Arc-disjoint strong spanning subdigraphs in compositions and products of digraphs
Yuefang Sun, Gregory Gutin, Jiangdong Ai

TL;DR
This paper investigates conditions under which compositions and products of digraphs can be decomposed into two strong, arc-disjoint subdigraphs, providing new characterizations and sufficient conditions for such decompositions.
Contribution
It offers new sufficient conditions and characterizations for good decompositions in digraph compositions and products, especially when the base digraphs are strong and semicomplete.
Findings
Strong product of two digraphs always has a good decomposition.
Cartesian powers of strong digraphs with cycle covers have good decompositions.
Characterizations for compositions with strong semicomplete digraphs.
Abstract
A digraph has a good decomposition if has two disjoint sets and such that both and are strong. Let be a digraph with vertices and let be digraphs such that has vertices Then the composition is a digraph with vertex set and arc set For digraph compositions , we obtain sufficient conditions for to have a good decomposition and a characterization of with a good decomposition when is a strong semicomplete digraph and each is an arbitrary digraph with at least two vertices. For digraph products, we prove the following: (a) if is an integerβ¦
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Graph Theory Research Β· Nuclear Receptors and Signaling Β· graph theory and CDMA systems
Arc-disjoint strong spanning subdigraphs in compositions and products of digraphs
Yuefang Sun1,111Yuefang Sun was supported by National Natural Science Foundation of China (No. 11401389). Gregory Gutin2,222Corresponding author. Gregory Gutin was partially supported by Royal Society Wolfson Research Merit Award. Jiangdong Ai2
1 Department of Mathematics, Shaoxing University
Zhejiang 312000, P. R. China, [email protected]
2 Department of Computer Science
Royal Holloway, University of London
Egham, Surrey, TW20 0EX, UK
[email protected], [email protected]
Abstract
A digraph has a good decomposition if has two disjoint sets and such that both and are strong. Let be a digraph with vertices and let be digraphs such that has vertices Then the composition is a digraph with vertex set and arc set
[TABLE]
For digraph compositions , we obtain sufficient conditions for to have a good decomposition and a characterization of with a good decomposition when is a strong semicomplete digraph and each is an arbitrary digraph with at least two vertices.
For digraph products, we prove the following: (a) if is an integer and is a strong digraph which has a collection of arc-disjoint cycles covering all vertices, then the Cartesian product digraph (the th powers with respect to Cartesian product) has a good decomposition; (b) for any strong digraphs , the strong product has a good decomposition.
Keywords: strong spanning subdigraph; decomposition into strong spanning subdigraphs; semicomplete digraph; digraph composition; Cartesian product; strong product.
AMS subject classification (2010): 05C20, 05C70, 05C76, 05C85.
1 Introduction
We refer the readers to [1, 2, 6] for graph theoretical notation and terminology not given here. A digraph is strongly connected (or strong) if there exists a path from to and a path from to in for every pair of distinct vertices of . A digraph is -arc-strong if is strong for every subset of size at most .
An out-branching (respectively, in-branching ) in a digraph is a connected spanning subdigraph of in which each vertex has precisely one arc entering (leaving) it and has no arcs entering (leaving) it. The vertex is the root of (respectively, ). Edmonds [9] characterized digraphs with have arc-disjoint out-branchings rooted at a specified vertex Furthermore, there exists a polynomial algorithm for finding arc-disjoint out-branchings from a given root if they exist (see p. 346 of [1]). However, if we ask for the existence of a pair of arc-disjoint branchings , such that the first is an out-branching rooted at and the latter is an in-branching rooted at , then the problem becomes NP-complete (see Section 9.6 of [1]). In connection with this problem, Thomassen [12] posed the following conjecture: There exists an integer so that every -arc-strong digraph contains a pair of arc-disjoint in- and out-branchings.
Bang-Jensen and Yeo generalized the above conjecture as follows.333Every strong digraph has an out- and in-branching. A digraph has a good decomposition if has two disjoint sets and such that both and are strong [4].
Conjecture 1.1
[5]** There exists an integer so that every -arc-strong digraph contains a pair of arc-disjoint strong spanning subdigraphs.
For a general digraph , it is a hard problem to decide whether has a decomposition into two strong spanning subdigraphs.
Theorem 1.1
[5]** It is NP-complete to decide whether a digraph contains a pair of arc-disjoint strong spanning subdigraphs.
