## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Chapter 1: Fundamental Concepts

Section 1.2: Paths, Cycles, and Trails

18

1.1.44. For each pair of 3-edge paths P = (u 0 , u 1 , u 2 , u 3 ) and Q = (v0 , v1 , v2 , v3 ) in the Petersen graph, there is an automorphism of the Petersen graph that turns P into Q . In the disjointness representation of the Petersen graph, suppose the pairs corresponding to the vertices of P are ab, cd , e f , gh , respectively. Since consecutive pairs are disjoint and the edges are unordered pairs, we may write the pairs so that a , b, c, d , e are distinct, f = a , g = b, and h = c. Putting the vertex names of Q in the same format AB , C D , E F , G H , we chose the isomorphism generated by the permutation of [5] that turns a , b, c, d , e into A , B , C , D , E , respectively. 1.1.45. A graph with 12 vertices in which every vertex has degree 3 and the only automorphism is the identity.

σ (51) = ea , and the only vertex not on (ab, cd , ea , bc, de) that is adjacent to ea is bd , we must have σ (24) = bd , as desired.

e • b• d• p • • q

•c • a

•x

u • w•

•v

a new vertex of degree 2, then G is edge-transitive but not vertex-transitive. Every edge consists of an old vertex and a new vertex. The n ! permutations of old vertices yield automorphism. Let x & y denote the new vertex on the path replacing the old edge x y ; note that x & y = y &x . The edge joining x and x & y is mapped to the edge joining u and u &v by any automorphism that maps x to u and y to v . The graph is not vertex-transitive, since x & y has degree 2, while x has degree n − 1. b) If G is edge-transitive but not vertex-transitive and has no isolated vertices, then G is bipartite. Let u v be an arbitrary edge of G . Let S be the set of vertices to which u is mapped by automorphisms of G , and let T be the set of vertices to which v is mapped. Since G is edge-transitive and has no isolated vertex, S ∪ T = V (G ). Since G is not vertex-transitive, S = V (G ). Together, these statements yield S ∩ T = ∅, since the composition of two automorphisms is an automorphism. By edge-transitivity, every edge of G contains one vertex of S and one vertex of T . Since S ∩ T = ∅, this implies that G is bipartite with vertex bipartition S , T . c) The graph below is vertex-transitive but not edge-transitive. A composition of left-right reﬂections and vertical rotations can take each vertex to any other. The graph has some edges on triangles and some edges not on triangles, so it cannot be edge-transitive.

• y

There are many ways to prove that an automorphism must ﬁx all the vertices. The graph has only two triangles (abc and u vw ). Now an automorphism must ﬁx p , since is the only vertex having no neighbor on a triangle, and also e, since it is the only vertex with neighbors on both triangles. Now d is the unique common neighbor of p and e. The remaining vertices can be ﬁxed iteratively in the same way, by ﬁnding each as the only unlabeled vertex with a speciﬁed neighborhood among the vertices already ﬁxed. (This construction was provided by Luis Dissett, and the argument forbidding nontrivial automorphisms was shortened by Fred Galvin. Another such graph with three triangles was found by a student of Fred Galvin.) 1.1.46. Vertex-transitivity and edge-transitivity. The graph on the left in Exercise 1.1.21 is isomorphic to the 4-dimensional hypercube (see Section 1.3), which is vertex-transitive and edge-transitive via the permutation of coordinates. For the graph on the right, rotation and inside-out exchange takes care of vertex-transitivity. One further generating operation is needed to get edge-transitivity; the two bottom outside vertices can be switched with the two bottom inside vertices. 1.1.47. Edge-transitive versus vertex-transitive. a) If G is obtained from K n with n ≥ 4 by replacing each edge of K n with a path of two edges through

• • • •

• •

**1.2. PATHS, CYCLES, AND TRAILS
**

1.2.1. Statements about connection. a) Every disconnected graph has an isolated vertex—FALSE. A simple 4-vertex graph in which every vertex has degree 1 is disconnected and has no isolated vertex. b) A graph is connected if and only if some vertex is connected to all other vertices—TRUE. A vertex is “connected to” another if they lie in a common path. If G is connected, then by deﬁnition each vertex is connected to every other. If some vertex x is connected to every other, then because a u , x -path and x , v -path together contain a u , v -path, every vertex is connected to every other, and G is connected.

