2 Trees 2.1 TREES An acyclic graph is one that contains no cycles. A tree is a connected acyclic graph. The trees on six vertices are shown in figure 2.1. Theorem 2.1 In a tree, any two vertices are connected by a unique path. Proof By contradiction. Let G be a tree, and assume that there are two distinct (u, v)-paths P, and P, in G. Since P, ¥ Pz, there is an edge e = xy of P, that is not an edge of P2. Clearly the graph (P:U P2)—e is connected. It therefore contains an (x, y)-path P. But then P+e is a cycle in the. acyclic graph G, a contradiction 0 The converse of this theorem holds for graphs without loops (exercise 2.41.1). Observe that all the trees on six vertices (figure 2.1) have five edges. In general we have: Theorem 2.2 If G is a tree, then e=v—1. Proof By induction on v. When v= 1, G=K, and e= LAY Figure 2.1. The trees on six vertices26 Graph Theory with Applications Suppose the theorem true for all trees on fewer than v vertices, and let G be a tree on v=2 vertices. Let uv € E. Then G—uv contains no (u, v)-path, since uv is the unique (u, v)-path in G. Thus G—uv is disconnected and so (exercise 1,6.8a) w(G—uv)=2. The components G; and G2 of G—uv, being acyclic, are trees. Moreover, each has fewer than v vertices. Therefore, by the induction hypothesis e(G)=v(G)-1 for i=1,2 Thus e(G)=e(G,)+ e(G2)+1 = »(G,)+ (G.)—1=»(G)-1 O Corollary 2.2 Every nontrivial tree has at least two vertices of degree one. Proof Let G be a nontrivial tree. Then d(v)=1 forall veV Also, by theorems 1.1 and 2.2, we have Yao) =26 =2y-2 It now follows that d(v)=1 for at least two vertices v O Another, perhaps more illuminating, way of proving corollary 2.2 is to show that the origin and terminus of a longest path in a nontrivial tree both have degree one (see exercise 2.1.2). Exercises 2.1.1 Show that if any two vertices of a loopless graph G are connected by a unique path, then G is a tree. 2.1.2 Prove corollary 2.2 by showing that the origin and terminus of a longest path in a nontrivial tree both have degree one. 2.1.3 Prove corollary 2.2 by using exercise 1.7.2. 2.1.4 Show that every tree with exactly two vertices of degree one is a path. 2.1.5 Let G be a graph with y—1 edges. Show that the following three statements are equivalent: (a) G is connected; (b) G is acyclic; (c) G is a tree. 2.1.6 Show that if G is a tree with A>k, then G has at least k vertices of degree one. 2.1.7. An acyclic graph is also called a forest. Show that (a) each component of a forest is a tree; (b) G is a forest if and only if e=v—w.Trees 27 2.1.8 A centre of G is a vertex u such that max d(u, v) is as small as vev possible. Show that a tree has either exactly oné centre or two, adjacent, centres. 2.1.9 Show that if G is a forest with exactly 2k vertices of odd degree, then there are k edge-disjoint paths P,, P2,..., Px in G such that E(G)=E(P,)UE(P,)U...UE(P,). 2.1.10* Show that a sequence (d;, do, ..., d,) of positive integers is a degree sequence of a tree if and only if Y d\=2(v-1). a Loy 2.1.11 Let T be an arbitrary tree on k+1 vertices. Show that if G is simple and 8=k then G has a subgraph isomorphic to T. 2.1.12 A saturated hydrocarbon is a molecule C,H, in which every carbon atom has four bonds, every hydrogen atom has one bond, and no sequence of bonds forms a cycle. Show that, for every positive integer m, C,H, can exist only if n =2m+2. 