TCSLib

138 Directed Graphs

138.1 Directed Graphs

Definition 138.1 Edge connectivity \(\kappa '(G)\)
✓
#

Edge connectivity \(\kappa \)’(G).

We … define the edge connectivity \(\kappa \)’(G) of G to be the minimum k for which G has a k-edge cut. If G is trivial, \(\kappa \)’(G) is defined to be zero. … G is said to be k-edge-connected if \(\kappa \)’(G) \(\ge \) k.

Definition 138.2 The bridge Digraph \(\to \) Quiver: Quiver.{0} V has Hom : V \(\to \) V \(\to \) Prop,…
✓
#

The bridge Digraph \(\to \) Quiver: Quiver.{0} V has Hom : V \(\to \) V \(\to \) Prop, exactly Digraph.Adj.

Definition 138.3 A directed path: a Quiver.Path with no repeated vertex
✓
#

A directed path: a Quiver.Path with no repeated vertex.

A directed walk in D is a finite non-null sequence W = (v\(_{0}\), a\(_{1}\), v\(_{1}\), …, a_k, v_k), whose terms are alternately vertices and arcs, such that, for i = 1, 2, …, k, the arc a\(_{i}\) has head v\(_{i}\) and tail v_{i-1}. … A directed trail is a directed walk that is a trail; directed paths, directed cycles and directed tours are similarly defined.

Definition 138.4 A directed cycle: a positive-length closed walk whose vertices (bar the repeated endpoint) are distinct
✓
#

A directed cycle: a positive-length closed walk whose vertices (bar the repeated endpoint) are distinct.

The book defines it only by analogy — directed paths, directed cycles and directed tours are similarly defined — so the unfolding here (a closed directed walk of positive length whose vertices, apart from the coinciding endpoints, are distinct) follows §1.4’s cycle.

Definition 138.5 \(v\) is reachable from \(u\)
✓
#

v is reachable from u.

If there is a directed (u, v)-path in D, vertex v is said to be reachable from vertex u in D.

Definition 138.6 \(D\) is diconnected: every vertex reaches every other
✓
#

D is diconnected: every vertex reaches every other.

Two vertices are diconnected in D if each is reachable from the other. As in the case of connection in graphs, diconnection is an equivalence relation on the vertex set of D. The subdigraphs D[V\(_{1}\)], D[V\(_{2}\)], …, D[V_m] induced by the resulting partition (V\(_{1}\), V\(_{2}\), …, V_m) of V(D) are called the dicomponents of D. A digraph D is diconnected if it has exactly one dicomponent.

Definition 138.7 \(d^{-}(v)\), the indegree
✓
#

d\(^{-}\)(v), the indegree.

The indegree d\(^{-}\)_D(v) of a vertex v in D is the number of arcs with head v.

Definition 138.8 \(d^{+}(v)\), the outdegree
✓
#

d\(^{+}\)(v), the outdegree.

…the outdegree d\(^{+}\)_D(v) of v is the number of arcs with tail v.

Definition 138.9 \(\varepsilon \), the number of arcs
✓
#

\(\varepsilon \), the number of arcs.

Throughout this chapter, D will denote a digraph and G its underlying graph. This is a useful convention; it allows us, for example, to denote the vertex set of D by V (since V = V(G)), and the numbers of vertices and arcs in D by \(\nu \) and \(\varepsilon \), respectively.

Definition 138.10 \(\delta ^{-}\), the minimum indegree
✓
#

\(\delta \)\(^{-}\), the minimum indegree.

We denote the minimum and maximum indegrees and outdegrees in D by \(\delta \)\(^{-}\)(D), \(\Delta \)\(^{-}\)(D), \(\delta \)\(^{+}\)(D) and \(\Delta \)\(^{+}\)(D), respectively.

Definition 138.11 \(\delta ^{+}\), the minimum outdegree
✓
#

\(\delta \)\(^{+}\), the minimum outdegree.

The same sentence as for \(\delta \)\(^{-}\): We denote the minimum and maximum indegrees and outdegrees in D by \(\delta \)\(^{-}\)(D), \(\Delta \)\(^{-}\)(D), \(\delta \)\(^{+}\)(D) and \(\Delta \)\(^{+}\)(D), respectively.