Thus there are four components.9. The union of an even cycle and an odd cycle that share one vertex is an Eulerian graph with an even number of vertices and an odd number of edges. the i th row and i th column vanish. In a bipartite graph. this counts every edge exactly once. .2. . vn .2. Given the drawing of the Petersen graph consisting of two disjoint 5-cycles and edges between them. i by one. 1.2. It has a spanning path 7. the total is also even. Since the summands are all even.11. v -walk of odd length. a) Every Eulerian bipartite graph has an even number of edges—TRUE. then having a neighbor in each such component yields d G (v) ≥ k . bacd (two are maximum paths). 1.10. cd (one is a maximum clique). following an Eulerian circuit and ending at the initial vertex requires taking an even number of steps. If G is connected. the graph has no edges and is Eulerian.2. b) K 4 has a trail that is not closed and is not a path. bd . The biclique K m . so it needs at least 5 trails. form paths consisting of one edge from each cycle and one edge joining them.2. a single vertex forms a closed trail that is not a cycle. 1.12. Maximal cliques: abc.13 are isolated. 1. Maximal independent sets: c.3.2. If G is connected and G − v has k components. The vertices and edges of a closed trail form an even graph. The minimum number of trails that decompose the Petersen graph is 5. .6. and Trails 20 c) The edge set of every closed trail can be partitioned into edge sets of cycles—TRUE. 1. b) Every Eulerian simple graph with an even number of vertices has an even number of edges—FALSE. and Theorem 1.2. . b• a• c • •d 1.3. . in the adjacency matrix it reduces positions i . Since G is connected.2: Paths. Let G be a bipartite graph.2.27 applies.n is Eulerian if and only if m and n are both even or one of them is 0. and Proposition 1. then from a ﬁxed vertex u we can walk to all other vertices.33 implies that ﬁve trails sufﬁce. 15}. and it also deletes the column for each edge incident to vi . and it must be in the opposite set if there is a u . When m or n is 0.5. Since every walk alternates between the partite sets.8. j and j . Proof 3. The Petersen graph has exactly 10 vertices of odd degree. then v has a neighbor in every component of G − v .6. 1. . then the trail uses an odd number of edges incident to v . Walks in K 4 . If v is a cut-vertex. repeat an edge. Hence the number of edges is a sum of even numbers. and the maximal path length is 11. Proof 1. and there is no effect on the rest of the matrix. Maximal paths: acb.4.2. Vertices 1.2. In the adjacency matrix. then there remains an incident edge at v on which to extend the trail. which forbids connected nontrivial graphs that are not cycles.4. c) The closed trails in K 4 that are not cycles are single vertices.5.9. The non-coprimality graph with vertex set {1. Effect on the adjacency and incidence matrices of deleting a vertex or edge.14. Statements about Eulerian graphs. Consider the vertex ordering v1 . 1.2. every cycle has even length. .2. If v is a vertex in a connected graph G .2. A vertex v must be in the same partite set as u if there is a u . 1. Proof 2. and hence dG (v) ≥ 2. Cycles. We can count the edges by summing the degrees of the vertices in one partite set.7. Deleting edge vi v j simply deletes the corresponding column of the incidence matrix. and a path can only leave a component of G − v via v .15. Assume that the graph has no loops. v walk of even length. then k ≥ 2. If v has even degree. abcd . a) K 4 has a walk that is not a trail. d) If a maximal trail in a graph is not closed.10. traverse a triangle and then one additional edge. . The remainder induce a single component. The paw. Hence v has a neighbor in each component. If an endpoint v is different from the other endpoint. ad (two are maximum independent sets). A closed trail has even vertex degrees. Every vertex has even degree. in K 4 this requires degrees 2 or 0. then there is a bipartition in which u and v are in the same partite set and another in which they are in opposite partite sets.19 Chapter 1: Fundamental Concepts Section 1. It vertices have degrees m and n (if both are nonzero). . A bipartite graph has a unique bipartition (except for interchanging the two partite sets) if and only if it is connected.8. which are all even if and only if m and n are both even. The graph is connected. If u and v are vertices in distinct components. then its endpoints have odd degree. No cut-vertex has degree 1. the vertices in one component of G − v must have paths in G to every other component of G − v . The Petersen graph does have a decomposition into ﬁve paths. Every Eulerian graph has even vertex degrees and decomposes into cycles. By convention. Deleting a vertex vi eliminates the i th row of the incidence matrix.

If P and Q have no common internal vertices. proving the statement for all pairs of vertices. If this is not a path. arriving from v on edge e. v -path. v -walk in W than W . If there is no vertex repetition other than ﬁrst vertex = last vertex. We can apply the induction hypothesis to the pair that is a pair of distinct paths joining the same endpoints. v ∈ V ( Q ). v -paths.14. v -walk W contained in W . which is a cycle. the edges of the u . Proof 2 (induction). Let P and Q be distinct u . W has an edge repeating twice in succession. If u = v . Proof 1 (induction on the length l of W ). let y be the last vertex that belongs to Q . v -subpath of P . and Trails 22 1. We use ordinary induction on the length l of the walk. x .or z . let this new trail be T and repeat the iterative step. then edges incident to v that belong to the same cycle cannot appear consecutively on an Eulerian circuit. v -walks with vertex lists u . Each u . If this new trail includes all of E (G ). Q are distinct.2. If u = v . then l ≥ 2. we may assume (by symmetry) that P has an edge e not belonging to Q . then following T . x . this can only happen upon a return to the original vertex v . v and u . since this subpath of Q contains no vertex of P between y and z . and together they form a cycle of length 2. The union of distinct u . 1. Proof 1 (extremality).2: Paths. then the rest of T . Proof 2. Since G − E (T ) is an even graph. v -path we construct all belong to W . z . the procedure must terminate. then their union is a cycle. Similarly deﬁne Q . Let G = P3 . Since it contains no cycle. and let W be a u . If w = v . the walk must take a step and return immediately on the same edge. y -subpath of Q to form a cycle. w -path P in W . u . v -walk W . v -path P . z -subpath of P combines with the y . then we have distinct edges consisting of u and v . with vertices u . Similarly. Within the portion of P after P traverses e. In each case. If P and Q have a common internal vertex w . then W traverses a cycle. by the induction hypothesis. Since P . v in order. Since a path in a simple graph is determined by its set of edges. Let W be the portion of W between the instances of such a repetition.2. This pair contains the edges of a cycle. v -path. If G consists of two edge-disjoint cycles sharing one common vertex v . Q distinct. If l = 2. If P and Q are u . then P contains a u . which in turn is contained in the union of P and Q . . Within the portion of P before P traverses e. The union of the edge sets of distinct u . A closed walk of length 1 is a loop. 1. we traverse T to reach a vertex v on T that is incident to an edge e not in T . The distinct u . the desired path has length 0. v paths.12. suppose l > 1. Initialize T to be a closed trail of length 0. the induction hypothesis guarantees a u . For the induction step. It can only terminate when an Eulerian circuit has been found. then it is an Eulerian circuit. x . Q are w. Q . v walk.21 Chapter 1: Fundamental Concepts Section 1. P . x . v -path.13. since W has none. which is a cycle if v = w or contains a nontrivial w. w -subpath of P and the w.2. this provides the basis. eventually the trail cannot be extended. Then P . The y . consider a shortest u . Since each successive trail is longer and G has ﬁnitely many edges. and the portion between the instances of one vertex can be removed to obtain a shorter u . which is forbidden. Otherwise. then some edge of W occurs twice in succession (once in each direction). Beginning from v along e. Let w be the ﬁrst repetition of a vertex along W . If W is a nontrivial closed walk that does not contain a cycle. Since W has length l − 1. then we extend P by the edge wv to obtain a u . v -paths contains a cycle. x . then let P . and this repetition also appears in W . then G need not have an Eulerian circuit in which e. Hence there is some other vertex repetition. The walk W is a shorter closed walk than W and contains no cycle. a single vertex. a) (induction). From the ﬁrst occurrence of w to the visit to v is a w. v -paths with total length less than l . Q distinct or P .2. Assume that G is a connected even graph. The vertices y . 1. the induction hypothesis is that walks of length less than l contain paths linking their endpoints. f that share a vertex. v do not contain a cycle in their union. then it has a repeated vertex. For the induction step. w -walk obtained by deleting wv from W . Thus we may assume l ≥ 2. we must have P . 1. and we stop. z exist. This u• w • •v b) (extremality) Given a u . f appear consecutively. traversing an arbitrary trail T not using edges of T . Cycles. If G is an Eulerian graph with edges e. Q are u . v -walks need not contain a cycle. Algorithm for Eulerian circuit. We use induction on the sum l of the lengths of the two paths. A u . completing a closed trail. By the induction hypothesis. v. because u . and let W be the u . If w = v and v is not on P . If w = v and v is on P . let z be the ﬁrst vertex that belongs to Q . for all vertex pairs simultaneously. suppose l > 2. v -walk of length l . If T is not all of G . v -walk contains a u .15. v -path. v -walk of length 1 is a u . let wv be the last edge of W . Induction step: l > 2. We convert the proof by extremality to an iterative algorithm. w -paths with total length less than l . Splice T into T by traversing T up to v . then P is the desired u . We are given l ≥ 1. Basis step: l = 2. v -path of length 1.2. P be the u .11.