Clearly, every digraph with a good decomposition is 2-arc-strong. Bang-Jensen and Yeo characterized the semicomplete digraphs with a good decomposition.
Theorem 1.2
[5]** A 2-arc-strong semicomplete digraph has a pair of arc-disjoint strong spanning subdigraphs if and only if is not isomorphic to , where is obtained from the complete digraph with four vertices by deleting a cycle of length four. Furthermore, a good decomposition of can be obtained in polynomial time when it exists.
The following result extends Theorem 1.2 to locally semicomplete digraphs.
Theorem 1.3
[4]** A 2-arc-strong locally semicomplete digraph has a pair of arc-disjoint strong spanning subdigraphs if and only if is not the second power of an even cycle.
Let be a digraph with vertices and let be digraphs such that has vertices Then the composition is a digraph with vertex set and arc set
[TABLE]
In this paper, we continue research on good decompositions in classes of digraphs and consider digraph compositions and products.
In Section 2, for digraph compositions , we obtain sufficient conditions for to have a good decomposition (Theorem 2.2) and a characterization of with a good decomposition when is a strong semicomplete digraph and each is an arbitrary digraph with at least two vertices (Theorem 2.3). Remarkably, in Theorem 2.3 as in Theorem 1.2, there are only a finite number of exceptional digraphs, which for Theorem 2.3 is three. Thus, as Theorems 1.2 and 1.3, Theorem 2.3 confirms Conjecture 1.1 for a special class of digraphs.
In Section 3, for digraph products, we prove the following: (a) if is an integer and is a strong digraph which arcs can be partitioned into cycles, then the Cartesian product digraph (the th powers with respect to Cartesian product) has a good decomposition (Theorem 3.4); (b) for any strong digraphs , the strong product has a good decomposition (Theorem 3.7). Necessary definitions of the digraph products are given in Section 3.
Simple examinations of our constructive proofs show that all our decompositions can be found in polynomial time.
We conclude the paper in Section 4, where we pose a number of open problems.
2 Compositions of digraphs
Let denote with all arcs deleted, where and let
Compositions of digraphs is a useful concept in digraph theory, see e.g. [1]. In particular, they are used in the Bang-Jensen-Huang characterization of quasi-transitive digraphs and its structural and algorithmic applications for quasi-transitive digraphs and their extensions, see e.g. [1, 2, 8].
Let us start from a simple observation, which will be useful in the proofs of the theorems of this section.
Lemma 2.1
Let If an induced subdigraph of with at least one vertex in each has a good decomposition, then so have and
Proof: For every let be the subdigraph of induced by where Without loss of generality, let and let have a decomposition into arc-disjoint strong spanning subdigraphs To extend this decomposition to for every and , add to the vertices and let them have the same in- and out-neighbors as
The following theorem gives sufficient conditions for a digraph composition to have a good decomposition. As in Theorem 1.2, will denote the digraph obtained from the complete digraph of order 4 by deleting a cycle of length 4.
Theorem 2.2
Let where Then has a good decomposition if at least one of the following conditions holds:
(a) is a 2-arc-strong semicomplete digraph and are arbitrary digraphs, but is not isomorphic to
(b) has a Hamiltonian cycle and either is even and for every or is odd and for every apart from one for which or is odd, for every and at least two distinct subdigraphs have arcs.
(c) If and all are strong digraphs of orders at least 2.
Proof: Part (a) If is not isomorphic to then we are done by Theorem 1.2 and Lemma 2.1. Now assume that is isomorphic to , but is not isomorphic to . Let the vertices of be and its arcs
[TABLE]
Since is not isomorphic to , at least one of has at least two vertices. Without loss of generality, let have at least two vertices. Consider the subdigraph of induced by Then has two arc-disjoint strong spanning subdigraphs: with arcs
[TABLE]
and with arcs
[TABLE]
It remains to apply Lemma 2.1 to obtain a good decomposition of
Part (b) Without loss of generality, assume that is a Hamiltonian cycle of Let
Case 1: is even and for every
The following arc sets induce arc-disjoint strong spanning subdigraphs of
[TABLE]
[TABLE]
It remains to apply Lemma 2.1.
Case 2: is odd, for every and at least two distinct subdigraphs have arcs.
