• 沒有找到結果。

Maximal sets of Hamilton cycles in Dn

N/A
N/A
Protected

Academic year: 2021

Share "Maximal sets of Hamilton cycles in Dn"

Copied!
5
0
0

加載中.... (立即查看全文)

全文

(1)

www.elsevier.com/locate/disc

Note

Maximal sets of Hamilton cycles in

Dn

Liqun Pu

a

, Hung-Lin Fu

b

, Hao Shen

c aDepartment of Mathematics, Zhengzhou University, Zhengzhou 450001, China bDepartment of Applied Mathematics, National Chiao Tung University, Hsin Chu 30050, Taiwan

cDepartment of Mathematics, Shanghai Jiao Tong University, Shanghai 200240, China Received 4 December 2006; received in revised form 21 June 2007; accepted 5 July 2007

Available online 17 August 2007

Abstract

In this paper, we prove that there exists a maximal set of m directed Hamilton cycles inDnif and only ifn/2mn − 1 for

n7.

© 2007 Elsevier B.V. All rights reserved.

Keywords: Directed Hamilton cycles; Complete directed graph; Maximal sets

1. Introduction

A Hamilton cycle in a graph T is a spanning cycle of T. If S is a set of edge-disjoint Hamilton cycles in T and ifE(S) is the set of edges occurring in the Hamilton cycles in S, then S is said to be maximal ifT − E(S) has no Hamilton cycle.

In paper[5], Hoffman et al. showed that there exists a maximal set S of m edge-disjoint Hamilton cycles in the complete graphKnif and only if(n + 3)/4m(n − 1)/2. In paper[1], Bryant et al. showed that there exists a maximal set S of m edge-disjoint Hamilton cycles in the complete bipartite graphKn,nif and only ifn/4 < mn/2. Later, Daven et al.[2]showed forn3 and p3, there exists a maximal set S of m Hamilton cycles in the complete multipartite graphKnp(p parts of size n) if and only if n(p − 1)/4mn(p − 1)/2, and m > n(p − 1)/4 if n is odd andp ≡ 1 (mod 4), except possibly if n is odd and m((n + 1)(p − 1) − 2)/4. Recently, Fu et al.[3]extended the results in[2]and proved that if(p − 1)/2mp − 1, then there exists a maximal set S of m Hamilton cycles inK2p− F , where F is a 1-factor of K2p.

In this paper, we will extend these results to directed graphs and study maximal sets of Hamilton cycles in the complete directed graphDn.

Throughout this paper, we will use the following notation and terminology. A digraph D is said to be r-regular if deg+(v) = deg(v) = r for each vertex v in D. Let Dnbe a complete directed graph of order n (without loops); it is clear thatDnis(n − 1)-regular. For disjoint sets A and B, define the directed complete bipartite graph DA,B as the

E-mail addresses:liqunpu@yahoo.com.cn,liqunpu@sina.com(L. Pu),hlfu@math.nctu.edu.tw(H. Fu),haoshen@sjtu.edu.cn(H. Shen). 0012-365X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.

(2)

graph with verticesV (DA,B) = A ∪ B and E(DA,B) = (A × B) ∪ (B × A). Note that |E(DA,B)| = 2mn. If |A| = m and|B| = n, then we often denote DA,B asDm,n. Obviously,Dm+n= Dm+ Dm,n+ Dn.

LetV (Dm,n) be the vertex set of Dm,nand|V (Dm,n)| be the number of vertices in V (Dm,n). Let E(Dm,n) be the edge set ofDm,nand|E(Dm,n)| be the number of edges in E(Dm,n).

LetAi be an ordered i-set defined onZt which can be expressed as slanted parentheses i.e.,Ai= (a1, a2, a3, . . . , ai−1, ai). Let Ai denote a path obtained fromAi,Ai = a1, a2, a3, . . . , ai−1, ai. Furthermore, for j ∈ Zt, letAi+ j = (a1+ j, a2+ j, a3+ j, . . . , ai−1+ j, ai+ j) and Ai+ j = a1+ j, a2+ j, a3+ j, . . . , ai−1+ j, ai+ j. Note that all the entries inAi + j and Ai + j are done modulo t.

