• 検索結果がありません。

Independent Set Neighborhood Union And Fractional Critical Deleted Graphs

N/A
N/A
Protected

Academic year: 2022

シェア "Independent Set Neighborhood Union And Fractional Critical Deleted Graphs"

Copied!
8
0
0

読み込み中.... (全文を見る)

全文

(1)

Independent Set Neighborhood Union And Fractional Critical Deleted Graphs

Yun Gao

y

, Mohammad Reza Farahani

z

, Wei Gao

x

Received 28 June 2016

Abstract

A graph G is called a fractional (k; n0; m)-critical deleted graph if any n0 vertices are removed from G the resulting graph is a fractional (k; m)-deleted graph. In this paper, we determine that for integers k 1, i 2, n0; m 0, n >4ki+n0+ 4m 4, and (G) k(i 1) +n0+ 2m, if

jNG(x1)[NG(x2)[ [NG(xi)j n+n0 2

for any independent subsetfx1; x2; : : : ; xigofV(G), thenGis a fractional(k; n0; m)- critical deleted graph. The bound for independent set neighborhood union con- dition is sharp.

1 Introduction

The fractional factor problem can be regarded as a relaxation problem of the cardinality matching. It has wide applications in …elds such as combinatorial polyhedron, network design and scheduling. For instance, in a communication network, several large data packets were to be sent to certain destinations via multiple channels. We can improve the e¢ ciency of this task by dividing the large data packets into small parcels. The available distribution of data packets can be considered as a fractional ‡ow problem.

Moreover, it can be looked upon as a fractional factor problem if the sources and destinations of a network are di¤erent.

In theoretical model, the whole network can be expressed as a graph in which each vertex represents a site and each edge denotes a channel between two sites. In this setting, the framework of data transmission problem is the existence of fractional factor in the graph corresponding to a network.

All graphs considered in this paper are …nite, loopless, and without multiple edges.

LetGbe a graph with vertex setV(G)and edge setE(G). For anyx2V(G), the degree and the neighborhood of xin G are denoted by dG(x) and NG(x), respectively. For

Mathematics Sub ject Classi…cations: 05C70.

yDepartment of Editorial, Yunnan Normal University, Kunming 650092, P. R. China

zDepartment of Applied Mathematics, Iran University of Science and Technology, Narmak, Tehran 16844, Iran

xSchool of Information Science and Technology, Yunnan Normal University, Kunming 650500, P.

R. China

268

(2)

S V(G), we denote byG[S]the subgraph ofGinduced byS, andG S =G[V(G)nS].

For two vertex-disjoint subsets S andT ofG, we use eG(S; T) to denote the number of edges with one end in S and the other end in T. We denote the minimum degree and the maximum degree ofGby (G)and (G), respectively. Thedistance dG(x; y) between two verticesxandyis de…ned to be the length of a shortest path connecting them. The notation and terminology used but unde…ned in this paper can be found in [1].

Let k 1 be an integer. A spanning subgraph F of G is called a k-factor if dF(x) = kfor each x2V(G). Leth:E(G)![0;1]be a function. If P

x2eh(e) =k for anyx2V(G), then we callG[Fh]afractionalk-factor ofGwith indicator function hwhereFh=fe2E(G) :h(e)>0g. Zhou [10] introduced the concept of a fractional (k; m)-deleted graph, that is, a graph G is called a fractional (k; m)-deleted graph if removing anymedges fromG, the resulting graph has a fractionalk-factor. A fractional (k; m)-deleted graph is simply called a fractionalk-deleted graph ifm= 1. A graphG is called afractional (k; n0)-critical graph if after deleting anyn0 vertices fromG, the resulting graph still has a fractionalk-factor.

A graphGis called afractional(k; n0; m)-critical deleted graph if after deleting any n0 vertices fromG, the resulting graph is still a fractional(k; m)-deleted graph.

In what follows, we always assume that n = jV(G)j. Yu et al. [8] determined a degree condition for the existence of a fractionalk-factor as follows.

THEOREM 1 ([8]). Letk 1be an integer, and letGbe a connected graph with n 4k 3 and (G) k. If

maxfdG(x); dG(y)g n 2

for each pair of nonadjacent vertices x; yofG, thenGhas a fractionalk-factor.

Liu and Zhang [7] obtained the following toughness condition for a graph to have fractionalk-factors.

THEOREM 2 ([7]). Letk 2be an integer. A graphGof ordernwithn k+ 1 has a fractionalk-factor ift(G) k 1k.

For fractional(k; m)-deleted graphs, the following known results are stated by Zhou.