Definition 138.12 The converse \(\breve{D}\): reverse every arc
✓
#

The converse \(\breve{D}\): reverse every arc.

The converse \(\breve{D}\) of D is the digraph obtained from D by reversing the orientation of each arc.

Definition 138.13 \(D\) is strict: loopless (Digraph already forbids parallel same-direction arcs)
✓
#

D is strict: loopless (Digraph already forbids parallel same-direction arcs).

A digraph is strict if it has no loops and no two arcs with the same ends have the same orientation.

Definition 138.14 \(D\) is an orientation of \(G\) — the chapter’s keystone gap
✓
#

D is an orientation of G — the chapter’s keystone gap.

With each digraph D we can associate a graph G on the same vertex set; corresponding to each arc of D there is an edge of G with the same ends. This graph is the underlying graph of D. Conversely, given any graph G, we can obtain a digraph from G by specifying, for each link, an order on its ends. Such a digraph is called an orientation of G.

Definition 138.15 \(D\) is a tournament: an orientation of the complete graph
✓
#

D is a tournament: an orientation of the complete graph.

An orientation of a complete graph is called a tournament. The tournaments on four vertices are shown in figure 10.4. Each can be regarded as indicating the results of games in a round-robin tournament between four players; for example, the first tournament in figure 10.4 shows that one player has won all three games and that the other three have each won one.

Definition 138.16 B&M’s \((S, T)\): arcs with tail in \(S\), head in \(T\)
✓
#

B&M’s (S, T): arcs with tail in S, head in T.

If S and T are subsets of V, we denote by (S, T) the set of arcs of D that have their tails in S and their heads in T.

Definition 138.17 \(D\) is \(k\)-arc-connected: every nonempty proper cut has \(\ge k\) outgoing arcs
✓
#

D is k-arc-connected: every nonempty proper cut has \(\ge \) k outgoing arcs.

A nontrivial digraph D is k-arc-connected if, for every nonempty proper subset S of V, |(S, \(\bar{S}\))| \(\ge \) k.

Definition 138.18 The induced subdigraph on \(S\) (kept on the same carrier to avoid subtype juggling)
✓
#

The induced subdigraph on S (kept on the same carrier to avoid subtype juggling).

A digraph D’ is a subdigraph of D if V(D’) \(\subseteq \) V(D), A(D’) \(\subseteq \) A(D) and \(\psi \)_{D’} is the restriction of \(\psi \)_D to A(D’). The terminology and notation for subdigraphs is similar to that used for subgraphs.

Definition 138.19 The list of arcs traversed by a directed walk
#

The list of arcs traversed by a directed walk.

Definition 138.20 A directed trail: no repeated arc

A directed trail: no repeated arc.

A directed trail is a directed walk that is a trail — i.e. one whose arcs are all distinct.

Definition 138.21 A directed Euler tour: a closed directed trail using every arc
✓

A directed Euler tour: a closed directed trail using every arc.

A directed Euler tour of D is a directed tour that traverses each arc of D exactly once.

Definition 138.22 \(D\) is unilateral: any two vertices are comparable by reachability
✓
#

D is unilateral: any two vertices are comparable by reachability.

A digraph D is unilateral if, for any two vertices u and v, either v is reachable from u or u is reachable from v.

Definition 138.23 The 0/1 adjacency matrix of \(D\)
✓
#

The 0/1 adjacency matrix of D.

Let v\(_{1}\), v\(_{2}\), …, v_\(\nu \) be the vertices of a digraph D. The adjacency matrix of D is the \(\nu \) \(\times \) \(\nu \) matrix A = [a_ij] in which a_ij is the number of arcs of D with tail v\(_{i}\) and head v\(_{j}\).

Definition 138.24 The condensation \(\hat{D}\): dicomponents contracted
✓
#

The condensation \(\hat{D}\): dicomponents contracted.

Let D\(_{1}\), D\(_{2}\), …, D_m be the dicomponents of D. The condensation \(\hat{D}\) of D is a directed graph with m vertices w\(_{1}\), w\(_{2}\), …, w_m; there is an arc in \(\hat{D}\) with tail w\(_{i}\) and head w\(_{j}\) if and only if there is an arc in D with tail in D\(_{i}\) and head in D\(_{j}\).

Definition 138.25 dirDist
✓
#

Directed distance d(u,v).

