Skip to main content
Heliyon logoLink to Heliyon
. 2020 Nov 20;6(11):e05561. doi: 10.1016/j.heliyon.2020.e05561

On the super edge-magic deficiency of some graphs

Vira Hari Krisnawati a,, Anak Agung Gede Ngurah b,, Noor Hidayat a, Abdul Rouf Alghofari a
PMCID: PMC7689529  PMID: 33294698

Abstract

A graph G is called super edge-magic if there exists a bijection f:V(G)E(G){1,2,,|V(G)|+|E(G)|}, where f(V(G))={1,2,,|V(G)|}, such that f(u)+f(uv)+f(v) is a constant for every edge uvE(G). Such a case, f is called a super edge magic labeling of G. A bipartite graph G with partite sets A and B is called consecutively super edge-magic if there exists a super edge-magic labeling f with the property that f(A)={1,2,,|A|} and f(B)={|A|+1,|A|+2,,|V(G)|}. The super edge-magic deficiency of a graph G, denoted by μs(G), is either the minimum nonnegative integer n such that GnK1 is super edge-magic or +∞ if there exists no such n. The consecutively super edge-magic deficiency of a bipartite graph G, denoted by μc(G), is either the minimum nonnegative integer n such that GnK1 is consecutively super edge-magic or +∞ if there exists no such n. In this paper, we study the super edge-magic deficiency of some graphs. We investigate the (consecutively) super edge-magic deficiency of forests with two components. We also investigate the super edge-magic deficiency of a 2-regular graph 2C3Cn and join product of K1,nPm with an isolated vertex.

Keywords: (Consecutively) super edge-magic graph, (Consecutively) super edge-magic deficiency, Forest with two components, 2-regular graph, Join product graph


(Consecutively) super edge-magic graph; (Consecutively) super edge-magic deficiency; Forest with two components; 2-regular graph; Join product graph

1. Introduction

Let G be a finite and simple graph having vertex set V(G) and edge set E(G), where p=|V(G)| and q=|E(G)|. A labeling of G is a bijection f:V(G)E(G){1,2,,p+q}. Weight of an edge xy under the labeling f is wf(xy)=f(x)+f(y)+f(xy). The labeling f is called an edge-magic labeling if wf(xy)=k for any xyE(G). In such a case, G is called an edge-magic graph and the constant k is called the magic constant of the labeling f. An edge-magic labeling f of G with additional property that f(V(G))={1,2,,p} is called a super edge-magic (SEM) labeling. Thus, a SEM graph is a graph that admits a SEM labeling. The concept of an edge-magic labeling was introduced by Kotzig and Rosa in 1970 [1]. Meanwhile, the terminology of a SEM labeling was introduced by Enomoto et al. in 1998 [2]. In 2001, Muntaner-Batle [3] introduced the concept of special SEM labeling of a bipartite graph. A special SEM labeling of a bipartite graph G with partite sets A and B is a SEM labeling f of G with the property that f(A)={1,2,,|A|} and f(B)={|A|+1,|A|+2,,p}. In 2007, Oshima [4] called such a labeling consecutively SEM labeling.

The following lemma proved by Figueroa-Centeno et al. [5] provides sufficient and necessary conditions for a SEM graph.

Lemma 1.1

[5] A graph G is SEM if and only if there exists a bijection f:V(G){1,2,,p} such that the set of all edge-sums S={f(x)+f(y):xyE(G)} consists of q consecutive integers. In this case, f extends to be a SEM labeling of G with magic constant k=p+q+min(S).

The next lemma proved by Enomoto et al. [2] gives sufficient condition for non-existence of a SEM labeling of a graph.

Lemma 1.2

[2] If G is a SEM graph then q2p3.

Moreover, Kotzig and Rosa [1] also proved that for every graph G there exists a nonnegative integer n such that GnK1 is an edge-magic graph. This fact leads them to introduce the concept of edge-magic deficiency of a graph. The edge-magic deficiency of a graph G, μ(G), is defined as the minimum nonnegative integer n such that GnK1 is an edge-magic graph. This concept motivated Figueroa-Centeno et al. [6] to introduce the concept of super edge-magic deficiency of a graph. The super edge-magic deficiency (SEMD) of a graph G, μs(G), is defined as either the minimum nonnegative n such that GnK1 is a SEM graph or +∞ if there exists no such n. In 2016, Ichishima et al. [7] give the notion of consecutively super edge-magic deficiency of a bipartite graph. The consecutively super edge-magic deficiency (consecutively SEMD) of a bipartite graph G, μc(G), is defined to be either the smallest nonnegative integer n with the property that GnK1 is consecutively SEM or +∞ if there exists no such n.