Let be arcs in two distinct subdigraphs and . We may assume that both end-vertices of and are in Observe that while (with arcs listed in (1)) is strong, (with arcs listed in (2)) forms two arc-disjoint cycles and We may assume that the tail (head) of () is in and and the head (tail) of () is in (otherwise, relabel vertices in and/or ). Thus, adding and to makes it strong. To obtain two arc-disjoint strong spanning subdigraphs of from , let every vertex for and have the same out- and in-neighbors as in
Case 3: is odd and for every apart from one for which
Without loss of generality, assume that and for all
First we consider the subcase in which , and Then has two arc-disjoint spanning subdigraphs and with arc sets
[TABLE]
[TABLE]
respectively. It is not hard to see that and are strong by constructing closed walks through all vertices.
Now we extend the previous subcase to that in which and for all First replace index 3 in every vertex of the form by in the two arc sets of the previous subcase. Then replace every arc of the form in by the path In , we replace by the path , replace by the path , replace by the path , and finally add the path .
Finally, we extend the previous subcase to the general one using Lemma 2.1.
Part (c) For , let be the subdigraph of induced by vertex set . Clearly, and and are strong.
Let be the spanning subdigraph of with arc set . Observe that is strong since and each are strong, and has a common vertex with each , where .
Let be the spanning subdigraph of with arc set . To see that is strong, we only need to find a strong subdigraph in which contains and for each pair of distinct vertices and in . We will consider two cases.
Case 1: .
Without loss of generality, we assume that and . We first consider the subcase that . Observe that there is at least one arc entering and one arc leaving in , and so there are two arcs, say and Β ( and ), with opposite directions between and in . Then by adding the arcs , and the vertices to , we obtain a strong subdigraph of which contains both and , as desired. For the case that , we just add the arcs , and the vertex to , and then obtain a strong subdigraph of which contains both and .
Case 2: .
Without loss of generality, we assume that and (if and exist). By Case 1 and the fact that is strong, we are done if . For the case that , by adding the arcs and the vertex to , we can obtain a strong subdigraph of which contains both and . With a similar argument, we can get the desired strong subdigraph for the case that .
Hence, we complete the argument and conclude that has a good decomposition.
We will use Theorem 2.2 to prove the following characterization for certain compositions , where is a strong semicomplete digraph. In the characterization, will stand for the digraph of order with no arcs. Also, and will denote the cycle and path with vertices, respectively.
Theorem 2.3
*Let be a strong semicomplete digraph on vertices and let be arbitrary digraphs, each with at least two vertices. Then has a good decomposition if and only if is not isomorphic to one of the following three digraphs: , *
Proof: Let us first prove the βonly ifβ part of the theorem, i.e. and do not have good decompositions. By Lemma 2.1, it suffices to show that neither nor has a good decomposition. The proof is by reductio ad absurdum.
Suppose that has a decomposition into two strong spanning subdigraphs Since has 13 arcs, without loss of generality, we may assume that is a Hamiltonian cycle of Since the arc of cannot be in a Hamiltonian cycle of , without loss of generality, let . Then the remaining arcs of form two disjoint cycles and and a single arc between them, a contradiction to the assumption that is strong.
Suppose that has a decomposition into two strong spanning subdigraphs Since has 16 arcs and has no Hamiltonian cycle, each of has 8 arcs. Since has only cycles of lengths 3 and 6 and is strong, without loss of generality, we may assume that consists of a cycle and a path Then consists of two cycles and and a path Observe that is not strong, a contradiction.
Now we will show the βifβ part of the theorem by reductio ad absurdum as well. Assume that is not isomorphic to either of the three digraphs, but has no good decomposition.
By Camionβs Theorem [7], has a Hamiltonian cycle . Thus, Conditions (b) of Theorem 2.2 are applicable. By the conditions, must be odd and for at least two distinct indexes , we have
Suppose Then there will be arcs between and in for every Recall Case 2 of Part (b) of the proof of Theorem 2.2. The arcs between and arcs can be used to make strong instead of arcs and used in Case 2 of Part (b) of the proof of Theorem 2.2. Thus, has a good decomposition, a contradiction. Hence, and, without loss of generality, and
Suppose that has opposite arcs. One of these arcs will not be on the Hamiltonian cycle of and will correspond to four or more arcs in Now recall Case 2 of Part (b) of the proof of Theorem 2.2. Two of the above-mentioned arcs can be used to make strong instead of arcs and used in Case 2 of Part (b) of the proof of Theorem 2.2. Thus, has a good decomposition, a contradiction. Hence,
Suppose that To get a contradiction, by Lemma 2.1 it suffices to show that has a decomposition into two strong spanning subdigraphs where consists of a cycle and two paths and and consists of two cycles and and two paths and Thus,
Now consider the case of and Since is not isomorphic to it has an arc in either or or , and by Conditions (b) of Theorem 2.2, only one of has an arc Without loss of generality, assume that if has an arc then , if has an arc then and if has an arc then Then has a decomposition into two spanning subdigraphs where consists of a cycle and a path and consists of two cycles and , a path and arc . Observe that both and are strong, a contradiction.