In a diconnected digraph D, the length of a shortest directed (u, v)-path is denoted by d_D(u, v) and is called the distance from u to v; the directed diameter of D is the maximum distance from any one vertex of D to any other.

Definition 138.26 dirDiameter
✓
#

Directed diameter.

…the directed diameter of D is the maximum distance from any one vertex of D to any other.

Definition 138.27 The associated digraph \(D(G)\): each edge becomes two opposite arcs. G.Adj is already symmetric,…
✓
#

The associated digraph D(G): each edge becomes two opposite arcs. G.Adj is already symmetric, so it is the arc relation.

The associated digraph D(G) of a graph G is the digraph obtained when each edge e of G is replaced by two oppositely oriented arcs with the same ends as e.

Definition 138.28 A primitive \(\mathbb {N}\)-matrix: some power is entrywise positive
✓
#

A primitive \(\mathbb {N}\)-matrix: some power is entrywise positive.

A real matrix R is called primitive if R\(^{k}\) > 0 for some k.

Definition 138.29 The de Bruijn digraph \(D_n\): vertices are \((n-1)\)-bit strings, arcs are left-shifts
✓
#

The de Bruijn digraph D_n: vertices are (n-1)-bit strings, arcs are left-shifts.

We define a digraph D_n as follows: the vertices of D_n are the (n-1)-digit binary numbers p\(_{1}\) p\(_{2}\) … p_{n-1} with p\(_{i}\) = 0 or 1. There is an arc with tail p\(_{1}\) p\(_{2}\) … p_{n-1} and head q\(_{1}\) q\(_{2}\) … q_{n-1} if and only if p_{i+1} = q\(_{i}\) for 1 \(\le \) i \(\le \) n-2; in other words, all arcs are of the form (p\(_{1}\) p\(_{2}\) … p_{n-1}, p\(_{2}\) p\(_{3}\) … p_n). In addition, each arc (p\(_{1}\) p\(_{2}\) … p_{n-1}, p\(_{2}\) p\(_{3}\) … p_n) of D_n is assigned the label p\(_{1}\) p\(_{2}\) … p_n.

Theorem 138.30 Counting the orientations of a finite simple graph
✓
#

Let \(G\) be a simple graph on a finite vertex set. Then the number of orientations of \(G\) is \(2^{\left\lvert E(G)\right\rvert }\), where \(\left\lvert E(G)\right\rvert \) denotes the number of edges of \(G\); that is, each edge may be directed in either of its two ways, and every such choice yields a distinct orientation.

Theorem 138.31 Degree sums count the arcs of a digraph
✓

Let \(D\) be a digraph on a finite vertex set \(V\), and write \(\varepsilon \) for its number of arcs. Then both the sum of the indegrees and the sum of the outdegrees over all vertices equal \(\varepsilon \):

\[ \sum _{v \in V} d^{-}(v) \; =\; \varepsilon \; =\; \sum _{v \in V} d^{+}(v). \]
Theorem 138.32 Acyclic digraphs have a source
✓

Let \(D\) be a digraph on a finite, nonempty vertex set that contains no directed cycle. Then \(D\) has a vertex \(v\) whose indegree vanishes, \(d^{-}(v) = 0\); that is, no arc of \(D\) has \(v\) as its head.

Theorem 138.33 Topological ordering of an acyclic digraph
✓

Let \(D\) be a digraph on a finite vertex set \(V\) with \(n = |V|\), and suppose \(D\) is acyclic: it contains no directed cycle, that is, no closed directed walk of positive length whose vertices, apart from the coinciding endpoints, are distinct. Then \(V\) admits a topological ordering: there is a bijection \(f \colon V \to \{ 0, 1, \dots , n-1\} \) such that \(f(u) \lt f(v)\) for every arc of \(D\) from \(u\) to \(v\).

Theorem 138.34 The converse is an involution
✓
#

Let \(D\) be a digraph on a vertex set \(V\). Reversing every arc twice recovers the original digraph, so the converse of the converse of \(D\) equals \(D\); that is, \(\breve{\breve{D}} = D\).

