459
Views
0
CrossRef citations to date
0
Altmetric
Articles

On the strong beta-number of galaxies with three and four components

, &

Abstract

The beta-number of a graph G is the smallest positive integer n for which there exists an injective function f:VG0,1,,n such that each uvEG is labeled |fufv| and the resulting set of edge labels is c,c+1,,c+|EG|1 for some positive integer c. The beta-number of G is + if there exists no such integer n. If c=1, then the resulting beta-number is called the strong beta-number of G. A galaxy is a forest for which each component is a star. In this paper, we establish a lower bound for the strong beta-number of an arbitrary galaxy under certain conditions. We also determine formulas for the (strong) beta-number and gracefulness of galaxies with three and four components. As corollaries of these results, we provide formulas for the beta-number and gracefulness of the disjoint union of multiple copies of the same galaxies if the number of copies is odd. Based on this work, we propose some problems and a new conjecture.

1 Introduction

All graphs considered in this paper are finite and undirected without loops or multiple edges. The vertex set of a graph G is denoted by VG, while the edge set is denoted by EG. The union G1G2 of two subgraphs G1 and G2 of a graph G is the subgraph with vertex set VG1VG2 and edge set EG1EG2. The union of any finite number of subgraphs is defined similarly.

For integers a and b with ab, the set xZ:axb will be denoted by writing a,b, where Z denotes the set of all integers. On the other hand, if a>b, then we treat a,b as the empty set. If such a situation appears in particular formulas for a given vertex labeling, then we ignore the corresponding portions of the formulas.

As a possible way of attacking graph decomposition problems, β-valuations were originated by Rosa Citation[1]. For a graph G of size q, an injective function f:V(G)0,q is called a β-valuation if each uvE(G) is labeled |f(u)f(v)| and the resulting edge labels are distinct. Such a valuation is now commonly known as a graceful labeling (the term was coined by Golomb Citation[2]) and a graph with a graceful labeling is called graceful. Graceful labelings have been the focus of many papers. For recent contributions to this subject and other types of labelings, the authors refer the reader to the survey by Gallian Citation[3].

The gracefulness, gracG, of a graph G was introduced by Golomb Citation[2] as the smallest positive integer n for which there exists an injective function f:VG0,n such that each uvEG is labeled |fufv| and the resulting edge labels are distinct. It is clear that if G is a graph of size q with gracG=q, then G is graceful. Thus, the gracefulness of a graph G is a parameter that measures how close G is to being graceful.

A number of authors have invented analogues of gracefulness. For instance, the beta-number and strong beta-number introduced in Citation[4] are such type of parameters. The beta-number, βG , of a graph G with q edges is the smallest positive integer n for which there exists an injective function f:VG0,n such that each uvEG is labeled |fufv| and the resulting set of edge labels is c,c+q1 for some positive integer c. The beta-number of G is + if there exists no such integer n. If c=1, then the resulting beta-number is called the strong beta-number of G and is denoted by βsG.

The following result observed in Citation[4] summarizes how the parameters discussed thus far are related.

Lemma 1

For every graph G of order p and size q, maxp1,qgracGβGβsG.

A galaxy is a forest for which each component is a star. In Citation[4] the authors determined the exact values for the (strong) beta-number for several classes of graphs including galaxies with two components, and proved that every nontrivial tree and forest has finite strong beta-number. Ichishima et al. Citation[5] studied the (strong) beta-number for forests with isomorphic components. This led them to conjecture that the (strong) beta-number and gracefulness of a forest of order p are either p1 or p . They further obtained the following result, which will prove to be useful later.

Theorem 1

If F is a forest of order p such that βF=p1, then βmF=mp1when m is odd.

In this paper, we present a lower bound for the strong beta-number of an arbitrary galaxy under certain conditions. We also determine formulas for the (strong) beta-number and gracefulness of galaxies with three and four components. As corollaries of these results, we provide formulas for the beta-number and gracefulness of the disjoint union of multiple copies of the same galaxies if the number of copies is odd. These results add credence to the mentioned conjectures, and lead us to propose some problems and a new conjecture on the strong beta-number of galaxies.

There are other kinds of parameters that measure how close a graph is to being graceful. For further knowledge on the (strong) beta-number of graphs and related concepts, the authors suggest that the reader consults the results found in Citation[6–10]. For the most recent advances on the mentioned conjectures, the authors also direct the reader to the papers Citation[5,11].