Many researchers have investigated the SEMD of some classes of graphs. The complete results on this subject can be seen in a dynamic survey of graph labeling by Gallian [8]. In this paper, we study the (consecutively) SEMD of forests with two components, where its components are non isomorphic combs, isomorphic a subdivision of K1,3 and K1,4, and union of a comb and a subdivision of K1,3 or K1,4. Moreover, we find SEMD of 2-regular graph 2C3Cn. By applying Ichishima's et al. [9] and Cichacz's et al. [10] methods to this result, we prove that some 2-regular graphs with a large order have zero SEMD, and then we also obtain the SEMD of union of cycles and paths. In addition, we study the SEMD of graph (K1,nPm)+K1 for any integer n1 and m3.

2. The SEMD of forests with two components

Figueroa-Centeno et al. [6] proved that the SEMD of all forests is finite. Next, the same authors [11] investigated the SEMD of some classes of forests with two components, such as PmK1,n,K1,nK1,m, and PmPn. Based on these results, they gave the following conjecture.

Conjecture 2.1

[11] If F is a forest with two components then μs(F)1.

Inspired by Conjecture 2.1, many researchers have investigated the SEMD of some forests with two components. Baig et al. [12] found the SEMD of union of combs and stars. Javed et al. [13] studied the SEMD of forests with two components consisting of combs, generalized combs, and stars. In particular, they found the SEMD of two isomorphic combs. Imran and Mukhtar [14] showed that forests with two components consisting of stars and subdivision of stars have zero SEMD. Krisnawati et al. [15] investigated SEMD of two non isomorphic of a subdivision of K1,3 and K1,4.

In this section, we study SEMD of disjoint union of two non isomorphic combs. We also continue Krisnawati's et al. [15] work to find SEMD of disjoint union of two isomorphic of a subdivision of K1,3 and K1,4. We prove that these graphs have zero SEMD. Some of these graphs also have zero consecutively SEMD. Additionally, we investigate the (consecutively) SEMD of union of a comb and a subdivision of K1,3 or K1,4.

A comb, Cbn with n1, is a tree consisting of path Pn+1, whose vertices are u0,u1,u2,,un, together with edges {uivi:1in}. We investigate zero (consecutively) SEMD of two non isomorphic combs in the following theorem.

Theorem 2.2

Let Gn,m=CbnCbm. For any n,m1 and nm, μs(Gn,m)=μc(Gn,m)=0.

Proof

Let Gn,m be a graph with vertex set and edge set as follows.

V(Gn,m)={ui:0in}{vi:1in}{wj:0jm}{xj:1jm}andE(Gn,m)={uiui+1:0in1}{uivi:1in}{wjwj+1:0jm1}{wjxj:1jm}.

Thus, Gn,m is a graph of order 2(n+m+1) and size 2(n+m).

For any n,m1 and nm, define a labeling f:V(Gn,m){1,2,,2(n+m+1)}. Without loss of generality, we assume n<m. We consider the proof based on the value of n.

Case 1. n is odd.

f(a)={i+2,ifa=uiandiis even,n+m+i+1,ifa=uiandiis odd,i+2,ifa=viandiis odd,n+m+i+1,ia=viandiis even,2n+m+j+2,ifa=wjandjis even,n+j+2,ifa=wjandjis odd,1jn,n+j+1,ifa=wjandjis odd,n+2jm,2n+m+j+2,ifa=xjandjis odd,n+j+2,ifa=xjandjis even,2jn1,1,ifa=xn+1,n+j+1,ifa=xjandjis even,n+3jm.

Case 2. n is even.

f(a)={n+m+i+2,ifa=uiandiis even,i+1,ifa=uiandiis odd,n+m+i+2,ifa=viandiis odd,i+1,ifa=viandiis even,n+j+2,ifa=wjandjis even,0jnn+j+1,ifa=wjandjis even,n+2jm,2n+m+j+2,ifa=wjandjis odd,n+j+2,ifa=xjandjis odd,1jn11,ifa=xn+1,n+j+1,ifa=xjandjis odd,n+3jm,2n+m+j+2,ifa=xjandjis even.