It remains to consider the case of Since is not isomorphic to , at least one of and has an arc. By Conditions (b) of Theorem 2.2, only one of and has an arc. Without loss of generality, assume that has an arc. Suppose that has two arcs. Then . Then we can use the arcs of to make strong instead of arcs and used in Case 2 of Part (b) of the proof of Theorem 2.2. Thus, has a good decomposition, a contradiction. Hence, if has an arc, it must have just one arc. This concludes our proof.
3 Products of digraphs
The Cartesian product of two digraphs and is a digraph with vertex set and arc set By definition, we know the Cartesian product is associative and commutative, and is strongly connected if and only if both and are strongly connected [10]. We define the th powers with respect to Cartesian product as .
In the argument of this section, we will use the following terminology and notation. Let and be two digraphs with and . For simplicity, we let for . We use to denote the subdigraph of induced by vertex set where , and use to denote the subdigraph of induced by vertex set where . Clearly, we have and . (For example, as shown in Figure 1, for and for .) For , and belong to the same digraph where ; we call the vertex corresponding to in ; for , we call the vertex corresponding to in . Similarly, we can define the subdigraph corresponding to some other subdigraph. For example, in Fig. 1(c), let Β be the path labelled 1 (2) in , then is called the path corresponding to in .
Lemma 3.1
For any integer , the product digraph can be decomposed into two arc-disjoint Hamiltonian cycles.
Proof: Let ; moreover and Let for and . Let be the subdigraph of which is a union of paths and the following arcs: Let be a spanning subdigraph of with . It is not hard to check that both and are Hamiltonian cycles of ; this completes the proof.
Note that deciding whether a digraph has a collection of arc-disjoint cycle covering all vertices of , can be done in polynomial time using network flows. Indeed, assign lower bound 1 and upper bound to every vertex in and lower bound 0 and upper bound 1 to every arc of . Observe that the resulting network has a feasible flow if and only if has a collection of arc-disjoint cycle covering all vertices of . Observe that the existence of a flow in a network with lower and upper bounds on vertices and arcs can be decided in polynomial time, see e.g. Chapter 4 in [1]. Moreover, we can compute such a flow in polynomial time (if it exists) and obtain the corresponding collection of cycles in The following lemma may be of independent interest.
Lemma 3.2
Let be a strong digraph of order at least two which has a collection of arc-disjoint cycle covering all its vertices. Then the product digraph can be decomposed into two arc-disjoint strong spanning subdigraphs. Moreover, these two arc-disjoint strong spanning subdigraphs can be found in polynomial time.
Proof: By the arguments in the paragraph before this lemma, we may assume that we are given a collection of arc-disjoint cycle covering all vertices of . For each , let denote the digraph with vertices and arcs . Now we will prove the lemma by induction on the number of cycles in the collection.
For the base step, by Lemma 3.1, we have that can be decomposed into two arc-disjoint strong spanning subdigraphs.
For the inductive step, we assume that can be decomposed into two arc-disjoint strong spanning subdigraphs and . We will construct two arc-disjoint strong spanning subdigraphs in .
If , then is a Hamiltonian cycle of , and we are done by Lemma 3.1. If , then is a strong spanning subdigraph of , and we are also done.
In the following argument, we assume that and . Without loss of generality, for the first copies of and in and , let , . We have . For the second copies of and in and , we will use βs rather than βs.
By Lemma 3.1, in , the subdigraph can be decomposed into two arc-disjoint strong spanning subdigraphs and . Observe that
[TABLE]
For , let be the subdigraph of corresponding to . For , let be the subdigraph of corresponding to . For , let be the subdigraph of corresponding to . For , let be the subdigraph of corresponding to .
Now let be a union of the following strong digraphs: , , and for all . Observe that is a strong spanning subdigraph of since has at least one common vertex with each of , and for all . Let be a spanning subdigraph of with . Observe that is the union of , , and for all . And has at least one common vertex with each of , and for all , thus is strong.