2 General result

Our first result of this paper provides a lower bound for the strong beta-number of an arbitrary galaxy. For the purpose of presenting this result, we denote the star with n+1 vertices by Sn.

Theorem 2

Let GSn1Sn2Snk. Ifn1n2nk is odd, and k2 or3(mod4), thenβsGm+k, wherem=n1+n2++nk.

Proof

We prove the logically equivalent contrapositive statement of the theorem, that is, we show that if βsGm+k1, then n1n2nk is even, or k0 or 1(mod4). Thus, assume that n1n2nk is odd (ni is odd for each i1,k), and we will then verify that k0 or 1(mod4). By our assumption, there exists an injective function f:VG0,m+k1 such that each uvEG is labeled |fufv| and the resulting set of edge labels is 1,m.

Now, define the galaxy G with VG=xi:i1,ki=1kyiji:ji1,niand EG=i=1kxiyiji:ji1,ni. Let fxi=ci for each i1,k. For each i1,k, let fyiji=aiji with aiji>ci (ji1,si), and let fyiji=biji with biji<ci (ji1,ti). Then m+k1m+k2=vVGfv=i=1kci+i=1kji=1siaiji+i=1kji=1tibiji and mm+12=uvEG|fufv|=i=1kji=1siaijici+i=1kji=1ticibiji. Summing the above two equations together with the fact that ni=si+ti (i1,k), we obtain m2+km+kk12=i=1kci+i=1kji=1si2aijici+i=1kji=1tici=i=1ksi+ti+1ci+2i=1kji=1siaijisici=i=1kni+1ci+2i=1kji=1siaijisici. Recall that ni is odd for each i1,k. This implies that i=1kni+1ci+2i=1kji=1siaijisiciis even; so m2+km+kk12 is even. It is straightforward to see that if k2 or 3(mod4), then m2+km+kk12 is odd. Therefore, k0 or 1(mod4), and the proof is completed. □

3 Results on galaxies with three components

In this section, we first prove the following theorem, which shows in this case that the bound given in Theorem 2 is sharp.

Theorem 3

For any positive integers n1,n2 andn3, βsSn1Sn2Sn3=n1+n2+n3+2if n1n2n3 is even,n1+n2+n3+3if n1n2n3 is odd.

Proof

For positive integers n1, n2 and n3, let GSn1Sn2Sn3 be the galaxy with VG=xi:i1,3i=13yiji:ji1,niand EG=i=13xiyiji:ji1,ni. It follows immediately from Lemma 1 that βsGn1+n2+n3+2.

To verify that βsGn1+n2+n3+2, suppose that n1n2n3 is even, and consider the following two cases for the vertex labeling f:VG0,n1+n2+n3+2.

Case 1. For n1=2k1, n2=2k2 and n3=2k3, where k1, k2 and k3 are positive integers, let fxi=i1if i1,2,2k1+2k2+2k3+2if i=3,fy1j1=k2+1+j1if j11,k1,k2+2k3+j1if j1k1+1,2k1,fy2j2=1+j2if j21,k2,2k1+2k3+1+j2if j2k2+1,2k2,fy3j3=k1+k2+1+j3if j31,2k31,2k1+k2+2k3+1if j3=2k3. Notice then that fy1j1:j11,k1=k2+2,k1+k2+1,fy1j1:j1k1+1,2k1=k1+k2+2k3+1,2k1+k2+2k3,fy2j2:j21,k2=2,k2+1,fy2j2:j2k2+1,2k2=2k1+k2+2k3+2,2k1+2k2+2k3+1,fy3j3:j31,2k31=k1+k2+2,k1+k2+2k3,fy3j3:j3=2k3=2k1+k2+2k3+1. This together with the values of fx1, fx2 and fx3 implies that f is a bijective function. Notice also that |fx1fy1j1|:j11,k1=k2+2,k1+k2+1,|fx1fy1j1|:j1k1+1,2k1=[k1+k2+2k3+1,2k1+k2+2k3],|fx2fy2j2|:j21,k2=1,k2,|fx2fy2j2|:j2k2+1,2k2=[2k1+k2+2k3+1,2k1+2k2+2k3],|fx3fy3j3|:j31,2k31=k1+k2+2,k1+k2+2k3,|fx3fy3j3|:j3=2k3=k2+1. Thus, |fufv|:uvEG=1,|EG|,since |EG|=2k1+2k2+2k3. Consequently, βsGn1+n2+n3+2 when n1, n2 and n3 is even.