2.2 CUT EDGES AND BONDS A cut edge of G is an edge e such that w(G—e)>(G). The graph of figure 2.2 has the three cut edges indicated. Theorem 2.3 An edge e of G is a cut edge of G if and only if e is contained in no cycle of G. Proof Let e be a cut edge of G. Since w(G—e)>w(G), there exist vertices u and v of G that are connected in G but not in G—e. There is therefore some (u, v)-path P in G which, necessarily, traverses e. Suppose that x and y are the ends of e, and that x precedes y on P. In G—e, u is connected to x by a section of P and y is connected to v by a section of P. If e were in a cycle C, x and y would be connected in G—e by the path C—e. Thus, u and v would be connected in G—e, a contradiction. Figure 2.2. The cut edges of a graph28 Graph Theory with Applications Conversely, suppose that e = xy is not a cut edge of G; thus, o(G-e)= (G). Since there is an (x, y)-path (namely xy) in G, x and y are in the same component of G. It follows that x and y are in the same component of G-—e, and hence that there is an (x, y)-path P in G—e. But then e is in the cycle P+e of G O Theorem 2.4 A connected graph is a tree if and only if every edge is a cut edge. Proof Let G be a tree and let e be an edge of G. Since G is acyclic, e is contained in no cycle of G and is therefore, by theorem 2.3, a cut edge of G. Conversely, suppose that G is connected but is not a tree. Then G contains a cycle C. By theorem 2.3, no edge of C can be a cut edge of G O A spanning tree of G is a spanning subgraph of G that is a tree. Corollary 2.4.1 Every connected graph contains a spanning tree. Proof Let G be connected and let T be a minimal connected spanning subgraph of G. By definition w(T) = 1 and w(T—e)>1 for each edge e of T. It follows that each edge of T is a cut edge and therefore, by theorem 2.4, that T, being connected, is a tree O Figure 2.3 depicts a connected graph and one of its spanning trees. Corollary 2.4.2 If G is connected, then e=v—1 Proof Let G be connected. By corollary 2.4.1, G contains a spanning tree T. Therefore e(G)=e(T)=r(T)-1=(G)-1 O Figure 2.3. A spanning tree in a connected graphTrees 29 (a) (b) Figure 2.4. (a) An edge cut; (b) a bond Theorem 2.5 Let T be a spanning tree of a connected graph G and let e be an edge of G not in T. Then T+e contains a unique cycle. Proof Since T is acyclic, each cycle of T+e contains e. Moreover, C is a cycle of T +e if and only if C—e is a path in T connecting the’ends of e. By theorem 2.1, T has a unique such path; therefore T+e contains a unique cycle O For subsets $ and S’ of V, we denote by [S, S’] the set of edges with one end in S and the other in S’. An edge cut of G is a subset of E of the form [5, 5], where S is a nonempty proper subset of V and $= V\S. A minimal nonempty edge cut of G is called a bond; each cut edge e, for instance, gives rise to a bond {e}. If G is connected, then a bond B of G is a minimal subset of E such that G—B is disconnected. Figure 2.4 indicates an edge cut and a bond in a graph. If H is a subgraph of G, the complement of H in G, denoted by H(G), is the subgraph G—E(H). If G is connected, a subgraph of the form T, where T is a spanning tree, is called a cotree of G. Theorem 2.6 Let T be a spanning tree of a connected graph G, and let e be any edge of T. Then (i) the cotree T contains no bond of G; (ii) T+e contains a unique bond of G. Proof (i) Let B be a bond of G. Then G—B is disconnected, and so cannot contain the spanning tree T. Therefore B is not contained in T. (ii) Denote by $ the vertex set of one of the two components of T-e. The edge cut B =[S, 5] is clearly a bond of G, and is contained in T + e. Now, for any beB, T-e+b is a spanning tree of G. Therefore every bond of G contained in T+e must include every such element b. It follows that B is the only bond of G contained in T+e O The relationship between bonds and cotrees is analogous to that between cycles and spanning trees. Statement (i) of theorem 2.6 is the analogue for30 Graph Theory with Applications bonds of the simple fact that a spanning tree is acyclic, and (ii) is the analogue of theorem 2.5. This ‘duality’ between cycles and bonds will be further explored in chapter 12 (see also exercise 2.2.10). Exercises 2.2.1 Show that G is a forest if and only if every edge of G is a cut edge. 2.2.2 Let G be connected and let e € E. Show that (a) ¢ is in every spanning tree of G if and only if ¢ is a cut edge of (b) ¢ is in no spanning tree of G if and only if e is a loop of G. 2.2.3 Show that if G is loopless and has exactly one spanning tree T, then G=T. 2.2.4 Let F be a maximal forest of G. Show that (a) for every component H of G, FOH is a spanning tree of H; (b) e(F)=v(G)—(G). 