THEOREM 3 ([10]). Letk 2 andm 0be two integers. Let Gbe a connected graph with

n 9k 1 p

2(k 1)2+ 2 + 2(2k+ 1)m and (G) k+m+(m+1)4k2 1. If

jNG(x)[NG(y)j 1

2(n+k 2)

for each pair of non-adjacent vertices x, y of G, thenG is a fractional(k; m)-deleted graph.

(3)

THEOREM 4 ([9]). Letk 1 andm 1be two integers. LetGbe a graph with n 4k 5 + 2(2k+ 1)m. If (G) n2, thenGis a fractional(k; m)-deleted graph.

Recently, Gao et al. gave the following result on the neighborhood union condition for fractional(k; n0; m)-critical deleted graphs.

THEOREM 5. Let k 2 and n0; m 0 be three integers, and let G be a graph withn 8k+n0+ 4m 7and (G) k+n0+m. If

jNG(x)[NG(y)j n+n0 2

for each pair of non-adjacent verticesx,y ofG, thenGis a fractional(k; n0; m)-critical deleted graph.

More su¢ cient conditions for graphs to have fractional factors can be found in Gao and Gao [3], Gao et al. [4], Gao and Wang [5], Zhou [10, 11, 12], and Zhou and Bian [13].

In this paper, we manifest the relationship between independent set neighborhood union condition and fractional (k; n0; m)-critical deleted graph. Furthermore, we will show that the bound for independent set neighborhood union is sharp in some sense.

Our main result is stated as follows.

THEOREM 6. Let Gbe a graph of ordern. Let k; i; n0; m be four integers with i 2,k 1 andm; n0 0. If (G) k(i 1) + 2m+n0,n >4ki+n0+ 4m 4, and

jNG(x1)[NG(x2)[ [NG(xi)j n+n0 2

for any independent subset fx1; x2; : : : ; xig ofV(G), thenGis a fractional (k; n0; m)- critical deleted graph.

Setn0 = 0in Theorem 6, then it becomes the following necessary independent set neighborhood union condition on fractional(k; m)-deleted graph.

COROLLARY 1. LetG be a graph of ordern. Let k; i; mbe three integers with i 2,k 1 andm 0. If (G) k(i 1) + 2m,n >4ki+ 4m 4, and

jNG(x1)[NG(x2)[ [NG(xi)j n 2

for any independent subset fx1; x2; : : : ; xig of V(G), then G is a fractional (k; m)- deleted graph.

Ifm= 0in Theorem 6, then we deduce the following corollary on the independent set neighborhood union condition of fractional (k; n0)-critical graph.

(4)

COROLLARY 2. LetG be a graph of ordern. Let k; i; n0 be three integers with i 2,k 1 andn0 0. If (G) k(i 1) +n0,n >4ki+n0 4, and

jNG(x1)[NG(x2)[ [NG(xi)j n+n0 2

for any independent subset fx1; x2; : : : ; xig of V(G), then G is a fractional (k; n0)- critical graph.

In order to prove our main result, we need the following lemma which present the necessary and su¢ cient condition of fractional (k; n0; m)-critical deleted graph.

LEMMA 1 ([2]). Letk 1and n0; m 0 be three integers, and letGbe a graph andH a subgraph ofGwithmedges. ThenGis a fractional(k; n0; m)-critical deleted graph if and only if

G(S; T) =kjSj+X

x2T

dG S(x) kjTj kn0+X

x2T

dH(x) eH(S; T);

for all disjoint subsetsS andT ofV(G)withjSj n0.

Since our result refers to independent set neighborhood union condition, we need new tricks compare to Gao et al. [6].

2 Proof of Theorem 6

Assume to the contrary that Gsatis…es the conditions of the Theorem 6, but is not a fractional (k; n0; m)-critical deleted graph. By Lemma 1 and noting the fact that P

x2TdH(x) eG(T; S) 2m, there exist disjoint subsetsS andT ofV(G)such that

G(S; T) =kjSj+X

x2T

dG S(x) kjTj kn0+ 2m 1: (1) We choose subsetsS andT such thatjTjis minimal. Obviously,T 6=;.

CLAIM 1. dG S(x) k 1 for anyx2T.

PROOF. IfdG S(x) k for some x2T, then the subsets S and T n fxg satisfy (1). This contradicts the choice ofS andT.