Since the number of trips is odd. The “adjacent-transposition graph” G n on permutations of [n ] is connected.18. e.17. If edge e in walk W does not lie on a cycle consisting of edges in W . the vertices with i at the end induce a copy Hi of G n −1 .) Proof 2 (direct u . bn . Each vertex is a permutation of [n ].16. Let x and y be the endpoints of e. note that for each i ∈ [n ]. For k ≥ 1. v -path in G . . Therefore. This means that the walk can only return to e at the endpoint from which it most recently left e. move element 1 to the front by adjacent interchanges. By the transitivity of the connection relation. We show that every permutation has a path to the identity permutation I = 1. y . 1. this yields for all u . as in Lemma 1. e. x walk that does not contain e. (Actually. and Trails 24 completes a cycle with e unless in fact P is the path of length 1 with edge e. every trip through e is x . G n is (n − 1)-regular. then G n ∼ = K 1 and G is connected (we can also start with n = 2). Basis step (l = 1): The edge e in a closed walk of length 1 is a loop and thus a cycle. that subwalk contains the edges of a cycle through e. Thus the parity of the . For n > 1. . We can interchange the last two positions to obtain a neighbor in Hi . Since a closed walk ends where it starts (that is. Proof 3 (induction on n ). showing that G n is connected shows that it is Eulerian if and only if n is odd. Proof 1 (induction on the length of W .2. y . Every vertex of G is connected to a vertex of H by a path formed by moving element n to the end. v ∈ V (G ).7). v -path in H that exists by the induction hypothesis. By symmetry. Note that since every permutation of [n ] has n − 1 adjacent pairs that can be transposed. The shortest closed walk has length 1. (Actually. we construct at u . The basis is as in Proof 3. Changing two coordinates changes the number of 1s in the name of the vertex by zero or by ±2. Hence there is an edge from each Hi to Hn . the subgraph H induced by the vertices having n at the end is isomorphic to G n −1 . . choose two appearances of some vertex (other than the beginning and end of the walk).6. . . e.2. the two types cannot alternate. there is a u .23 Chapter 1: Fundamental Concepts Section 1. Proof 1 (path to ﬁxed vertex). . e is a cut-edge of the subgraph H consisting of the vertices and edges in W . assume that G n −1 is connected. then W is a cycle. .6). and this is contained in W . the walk is a u . beginning a walk from u .2: Paths. Next ﬁnd b2 among a2 . n . (The part of G 4 used in the induction step appears below. . If n = 1. move it to the front by adjacent transpositions. and so on. in which case e repeats immediately in opposite directions. v -path in G . . This requires the traversals of e to alternate directions along e. This splits the walk into two closed walks shorter than W . Proof 3 (contrapositive). It reaches v when all positions have been “corrected”. we thus have a path from u to a vertex u ∈ V ( H ). . y or y . To create a v.2. If there is a vertex repetition. v -path). Adding e to this path completes a cycle with e consisting of edges in W . a path from v to a vertex v ∈ V ( H ). The element • 2314 • 2341 • 2431 • 4231 Proof 4 (induction on n ). then move 2 forward to position 2. there are two components in the graph G k whose vertex set is the set of binary k -tuples and whose edge set consists of the pairs that differ in exactly two places. Cycles. . For n > 1. this builds a path. Also. we may assume that this is x . b1 appears in u as some ai . v -path. one of them uses e an odd number of times. and transitivity of the connection relation again completes the proof. By the induction hypothesis. one step at a time.) Note that since we always bring the desired element forward.2. x . As we traverse the walk.2. I -path. By Lemma 1. Hn has vertices with i in position n − 1 whenever 1 ≤ i ≤ n − 1. . plus Lemma 1. e. v -path. in the same component of H − e). For u . then by our characterization of cut-edges. Proof 2 (parity ﬁrst. the number of traversals of e by W must be even. an and move it to position 2. By the transitivity of the connection relation. 1. . Iterating this procedure brings the elements of v toward the front. Hence some two successive trips through e have the same direction. which contains a path to I . . If edge e appears an odd number of times in a closed walk W . an and v = b1 . By the induction hypothesis. each such subgraph is connected. in order. . Induction step (l > 1): If there is no vertex repetition. This completes the proof of the induction step. Since each step is in exactly one of these subwalks. The portion of the walk between these two trips through e is a y . Let u = a1 .) 4132 • 1432 • 1342 • 1324 • 3142 4312 • • • • 3124 3412 3214 • 3241 • 3421 • 4321 • • 4123 • 1423 • 1243 • 1234 2413 2134 • • • • 4213 2143 1.2. This builds a walk to I . x . v ∈ V (G ) a u . . we never disturb the position of the elements that were already moved to their desired positions. . then W contains the edges of a cycle through e. . In G n . . while following a walk. x -path (that does not contain e. and a u . . it contains a y .