Theorem 138.35 Degrees swap under the converse digraph
✓

Let \(D\) be a digraph on a finite vertex set, and let \(\breve{D}\) be its converse, obtained by reversing the orientation of every arc. Then for each vertex \(v\) the indegree and outdegree interchange:

\[ d^{-}_{\breve{D}}(v) = d^{+}_{D}(v) \qquad \text{and}\qquad d^{+}_{\breve{D}}(v) = d^{-}_{D}(v), \]

where \(d^{-}\) counts the arcs with head \(v\) and \(d^{+}\) the arcs with tail \(v\).

Theorem 138.36 Reachability in the converse digraph
✓
#

Let \(D\) be a digraph with vertices \(u\) and \(v\), and let \(\breve{D}\) denote its converse, the digraph obtained by reversing the orientation of every arc. Then \(v\) is reachable from \(u\) in \(\breve{D}\) if and only if \(u\) is reachable from \(v\) in \(D\).

Theorem 138.37 Every finite acyclic digraph has a sink
✓

Let \(D\) be a digraph on a finite, nonempty vertex set \(V\). If \(D\) contains no directed cycle, then there is a vertex \(v \in V\) whose outdegree vanishes, \(d^{+}(v) = 0\).

Theorem 138.38 A directed path as long as the minimum degree
✓

Let \(D\) be a strict digraph — a loopless directed graph — on a finite, nonempty vertex set. Write \(\delta ^-(D)\) for the minimum indegree and \(\delta ^+(D)\) for the minimum outdegree, each taken over all vertices of \(D\). Then \(D\) contains a directed path — a directed walk in which no vertex is repeated — whose length (its number of arcs) is at least \(\max \{ \delta ^-(D),\, \delta ^+(D)\} \).

Theorem 138.39 A long directed cycle from large minimum degree
✓

Let \(D\) be a strict digraph on a finite, nonempty vertex set, and let \(k\) be a positive integer such that \(\max \{ \delta ^{-}(D),\delta ^{+}(D)\} =k\), where \(\delta ^{-}(D)\) and \(\delta ^{+}(D)\) denote the minimum indegree and minimum outdegree of \(D\). Then \(D\) contains a directed cycle of length at least \(k+1\).

Theorem 138.40 Powers of the adjacency matrix count directed walks
✓

Let \(D\) be a digraph on a finite vertex set \(V\), and let \(A\) be its \(0/1\) adjacency matrix over \(\mathbb {N}\), so \(A_{uv} = 1\) when there is an arc from \(u\) to \(v\) and \(A_{uv} = 0\) otherwise. Then for every \(k \in \mathbb {N}\) and all vertices \(u, v \in V\), the entry \((A^k)_{uv}\) equals the number of directed walks of length \(k\) from \(u\) to \(v\) in \(D\), that is, the number of sequences of \(k\) arrows \(u = w_0 \to w_1 \to \cdots \to w_k = v\).

Theorem 138.41 The condensation is acyclic
✓

Let \(D\) be a digraph on a finite vertex set, and let \(\hat{D}\) denote its condensation: the digraph whose vertices are the strongly connected components (dicomponents) of \(D\), with an arc from one component to another precisely when \(D\) has an arc whose tail lies in the first and whose head lies in the second. Then \(\hat{D}\) contains no directed cycle; that is, for every dicomponent \(c\) and every closed walk in \(\hat{D}\) from \(c\) to \(c\), that walk is not a directed cycle — it does not have positive length together with distinct vertices apart from its coinciding endpoints.

Theorem 138.42 Existence of a balanced orientation
✓

Let \(G\) be a finite simple graph. Then \(G\) admits an orientation \(D\) in which the in- and outdegrees at every vertex differ by at most one; that is, \(\left\lvert d^{+}(v) - d^{-}(v)\right\rvert \le 1\) for every vertex \(v\), where \(d^{+}(v)\) and \(d^{-}(v)\) denote the outdegree and indegree of \(v\) in \(D\).

Theorem 138.43 Gallai–Roy theorem: chromatic number forces a long directed path
✓

(Roy, 1967; Gallai, 1968.) Let \(D\) be a digraph on a finite, nonempty vertex set \(V\), and suppose the underlying simple graph of \(D\) has chromatic number \(k\). Then \(D\) contains a directed path — a directed walk \(v_0, v_1, \dots , v_\ell \) whose consecutive vertices are joined by arcs \(v_{i-1} \to v_i\) and whose vertices are pairwise distinct — of length at least \(k - 1\), that is, with \(\ell \ge k - 1\) arcs.