Let d1 = minfdG S(x)jx2 Tg and choose x1 2 T such that dG S(x1) =d1. If z 2andT n([zj=11NT[xj])6=;, let

dz= minfdG S(x)jx2T n([zj=11NT[xj])g

and choose xz 2T n([zj=11NT[xj]) such that dG S(xz) = dz. So, we get a sequence such that0 d1 d2 d k 1and an independent setfx1; x2; : : : ; x g T.

(5)

CLAIM 2. jTj k(i 1) + 1.

PROOF. Assume thatjTj k(i 1). ThenjSj+d1 dG(x1) (G) k(i 1) + n0+ 2m. By (1) and 0 d1 k 1, we have

kn0+ 2m 1 kjSj+X

x2T

dG S(x) kjTj kjSj+d1jTj kjTj

= kjSj+ (d1 k)jTj

k(k(i 1) d1+n0+ 2m) + (d1 k)(i 1)k

= k2(i 1) +d1(k(i 1) k) k2(i 1) +kn0+ 2km kn0+ 2m:

This produces a contradiction.

SincedG S(x) k 1 andjTj k(i 1) + 1, we get i. Thus, we can choose an independent setfx1; x2; : : : ; xig T.

In view of the condition of the theorem, we deduce n+n0

2 jNG(x1)[NG(x2)[ [NG(xi)j jSj+ Xi

j=1

dj

and

jSj n+n0 2

Xi

j=1

dj: (2)

Noting that

jNT[xj]j NT[xj]\([jz=11NT[xz]) 1; j= 2;3; : : : ; i 1 and

j [jz=1NT[xz]j Xj

z=1

jNT[xz]j Xj

z=1

(dG S(xz) + 1) = Xj

z=1

(dz+ 1); j= 1;2; : : : ; i:

Hence, we infer

kjSj+X

x2T

dG S(x) kjTj

kjSj kjTj+d1jNT[x1]j+d2(jNT[x2]j jNT[x2]\NT[x1]j) + +di 1(jNT[xi 1]j NT[xi 1]\([ij=12NT[xj]) ) +di(jTj j [ij=11NT[xj])jj)

kjSj+ (d1 di)jNT[x1]j+

i 1

X

j=2

dj+ (di k)jTj di

i 1

X

j=2

jNT[xj]j

(6)

= kjSj+ (d1 di)(d1+ 1) +

i 1

X

j=2

dj+ (di k)jTj di

i 1

X

j=2

(dj+ 1)

= kjSj+d21+

i 1

X

j=1

dj+ (di k)jTj di

i 1

X

j=1

(dj+ 1);

which implies

(n jSj jTj)(k di) kjSj+X

x2T

dG S(x) kjTj 2m kn0+ 1

kjSj+d21+

i 1

X

j=1

dj+ (di k)jTj di

i 1

X

j=1

(dj+ 1) 2m kn0+ 1:

Equivalently,

0 n(k di) (2k di)jSj+di

i 1

X

j=1

dj

i 1

X

j=1

dj+di(i 1) d21+ 2m kn0 1: (3) By (2), (3),d1 d2 di k 1andn >4ki+n0+ 4m 4, we yield

0 n(k di) (2k di)(n+n0 2

Xi

j=1

dj) +di i 1

X

j=1

dj i 1

X

j=1

dj

+di(i 1) d21+ 2m+kn0 1

= n

2di+ 2k Xi

j=1

dj di Xi

j=1

dj+di

i 1

X

j=1

dj

i 1

X

j=1

dj

+di(i 1) d21+ 2m+n0di

2 1

= n

2di+ ((2k 1)d1 d21) + (2k 1)

i 1

X

j=2

dj

+di(2k+i 1) d2i + 2m+n0di

2 1

n

2di+ (2k 1)di+ (2k 1)

i 1

X

j=2

di

+di(2k+i 1) d2i + 2m+n0di

2 1

= n

2di+ 2kidi d2i + 2m+n0di

2 1:

Ifdi>0, then0<2di d2i 1 0sincen >4ki+n0+ 4m 4and2m(1 di) 0, a contradiction.

(7)

Ifdi= 0, thend1= =di= 0. By (2), we havejSj n+n2 0 and jTj n jSj

n n0

2 . SincedG S(T) P

x2TdH(x) eG(T; S), we have kjSj+X

x2T

dG S(x) kjTj (X

x2T

dH(x) eG(T; S)) kn0

k n+n0

2 k n n0

2 + (dG S(T) X

x2T

dH(x) +eG(T; S)) kn0 0;

also a contradiction. This completes the proof of the theorem.

3 Sharpness

Theorem 6 is best possible, in some extent, on the conditions. Actually, we can con- struct some graphs such that the independent set neighborhood union condition in Theorem 6 can’t be replaced byjNG(x1)[NG(x2)[ [NG(xi)j n+n2 0 1.