Case 2. For n1=k1, n2=2k2 and n=2k31, where k1, k2 and k3 are positive integers, let fxi=i1if i1,2,k1+2k2+2k3+1if i=3,fy1j1=k2+k3+j1if j11,k1,fy2j2=1+j2if j21,k2,k1+2k3+j2if j2k2+1,2k2,fy3j3=k2+1+j3if j31,k31,k1+k2+1+j3if j3k3,2k31, Notice then that fy1j1:j11,k1=k2+k3+1,k1+k2+k3,fy2j2:j21,k2=2,k2+1,fy2j2:j2k2+1,2k2=k1+k2+2k3+1,k1+2k2+2k3,fy3j3:j31,k31=k2+2,k2+k3,fy3j3:j3k3,2k31=k1+k2+k3+1,k1+k2+2k3. This together with the values of fx1, fx2 and fx3 implies that f is a bijective function. Notice also that |fx1fy1j1|:j11,k1=k2+k3+1,k1+k2+k3,|fx2fy2j2|:j21,k2=1,k2,|fx2fy2j2|:j2k2+1,2k2=[k1+k2+2k3,k1+2k2+2k31],|fx3fy3j3|:j31,k31=[k1+k2+k3+1,k1+k2+2k31],|fx3fy3j3|:j3k3,2k31=k2+1,k2+k3. Thus, |fufv|:uvEG=1,|EG|,since |EG|=k1+2k2+2k31. Consequently, βsGn1+n2+n3+2 when n1 is arbitrary, n2 is even and n3 is odd. It is now immediate that βsG=n1+n2+n3+2 when n1n2n3 is even.

Next, suppose that n1n2n3 is odd. Then, by Theorem 2, βsGn1+n2+n3+3. It only remains to establish that βsGn1+n2+n3+3 when n1n2n3 is odd. Thus, let n1=2k11, n2=2k21, n3=2k31, where k1, k2 and k3 are positive integers, and consider the vertex labeling g:VG0,n1+n2+n3+3 such that gxi=i1if i1,2,2k1+2k2+2k3if i=3,gy1j1=k2+k3+j1if j11,2k11,gy2j2=1+j2if j21,k2,2k1+2k31+j2if j2k2+1,2k21,gy3j3=k2+1+j3if j31,k31,2k1+k2+j3if j3k3,2k31. Notice then that gy1j1:j11,2k11=k2+k3+1,2k1+k2+k31,gy2j2:j21,k2=2,k2+1,gy2j2:j2k2+1,2k21=2k1+k2+2k3,2k1+2k2+2k32,gy3j3:j31,k31=k2+2,k2+k3,gy3j3:j3k3,2k31=2k1+k2+k3,2k1+k2+2k31. This together with the values of gx1, gx2 and gx3 implies that g is an injective function. Notice also that |gx1gy1j1|:j11,2k11=[k2+k3+1,2k1+k2+k31],|gx2gy2j2|:j21,k2=1,k2,|gx2gy2j2|:j2k2+1,2k21=[2k1+k2+2k3,2k1+2k2+2k33],|gx3gy3j3|:j31,k31=[2k1+k2+k3,2k1+k2+2k31],|gx3gy3j3|j3k3,2k31=k2+1,k2+k3. Thus, |gugv|:uvEG=1,|EG|,since |EG|=2k1+2k2+2k33. Consequently, βsGn1+n2+n3+3 when n1n2n3 is odd. The desired result now follows. □

With the aid of Theorem 3, it is now possible to determine the beta-number of galaxies with three components.

Theorem 4

For any positive integers n1,n2 andn3, βSn1Sn2Sn3=n1+n2+n3+2.

Proof

For positive integers n1, n2 and n3, let GSn1Sn2Sn3 be the galaxy defined as in the proof of the preceding theorem. In light of Theorem 3 and Lemma 1, it suffices to verify that βGn1+n2+n3+2 when n1n2n3 is odd. Thus, let ni=2ki1, where ki is a positive integer for each i1,3, and consider the vertex labeling f:VG0,n1+n2+n3+2 such that fxi=0if i=1,2k1+2k2+2k34+iif i2,3,fy1j1=2k1+2k2+k32if j1=1,k2+k32+j1if j12,k1,k2+k31+j1if j1k1+1,2k11,fy2j2=k3+j2if j21,k21,k1+k2+k31if j2=k2,2k1+k32+j2if j2k2+1,2k21,fy3j3=1+j3if j31,k31,1if j3=k3,2k1+2k22+j3if j3k3+1,2k31. It remains to observe that f leads us to conclude that βGn1+n2+n3+2, which yields the desired result. □

The next result follows immediately from Theorems 1 and 4.