2.2.5 Show that G contains at least e—v+w distinct cycles. 2.2.6 Show that (a) if each degree in G is even, then G has no cut edge; (0) if G is a k-regular bipartite graph with k =2, then G has no cut edge. 2.2.7 Find the number of nonisomorphic spanning trees in the following graphs: 2.2.8 Let G be connected and let_S be a nonempty proper subset of V. Show that the edge cut [S, S] is a bond of G if and only if both G[S] and G[S] are connected. 2.2.9 Show that every edge cut is a disjoint union of bonds. 2.2.10 Let B: and Bz be bonds and let C; and C; be cycles (regarded asTrees 31 sets of edges) in a graph. Show that (a) B, AB, is a disjoint union of bonds; (b) C. AC, is a disjoint union of cycles, where A denotes symmetric difference; (c) for any edge e, (B,UB.)\{e} contains a bond; (d) for any edge e, (C:UC:)\{e} contains a cycle. 2.2.11 Show that if a graph G contains k edge-disjoint spanning trees then, for each partition (Vi, V2,..., Va) of V, the number of edges which have ends in different parts of the partition is at least k(n-1). (Tutte, 1961 and Nash-Williams, 1961 have shown that this necessary condition for G to contain k edge-disjoint spanning trees is also sufficient.) 2.2.12* Let S be an n-element set, and let of ={Ai, A2,..., As} be a family of n distinct subsets of S. Show that there is an element x € S such that the sets A, U{x}, A2U{x},..., AsU{x} are all distinct. 2.3. CUT VERTICES A vertex v of G is a cut vertex if E can be partitioned into two nonempty subsets E, and E, such that G[E,] and G[E,] have just the vertex v in common. If G is loopless and nontrivial, then v is a cut vertex of G if and only if o(G—v)>w(G). The graph of figure 2.5 has the five cut vertices indicated. Theorem 2.7 A vertex v of a tree G is a cut vertex of G if and only if d(v)>1. Proof If d(v)=0, G=K, and, clearly, v is not a cut vertex. Figure 2.5. The cut vertices of a graph32 Graph Theory with Applications If d(v)=1, G—v is an acyclic graph with v(G—v)—1 edges, and thus (exercise 2.1.5) a tree. Hence w(G—v)=1=(G), and v is not a cut vertex of G. If d(v)>1, there are distinct vertices u and w adjacent to v. The path uvw is a (u, w)-path in G. By theorem 2.1 uow is the unique (u, w)-path in G. It follows that there is no (u, w)-path in G—, and therefore that w(G—v)> 1=«(G). Thus v is a cut vertex of G 0 Corollary 2.7 Every nontrivial loopless connected graph has at least two vertices that are not cut vertices. Proof Let G be a nontrivial loopless connected graph. By corollary 2.4.1, G contains a spanning tree T. By corollary 2.2 and theorem 2.7, T has at least two vertices that are not cut vertices. Let v be any such vertex. Then o(T-v)=1 Since T is a spanning subgraph of G, T—v is a spanning subgraph of G—v and therefore o(G-0v)so(T-v) It follows that w(G—v) = 1, and hence that v is not a cut vertex of G. Since there are at least two such vertices v, the proof is complete 0 Exercises 2.3.1 Let G be connected with v=3. Show that (a) if G has a cut edge, then G has a vertex v such that w(G—v)> «(G); (b) the converse of (a) is not necessarily true. 2.3.2 Show that a simple connected graph that has exactly two vertices which are not cut vertices is a path. 2.4 CAYLEY'S FORMULA There is a simple and elegant recursive formula for the number of spanning trees in a graph. It involves the operation of contraction of an edge, which we now introduce. An edge e of G is said to be contracted if it is deleted and its ends are identified; the resulting graph is denoted by G - e. Figure 2.6 illustrates the effect of contracting an edge. It is clear that if e is a link of G, then v(G-e)=v(G)-1 — e(G-e)=e(G)-1 and