: Discrete Maths, Solutions to Problem Set 1. These solutions, as well as the corresponding problems, are available at

Course Introduction / Review of Fundamentals of Graph Theory

Graph Theory Day Four

Advanced Combinatorial Optimization September 17, Lecture 3. Sketch some results regarding ear-decompositions and factor-critical graphs.

1. Chapter 1, # 1: Prove that for all sets A, B, C, the formula

5 Matchings in Bipartite Graphs and Their Applications

Solutions to Problem Set 2

Discrete Mathematics for CS Spring 2008 David Wagner Note 13. An Introduction to Graphs

Two Characterizations of Hypercubes

{ 1} Definitions. 10. Extremal graph theory. Problem definition Paths and cycles Complete subgraphs

by conservation of flow, hence the cancelation. Similarly, we have

Maximal Monochromatic Geodesics in an Antipodal Coloring of Hypercube

The Six Color Theorem

A generalization of Mader s theorem

2. CONNECTIVITY Connectivity

Some Upper Bounds for Signed Star Domination Number of Graphs. S. Akbari, A. Norouzi-Fard, A. Rezaei, R. Rotabi, S. Sabour.

On the Relationships between Zero Forcing Numbers and Certain Graph Coverings

Graph theory - solutions to problem set 1

Math 454 Final Exam, Fall 2005

AMS /672: Graph Theory Homework Problems - Week V. Problems to be handed in on Wednesday, March 2: 6, 8, 9, 11, 12.

Introduction to Graph Theory

CS 441 Discrete Mathematics for CS Lecture 26. Graphs. CS 441 Discrete mathematics for CS. Final exam

CSE 331: Introduction to Algorithm Analysis and Design Graphs

v V Question: How many edges are there in a graph with 10 vertices each of degree 6?

Week 9-10: Connectivity

V10 Metabolic networks - Graph connectivity

Discrete Mathematics and Probability Theory Fall 2009 Satish Rao,David Tse Note 8

Generalized Pebbling Number

Problem Set 2 Solutions

BIL694-Lecture 1: Introduction to Graphs

Vertex coloring, chromatic number

Math 485, Graph Theory: Homework #3

The vertex set is a finite nonempty set. The edge set may be empty, but otherwise its elements are two-element subsets of the vertex set.

Module 11. Directed Graphs. Contents

FOUR EDGE-INDEPENDENT SPANNING TREES 1

Paths, Flowers and Vertex Cover

Introduction III. Graphs. Motivations I. Introduction IV

Vertex coloring, chromatic number

Transcription:

Ma/CS 6b Class 5: Graph Connectivity By Adam Sheffer Communications Network We are given a set of routers and wish to connect pairs of them to obtain a connected communications network. The network should be reliable a few malfunctioning routers should not disable the entire network. What condition should we require from the network? That after removing any k routers, the network remains connected. 1

k-connected Graphs An graph G = (V, E) is said to be k-connected if V > k and we cannot obtain a non-connected graph by removing k 1 vertices from V. Is the graph in the figure 1-connected? Yes. -connected? Yes. 3-connected? No! Connectivity Which graphs are 1-connected? These are the connected graphs ( V > 1). The connectivity of a graph G is the maximum k such that G is k-connected. What is the connectivity of the complete graph K n? n 1. The graph in the figure has a connectivity of.

Hypercube 1 A hypercube is a generalization of the cube to any dimension. In d-dimensions, we consider the vertices of the hypercube as the points whose coordinates are 0 s and 1 s. Two vertices are adjacent if they differ in a single coordinate. (0,1) (1,1) 0 (0,0) (1,0) Hypercube Properties 1 The points of a d-dimensional hypercube correspond to the d-dimensional vectors with 0-1 entries. Thus, it has d vertices. Two vertices are adjacent if they have d 1 common coordaintes. Thus, there are d 1 d edges. (0,1) (1,1) 0 (0,0) (1,0) 3

Hypercube graphs The hypercube graph Q d corresponds to the d-dimensional hypercube. A vertex for each vertex of the hypercube. An edge for each edge of the hypercube. The graph Q d has d vertices and d 1 d edges. It is d-regular. 4d-hypercube Hypercube graph Properties Problem. Is Q d a bipartite graph? Answer. Yes! On one side we place every vertex with an even number of 1-coordinates. On the other, the vertices with an odd number of 1-coordintes. 4d-hypercube 4

More Properties Problem. What is the connectivity of Q d? Answer. We can disconnect Q d by removing the d neighbors of any vertex of Q d. So the connectivity is at most d. We prove by induction that it is exactly d. Q 4 Proof by Induction We prove by induction that the connectivity of Q d is d. Induction basis. Easy to check for d = 1,. Induction step. We can consider Q d as two copies Q, Q of Q d 1 with a perfect matching between their vertices. Q Q Q d 5