the edges don’t change. A graph is connected if and only if for every partition of its vertices into two nonempty sets. choose u ∈ S and v ∈ T . s ). q such that p (l / k ) + q (n / k ) = 1. Thus for the partition of V (G ) into V ( H ) and V (G ) − V ( H ) there is no edge with endpoints in both sets. the graph G − v is connected. G k consists of two copies of G k −1 plus additional edges.2. Since a graph and its complement cannot both be disconnected. Let l = gcd(r. r. 1. let v be a cut-vertex in a self-complementary graph G . Adding y to S produces a larger set within the same equivalence class. When r = s = 2 and n is odd. The even piece in the 0-copy has 0 · · · 000. If v is a cut-vertex of a simple graph G . Let k = gcd(n . Since they differ in an even number of places. To show that G k has at most two components. Rewriting this as p l = k − q n means that if we use p of the paths that add l .2. then v is not a cut-vertex of G . Cycles. This is true because each change of a bit in obtaining one label from the other switches the parity. a vertex other than u must be in the same partite set of the spanning biclique of G − u as the vertices not in the same component as u in G − v . If S = V (G ). v -path P . Since each vertex of G − v is nonadjacent to all vertices in the other components of G − v . since these vertices all agree in the new place. there is an edge with endpoints x ∈ S and y ∈ / S . . Since there are integers (one positive and one negative) such that pa + qb = 1. 1. Proof 2 (algorithmic approach). Start with one vertex in S . We conclude that v has degree at most 1 in G . Let G be a connected graph. We use induction on k . By hypothesis. r. if G is disconnected. there cannot be an edges with one endpoint in V ( H ) and the other endpoint outside. Since k divides n . This procedure ends only when there are no more vertices outside S .22. By the induction hypothesis. which is adjacent to 0 · · · 001 in the even piece of the 1-copy.25 Chapter 1: Fundamental Concepts Section 1.20. there exist integers p . . Basis step (k = 1): G 1 has two components. each subgraph has two components. each an isolated vertex. there is an edge with endpoints in both sets. Since G is connected. 1. T . then its neighbor is a cut-vertex ( K 2 is not self-complementary). Thus the four pieces reduce to (at most) two components in G k . s ). Note: The text gives the vertex set incorrectly. which is adjacent to 0 · · · 011 in the odd piece of the 1-copy. While S does not include all vertices. let S be the set of all vertices that can be reached from x via paths. eventually accumulating all vertices. the congruence classes modulo n fall into congruence classes modulo k in a well-deﬁned way. then let H be a component of G . r. the simple graph G with vertex set Zn and edge set {i j : | j − i | ∈ {r. It sufﬁces to build a path from i to i + k . which makes at least k components. s ) components. consider the partition into S and V (G ) − S . Proof 2. Hence only v can be in the other partite set.21. then we will have moved from i to i + k (mod n ). Within a copy. P has an edge from S to T . we prove that any two vertices with the same parity lie on a path. G has a vertex of degree 1. Proof 1 (contrapositive). G has an edge with endpoints . in which case all of G is in the same equivalence class. moving p edges with difference +r and q edges with difference +s achieves a change of +l .19. so G has only one component. In particular. Necessity. Thus all vertices in a component lie in the same congruence class modulo k . The graph G − v has a spanning biclique. Given a vertex x ∈ V (G ). because there is no edge from an k -tuple with an even number of 1s to an k -tuple with an odd number of 1s. Since G is self-complementary. 1. Now. so G has a vertex of degree at most 1. Vk . . using the transitivity of the connection relation. Hence v is not a cut-vertex of G . As above. . then for some partition there is no edge across. and let a = r / l and b = s / l . and Trails 26 number of 1s remains the same along every edge. we can change two places at a time to travel from u to v along a path in G k . . We thus have a path from i to i + l . If there is a vertex of degree 1. k = gcd(l . G has a u . We show that if G is not connected. If u and v are vertices with the same parity. Proof 3 (extremality).2. Vk be the vertex sets of the components of G − v . For the converse. Induction step (k > 1): when k > 1. s }} has gcd(n . We grow a set of vertices that lie in the same equivalence class of the connection relation. and v has degree at least n − 2. After its last vertex in S . s ∈ N. there are several approaches. it is necessary to go up to n ≡ 0 to switch from odd to even. then they differ in an even number of places. Then G contains the complete multipartite graph with partite sets V1 . Given a partition of V (G ) into nonempty sets S . Proof 1. The two copies are obtained by appending 0 to all the vertex names in G k −1 or appending 1 to them all. . Sufﬁciency. n ). To show that there are only k components. Since this includes all vertices of G − v .2. . meaning a complete bipartite subgraph that contains all its vertices.2: Paths. For n . Let V1 . The odd piece in the 0-copy has 0 · · · 010. we show that all vertices with indices congruent to i (mod k ) lie in one component (for each i ). note that k ≥ 2. . for each i . All neighbors of vertex i differ from i by a multiple of k . where “parity” means parity of the number of 1s. In each. A self-complementary graph has a cut-vertex if and only if it has a vertex of degree 1. Since H is a maximal connected subgraph of G and the connection relation is transitive. This implies that G k has at least two components. also G must have a vertex u such that G − u has a spanning biclique.