w(G-e)=(G) Therefore, if T is a tree, so too is T-e. We denote the number of spanning trees of G by 1(G).Trees 33 y e Ye ea] lee Ma 63 ¥3 G Figure 2.6. Contraction of an edge Theorem 2.8 If e is a link of G, then r(G)=7(G—e)+7(G-e). Proof Since every spanning tree of G that does not contain ¢ is also a spanning tree of G—e, and conversely, 7(G—e) is the number of spanning trees of G that do not contain e. Now to each spanning tree T of G that contains e, there corresponds a spanning tree T-e of G-e. This correspondence is clearly a bijection (see figure 2.7). Therefore +(G -e) is precisely the number of spanning trees of G that contain e. It follows that 7(G)=7(G—e)+7(G-e) O Figure 2.8 illustrates the recursive calculation of 7(G) by means of theorem 2.8; the number of spanning trees in a graph is represented symbolically by the graph itself. Although theorem 2.8 provides a method of calculating the number of spanning trees in a graph, this method is not suitable for large graphs. Fortunately, and rather surprisingly, there is a closed formula for T(G) which expresses 1(G) as a determinant; we shall present this result in chapter 12. In the special case when G is complete, a simple formula for +(G) was discovered by Cayley (1889). The proof we give is due to Priifer (1918). Figure 2.7NTT) (LIA Wo) [IZ 2) I) (oz) “LILA 0) Td 02Trees 35 Theorem 2.9 +(K,)=n"?. Proof Let the vertex set of K, be N={I1,2,...,n}. We note that n"”? is the number of sequences of length n—2 that can be formed from N. Thus, to prove the theorem, it suffices to establish a one-one correspondence between the set of spanning trees of K, and the set of such sequences. With each spanning tree T of K,, we associate a unique sequence (th, tay. +5 ta-2) as follows. Regarding N as an ordered set, let s, be the first vertex of degree one in T; the vertex adjacent to s, is taken as t,. We now delete s; from T, denote by sz the first vertex of degree one in T—s:, and take the vertex adjacent to s2 as t2. This operation is repeated until t,-2 has been defined and a tree with just two vertices remains; the tree in figure 2.9, for instance, gives rise to the sequence (4, 3, 5, 3, 4, 5). It can be seen that different spanning trees of K, determine difference sequences. <— (4,3,5,3,4,5) Figure 2.9 The reverse procedure is equally straightforward. Observe, first, that any vertex v of T occurs dr(v)—1 times in (th, t2,..., ta-2). Thus the vertices of degree one in T are precisely those that do not appear in this sequence. To reconstruct T from (th, t2,..., t-2), we therefore proceed as follows. Let s be the first vertex of N not in (t1, f,..., tn-2); join s, to t,. Next, let s2 be the first vertex of N\{s;} not in (t2, .. 2), and join s2 to f2. Continue in this way until the n—2 edges siti, Sota, ..., Sn-2te-2 have been determined. T is now obtained by adding the edge joining the two remaining vertices of N\{s1, S2,---5S0-2}. It is easily verified that different sequences give rise to different spanning trees of K,. We have thus established the desired one- one correspondence (0 Note that n"”? is not the number of nonisomorphic spanning trees of Ka, but the number of distinct spanning trees of K,; there are just six nonisomorphic spanning trees of Ke (see figure 2.1), whereas there are 6* = 1296 distinct spanning trees of Ke.36 Graph Theory with Applications Exercises 2.4.1 Using the recursion formula of theorem 2.8, evaluate the number of spanning trees in Kas. 2.4.2" A wheel is a graph obtained from a cycle by adding a new vertex and edges joining it to all the vertices of the cycle; the new edges are called the spokes of the wheel. Obtain an expression for the number of spanning trees in a wheel with n spokes. 2.4.3. Draw all sixteen spanning trees of Ku. 2.4.4 Show that if e is an edge of K,, then t(K,—e)=(n—2)n"°. 