Corollary 1

For any positive integers n1,n2 andn3, βmSn1Sn2Sn3=mn1+n2+n3+31,where m is odd.

The following result is now obtained from Theorem 4 and Lemma 1.

Corollary 2

For any positive integers n1,n2 andn3, gracSn1Sn2Sn3=n1+n2+n3+2.

The following result is a simple consequence of Corollary 1 and Lemma 1.

Corollary 3

For any positive integers n1,n2 andn3, gracmSn1Sn2Sn3=mn1+n2+n3+31,where m is odd.

4 Results on galaxies with four components

In this section, we present formulas for the (strong) beta-number of galaxies with four components and related results. We start with the following theorem.

Theorem 5

For any positive integers n1,n2,n3 andn4, βsSn1Sn2Sn3Sn4=n1+n2+n3+n4+3.

Proof

For positive integers n1, n2, n3 and n4, let GSn1Sn2Sn3Sn4 be the galaxy with VG=xi:i1,4i=14yiji:ji1,niand EG=i=14xiyiji:ji1,ni. In light of Lemma 1, it suffices to show that βsGn1+n2+n3+n4+3. Thus, consider the following cases for the vertex labeling f:VG0,n1+n2+n3+n4+3.

Case 1. For ni=2ki, where ki is a positive integer for each i1,4, let fxi=i1if i1,3,2k1+2k2+2k3+2k4+3if i=4,fy1j1=k2+k3+2+j1if j11,k1,k2+k3+2k4+j1if j1k1+1,2k1,fy2j2=k3+2+j2if j21,k2,2k1+k3+2k4+1+j2if j2k2+1,2k2,fy3j3=2+j3if j31,k3,2k1+2k2+2k4+2+j3if j3=k3+1,2k3,fy4j4=k1+k2+k3+2+j4if j41,2k42,2k1+k2+k3+2k4+1if j4=2k41,2k1+2k2+k3+2k4+2if j4=2k4. Case 2. For n1=n2=1, n3=k3 and n4=k4, where k3 and k4 are positive integers, assume, without loss of generality, that k3k4 and let fxi=0if i=1,iif i2,3,2k4+5if i=4,fy1j1=1,fy2j2=2k4+4,fy3j3=3+2j3if j31,k4,k4+5+j3if j3k4+1,k3,fy4j4=2k4+42j4if j41,k4 and k3>k4. Case 3. For n1=2k11, n2=2k21, n3=k3 and n4=k4, where ki is a positive integer for each i1,4 with k12, let fxi=2if i=1,2k1+2k2+k3+2if i=2,2i5if i3,4,fy1j1=0if j1=1,3+j1if j12,k11,2k2+k3+2+j1if j1k1,2k11,fy2j2=k1+2+j2if j21,k21,k1+k3+2+j2if j2k2,2k21,fy3j3=k1+k2+1+j3if j31,k3,fy4j4=4if j4=1,2k1+2k2+k3+1+j4if j42,k4. Case 4. For n1=1, n2=2k2, n3=2k3 and n4=k4, where ki is a positive integer for each i2,4 with k42, let fxi=2i2if i1,2,2k2+2k3+6if i=3,3if i=4,fy1j(1)=k2+k3+3,fy2j2=1if j2=1,2+j2if j22,k2,2k3+5+j2if j2k2+1,2k2,fy3j3=k2+2+j3if j31,k3,k2+5+j3if j3k3+1,2k3,fy4j4=k2+k3+3+j4if j41,2,2k2+2k3+4+j4if j43,k4. Case 5. For n1=2k11, n2=2k2, n3=2k3 and n4=k4, where ki is a positive integer for each i1,4 with k12, let fxi=3iif i1,2,2k1+2k2+2k3+3if i=3,3if i=4,fy1j1=0if j1=1,3+j1if j12,k11,2k2+2k3+3+j1if j1k1,2k11,fy2j2=k1+2+j2if j21,k2,k1+2k3+1+j2if j2k2+1,2k2,fy3j3=k1+k2+2+j3if j31,2k31,k1+2k2+2k3+2if j3=2k3,fy4j4=4if j4=1,2k1+2k2+2k3+2+j4if j42,k4. Therefore, f leads us to conclude that βsGn1+n2+n3+n4+3 for any positive integers n1, n2, n3 and n4. This completes the proof. □

As a simple consequence of the preceding theorem and Lemma 1, we have the following two results.