2. Maximal sets of Hamilton cycles inDn

Before proving the main result, we need the following lemma. Lemma 2.1. For an integer n,n7, Dnhas a Hamilton decomposition.

Proof. Dncan be regarded as twoKnby orienting properly. When n is odd,Knhas a Hamilton decomposition[7]. When n is even, Tillson[8]gives the proof. 

The following lemma plays an important part to prove our main theorem.

Lemma 2.2. Let D be an r-regular directed graph on n vertices. Ifr n/2, then D contains a directed Hamilton

cycle.

In fact, this result is a special case of the following theorem which is obtained by Ghouila-Houri[4].

Theorem 2.1 (Ghouila-Houri[4]). Let +(D) and (D) denote the minimum outdegree and minimum indegree in

D, respectively. If D is a strict digraph (without cycles of length 2), and min{+(D), (D)}|V (D)|/2, then D is Hamiltonian.

On the other hand, ifD is regular strict digraph, then the condition for D to be Hamiltonian can be weaken slightly. The following result was obtained by Jackson.

Lemma 2.3 (Jackson[6]). Every strict digraph of minimum in-degree and out-degree k 2, on at most 2k +2 vertices,

is Hamiltonian.

Now, we are ready to prove our main result. In order to construct maximal sets of Hamilton cycles. By Lemma 2.3, we conclude that the minimum number m in a maximal set of m directed Hamilton cycles inDnisn/2. Therefore, it suffices to construct a maximal set of m directed Hamilton cycles for eachn/2mn − 1 for n7. In order to do that, the following lemma is essential. Mainly, we make an arrangement of directed differences in order that we can construct maximal sets of Hamilton cycles systematically.

Lemma 2.4. For positive integers l and m where m5 and 1l m − 1, there exists an ordered set Am−l =

(a1, a2, . . . , am−l) ⊆ Zmsuch thatS = {(at+1− at) (mod m) | t = 1, 2, . . . , m − l − 1} is an (m − l − 1)-subset

of{1, 2, . . . , m − 1} (all directed differences are distinct).

Proof. We divide the proof into two cases.

Case 1: When m is even.

Whenl = 1, we delete the last element from the following ordered set (0, 1, m − 1, 2, m − 2, 3, m − 3, . . . , m/2 − 2, m/2 + 2, m/2 − 1, m/2 + 1, m/2) to get Am−1, i.e., Am−1=0, 1, m − 1, 2, m − 2, 3, m − 3, . . . ,m 2 − 2, m 2 + 2, m 2 − 1, m 2 + 1  .

(3)

1 3 0 4 1 2 0 6 5 7 5 2 3 4 Fig. 1. 1 3 0 5 1 3 5 0 7 2 4 2 4 6 8 6 Fig. 2.

Whenl = 2, we delete the last element in Am−1and get Am−2 =  0, 1, m − 1, 2, m − 2, 3, m − 3, . . . ,m 2 − 2, m 2 + 2, m 2 − 1  . We give examples inFig. 1(i) (m = 6, l = 1) andFig. 1(ii) (m = 8, l = 1).

Note that inFig. 1(i),A5= (0, 1, 5, 2, 4), in which 3 is deleted and inFig. 1(ii),A7= (0, 1, 7, 2, 6, 3, 5), in which

4 is deleted.

We generalize the above method and get

Am−l= ⎧ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎩  0, 1, m − 1, 2, m − 2, 3, m − 3, 4, . . . ,m 2 − l 2+ 2 ,m 2 +  l 2 + 2 ,m 2 −  l 2 + 1 ,m 2 +  l 2+ 1 ,m 2 − l 2 if l ≡ 0 (mod 2); and  0, 1, m − 1, 2, m − 2, 3, m − 3, 4, . . . ,m 2 −  l + 1 2 + 2 ,m 2 +  l + 1 2 + 2 ,m 2 −  l+1 2 + 1  ,m 2 +  l + 1 2 + 1 ,m 2 −l + 1 2 , m 2 + l + 1 2 if l ≡ 1 (mod 2). Case 2: When m is odd.