Hence, we complete the inductive step and conclude that can be decomposed into two arc-disjoint strong spanning subdigraphs. Moreover, by the above argument, these subdigraphs can be found in polynomial time.
Lemma 3.3
For any two strong digraphs and , if contains a pair of arc-disjoint strong spanning subdigraphs, then the product digraph can be decomposed into two arc-disjoint strong spanning subdigraphs.
Proof: Let , and contain two arc-disjoint strong spanning subdigraphs and . For , let be the subdigraph of corresponding to . Let be a union of and for all , and be a subdigraph of with and . It is not hard to verify that both and are strong spanning subdigraphs of . This completes the proof.
By the definition of , associativity of the Cartesian product, and Lemmas 3.2 and 3.3, we can obtain the following result on for any integer .
Theorem 3.4
Let be a strong digraph of order at least two which has a collection of arc-disjoint cycle covering all its vertices and let be an integer. Then the product digraph can be decomposed into two arc-disjoint strong spanning subdigraphs. Moreover, for any fixed integer , these two subdigraphs can be found in polynomial time.
The strong product of two digraphs and is a digraph with vertex set and arc set By definition, is a spanning subdigraph of , and is strongly connected if and only if both and are strongly connected [10]. In the following argument, we will still use the terminology and notation introduced earlier in this section, since is a spanning subdigraph of .
Lemma 3.5
For any two integers , the product digraph can be decomposed into two arc-disjoint strong spanning subdigraphs.
Proof: Let and Let be the spanning subdigraph of which is the union of for and the following additional arcs: . Observe that is strong. Let be a spanning subdigraph of with . To see that is strong, observe that it contains for and arcs
We will use the following decomposition of strong digraphs.
An ear decomposition of a digraph is a sequence , where is a cycle or a vertex and each is a path, or a cycle with the following properties:
Β and are arc-disjoint when .
Β For each : let denote the digraph with vertices and arcs . If is a cycle, then it has precisely one vertex in common with . Otherwise the end vertices of are distinct vertices of and no other vertex of belongs to .
Β .
The following result is well-known, see e.g. [1].
Theorem 3.6
Let be a digraph with at least two vertices. Then is strong if and only if it has an ear decomposition. Furthermore, if is strong, every cycle can be used as a starting cycle for an ear decomposition of , and there is a linear-time algorithm to find such an ear decomposition.
Theorem 3.7
For any strong digraphs and with orders at least 2, the product digraph can be decomposed into two arc-disjoint strong spanning subdigraphs. Moreover, these two arc-disjoint strong spanning subdigraphs can be found in polynomial time.
Proof: By Theorem 3.7 has an ear decomposition and has an ear decomposition , such that is a cycle of and is a cycle of by Theorem 3.6. Let denote the subdigraph of with vertices and arcs and let denote the subdigraph of with vertices and arcs
We will prove the theorem by induction on For the base step, by Lemma 3.5, we have that can be decomposed into two arc-disjoint strong spanning subdigraphs. For the inductive step, we assume that and can be decomposed into two arc-disjoint strong spanning subdigraphs and .
Since strong product is a commutative operation, without loss of generality it suffices to prove that can be decomposed into two arc-disjoint strong spanning subdigraphs. Let , and Let be the subdigraph of corresponding to for We will consider two cases.
Case 1: is a cycle.
Let Observe that every for shares vertex with Thus, the union of and for is a strong spanning subdigraph of Let and
Observe that contains , for and Thus, is strong.
Case 2: is a path.
Let where Let be the union of and for Observe that is a spanning subdigraph of and strong since every for shares its end-vertices with Let and Observe that contains , for and Thus, is strong.
Hence, we complete the inductive step and conclude that can be decomposed into two arc-disjoint strong spanning subdigraphs. Furthermore, by Theorem 3.6, the proof of Lemma 3.5, and the argument of this theorem, we can conclude that these two strong spanning subdigraphs can be found in polynomial time.
The lexicographic product of two digraphs and is a digraph with vertex set and arc set [10]. By definition, is a spanning subdigraph of , so the following result holds by Theorem 3.7: For any strong connected digraphs and with orders at least 2, the product digraph can be decomposed into two arc-disjoint strong spanning subdigraphs. Moreover, these two arc-disjoint strong spanning subdigraphs can be found in polynomial time. In fact, we can get a more general result.