It is not hard to verify that the set of all edge-sums generated by f is S={n+m+4,n+m+5,,3(n+m+1)}. By Lemma 1.1, f extends to a SEM labeling of Gn,m with magic constant k=5(n+m)+6. Hence, for any n,m1 and nm, μs(Gn,m)=0.

Next, let A and B be partite sets of Gn,m.

For odd n,

A={ui:i0(mod2)}{vi:i1(mod2)}{wj:j1(mod2)}{xj:j0(mod2)}andB={ui:i1(mod2)}{vi:i0(mod2)}{wj:j0(mod2)}{xj:j1(mod2)}.

For even n,

A={ui:i1(mod2)}{vi:i0(mod2)}{wj:j0(mod2)}{xj:j1(mod2)}andB={ui:i0(mod2)}{vi:i1(mod2)}{wj:j1(mod2)}{xj:j0(mod2)}.

Since f(A)={1,2,,n+m+1} and f(B)={n+m+2,n+m+3,,2(n+m+1)}, so Gn,m is a consecutively SEM graph. Thus, for any n,m1 and nm, μc(Gn,m)=0. □

As an illustration of the proof of Theorem 2.2, see the Fig. 1.

Figure 1.

Figure 1

The (consecutively) SEM labeling of Cb3 ∪ Cb6 and Cb4 ∪ Cb6.

A subdivision of a star K1,r, denoted by T(n1,n2,,nr), where ni1, 1ir, is a graph obtained by inserting ni1 vertices to each edge of the star K1,r. Now, we study (consecutively) SEMD of disjoint union of two isomorphic of subdivision of K1,3 and K1,4 in the following theorems.

Theorem 2.3

Let Gn=2T(n,n1,n1). For any even n2, μs(Gn)=μc(Gn)=0 and for any odd n3, μs(Gn)1.

Proof

Let Gn be a graph with

V(Gn)={y,y1,j,y2,l,y3,l:1jn,1ln1}{z,z1,j,z2,l,z3,l:1jn,1ln1}andE(Gn)={yyi,1,zzi,1:1i3}{y1,jy1,j+1,zi,jzi,j+1:1jn1}{y2,ly2,l+1,y3,ly3,l+1,z2,lz2,l+1,z3,lz3,l+1:1ln2}.

Thus, Gn is a graph of order 6n2 and size 6n4.

Let n2 be an even integer. Define a labeling f:V(Gn){1,2,,6n2} as follows.

f(y)=12(7n),f(z)=12(11n4).
f(a)={12(nj+1),ifa=y1,jandjis odd,12(7nj),ifa=y1,jandjnis even,6n2,ifa=y1,n,12(n+l+1),ifa=y2,landlis odd,12(7n+l),ifa=y2,landlis even,12(3nl+1),ifa=y3,landlis odd,12(9nl2),ifa=y3,landlis even,12(3n+j+1),ifa=z1,jandjis odd,12(9n+j4),ifa=z1,jandjis even,12(5nl+1),ifa=z2,landlis odd,12(11nl4),ifa=z2,landlis even,12(5n+l+1),ifa=z3,landlis odd,12(11n+l4),ifa=z3,landlis even.

The set of all edge-sums generated by f is S={3n+2,3n+3,,9n3}. By Lemma 1.1, f extends to a SEM labeling of Gn with magic constant k=15n4. Hence, for any even n2, μs(Gn)=0.

Moreover, let A and B be partite sets of Gn, where

A={y1,j,y2,l,y3,l:j,l1(mod2)}{z1,j,z2,l,z3,l:j,l1(mod2)}andB={y,y1,j,y2,l,y3,l:j,l0(mod2)}{z,z1,j,z2,l,z3,l:j,l0(mod2)}.

It is easy to see that f(A)={1,2,,3n} and f(B)={3n+1,3n+2,,6n2}. So, Gn is a consecutively SEM graph. Thus, μc(Gn)=0 for any even n2.