Completing the Proof S a minimum disconnecting set of Q d. To complete the proof, we claim that S = d. If after removing S both Q and Q connected, it must disconnect Q from Q, and must be of size at least d 1 d. By the induction hypothesis, disconnecting Q (or Q ) requires removing d 1 vertices of Q. To disconnect Q in Q d, we must remove at least one additional vertex from Q. k-edge Connected Graphs A graph G = (V, E) is said to be k-edgeconnected if V > 1 and we cannot obtain a non-connected graph by removing at most k 1 edges from E. Is the graph in the figure 1-edge-connected? Yes. -edge-connected? Yes. 3-edge-connected? Yes. 6

k-edge-connectivity Which graphs are 1-edge-connected? These are the connected graphs ( V > 1). The edge-connectivity of a graph G is the maximum k such that G is k-edge-connected. What is the edge-connectivity of the complete graph K n? n 1. The graph in the figure has an edge-connectivity of 3. Edge Connectivity and Minimum Degree What is the relation between the minimum degree and the edge-connectivity of a graph? Can they be equal? Yes Can the edge connectivity be smaller? Yes Can the minimum degree be smaller? 7

Edge Connectivity and Minimum Degree (cont.) Claim. In any graph, the edge connectivity is at most the minimum degree. Proof. Consider a vertex v of minimum degree, and denote this degree as d. By removing the d edges that are adjacent to v, we disconnect the graph. Connectivity and Edge-Connectivity What is the relation between the connectivity and the edge-connectivity of a graph? Can they be equal? Yes Can the connectivity be smaller? Yes Can the edge-connectivity be smaller? 8

Connectivity and Edge-Connectivity Claim. In any graph G = V, E, the connectivity is at most the edge-connectivity. Proof. Let F E be a minimum set of edges whose removal disconnects G. We need to find a set of at most F vertices that disconnects G. We divide the analysis into two cases: A vertex v V not adjacent to any edge of F. Every vertex of V is adjacent to an edge of F. Analysis of Case 1 Assume that a vertex v V is not adjacent to any edge of F. By definition, the removal of F disconnects G. Let C be the connected component in G F that contains v. Removing the vertices V in C that are adjacent to an edge of F disconnects v from the vertices of G C. Since F is minimal, each edge of F has at most F v one endpoint in C, so V F. 9

Analysis of Case Assume that every vertex of V is adjacent to an edge of F. If G is a complete graph, the connectivity is V 1 and edge-connectivity is V 1. Analysis of Case (cont.) Assume that every vertex of V is adjacent to an edge of F. If G is not complete, there exists v V such that deg v < V 1. Removing the set V of neighbors of v disconnects G. C the connected component of v in G F. Every vertex of V is either in C or connected to v by an edge of F. Since no edge of F is between two vertices of C, we have F V F. v 10

Connectivity in 3-regular Graphs Claim. If G = V, E is a connected 3-regular graph, then the connectivity and the edge-connectivity of G are equal. Proof k v (vertex) connectivity. k e edge connectivity. We already know that k v k e. Thus, it suffices to find a set of k v edges that disconnects G. S a minimum set of k v vertices whose removal disconnects G into C 1 and C (which might not be connected themselves) 11

Proof (cont.) Every vertex of v S is connected to both C 1 and C, since S is a minimum disconnecting set. Due to 3-regularity, either v is adjacent to a single vertex of C 1, or to a single vertex of C. That is, by removing a single edge, we can disconnect v from either C 1 or from C. By removing such an edge from each v S, we obtain a disconnecting set of k v edges. Recap We proved that in every graph: Edge Connectivity connectivity Minimum degree This implies that small minimum degree implies small connectivity. Does large minimum degree imply large connectivity. No! 1

Highly Connected Subgraphs Recall. The average degree of a graph G = V, E is deg G = 1 V v V deg v = E V. Theorem (Mader `7). Let k be a positive integer and let G = V, E be a graph such that deg G 4k. Then there exists a k + 1 -connected subgraph H G with minimum degree > deg(g)/. Proof Set d = deg G 4k. Let H = (V, E ) be a subgraph of G such that V k and E > d V k. Among the subgraphs that satisfy the above, we take as H one that minimizes V. Such subgraphs exist since G is one: Since d 4k, there exists a vertex of degree at least 4k, and thus V 4k + 1. E = E V V = d V > d V k. 13

Proof () Set d = deg G 4k. H = (V, E ) subgraph of G such that V k and E > d V k. If V = k, we get the contradiction E > d V k = d k k > k = V E. Proof (3) H = (V, E ) the minimal subgraph of G such that V k and E > d V k. We actually have V k + 1. Removing a vertex of degree d/ decreases d V k by d and E by at most d. That is, both V k and E > d k remains valid, contradicting the minimality of H. V 14

Proof (4) H = (V, E ) the minimal subgraph of G such that V k and E > d V k. We proved that every vertex of V is of degree larger than d/ in H. We need to show that H is (k + 1)-connected. Assume for contradiction that the removal of a set S of k vertices disconnects H. We can thus partition the vertices of V S into the subsets V 1, V with no edges between them. Proof (5) Assume for contradiction that the removal of a set S of k vertices whose removal disconnects H. Partition the vertices of V S into subsets V 1, V with no edges between them. H 1 the subgraph induced by V 1 S. Consider a vertex v V 1 and notice that all the edges that are adjacent to it remain in H 1. Since deg v > d k, we have V 1 S > k. S V 1 V 15

Illustration V 1 V S H 1 H S S Proof (6) The subgraph H 1 has more than k vertices. In order not to contradict the minimality of H, the number of edges of H 1 must be d V 1 S k. Similarly, the number of edges of H must be d V S k. E d V 1 S k + d V S k = d V k 16

The End 17