Whenl = 1, we can get Am−1fromZm\{(m + 1)/4}.

Am−1 = (0, 1, m − 1, 2, m − 2, 3, . . . , (m + 1)/4 − 1, m − (m + 1)/4 + 1, (m + 1)/4 + 1, m − (m + 1)/4, (m + 1)/4 + 2, m − (m + 1)/4 − 1, . . . , (m + 1)/2 + 2, (m + 1)/2 − 1, (m + 1)/2 + 1, (m + 1)/2).

Similarly, we can getAm−2by deleting the last element inAm−1. We give examples inFig. 2(i) (m = 7, l = 1) andFig. 2(ii) (m = 9, l = 1).

Note that inFig. 2(i),A6= (0, 1, 6, 3, 5, 4), in which 2 is deleted and inFig. 2(ii),A8= (0, 1, 8, 2, 7, 4, 6, 5), in

which 3 is deleted.

(4)

Thus, we have Am−l= ⎧ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎩  0, 1, m − 1, 2, m − 2, 3, m − 3, 4, . . . , m + 1 4 − 1, m +1 − m + 1 4 , m + 1 4 + 1, m − m + 1 4 , m + 1 4 +2, m − m + 1 4 − 1, . . . ,m + 1 2 − l 2 + 1 ,m + 1 2 +  l 2 + 1 ,m + 1 2 − l 2 ,m + 1 2 + l 2 if l ≡ 0 (mod 2); and  0, 1, m − 1, 2, m − 2, 3, m − 3, 4, . . . , m + 1 4 − 1, m + 1m + 1 4 , m + 1 4 + 1, m − m + 1 4 , m + 1 4 +2, m − m + 1 4 − 1, . . . ,m + 1 2 + l + 1 2 +1,m + 1 2 −  l − 1 2 + 1 ,m + 1 2 + l + 1 2 ,m + 1 2 − l − 1 2 if l ≡ 1 (mod 2).  With the above preparations, we are now in a position to prove the main theorem of this paper.

Theorem 2.2. For two positive integersm, n, n7, there exist maximal sets of m Hamilton cycles in Dnif and only if n/2mn − 1.

Proof. Suppose thatm < n/2, and let S be a set of m edge-disjoint directed Hamilton cycles in Dn. LetG(S) be the graph consisting of the union of the m Hamilton cycles in S. Then G(S)c is an r-regular directed graph with r = n − m − 1n/2 except the case n is odd, m = (n − 1)/2. So by Lemma 2.2, G(S)ccontains a Hamilton cycle and S is not a maximal set. When n is odd,m = (n − 1)/2, G(S)cis an(n − 1)/2-regular directed graph of order n, by Lemma 2.3,G(S)chas at least one Hamilton cycle. Thusmn/2.

Now we shall show whenn/2mn − 2, we can construct maximal sets of m Hamilton cycles in Dn.

When integerl > 0, let l = n − m − 1. Otherwise let integer l = 0. By Lemma 2.4, we can choose ordered set Am−landbi (1i l) from Zmand thusZmcan be partitioned intoAm−land everything else not inAm−l. That is, Zm= Am−l∪ {b1, b2, . . . , bl} where Am−l= (a1, a2, . . . , am−l) and Am−l∩ {b1, b2, . . . , bl} = ∅.

SinceDn = Dn−m+ Dm,n−m+ Dm, let V (Dm,n−m) = {i0|i ∈ Zn−m} ∪ {i|i ∈ Zm}. Further, {i|i ∈ Zm} can be partitioned asAm−l∪ {b1, b2, . . . , bl}. Now, we obtain a set S of directed Hamilton cycles with S = {(00, (b1+ j), 10, (b2+ j), . . . , (n − m − 2)0, (bn−m−1+ j), (n − m − 1)0, (A∗m−l+ j))|j ∈ Zm} from Dm+ Dm,n−m. Note that the entries of{bt+ j|1t n − m − 1} and Am−l+ j are done modulo m. Obviously, E(Dm,n−m) ⊆ E(S) and Dn− E(S) is disconnected. Thus S is a maximal set of Hamilton cycles in Dnwhenn/2mn − 2.