2.4.5 (a) Let H be a graph in which every two adjacent vertices are joined by k edges and let G be the underlying simple graph of H. Show that 7(H)=k"'r(G). (b) Let H be the graph obtained from a graph G when each edge of G is replaced by a path of length k. Show that 7(H)= ke-"™4(G). (c) Deduce from (b) that t(K2,)=n2""". APPLICATIONS 2.5 THE CONNECTOR PROBLEM A railway network connecting a number of towns is to be set up. Given the cost ¢; of constructing a direct link between towns »; and vj, design such a network to minimise the total cost of construction. This is known as the connector problem. By regarding each town as a vertex in a weighted graph with weights w(v,0;) = cy, it is clear that this problem is just that of finding, in a weighted graph G, a connected spanning subgraph of minimum weight. Moreover, since the weights represent costs, they are certainly non-negative, and we may therefore assume that such a minimum-weight spanning subgraph is a spanning tree T of G. A minimum-weight spanning tree of a weighted graph will be called an optimal tree; the spanning tree indicated in the weighted graph of figure 2.10 is an optimal tree (exercise 2.5.1). We shall now present a good algorithm for finding an optimal tree in a nontrivial weighted connected graph, thereby solving the connector problem. Consider, first, the case when each weight w(e)=1. An optimal tree is then a spanning tree with as few edges as possible. Since each spanning tree of a graph has the same number of edges (theorem 2.2), in this special case we merely need to construct some spanning tree of the graph. A simpleTrees 37 1 Figure 2.10. An optimal tree in a weighted graph inductive algorithm for finding such a tree is the following: 1. Choose a link e:. 2. If edges e1,@2,...,¢, have been chosen, then choose ei: from E\{e1, e2,...,¢} in such a way that G[{es, €2,..., ei:}] is acyclic. 3. Stop when step 2 cannot be implemented further. This algorithm works because a maximal acyclic subgraph of a connected graph is necessarily a spanning tree. It was extended by Kruskal (1956) to solve the general problem; his algorithm is valid for arbitrary real weights. Kruskal’s Algorithm 1. Choose a link e: such that w(e,) is as small as possible. 2. If edges 1, €2,..., have been chosen, then choose an edge ei: from E\{e,, e2,..., ei} in such a way that (i) Ger, e2,..., eses}] is acyclic; (ii) w(e+1) is as small as possible subject to (i). 3. Stop when step,2 cannot be implemented further. As an example, consider the table of airline distances in miles between six of the largest cities in the world, London, Mexico City, New York, Paris, Peking and Tokyo: MC NY Pa Pe T 5558 3469 214 5074 5959 — 2090 5725 7753 7035 2090 — 3636 6844 6757 5725 3636 — 5120 6053 7753 6844 5120 - 1307 7035 6757 6053 1307 aeyTrees 39 This table determines a weighted complete graph with vertices L, MC, NY, Pa, Pe and T. The construction of an optimal tree in this graph is shown in figure 2.11 (where, for convenience, distances are given in hundreds of miles). Kruskal’s algorithm clearly produces a spanning tree (for the same reason that the simpler algorithm above does). The following theorem ensures that such a tree will always be optimal. Theorem 2.10 Any spanning tree T*= G[fe,, e2,..., €-s}] constructed by Kruskal’s algorithm is an optimal tree. Proof By contradiction. For any spanning tree T of G other than T*, denote by f(T) the smallest value of i such that e, is not in T. Now assume that T* is not an optimal tree, and let T be an optimal tree such that f(T) is as large as possible. Suppose that f(T) = k; this means that ¢1, e2,..., €x-1 are in both T and T*, but that e, is not in T. By theorem 2.5, T+ e, contains a unique cycle C. Let ek be an edge of C that is in T but not in T*. By theorem 2.3, e; is not a cut edge of T+e,. Hence T’=(T + e,)—et is a connected graph with v—1 edges, and therefore (exercise 2.1.5) is another spanning tree of G. Clearly w(T’) = w(T) + w(ex)— (ex) (2.1) Now, in Kruskal’s algorithm, ex was chosen as an edge with the smallest weight such that G[fe1, e2,..., ex}] was acyclic. Since G[{ei,e2,..