Corollary 4

For any positive integers n1,n2,n3 andn4, βSn1Sn2Sn3Sn4=n1+n2+n3+n4+3.

Corollary 5

For any positive integers n1,n2,n3 andn4, gracSn1Sn2Sn3Sn4=n1+n2+n3+n4+3.

Applying Theorem 1 together with Corollary 4, we obtain the following result.

Corollary 6

For any positive integers n1,n2,n3 andn4, βmSn1Sn2Sn3Sn4=mn1+n2+n3+n4+41,where m is odd.

This result also has rather immediate corollary.

Corollary 7

For any positive integers n1,n2,n3 andn4, gracmSn1Sn2Sn3Sn4=mn1+n2+n3+n4+41,where m is odd.

5 Conclusions

Our results in this paper ensure that the bounds for the three parameters given in Lemma 1 take the same values. On the other hand, if G is a graceful graph of order p and size q, then βsG=βG=gracG=qp1.However, if G is a forest, then q<p1, and the inequality βsGβG may be strict as Theorems 3 and 4 indicate. Thus, we propose the following problem.

Problem 1

Find a sufficient condition (in terms of forbidden subgraphs) for a forest F to have βsF=βF.

The inequality βGgracG may be strict when G is a forest. To see this, let m be a positive integer, and define the forest FmP2 with VF=xi:i1,myi:i1,m and EF=xiyi:i1,m.Then the injective function f:VF0,2m1 such that fxi=i1 (i1,m) and fyi=2mi (i1,m)provides that |fufv|:uvEF=|fxifyi|:i1,m=2mi+1:i1,m, which is a set of distinct integers. This implies that gracF2m1 for every positive integer m. This together with Lemma 1 implies that gracF=2m1 for every positive integer m. However, it is known from Citation[5] that if m2(mod4), then βF=2m. From these observations, we arrive at the following problem.

Problem 2

Find a sufficient condition (in terms of forbidden subgraphs) for a forest F to have βF=gracF.

We have seen that the inequalities βsGβG and βGgracG are strict when G is a forest. This leads us to propose the following problem.

Problem 3

Find a sufficient condition (in terms of forbidden subgraphs) for a forest F to have βsF=βF=gracF.

Finally, the results in this paper and author’s computation of βsSn1Sn2 in Citation[7] lead us to propose the following conjecture.

Conjecture 1

For positive integers n1,n2, …,nk, letGSn1Sn2Snk. Then βsG=m+k1ifn1n2nkis even, or k0 or 1(mod4),m+kifn1n2nkis odd, and k2 or 3(mod4).where m=n1+n2++nk.

References

  • RosaA., On certain valuations of the vertices of a graph Theory of Graphs (Internat. Symposium, Rome, 1966)1967Gordon and BreachNY and Dunod Paris349–355
  • GolombS.W., How to number a graphReadR.C.Graph Theory and Computing1972Academic PressNew York23–37
  • GallianJ.A., A dynamic survey of graph labeling Electron. J. Combin.2018 #DS6
  • IchishimaR.Muntaner-BatleF.A.OshimaA., The measurements of closeness to graceful graphs Australas. J. Combin. 62 3 2015 197–210
  • IchishimaR.LópezS.C.Muntaner-BatleF.A.OshimaA., On the beta-number of forests with isomorphic components Discuss. Math. Graph Theory 382018 683–701
  • BarrientosC., The gracefulness of unions of cycles and complete bipartite graphs J. Combin. Math. Combin. Comput. 522005 69–78
  • IchishimaR.Muntaner-BatleF.A.OshimaA., On the beta-number of the joins of graphs Australas. J. Combin. 69 3 2017 402–409
  • IchishimaR.OshimaA., Bounds on the gamma-number of graphs Util. Math. 1092018 313–325
  • OshimaA.IchishimaR.Muntaner-BatleF.A., New parameters for studying graceful properties of graphs Electron. Notes Discrete Math. 602017 3–10
  • SinghT.PereiraJ.ArumugamS., A new measure for gracefulness of graphs Electron. Notes Discrete Math. 482015 275–280
  • IchishimaR.OshimaA., On the beta-number of linear forests with an even number of components AKCE Int. J. Graphs Comb. 15 3 2018 238–241