Now, to show μs(Gn)1 for any odd n3, let us consider the graph Hn=GnK1, where V(Hn)=V(Gn){w} and E(Hn)=E(Gn). Define a labeling g:V(Hn){1,2,,6n1} as follows.

g(y)=12(7n+1),g(z)=6n2,g(w)=12(11n3).
g(a)={12(nj+2),ifa=y1,jandjis odd,12(7nj+1),ifa=y1,jandjis even,12(n+l+2),ifa=y2,landlis odd,12(7n+l+1),ifa=y2,landln1is even,6n1,ifa=y2,n1,12(3nl),ifa=y3,landlis odd,12(9nl1),ifa=y3,landlis even,12(4nj+1),ifa=z1,jandjis odd,12(10nj2),ifa=z1,jandjis even,12(4n+l+1),ifa=z2,landlis odd,12(10n+l4),ifa=z2,landlis even,3n+l3,ifa=z3,l,lis odd,andn=5,3n1,ifa=z3,1andn5,12(12n+l11),ifa=z3,l,l=3,5,andn5,12(11n+l8),ifa=z3,l,l7is odd,27,ifa=z3,2andn=5,14,ifa=z3,4andn=5,3n,ifa=z3,2andn5.

To label the rest of vertices, we consider the following cases.

  • (i)
    For n3(mod4) and n>5, set
    g(z3,l)={12(6nl2)forl=4r,1rn34,12(6nl+2)forl=4r+2,1rn34.
  • (ii)
    For n1(mod4) and n>5, set
    g(z3,l)={3nl+2forl=2r+2,1rn14,12(7n2l+1)forl=12(n+4r+3),1rn54.

The set of all edge-sums generated by g is S={3n+2,3n+3,,9n3}. By Lemma 1.1, g extends to a SEM labeling of Hn with magic constant k=15n3. Hence, for any odd n3, μs(Gn)1. □

Theorem 2.4

Let Gn=2T(n,n1,n1,n). For any n2, μs(Gn)=0.

Proof

The graph Gn is a graph of order 8n2 and size 8n4. Let

V(Gn)={y,y1,j,y2,l,y3,l,y4,j:1jn,1ln1}{z,z1,j,z2,l,z3,l,z4,j:1jn,1ln1}andE(Gn)={yyi,1,zzi,1:1i4}{yi,jyi,j+1,zi,jzi,j+1:i=1,4;1jn1}{yi,lyi,l+1,zi,lzi,l+1:i=2,3;1ln2}.

Define a labeling f:V(Gn){1,2,,8n2} as follows.

f(y)=5n+1,
f(a)={12(j+1),ifa=y1,jandjis odd,12(8n+j+2),ifa=y1,jandjis even,12(2nl+1),ifa=y2,landlis odd,12(10nl+2),ifa=y2,landlis even,12(2n+l+1),ifa=y3,landlis odd,12(10n+l+2),ifa=y3,landlis even,12(4nj+1),ifa=y4,jandjis odd,12(12nj+2),ifa=y4,jandjis even.

To label the vertices of the second component of Gn, we consider two following cases.

Case 1. n is even.

For n=4, set f(z)=27 and

f(z1,1)=10,f(z1,2)=26,f(z1,3)=9,f(z1,4)=25,f(z2,1)=11,f(z2,2)=28,f(z2,3)=12,f(z3,1)=14,f(z3,2)=29,f(z3,3)=13,f(z4,1)=17,f(z4,2)=16,f(z4,3)=30,f(z4,4)=15.

For n4, set f(z)=7n and

f(a)={12(4n+j+1),ifa=z1,jandjis odd,12(12n+j),ifa=z1,jandjis even,12(6nl+1),ifa=z2,landlis odd,12(14nl),ifa=z2,landlis even,12(6n+l+1),ifa=z3,landlis odd,12(14n+l),ifa=z3,landlis even,4n,ifa=z4,1,12(16n+j9),ifa=z4,jandj=3,5,12(15n+j7),ifa=z4,jandj7is odd,4n+1,ifa=z4,2.

For even j4, label z4,j as follows.

  • (i)
    For n2(mod4),
    f(z4,j)={12(8nj)forj=4r,1rn24,12(8nj+4)forj=4r+2,1rn24.
  • (ii)
    For n0(mod4),
    f(z4,j)={4nj+3forj=2r+2,1rn4,12(9n2j+4)forj=12(n+4r+4),1rn44.

Case 2. n is odd.

Set f(z)=7n and

f(a)={12(4n+j+1),ifa=z1,jandjis odd,12(12n+j),ifa=z1,jandjis even,12(6nl+1),ifa=z2,landlis odd,12(14nl),ifa=z2,landlis even,4n+1,ifa=z3,1,12(16nl1),ifa=z3,landl3is odd,12(8nl+2),ifa=z3,landl3is even,12(6n+j+1),ifa=z4,jandjis odd,12(14n+j),ifa=z4,jandjis even.