Theorem 138.44 Orientation with short directed paths for a k-colourable graph
✓

Let \(G\) be a simple graph on a finite vertex set that admits a proper colouring with \(k\) colours. Then \(G\) has an orientation \(D\) — a digraph whose underlying graph is \(G\) and in which no pair of vertices is joined by arcs in both directions — such that every directed path in \(D\) (a directed walk that repeats no vertex) has length at most \(k-1\), that is, passes through at most \(k\) vertices.

Theorem 138.45 Rédei’s theorem: every tournament has a directed Hamilton path
✓

Let \(D\) be a tournament on a finite, nonempty vertex set \(V\) — a directed graph with no loops in which, for every pair of distinct vertices \(u\) and \(v\), exactly one of the arcs \(u \to v\) and \(v \to u\) is present. Then \(D\) has a directed Hamilton path: there exist vertices \(u, v\) and a directed path from \(u\) to \(v\) that visits every vertex of \(V\) exactly once.

Theorem 138.46 Existence of a quasi-kernel (Chvátal–Lovász)
✓
#

(Chvátal and Lovász, 1974.) Let \(D\) be a loopless digraph on a finite vertex set \(V\); that is, no vertex has an arc to itself. Then \(D\) admits an independent set \(S \subseteq V\) — no two distinct vertices of \(S\) are joined by an arc of \(D\) in either direction — such that every vertex \(v \notin S\) is reachable from some vertex \(u \in S\) by a directed path of length one or two: either \(u \to v\) is an arc of \(D\), or there is a vertex \(w\) with arcs \(u \to w\) and \(w \to v\).

Theorem 138.47 Every finite tournament has a king
✓
#

Let \(D\) be a tournament on a finite, nonempty vertex set \(V\); that is, for each pair of distinct vertices exactly one of the two directed edges between them is present, and no vertex has an edge to itself. Write \(u \to v\) when the edge from \(u\) to \(v\) is present. Then \(D\) has a king: a vertex \(u\) such that every other vertex \(v\) is reached from \(u\) by a directed path of length at most two — that is, either \(u \to v\), or there is a vertex \(w\) with both \(u \to w\) and \(w \to v\).

Theorem 138.48 Reversing one arc to make a tournament diconnected
✓

Let \(D\) be a tournament on a finite, nonempty vertex set \(V\). Then either \(D\) is diconnected, or there is an arc \(x \to y\) of \(D\) whose reversal makes it so: deleting the arc \(x \to y\) and inserting the arc \(y \to x\), while leaving every other arc unchanged, yields a diconnected digraph.

Theorem 138.49 Unilateral digraphs and spanning directed walks
✓

Let \(D\) be a digraph on a finite, nonempty vertex set \(V\). Then \(D\) is unilateral if and only if there exist vertices \(u, v \in V\) and a directed walk from \(u\) to \(v\) — a finite sequence of edges, each traversed in the direction it points — whose vertices include every vertex of \(V\). In other words, \(D\) is unilateral exactly when it admits a spanning directed walk.

Theorem 138.50 Inserting a vertex into a non-extendable directed path in a tournament
✓

Let \(D\) be a tournament on a finite vertex set \(V\) — an orientation of the complete graph on \(V\). Let \(v_1, v_2, \dots , v_k\) be a directed path in \(D\): a nonempty list of distinct vertices such that \(v_j \to v_{j+1}\) is an arc of \(D\) for every \(1 \le j \lt k\). Suppose this path cannot be extended at either end, meaning that for every vertex \(w\) not among \(v_1, \dots , v_k\) there is no arc from \(w\) to \(v_1\) and no arc from \(v_k\) to \(w\). Then for every vertex \(v\) not on the path there is an index \(i\) with \(1 \le i \lt k\) such that \(v_i \to v\) and \(v \to v_{i+1}\) are both arcs of \(D\); that is, \(v\) can be inserted between two consecutive vertices of the path.