This describes all changes in order except for the switch of ar and as itself. y of nonadjacent neighbors. Thus the induction hypothesis implies that G − v is bipartite. Thus the permutations with an even number of inversions form an independent set. y } induces P3 . one of which must contain at least half the vertices (though it need not be a maximum independent set). the argument applies to all subgraphs of a bipartite graph. Alternatively. and the condition is necessary. then combining bipartitions of the components of G − e so that the endpoints of e are in opposite sets produces a bipartition of G . etc. This contradicts the choice of S . this cycle has even length.23. Let G be such a graph. If G is not complete. If the neighborhood of v is not a clique. using any two sets covering V (G ). . This subgraph H has no independent set containing at least half its vertices. G contains an odd cycle H .27 Chapter 1: Fundamental Concepts Section 1. Let Q be a shortest x . G may have edges that belong to no induced subgraph isomorphic to P3 . • in G − e. w. x } induces P3 . If G is connected and not complete. it has length at least 2. v -path formed by extending an x . No pairs involving elements that are before r or after s have their order changed.2. Hence we can specify the bipartition X i . Thus for each such k the number of inversions changes twice and retains the same parity.2. 1. Every bipartite graph has a vertex partition into two independent sets. then v has a pair x . Considering several cases (common neighbor or not. then every vertex of G belongs to some induced subgraph isomorphic to P3 . there is some edge between a neighbor w of v and a vertex x that is not a neighbor of v . the transitivity of the connection relation implies that G is connected. since x is not a neighbor of v . For the converse. then it is a complete graph. Since there are paths from x to all other vertices. e • • • 1. with two edges. Inductive proof that every graph G with no odd cycles is bipartite. then G has two nonadjacent vertices. y -path. If e is a cut-edge. then edges from two components yield four vertices that induce a subgraph with two edges. then interchanging ar and as changes the order of ar and ak . Hence the bipartition X . Adding e completes a cycle with a u .2. so in fact S is all of V (G ). A graph G is bipartite if and only if for every subgraph H of G . v -path . w in both parts of the bipartition of G i . Induction step (e(G ) > 0): Discarding an edge e introduces no odd cycles. G k be the components of G − v . If v has neighbors u . and also it changes the order of ak and as . One can also use cases according to whether v is adjacent to all other vertices or not. We now have a bipartition of G by letting X = X i and Y = {v } ∪ ( Yi ). then let u and v be its endpoints. then the edges u v and vw and a shortest u . We claim that each transposition changes the parity of the number of inversions. then since G is not complete there is some vertex y outside the set S consisting of v and its neighbors. Proof 2 (induction on n (G )). 1. If r < k < s . The “transposition graph” on permutations of [n ] is bipartite. and the parity of the number of inversions changes. u path along the edge u v . Basis step (n (G ) = 1): A graph with one vertex and no odd cycles has no loop and hence no edge and is bipartite. and Trails 28 u ∈ S and v ∈ / S . Basis step (e(G ) = 0): Every graph with no edges is bipartite. because every set consisting of at least half the vertices in an odd cycle must have two consecutive vertices on the cycle.26. and therefore each edge in the graph joins vertices with opposite parity. j such that i < j and ai > a j (these are called inversions). .2. In the graph below. . w -path in G i form a cycle of odd length. Thus the total number of changes is odd. Y be a bipartition of G − e. a) If a connected simple graph G is not a complete graph. by hypothesis. Any three successive vertices on Q induce P3 . This forces u and v to be in opposite sets in the bipartition X . If e is not a cut-edge of G . as do those with an odd number of inversions.25. each has a bipartition. one can use proof by contradiction. Cycles. . Since every subgraph of a bipartite graph is bipartite. e lies in no such subgraph. and thus the graph is bipartite.2. Now there is an x .24. Induction step (n (G ) > 1): When we discard a vertex v . 1. Y of G − e is also a bipartition of G . with r < s . Yi of G i so that X i contains all neighbors of v in G i . If the neighborhood of v is a clique. Let v be a vertex of G . This is a bipartition. 1. By the characterization of bipartite graphs.27. suppose that G is not bipartite. there is an independent set containing at least half of the vertices of H .2: Paths. {x . Consider the transposition that interchanges the elements in position r and position s . If a simple graph with no isolated vertices has no induced subgraph with exactly two edges. b) When a connected simple graph G is not a complete graph. The two cases are similar to those above. The partite sets are determined by the parity of the number of pairs i . If G is disconnected. we introduce no odd cycles. v. Now the set {v. Proof 1 (induction on e(G )). Thus the induction hypothesis implies that G − e is bipartite. Let G 1 . and let X . Since G is connected. Y .) always yields an induced subgraph with two edges. then G has nonadjacent vertices x and y .

The 5-cycles (b. (a . h ) and (b. e. The two odd cycles in bold have only the central edge in common. this is the (i . When k = 1. then that edge must belong to all odd cycles in the graph. Another bipartite subgraph with 10 edges is obtained by deleting edges de and c f . then v. they are two subgraphs. j )th entry in the k th power of the adjacency matrix A is the number of (vi . G has no odd cycle and is bipartite. Since G has no C 3 .) It remains to show that we must delete at least two edges to obtain a bipartite subgraph.2. For 1 ≤ i ≤ k . which we write as ai(. every (vi . Assign vertex v the code (a1 . h } and Y = {a . By the deﬁnition of ma=1 ai . there are at most 2 k of them. x . Now S ∪ {x } is an independent set. and Trails 30 1. We can do this with (at most) one edge if and only if all the odd cycles have a common edge. if there is a single edge lying in all odd cycles. g } and Y = {a . since ar. a) The graph below has a unique largest bipartite subgraph. they lie in opposite partite sets in some G i . it must be bh . and ai = 1 if v ∈ Yi or v ∈ / X i ∪ Yi . otherwise. just as the Petersen graph has ten claws. Choose a vertex x .31. S ∪ {x }. a) In a simple graph G . the largest bipartite subgraph has 10 edges. z is an induced P4 . (The proof allows loops and multiple edges and applies without change for digraphs. Deleting edges bh and ag yields an X .32. .2. the diagonal entries of the matrix A r are all 0. the bipartition is X = {b. where ai = 0 if v ∈ X i . if G is bipartite. Yi . Cycles. a) Construction when n ≤ 2k . we must delete enough edges to break all odd cycles. Let S = V (G ) − N ( X ) − {x }. v j )-walks of length k by which vertex (k −1) n appears one step before v j yields r ar. N (x ) is independent. then G has no closed walks of odd length through any vertex. However. ai. j . the (i . b) A simple graph G with adjacency matrix A is bipartite if and only if. . By part (a). e. . Since every two vertices are adjacent and the edge joining them must be covered in the union. Given n ≤ 2k . d . Powers of the adjacency matrix. encode the vertices of K n as distinct binary k -tuples. where X i is the set of vertices whose codes have 0 in position i . j )th entry in A k . v j )-paths of length k that arrive via vr on the last step is ai(. When k > 1. each component has a maximal trail. in particular. . c. and at least two edges must be deleted. ak ). Hence all of S is adjacent to all of N (x ). d . a . c. vr )-walks of length k − 1 is entry (i . (Although these two subgraphs are isomorphic. x -path contains an induced P4 . so n ≤ 2k . c.30. When loops are present. g ) is another 5-cycle that does not contain this. We have proved that G is a biclique with bipartition N (x ). w. b) Upper bound. The number of (vi . and it will not be an Eulerian circuit unless the 1.2.i counts the closed walks of length r beginning at vi . v j )-walks of length k in G . since every closed odd walk contains an odd cycle.2. so no other edge belongs to all odd cycles. G k . . r ) in A k −1 . f . a walk is a list of edge traversals). c. f . Since the codes are binary k -tuples. Therefore the codes assigned to the vertices are distinct. note that there is no choice of “direction” on a loop. We use induction on k . and Yi is the set of vertices whose codes have 1 in position i . h }. Y -bigraph with X = {b. e. K n is the union of k bipartite graphs if and only if n ≤ 2k (without using induction). B • • A B • A • • B A • • B • A b) In the graph below.29 Chapter 1: Fundamental Concepts Section 1.kr−1) ar. f . 1. j counts the edges (walks of length 1) from i to j . j is the number of edges from vr to v j that can complete the walk. Let G i be the complete bipartite subgraph with bipartition X i . v j )-walk of length k has a unique vertex vr reached one step before the end at v j . Given that K n is a union of bipartite graphs G 1 . Deleting the central edge leaves a bipartite subgraph. since the indicated sets A and B are independent in that subgraph. g . not one. h . d . f . let X i . Every v ∈ S has a neighbor in N (x ). Ar i . . By the induction hypothesis. When an even graph has more than one component. since G has no C 3 . Counting the (vi . j . 1. for each odd integer r . “Every maximal trail in an even graph is an Eulerian circuit”— FALSE.2: Paths. the number of (vi . h ) have only the edge bh in common. then G has no odd cycle and hence no closed odd walk. Conversely. since a bipartite subgraph has no odd cycles. By the characterization of bipartite graphs. A connected simple graph not having P4 or C3 as an induced subgraph is a biclique.r trix multiplication. a shortest v. Therefore no edge lies in all odd cycles.2. If deleting one edge makes a graph bipartite. . Since every two vertex codes differ in some position. G 1 ∪ · · · ∪ G k = K n . g }. If v ∈ S is adjacent to w but not z in N (x ). If this is always 0. we deﬁne a code for each vertex. Therefore. obtained by deleting the central edge. Yi be a bipartition of G i .28.kr−1) . . b • • a d • • c f • • e h • • g 1. and it is not unique.29. .