For all cases, the set of all edge-sums generated by the labeling f is S={4n+3,4n+4,,12n2}. By Lemma 1.1, f extends to a SEM labeling of Gn with magic constant k=20n3. Therefore, for any n2, μs(Gn)=0. □

To clarify the proof of Theorem 2.3, Theorem 2.4, see Fig. 2.

Figure 2.

Figure 2

The (consecutively) SEM labeling of 2T(6,5,5) and 2T(7,6,6,7).

Next, we give the upper bound of (consecutively) SEMD of Cbn1T(n,n1,n1) and Cbn1T(n,n1,n1,n).

Theorem 2.5

Let Gn=Cbn1T(n,n1,n1). For any n2, μs(Gn)=μc(Gn)n22.

Proof

Let Gn be a graph having vertex and edge sets as follows.

V(Gn)={ui:0in1}{vi:1in1}{y,y1,r,y2,s,y3,s:1rn,1sn1}andE(Gn)={uiui+1:0in2}{uivi:1in1}{yyj,1,y1,ry1,r+1,y2,sy2,s+1,y3,sy3,s+1:1j3,1rn1,1sn2}.

For any n2, let Hn=Gnn22K1 be a graph with

V(Hn)=V(Gn){zt:1tn22}andE(Hn)=E(Gn).

Thus, Hn has 11n62 vertices and 5n4 edges.

For n=2, label (u0,u1,v1) and (y,y1,1,y1,2,y2,1,y3,1) with (6,1,7) and (8,4,5,2,3), respectively. These vertex labelings can be extended to a SEM labeling of H2 with magic constant k=21.

For n3, define a labeling f:V(Hn){1,2,,11n62} as follows.

f(ui)={3n+i1for even i,ifor odd i.f(vi)={3n+i1for odd i,ifor even i.
f(a)={12(2n+r1),ifa=y1,randris odd,12(4ns1),ifa=y2,sandsis odd,12(4n+s1).ifa=y3,sandsis odd.

Case 1. n is even.

f(y)=5n1, f(z1)=4n1, f(zt)=12(5n+2t2) for t=2,3,,n22, and

f(b)={12(8n+r2),ifb=y1,randris even,12(10ns2),ifb=y2,sandsis even12(10n+s2),ifb=y3,sandsn2is even,12(5n),ifb=y3,n2.

Case 2. n is odd.

f(y)=5n2, f(zt)=12(5n+2t1) for t=1,2,,n22, and

f(b)={12(8n+r4),ifb=y1,randris even12(10ns4),ifb=y2,sandsis even12(10n+s4),ifb=y3,sandsn1is even,12(5n1),ifb=y3,n1.

For all cases, it can be verified that the set of all edge-sums generated by the labeling f is S={3n,3n+1,,8n5}. By Lemma 1.1, f extends to a SEM labeling of Hn with magic constant k=27n142. Hence, μs(Gn)n22 for any n2.

Furthermore, let A and B be partite sets of Hn where

A={ui:i1(mod2)}{vi:i0(mod2)}{y1,r,y2,s,y3,s:r,s1(mod2)}andB={ui:i0(mod2)}{vi:i1(mod2)}{y,y1,r,y2,s,y3,s:r,s0(mod2)}{zt:1tn22}.

Since for any n2, f(A)={1,2,,5n22} and f(B)={5n2,5n+22,,11n62}, then Hn is a consecutively SEM graph. Thus, μc(Gn)n22. □

Theorem 2.6

Let Gn=Cbn1T(n,n1,n1,n). For any n2,

μs(Gn){n1,ifnis even,n2,ifnis odd.

Proof

Define the vertex and edge sets of Gn as follows.

V(Gn)={ui:0in1}{vi:1in1}{y,y1,r,y2,s,y3,s,y4,r:1rn,1sn1}andE(Gn)={uiui+1:0in2}{uivi:1in1}{yyj,1,y1,ry1,r+1,y2,sy2,s+1,y3,sy3,s+1,y4,ry4,r+1:1j4,1rn1,1sn2}.

For any even n2, let Hn=Gn(n1)K1 be a graph with

V(Hn)=V(Gn){zt:1tn1}andE(Hn)=E(Gn).