LetG1=Kkt+n0be a complete graph,G2= (kt+1)K1be a graph consisting ofkt+1 isolated vertices, and G=G1_G2, wheret is su¢ ciently large (i.e.,t > 2ki+2mk 2+ n0 2k1 for somei. Thus, (G) k(i 1) +n0+ 2m, andn >4ki+n0+ 4m 4). Then n=jG1j+jG2j= 2kt+n0+ 1, and for any independent setfx1; x2; : : : ; xig V(G2), we have

n+n0

2 >jNG(x1)[NG(x2)[ [NG(xi)j=kt+n0> n+n0

2 1:

LetS=V(G1), andT =V(G2). Then

kjSj kjTj+dG S(T) X

x2T

dH(x) eG(T; S)

! kn0

= kjSj kjTj kn0 =k(kt+n0) k(kt+ 1) kn0= k <0:

Hence,Gis not a fractional (k; n0; m)-critical deleted graph.

Acknowledgment. The research is partially supported by NSFC (no. 11401519).

The authors are grateful to the anonymous referee for a careful checking of the details and for helpful comments that improved this paper.

References

[1] J. A. Bondy and U. S. R. Murty, Graph Theory, Springer, Berlin, 2008.

[2] W. Gao, Some Results on Fractional Deleted Graphs, Doctoral Dissertation of Soochow University, 2012.

(8)

[3] W. Gao and Y. Gao, Toughness condition for a graph to be a fractional(g; f; n)- critical deleted graph, The Scienti…c World J. Vol. 2014, Article ID 369798, 7 pages, http://dx.doi.org/10.1155/2014/369798.

[4] W. Gao, L. Liang, T. W. Xu and J. X. Zhou, Tight toughness condition for fractional(g; f; n)-critical graphs, J. Korean Math. Soc., 51(2014), 55–65.

[5] W. Gao and W. F. Wang, Toughness and fractional critical deleted graph, Utilitas Math., 98(2015), 295–310.

[6] Y. Gao, M. R. Farahani and W. Gao, A neighborhood union condition for frac- tional (k; n0; m)-critical deleted graphs, Transactions on Combinatorics, 2016, In press.

[7] G. Liu and L. Zhang, Toughness and the existence of fractionalk-factors of graphs, Discrete Math., 308(2008), 1741–1748.

[8] J. Yu, G. Liu, M. Ma and B. Cao, A degree condition for graphs to have fractional factors, Adv. Math. (China), 35(2006), 621–628.

[9] S. Z. Zhou, A minimum degree condition of fractional (k; m)-deleted graphs, Comptes Rendus Math., 347(2009), 1223–1226.

[10] S. Z. Zhou, A neighborhood condition for graphs to be fractional (k; m)- deleted graphs, Glasgow Math. J., 52(2010), 33–40.

[11] S. Z. Zhou, A su¢ cient condition for a graph to be a fractional(f; n)-critical graph, Glasgow Math. J., 52(2010), 409–415.

[12] S. Z. Zhou, A su¢ cient condition for graphs to be fractional(k; m)-deleted graphs, Appl. Math. Lett., 24(2011), 1533–1538.

[13] S. Z. Zhou and Q. Bian, An existence theorem on fractional deleted graphs, Period.

Math. Hung., 71(2015), 125–133.

参照

関連したドキュメント

Fortunato, Abstract critical point theorems and applica- tions to some nonlinear problems with “strong” resonance at infinity, Nonlinear Anal.. 7

In the recent years, problems involving Kirchhoff type operators have been studied in many papers, we refer to [2, 9, 11, 15, 17, 19, 20] in which the authors have used

Some authors have used fixed point the- orems to show the existence of positive solutions to boundary value problems for ordinary differential equations, difference equations,

Abstract. In this paper, we study the existence of periodic solutions to a class of functional differential system. By using Schauder , s fixed point theorem, we show that the

Inequality (4.15) means that the error produced by considering weak solutions of (2.7) in two different domains, with conductivity function verifying (4.3), is proportional to

However, the method of upper and lower solutions for the existence of solution is less developed and hardly few results can be found in the literature dealing with the upper and

Its layer-to-layer transfer matrix is a polynomial of two spectral parameters, it may be re- garded in terms of quantum groups both as a sum of sl(N) transfer matrices of a chain

Every one-dimensional connected component of CR(ϕ), may be seen, from topological point of view, as finite graph whose vertices are stationary points, and edges are orbits