A digraph is Hamiltonian decomposable if it has a family of Hamiltonian dicycles such that every arc of the digraph belongs to exactly one of the dicycles. Ng [11] gives the most complete result among digraph products.
Theorem 3.8
[11]** If and are Hamiltonian decomposable digraphs, and is odd, then is Hamiltonian decomposable.
Theorem 3.8 implies that if and are Hamiltonian decomposable digraphs, and is odd, then can be decomposed into two arc-disjoint strong spanning subdigraphs. It is not hard to extend this result as follows: for any strong digraphs and of orders at least 2, if contains arc-disjoint strong spanning subdigraphs, then the product digraph can be decomposed into arc-disjoint strong spanning subdigraphs.
4 Open Problems
We have characterized digraphs where is strong semicomplete and every is arbitrary with at least two vertices, which have a good decomposition. It is a natural open problem to extend the characterization to all such digraphs, where some βs can have just one vertex. Of course, the extended characterization would generalize also Theorem 1.2.
A digraph is quasi-transitive, if for any triple of distinct vertices of , if and are arcs of then either or or both are arcs of For a recent survey on quasi-transitive digraphs and their generalizations, see a chapter [8] by Galeana-SΓ‘nchez and HernΓ‘ndez-Cruz. Bang-Jensen and Huang [3] proved that a quasi-transitive digraph is strong if and only if where is a strong semicomplete digraph and each is a non-strong quasi-transitive digraph or has just one vertex. Thus, a special case of the above problem is to characterize strong quasi-transitive digraphs with a good decomposition. This would generalize Theorem 1.2 as well.
We believe that these characterizations will confirm Conjecture 1.1 for the classes of quasi-transitive digraphs and digraphs where is strong semicomplete. In the absence of the characterizations, it would still be interesting to confirm the conjecture at least for quasi-transitive digraphs.
In Lemma 3.2, we show that contains a pair of arc-disjoint strong spanning subdigraphs when . However, the following result implies Lemma 3.2 cannot be extended to the case that , since it is not hard to show that the Cartesian product digraph of any two cycles has a pair of arc-disjoint strong spanning subdigraphs if and only if it has a pair of arc-disjoint Hamiltonian cycles.
Theorem 4.1
[13]** The Cartesian product is Hamiltonian if and only if there are non-negative integers for which and .
However, Lemma 3.2 could hold for the case that if we add other conditions. As shown in Lemma 3.3, we know contains a pair of arc-disjoint strong spanning subdigraphs when one of and contains a pair of arc-disjoint strong spanning subdigraphs. So the following open question is interesting: for any two strong digraphs and , neither of which contain a pair of arc-disjoint strong spanning subdigraphs, under what condition the product digraph contains a pair of arc-disjoint strong spanning subdigraphs?
Furthermore, we may also consider the following more challenging question: under what conditions the product digraph has more (than two) arc-disjoint strong spanning subdigraphs?
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] J. Bang-Jensen and G. Gutin, Digraphs: Theory, Algorithms and Applications, 2nd Edition, Springer, London, 2009.
- 2[2] J. Bang-Jensen and G. Gutin, Basic Terminology, Notation and Results, in Classes of Directed Graphs (J. Bang-Jensen and G. Gutin, eds.), Springer, 2018.
- 3[3] J. Bang-Jensen and J. Huang, Quasi-transitive digraphs, J. Graph Theory, 20(2), 1995, 141β161.
- 4[4] J. Bang-Jensen and J. Huang, Decomposing locally semicomplete digraphs into strong spanning subdigraphs, J. Combin. Theory Ser. B, 102, 2012, 701β714.
- 5[5] J. Bang-Jensen and A. Yeo, Decomposing k π k -arc-strong tournaments into strong spanning subdigraphs, Combinatorica 24(3), 2004, 331β349.
- 6[6] J.A. Bondy and U.S.R. Murty, Graph Theory, Springer, Berlin, 2008.
- 7[7] P. Camion, Chemins et circuits hamiltoniens des graphes complets, Comptes Rendus de lβAcadΓ©mie des Sciences de Paris, 249, 1959, 2151β 2152.
- 8[8] H. Galeana-SΓ‘nchez and C. HernΓ‘ndez-Cruz, Quasi-transitive digraphs and their extensions, in Classes of Directed Graphs (J. Bang-Jensen and G. Gutin, eds.), Springer, 2018.