Next, define a labeling f:V(Hn){1,2,,7n3} as follows.

f(ui)={4n+ifor even i,ifor odd i.f(vi)={4n+ifor odd i,ifor even i.
f(zt)=3n+tfort=1,2,,n1.

For n=4, set f(y)=22 and

f(y1,1)=5,f(y1,2)=21,f(y1,3)=4,f(y1,4)=20,f(y2,1)=6,f(y2,2)=23,f(y2,3)=7,f(y3,1)=9,f(y3,2)=24,f(y3,3)=8,f(y4,1)=12,f(y4,2)=11,f(y4,3)=28,f(y4,4)=10.

For n4, set f(y)=6n1 and

f(a)={12(2n+r1),ifa=y1,randris odd,12(10n+r2),ifa=y1,randris even,12(4ns1),ifa=y2,sandsis odd,12(12ns2),ifa=y2,sandsis even,12(4n+s1),ifa=y3,sandsis odd,12(12n+s2),ifa=y3,sandsis even,3n1,ifa=y4,1,12(14n+r11),ifa=y4,randr=3,5,12(13n+r9),ifa=y4,randr7is odd,3n,ifa=y4,2.

For even r4, label y4,r as follows.

  • (i)
    For n2(mod4),
    f(y4,r)={12(6nr2)forr=4l,1ln24,12(6nr+2)forr=4l+2,1ln24.
  • (ii)
    For n0(mod4),
    f(y4,r)={3nr+2forr=2l+2,1ln4,12(7n2r+2)forr=12(n+4l+4),1ln44.

It can be checked that the set of all edge-sums is S={4n+1,4n+2,,10n4}. Hence by Lemma 1.1, f extends to a SEM labeling of Hn with magic constant k=17n6. Therefore, μs(Gn)n1 for any even n2.

Now, for any odd n3, let Hn=Gn(n2)K1 be a graph with

V(Hn)=V(Gn){zt:1tn2}andE(Hn)=E(Gn).

Next, define a labeling g:V(Hn){1,2,,7n4} as follows.

g(ui)={4n+i2for even i,ifor odd i.g(vi)={4n+i2for odd i,ifor even i.
g(y)=6n3andg(zt)=3n+t1fort=1,2,,n2.
g(a)={12(2n+r1),ifa=y1,randris odd,12(10n+r6),ifa=y1,randris even,12(4ns1),ifa=y2,sandsis odd,12(12ns6),ifa=y2,sandsis even,12(4n+s1),ifa=y3,sandsis odd,12(12n+s6),ifa=y3,sandsis even,3n1,ifa=y4,1,12(14nr5),ifa=y4,randr3is odd,12(6nr2),ifa=y4,randris even.

Under the labeling g, the set of all edge-sums is S={4n1,4n,,10n6}. Hence by Lemma 1.1, g extends to a SEM labeling of Hn with magic constant k=17n9. Therefore, μs(Gn)n2 for any odd n. □

Fig. 3 shows an illustration of the proof of Theorem 2.5, Theorem 2.6.

Figure 3.

Figure 3

The (consecutively) SEM labeling of Cb7 ∪ T(8,7,7)∪3K1 and the SEM labeling of Cb6 ∪ T(7,6,6,7)∪5K1.

3. The SEMD of 2C3Cn and related graphs

Enomoto et al. [2] stated that a cycle Cn is SEM if and only if n is odd. Figueroa-Centeno et al. [6] investigated the SEMD of the cycle Cn. They also proved the following result.

Theorem 3.1

[6] If G is a graph of size q2(mod4) such that every vertex in V(G) has even degree then μs(G)=+.

In 2005, Figueroa-Centeno et al. [11] investigated the SEMD of 2Cn, 3Cn, 4Cn, and conjectured that

μs(mCn)={0,ifmn1(mod2),1,ifmn0(mod4),+,ifmn2(mod4).

In 2009, Holden et al. [16] showed that C4(2t1)C3 for any integer t3, C5(2t)C3 for any integer t3, and C7(2t)C3 for any integer t1 have a strong vertex-magic labeling, which is equivalent to a SEM labeling. Based on their results, they proposed the following conjecture.

Conjecture 3.2

[16] A 2-regular graph of odd order is SEM if and only if it is not one of C3C4, 3C3C4, or 2C3C5.