The assumption of connectedness is necessary. This can end only by arriving at the initial vertex along the edge paired with the initial edge. then there are two trails with a common vertex. Cycles. 1. which contradicts the hypothesis that T is a maximal trail. and a maximal trail in the even graph G − E (T ) that starts at e can be inserted to enlarge T .) 1. the return to A does not end the trail. deleting an edge x y reduces k . then G is Eulerian or every vertex of G is odd. Thus changing the pairing at v combines these two trails into one trail and leaves the other trails unchanged. and we have one trail. This trail is an Eulerian circuit. and at v we have edges from two different trails. obtain an Eulerian circuit. The trail decomposition of G guaranteed by the induction hypothesis has one trail ending at x and no trail ending at y . Furthermore. and because the reversal of a circuit C is in the same class as C . (Distinct ways of making the choices yields a distinct pairing at some vertex. The proof of the corrected statement is essentially that of Theorem 1. The graph G = G − E ( P ) has 2k − 2 odd vertices. This leads us through another neighbor of x to y . When k > 1. then it is incident to a missing edge e. Any component not having odd vertices has an Eulerian circuit that contains a vertex of P . An equivalence class of circuits is characterized by the pairing of edges at each vertex corresponding to visits through that vertex. •c •d . there is no choice in assembling this trail. then we consider them equivalent. think of making choices while following a circuit. We have shown that if the number of trails in the decomposition exceeds one.) Given edges from trails A and B at v . If e(G ) = 1. since deleting E ( P ) changes degree parity only at the ends of P .33. then we can obtain a decomposition with fewer trails be changing the pairing. If G has no even vertex adjacent to an odd vertex. Now we cannot use the edge ya or the edge just used. Proof 1 (induction on k ). The graph below has 6 equivalence classes of Eulerian circuits. we have used the desired number of trails to partition E (G ). and we can add x y as a trail of length one to the decomposition of G − x y guaranteed by the induction hypothesis. If G has an even vertex x adjacent to an odd vertex y .35. If two Eulerian circuits follow the same circular arrangement of edges. Algorithm for Eulerian circuits. because it would produce two edge-disjoint 4-cycles instead of a single trail. y • a• b• • x Alternatively. so every edge appears in exactly one such trail. 1. so two choices remain. the pairing decomposes G into closed trails. The only restriction is that the pairings must yield a single closed trail.34.31 Chapter 1: Fundamental Concepts Section 1. For each of the three ways to make the initial choice. and delete the added edge.2. Given a pairing at one 4-valent vertex below. there is a forbidden pairing at the other. The other two choices are okay. there was a choice of two later. After x we can follow one of 3 choices. We now count the choices made to determine the circuit. change the pairing by taking a pair in A and a pair in B and switching them to make two pairs that pair an edge of A with an edge of B . this completes a closed trail. and it must end since the graph is ﬁnite. we splice it into P to avoid having an additional trail. The induction hypothesis applies to each component of G that has odd vertices. Repeating the argument produces a decomposition using one closed trail. This determines the rest of the circuit. When k = 1. Therefore. In this case. If a maximal trail T is not an Eulerian circuit. but rather the trail continues and follows B before returning to the original edge. At the point where the ﬁrst edge would be repeated.32. (A shortest path connecting vertices in two of the trails ﬁrst leaves the ﬁrst trail at some vertex v . let P be a path connecting two odd vertices. Proof 2 (induction on e(G )). since G is connected. In total. use it to exit. Start along some edge. a 4-valent vertex has three possible pairings. differing only in the starting edges or the direction. Add x y to the trail ending at x to obtain the desired decomposition of G . At each arrival at a vertex. we add an edge between the two odd vertices. The edges of a connected graph with 2k odd vertices can be partitioned into k trails if k > 0. then G = K 2 . so there are 3 · 2 = 6 ways to specify distinct classes of circuits. Now when A is followed from v . Let G be a connected even graph. and Trails 32 other components have no edges. stop. A 2-valent vertex has exactly one such pairing. we may follow a canonical representative of the class from a along ax .2. At each vertex partition the incident edges into pairs (each edge appears in a pair at each endpoint). The added hypothesis needed is that the graph is connected. Thus the answer is 3 · 2 = 6. then G = G − x y has the same number of odd vertices as G . If there is more than one trail in the decomposition.2: Paths. Because each circuit uses each edge. because the conclusion is not true for G = H1 + H2 when H1 has some odd vertices and H2 is Eulerian. there is an edge paired with the entering edge.2.2.