Theorem 138.51 Chvátal–Komlós monotone directed path theorem
✓
#

Let \(D\) be a digraph on a finite vertex set \(V\), let \(f : V \to \mathbb {R}\) be a real-valued function on the vertices, and let \(m\) and \(n\) be natural numbers. Write \(\chi (D)\) for the chromatic number of the underlying simple graph of \(D\) (in which two vertices are joined whenever there is an arc between them in either direction). If \(\chi (D) \gt mn\), then at least one of the following holds: either \(D\) contains a directed path \(u_0 \to u_1 \to \cdots \to u_m\) on \(m+1\) distinct vertices along which \(f\) is nondecreasing, \(f(u_0) \le f(u_1) \le \cdots \le f(u_m)\); or \(D\) contains a directed path \(v_0 \to v_1 \to \cdots \to v_n\) on \(n+1\) distinct vertices along which \(f\) is strictly decreasing, \(f(v_0) \gt f(v_1) \gt \cdots \gt f(v_n)\).

Theorem 138.52 The Erdős–Szekeres theorem on monotone subsequences
✓
#

Let \(m, n, N\) be natural numbers with \(mn \lt N\), and let \(g_0, g_1, \dots , g_{N-1}\) be a sequence of \(N\) distinct integers. Then the sequence has an increasing subsequence of length \(m\) — indices \(i_1 \lt i_2 \lt \dots \lt i_m\) with \(g_{i_1} \lt g_{i_2} \lt \dots \lt g_{i_m}\) — or a decreasing subsequence of length \(n\) — indices \(j_1 \lt j_2 \lt \dots \lt j_n\) with \(g_{j_1} \gt g_{j_2} \gt \dots \gt g_{j_n}\).

Theorem 138.53 Orientation with directed paths bounded by maximum degree
✓

Let \(G\) be a finite simple graph with maximum degree \(\Delta (G)\). Then \(G\) admits an orientation \(D\) — a digraph on the same vertex set whose underlying graph is \(G\) and in which no two vertices are joined by arcs in both directions — such that every directed path in \(D\), that is, every directed walk that repeats no vertex, has length at most \(\Delta (G)\), where the length of a path is the number of arcs it traverses.

Theorem 138.54 Moon’s theorem on vertex-pancyclicity of diconnected tournaments
✓

Let \(D\) be a diconnected tournament on a finite vertex set with at least three vertices, and let \(u\) be any vertex of \(D\). Then for every integer \(k\) with \(3 \le k \le \left\lvert V(D)\right\rvert \) there is a directed cycle of length \(k\) passing through \(u\).

Theorem 138.55 Ghouila–Houri theorem for directed Hamilton cycles
✓

Let \(D\) be a loopless finite digraph on a vertex set of size \(n \gt 2\), and suppose that every vertex \(v\) satisfies \(d^{-}(v) \ge n/2\) and \(d^{+}(v) \ge n/2\), where \(d^{-}(v)\) and \(d^{+}(v)\) denote the indegree and outdegree of \(v\). Then \(D\) contains a directed cycle of length \(n\); that is, a closed directed walk of positive length whose vertices, apart from the coinciding endpoints, are distinct, and which passes through every vertex of \(D\).

Theorem 138.56 Existence of a directed Euler tour
✓

Let \(D\) be a digraph on a finite, nonempty vertex set \(V\) whose underlying undirected graph is connected. Then \(D\) admits a directed Euler tour—a closed directed trail that traverses every arc of \(D\) exactly once—if and only if every vertex \(v \in V\) is balanced, meaning its outdegree and indegree agree, \(d^{+}(v) = d^{-}(v)\).

Theorem 138.57 Arc-disjoint directed paths from an excess to a deficit vertex

Let \(D\) be a finite digraph, let \(x \neq y\) be two of its vertices, and let \(l\) be a natural number. Suppose that \(d^{+}(x) - d^{-}(x) = l\) and \(d^{-}(y) - d^{+}(y) = l\), and that every other vertex \(v\) (that is, \(v \neq x\) and \(v \neq y\)) satisfies \(d^{+}(v) = d^{-}(v)\), where \(d^{+}\) and \(d^{-}\) denote outdegree and indegree. Then \(D\) contains \(l\) pairwise arc-disjoint directed paths from \(x\) to \(y\).

Theorem 138.58 Odd cycle forces an odd directed cycle in a diconnected digraph
✓

Let \(D\) be a diconnected digraph on a finite vertex set, so that every vertex is reachable from every other along directed paths. Suppose the underlying simple graph of \(D\), obtained by forgetting the orientation of each arc, contains a cycle of odd length. Then \(D\) contains a directed cycle of odd length.