As form = n − 1, Dncan be decomposed into directed Hamilton cycles by Lemma 2.1. Thus, we conclude the proof. 

Example 1. Whenn = 8, 4m7, we give a construction of maximal sets of Hamilton cycles for m ∈ {4, 5, 6} and the casem = 7 can be settled by applying Lemma 2.1:

m = 4, the set is {(00, j, 10, 1 + j, 20, 2 + j, 30, 3 + j)|j ∈ Z4}. m = 5, the set is {(00, j, j + 1, j + 4, 10, 2 + j, 20, 3 + j)|j ∈ Z5}. m = 6, the set is {(00, j, j + 1, j + 5, j + 2, j + 4, 10, 3 + j)|j ∈ Z6}.

(5)

Example 2. Whenn = 9, 5m8, we give a construction of maximal sets of Hamilton cycles for m ∈ {5, 6, 7} and the casem = 8 can be settled by applying Lemma 2.1:

m = 5, the set is {(00, j, j + 1, 10, j + 4, 20, 3 + j, 30, 2 + j)|j ∈ Z5}. m = 6, the set is {(00, j, j + 1, j + 5, j + 2, 10, 3 + j, 20, 4 + j)|j ∈ Z6}. m = 7, the set is {(00, j, j + 1, j + 6, j + 3, j + 5, j + 4, 10, 2 + j)|j ∈ Z7}. Acknowledgements

The authors wish to thank the referees for their valuable comments and suggestions, and in particular, one of the referees for his/her effort to provide assistance in rewriting the paper.

References

[1]D.E. Bryant, S. El-Zanati, C.A. Rodger, Maximal sets of Hamilton cycles inKn,n, J. Graph Theory 33 (2000) 25–31.

[2]M. Daven, J.A. MacDougall, C.A. Rodger, Maximal sets of Hamilton cycles in complete multipartite graphs, J. Graph Theory 43 (1) (2003) 49–66.

[3]H.L. Fu, S.L. Logan, C.A. Rodger, Maximal sets of Hamilton cycles inK2p− F , Discrete Math., to appear.

[4]A. Ghouila-Houri, Une condition suffisante d’existence d’un circuit Hamiltonien, C. R. Acad. Sci. Paris 251 (1960) 495–497. [5]D.G. Hoffman, C.A. Rodger, A. Rosa, Maximal sets of 2-factors and Hamilton cycles, J. Combin. Theory (B) 57 (1993) 69–76. [6]B. Jackson, Long paths and cycles in oriented graphs, J. Graph Theory 5 (1981) 145–157.

[7]E. Lucas, Recreations Mathématiques, vol. 2, Gauthiers-Villars, Paris, 1892.

參考文獻

相關文件

Then End L (φ) is a maximal order in D. conjugacy classes) of maximal orders in D correspond bijectively to the orbits of isomorphism classes of supersingular Drinfeld A-modules

It’s easy to check that m is a maximal ideal, called the valuation ideal. We can show that R is a

Because both sets R m  and L h i ði ¼ 1; 2; :::; JÞÞ are second-order regular, similar to [19, Theorem 3.86], we state in the following theorem that there is no gap between

We can therefore hope that the exact solution of a lower-dimensional string will provide ideas which could be used to make an exact definition of critical string theory and give

strongly monotone or uniform P -function to obtain property of bounded level sets, see Proposition 3.5 of Chen and Pan (2006).. In this section, we establish that if F is either

Optim. Humes, The symmetric eigenvalue complementarity problem, Math. Rohn, An algorithm for solving the absolute value equation, Eletron. Seeger and Torki, On eigenvalues induced by

In particular, the parabolic second-order directional differentiability of projec- tion operator was used to establish the expression of second-order tangent sets, which plays

Now we assume that the partial pivotings in Gaussian Elimination are already ar- ranged such that pivot element a (k) kk has the maximal absolute value... The growth factor measures