14 every edge is a cut-edge. Deleting all the edges thus produces at least n + 1 components. The number of non-paths in this list of trails is even. Cycles. Let x . two of these paths have lengths with the same parity. and R is a path of length at least m /2 + m /2 + l = m + l . Let x be the ﬁrst vertex of P that is in P . it has a shortest path R between V ( P ) and V ( Q ). y -path in P . y -path in P . then let w be the ﬁrst instance of a vertex repetition on T . Since R is a shortest path. The trails we wish to delete thus come in pairs. This deﬁnes an involution under which the ﬁxed points are the u . b) If v is a vertex of odd degree in a graph G . then P and P together contain a u . This changes the number of trails in the list by an even number. then G − v has n − 1 vertices and at least n − 1 edges. If our graph G has a vertex v with degree at most 1. so the sum is even. An endpoint v of a maximal path P has at least three neighbors on P . The union of these two paths is an even cycle. If e = u v . Since the maximum path length is m . w -path. v -trails that visit v only at the end. Thus r = r . so the sum is odd. For a trail T that has not yet reached v . Let c(e) denote the number of cycles containing e. We prove that the same conclusion holds whenever u and v are the only vertices of odd degree in a graph H . For larger graphs. The connection relation is transitive. v -paths in G − u v . then G has an odd number of u . we have l = 0. v -paths. The process ends when all trails in the list end at v . y such that 1) G − {x . then G has a vertex v of odd degree.2. then every vertex of G has degree at least 2. Every n -vertex graph with at least n edges contains a cycle. Let m be the common length of P and Q . we obtain another trail T in the list. For the induction step. suppose that n > 1.37.2.25 guarantees that G contains a cycle.39. there are an odd number of ways to extend T by one edge. Sufﬁciency: We observe the contrapositive. Let x be a endpoint of a longest path P in G . the total number of these trails is odd.2. Following P from u to x and then P from x to w yields a u . These paths share only their endpoints.2. Summing c(e) over edges incident to v counts each cycle through v exactly twice. since both contain v . Since G is connected.2. Necessity: By part (a). since no vertex of P before x belongs to P . Q . and P and Q have a common vertex. If P and Q are disjoint. Every extension of every trail from u in G eventually reaches v . v -paths in G − u v is odd. A connected graph with at least three vertices has two vertices x . 1. we show that there are an odd number of such trails starting with each edge e incident to u . Since the list always has odd size. The portion P of P from r to the farther endpoint has length at least m /2. z be three such neighbors of v in order on P . Let l be the length of R . so the number of these cycles is odd. 1. w path. Since there are an odd number of terms in the sum. Proof 1 (induction on n ). then there is one such trail. which is a cycle.40. If G has no cycle. Proof 2 (use of cut-edges). If G is not Eulerian.41. A graph with one vertex that has an edge has a loop. This is immediate if H has only the edge u v . and the union of P . a) If G is loopless and Eulerian and G = G − u v . and the edge v z followed by the z . in which w and v are the only vertices of odd degree. Otherwise. y -paths: the edge v y . At least one vertex of P is in P . and this remains true as edges are deleted. which is impossible. the ﬁrst instance of a vertex repetition is again w . then by Theorem 1. y } is connected and 2) x . The graph consisting of two edge-disjoint paths of length 2k sharing their midpoint is connected and hence shows that P and Q need not have a common edge. we apply the induction hypothesis to H − e. and Trails 34 1. then P and Q have a common vertex. We maintain a list of trails from u . Now Lemma 1. c) A nontrivial connected graph is Eulerian if and only if every edge belongs to an odd number of cycles. It sufﬁces to show that if P is a u . then some edge incident to v lies in an even number of cycles. then G has a cycle of even length. Alternative characterization of Eulerian graphs. By the pigeonhole principle. 1. Proof 2 (induction and stronger result). We replace T in the list by these extensions. when e = u w with w = v . If T is such a trail that is not a path. Proof 1 (exhaustive counting and parity). some edge incident to v lies in an even number of cycles. G − v contains a cycle. y are adjacent or have a common neighbor. If P and Q are two paths of maximum length in a connected graph G . so the union of any two is a cycle. For T . By traversing the edges between the ﬁrst two occurrences of w in the opposite order.2: Paths. regardless of whether they are adjacent. Consider three v. and let v be . and thus T = T . The portion Q of Q from r to the farther endpoint has length at least m /2. The number of choices for the ﬁrst edge is odd.38. If G has no vertex of degree at most 1.36. R has no internal vertices in P or Q . The cycles through u v in G correspond to the u . c(e) must be even for some e incident to v . then P and Q are disjoint.2. y . If G is a loopless graph and δ (G ) ≥ 3. By the induction hypothesis. w path. Let the endpoints of R be r ∈ V ( P ) and r ∈ V ( Q ). so there are an even number of them. 1. and this cycle appears also in G .33 Chapter 1: Fundamental Concepts Section 1. v -path and P is a v. the edge v x followed by the x . because a maximal trail ends only at a vertex of odd degree. • v • x • y • • z • 1.2. By part (b). the number of u .2.