Theorem 138.59 Diconnection equals 1-arc-connectedness
✓

Let \(D\) be a finite digraph whose vertex set has at least two vertices. Then \(D\) is diconnected — every vertex is reachable from every other along a directed path — if and only if \(D\) is \(1\)-arc-connected; that is, if and only if for every nonempty proper subset \(S\) of the vertex set there is at least one arc of \(D\) with tail in \(S\) and head outside \(S\).

Theorem 138.60 Arc-connectivity of the associated digraph equals edge connectivity
✓

Let \(G\) be a finite simple graph on a vertex set with at least two vertices, let \(D(G)\) be its associated digraph (each edge of \(G\) replaced by the two oppositely oriented arcs joining its ends), and let \(k\) be a natural number. Then \(D(G)\) is \(k\)-arc-connected if and only if \(k \le \kappa '(G)\), where \(\kappa '(G)\) denotes the edge connectivity of \(G\).

Theorem 138.61 Every vertex of the de Bruijn digraph has indegree and outdegree two
✓

Let \(n \ge 2\) be an integer, and let \(v\) be any vertex of the de Bruijn digraph \(D_n\). Then both the indegree and the outdegree of \(v\) equal \(2\); that is, \(d^{-}(v) = 2\) and \(d^{+}(v) = 2\).

Theorem 138.62 Connectivity of the de Bruijn digraph
✓
#

Fix an integer \(n \ge 2\), and let \(D_n\) be the de Bruijn digraph whose vertices are the binary strings of length \(n-1\) and whose arcs run from each string \(p_1 p_2 \cdots p_{n-1}\) to every string of the form \(p_2 p_3 \cdots p_{n-1} b\) with \(b \in \{ 0,1\} \). Form the underlying undirected simple graph on the same vertex set, in which two distinct strings are adjacent precisely when an arc of \(D_n\) joins one to the other in either direction. Then this graph is connected: its vertex set is nonempty and any two vertices are linked by a path.

Theorem 138.63 Existence of a directed Euler tour in the de Bruijn digraph

For every integer \(n \ge 2\), the de Bruijn digraph \(D_n\) admits a directed Euler tour: there exist a vertex \(u\) and a closed directed walk based at \(u\) that traverses each arc of \(D_n\) exactly once.

Theorem 138.64 Robbins’ theorem on diconnected orientations
✓

Let \(G\) be a finite simple graph on a nonempty vertex set, and suppose its edge connectivity satisfies \(\kappa '(G) \ge 2\). Then \(G\) has a diconnected orientation: one can assign a direction to each edge of \(G\), forming a digraph \(D\) whose underlying graph is \(G\), so that for every ordered pair of vertices \(u, v\) there is a directed path in \(D\) from \(u\) to \(v\).

Theorem 138.65 A \(k\)-arc-connected orientation of a \(2k\)-edge-connected Eulerian graph
✓

Let \(G\) be a finite graph on a nonempty vertex set that admits an Eulerian trail — a walk traversing every edge of \(G\) exactly once — and let \(k\) be a natural number with \(2k \le \kappa '(G)\). Then \(G\) has an orientation that is \(k\)-arc-connected.

Theorem 138.66 Positivity of a diameter-shifted power of a tournament’s adjacency matrix
✓

Let \(D\) be a diconnected tournament on a finite vertex set \(V\) with \(\nu = \left\lvert V\right\rvert \ge 5\), and let \(A\) be its \(\nu \times \nu \) adjacency matrix over \(\mathbb {N}\), so that \(A_{uv} = 1\) when there is an arc from \(u\) to \(v\) and \(A_{uv} = 0\) otherwise. Writing \(d\) for the directed diameter of \(D\), every entry of the matrix power \(A^{\, d+3}\) is strictly positive; that is, \(\bigl(A^{\, d+3}\bigr)_{ij} \gt 0\) for all \(i, j \in V\).

Theorem 138.67 Primitivity of a tournament’s adjacency matrix
✓

Let \(D\) be a tournament on a finite vertex set \(V\) (an orientation of the complete graph on \(V\)), and let \(A\) be its \(0/1\) adjacency matrix. Then \(A\) is primitive — that is, some power \(A^{k}\) has all entries positive — if and only if \(D\) is diconnected (every vertex is reachable from every other) and \(V\) has at least four vertices.