Math 471 SOLUTIONS: Assignment #9 Fall 2014 CHAPTER 11. Problem 11. Label the four graphs from left to right as A, B, C, and D. For each graph A-C, label the vertices as 1,2,3,4, starting with 1 in the upper left corner and proceeding to label the subsequent vertices CLOCKWISE around the squares. Graphs A and B are NOT ISOMORPHIC, since A has one vertex of degree 3 while B has two vertices of degree 3. Graphs A and C are ISOMORPHIC, and the isomorphism sends 1 3, 2 4, 3 1, and 4 2. Graphs A and D are NOT ISOMORPHIC, since D has vertices of degree 4, while A does not. Graphs B and C are NOT ISOMORPHIC, since B has two vertices of degree 3 while C has only one such. Graphs B and D are NOT ISOMORPHIC, since B has only one vertex of degree 4 while D has two vertices of degree 4. Graphs C and D are NOT ISOMORPHIC since C has a vertex of degree 5 while D does not. Problem 12. Neither of the first two graphs contains a cycle of length 4, while the last one does. So the last graph cannot be isomorphic to either of the first two. However, the first two graphs are isomorphic. We re-draw them here with vertices labeled: E A 1 F 7 G J I D A 10 B9 C6 8 10 B 3 6 H 9 C One isomorphism that works is 2 5 4 D5 E7 F8 G3 H1 I4 J2 There are many other possibilities, of course. Problem 13. Suppose that there is a walk from vertex x to vertex y in a graph G. If the walk does not visit any vertex more than once, then it is already a path and we have a path from x to y. If, on the other hand, some vertex v along the walk is visited more than once, we remove all edges of the walk that occur between the first visit to v and the last visit to v. This will leave us with a shorter walk that visits vertex v only once. Repeat this process with any other vertices that are visited more than once by the walk. The result is a path from x to y. Problem 19. First note that if G’ is connected, then so is G, because G is obtained from G’ by adding edges and/or loops between the existing vertices. If there was already a walk between any two vertices in G’, then this walk still exists after additional edges and/or loops are added to G’ to get G. Conversely, deleting multiple edges and loops from the graph of G does not destroy the existence of a walk between any two given vertices. If there was a walk from vertex x to vertex y in G, then that walk is still available in G’, because the walk from vertex x to vertex y can be achieved without using loops and without using more than one edge connecting any two vertices of the graph. Note that if G is planar, then since G’ is obtained by deleting edges and/or loops from G, then G’ is obviously planar as well (removal of edges and loops from G does not create unavoidable crossings!). Conversely, given G’ planar (without any crossings), we can always add loops and/or multiple edges between vertices that are already connected without creating any crossings. For instance, we can make small, concentric loops around any vertex we want that do not interfere with other edges in the graph. Multiple edges can be added that are extremely close and parallel to already existing edges so that they also do not create crossings. Problem 20. We give a proof by contrapositive. Suppose that G is a graph of order n that is not connected. We will prove that the number of edges of G is at most (n-1)(n-2)/2. Since G is not connected, we can separate the vertex set V into two parts, X and Y, such that no vertex from X is connected to any vertex in Y. Suppose that |X| = k and |Y| = n-k for some positive integer k. The maximum number of edges connecting vertices in X is C(k,2) and the maximum number of edges connecting vertices in Y is C(n-k,2). Thus, the maximum number of edges that such a graph G could possess is C(k,2) + C(n-k,2) = 0.5k(k-1) + 0.5(n-k)(n-k-1) = 0.5[k2 – k + n2 – 2nk – n + k2 +k] = 0.5[2k2 + n2 – 2nk – n] = k2 – nk + C(n,2) = C(n,2) – k(n-k) ≤ C(n,2) – (n-1), since k(n-k) is smallest when k=1. But C(n,2)-(n-1) = (n-1)[n/2 – 1] = (n-1)(n-2)/2. This is what we needed to prove. To get a disconnected graph of order n with (n-1)(n-2)/2 edges, simply let one vertex be isolated (degree 0) and form the complete graph Kn-1 on the remaining n-1 vertices. Problem 21. If G is already connected, then adding new edges to G will preserve the connectedness of the graph, since one can still take a walk between any two vertices in the graph by following the edges of the previously existing walk in G. Therefore, G* is connected. Conversely, suppose that G* is connected. We wish to show that G is connected. Note that all vertices of G* have even degree. Therefore, by Theorem 11.2.2, G* has a closed Eulerian trail. The edge {x,y} is exactly one edge of the closed Eulerian trail, so if we remove this edge, we can still take a walk from x to y by walking around the Eulerian trail in the opposite direction (the long way around!). Therefore, removal of {x,y} has not prohibited us from walking between edges x and y, and thus, all vertices of G can still be joined along a walk. Therefore, G is connected. Problem 29. The first graph has four vertices of odd degree, so it has neither an open nor closed Eulerian trail. The second graph has all vertices of even degree, so it has a closed Eulerian trail (by Theorem 11.2.2). We reproduce the figure below, with vertices labeled: D A B C E F G I J H K One example of a closed Eulerian trail would be A B C H C G H K J K G B A D I D E I J G F J E F B E A. Problem 30. Recall that the degree of each vertex of Kn has degree n-1. For a graph to have a closed Eulerian trail, every vertex must have even degree. Thus, n-1 must be even. Hence, n must be odd. The only complete graph that has an open Eulerian trail is K2, since this is the only complete graph with exactly two vertices of odd degree (both of them have degree 1). Problem 34. Omitted.