-,e-1, eM] is a subgraph of T, it is also acyclic, We conclude that w(ei) = wer) (2.2) Combining (2.1) and (2.2) we have w(T)k = f(T) contradicting the choice of T. Therefore T = T*, and T* is indeed an optimal tree 0 A flow diagram for Kruskal’s algorithm is shown in figure 2.12. The edges are first sorted in order of increasing weight (box 1); this takes about e log & computations (see Knuth, 1973). Box 2 just checks to see how many edges have been chosen. (S is the set of edges already chosen and i is their number.) When i=v—1, S={e1, 2, ..., &-1} is the edge set of an optimal tree T* of G. In box 3, to check if G[S U{a}] is acyclic, one must ascertain whether the ends of a; are in different components of the forest G[S] or not. This can be achieved in the following way. The vertices are labelled so that, at any stage, two vertices belong to the same component of G[S] if and only40 Graph Theory with Applications Sort edges in order of increasing weight 24,025 ++ +5 Og Sule. iste itty Figure 2.12. Kruskal’s algorithm if they have the same label; initially, vertex v, is assigned the label |, 11. With this labelling scheme, G[S U{a}] is acyclic if and only if the ends of a; have different labels. If this is the case, a; is taken as ew; otherwise, a, is discarded and a,.1, the next candidate for ¢..,, is tested. Once +1 has been added to S, the vertices in the two components of G[S] that contain the ends of e;,, are relabelled with the smaller of their two labels. For each edge, one comparison suffices to check whether its ends have the same or different labels; this takes « computations. After edge e;., has been added to S, the relabelling of vertices takes at most v comparisons; hence, for all v— 1 edges €1,€2,...,€-1 we need v(v—1) computations. Kruskal’s algorithm is therefore a good algorithm. Exercises 2.5.1 Show, by applying Kruskal’s algorithm, that the tree indicated in figure 2.10 is indeed optimal.Trees 41 2.5.2 Adapt Kruskal’s algorithm to solve the connector problem with preas- signments: construct, at minimum cost, a network linking a number of towns, with the additional requirement that certain selected pairs of towns be directly linked. 2.5.3 Can Kruskal’s algorithm be adapted to find (a) a maximum-weight tree in a weighted connected graph? (b) a minimum-weight maximal forest in a weighted graph? If so, how? 2.5.4 Show that the following Kruskal-type algorithm does not necessarily yield a minimum-weight spanning path in a weighted complete graph: 1. Choose a link e, such that w(e,) is as small as possible. 2. If edges e1, €2,..., & have been chosen, then choose an edge e+: from E\{e:, é2,..., ei} in such a way that (i) Glfer, e2,..., es}] is a union of disjoint paths; (ii) w(eus) is as small as possible subject to (i). 3. Stop when step 2 cannot be implemented further. 2.5.5 The tree graph of a connected graph G is the graph whose vertices are the spanning trees T,, Ts,..., T, of G, with T, and T, joined if and only if they have exactly v—2 edges in common. Show that the tree graph of any connected graph is connected. REFERENCES Cayley, A. (1889). A theorem on trees. Quart, J. Math., 23, 376-78 Knuth, D. E. (1973). The Art of Computer Programming, vol. 3: Sorting and Searching, Addison-Wesley, Reading, Mass., p. 184 Kruskal, J. B. Jr. (1956). On the shortest spanning subtree of a graph and the traveling salesman problem. Proc. Amer. Math. Soc., 7, 48-50 Nash-Williams, C. St. J. A. (1961). Edge-disjoint spanning trees of finite graphs. J. London Math. Soc., 36, 445-50 Priifer, H. (1918), Neuer Beweis eines Satzes iiber Permutationen. Arch. Math. Phys., 27, 742-44 Tutte, W. T. (1961). On the problem of decomposing a graph into n connected factors. J. London Math. Soc., 36, 221-30