let x . For each way to pick these vertices. The vertices two apart on a 4-cycle must differ in two coordinates. there is a single such path when e(G ) = 2. z }. Modeling each opposed pair of edges by a single unoriented edge yields a 3-regular graph with 9 vertices. 1. The assumption of connectedness is necessary. and is essentially just this counting argument. If we can partition E (G ) into smaller connected subgraphs of even size. For each vertex. there are k − 1 ways to pick a third vertex completing P3 . VERTEX DEGREES & COUNTING 1. The sending of a valentine can be represented as a directed edge from the sender to the receiver. One way to do this is to partition E (G ) into connected subgraphs of even size and use P . then u v belongs to at least k triangles. {z . since the conclusion does not hold for a graph having components with an odd number of edges. The edges of a connected simple graph with 2k edges can be partitioned into paths of length 2. This is the same as showing that u and v have at least k common neighbors. otherwise. w } induce C 4 . call it w . then {z . let P = (x .3. Each odd component contains at least one of {x . 1. and suppose | S ∩ T | = j .43. then the graph has x → y if and k k 1. If each student sends to and receives cards from the same people. so n ≥ | S | + |T | − j = n + k − j . since an edge deletion increases the number of components by at most one). The degree of a vertex in a component of G is the same as its degree in G . The product of these factors counts each P3 twice. Thus the total number of P3 ’s is 2 2 . Every vertex of G appears in S or T or none or both. v. since otherwise we could build a longer path. only if y → x .3. since Q k has no multiple edges. Hence it is possible to add one of x y to one odd component and yz to the other odd component to obtain a partition of G into smaller connected subgraphs. For k each vertex of Q k . x .3.3. In this case.2. v. This is impossible.3. If x has a nonneighbor y . y . For e(G ) > 2. Proof 1. and let G = G − {x y . If z ↔ w . y . Otherwise. some neighbor z of x is not adjacent to v .2.3: Vertex Degrees and Counting 36 its neighbor on P . There are 2k ways to pick the ﬁrst vertex. Note that P has at least three vertices. The graph below is isomorphic to Q 4 . since k k Q k is a simple k -regular graph.5. since every graph has an even number of vertices of odd degree. then those components are graphs with odd degree sum. let y = v . w } induce P4 . 0010 • 0110• 1010 • 1110 • 0011 • •0111 1111 • 1011 • 1.3. v. Hence we are ﬁnished unless G has two components of odd size (G cannot have more than three components. Consider a vertex x of maximum degree. a component cut off from P − x − v in G − x − v has at most one vertex. If the vertices of odd degree are in different components. let y = w . yz }.4. since we build it from each end.3. If d (u ) + d (v) = n + k for an edge u v in a simple graph on n vertices. the 3vertex path. Since d (v) ≤ d (x ). • 1000 • 1100 0100• • 0000 • 1101 • 1001 •0101 • 0001 1. w be the begining of a shortest path to y (w may equal y ).1. A graph having exactly two vertices of odd degree must contain a path from one to the other. In a class with nine students where each student sends valentine cards to three others. A connected simple graph having no 4-vertex induced subgraph that is a path or a cycle has a vertex adjacent to every other vertex.2. there are 2 ways to choose two distinct neighbors. k k −2 Q k has 2 2 copies of C 4 . We use induction on e(G ). The k -dimensional cube Q k has 2 2 copies of P3 . The vertex w must be adjacent to v . To specify a particular subgraph isomorphic to P3 . 1. Proof 2. Hence the vertices of a 4-cycle . We can alternatively choose the starting vertex and the next two. it is not possible that each student sends to and receives cards from the same people. Common neighbors are counted twice. Proof 1 (direct counting). then we can apply the induction hypothesis to each piece and combine the resulting decompositions. If G − x − v is connected. Their two common neighbors each differ from each in exactly one of these coordinates.42. x . (Almost every proof of this using induction or contradiction does not need it.) 1. there are k ways to pick a neighbor. Let S be the neighbors of u and T the neighbors of v . Thus the total number of them is 2k k (k − 1)/2.35 Chapter 1: Fundamental Concepts Section 1. It sufﬁces to label the vertices with the names of the vertices in Q 4 so that vertices are adjacent if and only if their labels differ in exactly one place. we can specify the middle vertex and its two neighbors. Hence j ≥ k . Thus x must have no nonneighbor. z ) be an arbitrary path of length two in G .

- Graph Theory With Applications
- 16
- Glossary of Order Theory
- BC0052-Fall Drive Assignments 2012
- Zheng-Xu He- Rigidity of infinite disk patterns
- Parallel Algorithm
- Graph Theory Problems in Full Bloom
- Thesis 4-26-09
- Sunbelt 2016 R Network Visualization Handout
- Syllabus KKR BTech CSE
- 3semassigjan09
- Undergraduate Thesis Slide - The Matrix-Tree Theorem and some related problems.
- ch7.ppt
- Casarotto
- Spectrum Scheduling using Markov Chain Model in Cognitive Radio Networks
- Network Operations
- j 011326672
- Nr 211502 Applied Systems Engineering
- On Theexistenceofinfinitely
- Survey
- Spanning Tree 1
- Upper Bounds for the Roman Domination Subdivision Number of a Graph
- Chapter 4
- database authorization study
- Robot Localization Simulator Thesis
- Ford
- Paper9
- Math Worksheet-Graphing of Functions
- Descrete Math Assighment
- Graphs 1

Skip carousel

- A NEW S-(a,d) ANTIMAGIC LABELING OF A CLASS OF CIRCULANT GRAPHS
- tmp3C86.tmp
- Discrete Mathematics MCQ'S
- tmpC906
- A New S-(a,d) Antimagic Labelling of a Class of Generalized Petersen Graphs
- Discrete Mathematics
- tmp20DC.tmp
- tmp30E5.tmp
- tmpD746
- Discrete Mathematics MCQ'S
- tmp4359.tmp
- A Survey
- tmp7F8
- tmp2EBB.tmp
- Context Aware Decision Support Mobile Application for Recommendation in the Domain of Hotels
- Analysis and Design Algorithem
- Tmp 1387
- Algorithm and Data Structure
- tmp9F1C
- A Survey on Subgraph Matching Algorithm for Graph Database
- tmpDE4.tmp
- tmp6B5B.tmp
- tmpE648
- tmp3C0A.tmp
- tmpBF3
- tmp18A.tmp
- tmpB0E6
- Network Bucket Testing

Sign up to vote on this title

UsefulNot usefulRead Free for 30 Days

Cancel anytime.

Close Dialog## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Close Dialog## This title now requires a credit

Use one of your book credits to continue reading from where you left off, or restart the preview.

Loading