Motivated by Conjecture 3.2, Figueroa-Centeno et al. [17] showed that some 2-regular graphs with two components are SEM. They proved that C3Cn is SEM if and only if n4 is even, C4Cn is SEM if and only if n5 is odd, C5Cn is SEM if and only if n4 is even, and CmCn is SEM for any even m4 and odd n12(m+4). Ichishima and Oshima [18] determined the SEMD of CmCn for even m and n, and for arbitrary n when m=3,4,5 and 7. In this section, we investigate the SEMD of 2C3Cn.

Theorem 3.3

The SEMD of a 2-regular graph 2C3Cn is given by

μs(2C3Cn)={0,ifn1,3(mod4)andn5,1,ifn2(mod4),+,ifn0(mod4).

and μs(2C3C5)2.

Proof

First, let 2C3Cn be a graph with vertex and edge sets

V(2C3Cn)={u1,u2,u3}{v1,v2,v3}{wj:1jn}and E(2C3Cn)={u1u2,u2u3,u3u1}{v1v2,v2v3,v3v1}{wjwj+1,wnw1:1jn1}.

Next, we show that μs(2C3Cn)=0 for n1(mod4). Let n=4r+1 where r2. Define a labeling f:V(2C3Cn){1,2,,4r+7} as follows.

f(u1)=2r,f(u2)=2r+2,f(u3)=2r+3,f(v1)=3r+4,f(v2)=3r+5,f(v3)=3r+6,
f(wj)={ifor j=2i1 and 1ir1,2r+i+3for j=2i and 1ir1,3r+3for j=2r1.

To label wj, for 2rj4r+1, we consider the following cases.

Case 1. r2 is even.

f(wj)={r+2i1for j=2r+4i4 and 1ir2,3r+2i+6for j=2r+4i3 and 1i12(r2),r+2i2for j=2r+4i2 and 1i12(r2),3r+2i+5for j=2r+4i1 and 1i12(r2),4ri+8for j=4r+2i5 and 1i3,2r3i+4for j=4r+2i4 and 1i2.

Case 2. r3 is odd.

f(wj)={r+2i1for j=2r+4i4 and 1i12(r1),r+2i2for j=2r+4i2 and 1i12(r+1),3r+2i+7for j=2r+4i1 and 1i12(r1),3r+2i+6for j=2r+4i+1 and 1i12(r3),4r2i+9for j=4r+2i3 and 1i2,3r+7for j=2r+12r+1for j=4r2.

It can be checked that for all cases, the set of all edge-sums is S={2r+5,2r+6,,6r+11}. By Lemma 1.1, f extends to a SEM labeling of 2C3Cn with magic constant k=10r+19. Hence, μs(2C3Cn)=0 for n1(mod4).

Next, we prove that μs(2C3Cn)=0 for n3(mod4). Let n=4r+3, r0. For r=0, label (u1,u2,u3),(v1,v2,v3), and (w1,w2,w3) with (1,5,9),(2,6,7), and (3,4,8), respectively. For r1, define a labeling f:V(2C3Cn){1,2,,4r+9} as follows.

f(u1)=2r+2,f(u2)=2r+3,f(u3)=2r+4,f(v1)=3r+5,f(v2)=3r+6,f(v3)=3r+7,
f(wj)={ifor j=2i1and1i2r+1,2r+i+5for j=2iand1ir1,2r+i+8for j=2iandri2r+1,2r+5for j=n=4r+3.

It is not hard to verify that the set of all edge-sums is S={2r+6,2r+7,,6r+14}. Hence by Lemma 1.1, f extends to a SEM labeling of 2C3Cn with magic constant k=10r+24. Therefore, μs(2C3Cn)=0 for n3(mod4).

Now, let G=2C3CnK1 for n=4r+2, r1. Define a labeling f:V(G){1,2,,4r+9} as follows.

f(u1)=2r+4,f(u2)=2r+5,f(u3)=4r+8,f(z)=3r+7,

where z is the vertex of K1. For the other vertices, we consider the following cases.

Case 1. For any odd r1, set

f(v1)=r+2,f(v2)=r+4,f(v3)=3r+8,and
f(wj)={ifor j=2i1 and 1i2,2r+7for j=2,4r+9for j=4,2r2i+4for j=4i+1 and 1ir,4r2i+8for j=4i+2 and 1i12(r1),2r2i+5for j=4i+3 and 1i12(r1),4r2i+9for j=4i+4 and 1i12(r1),3r2i+7for j=2r+4i and 1i12(r+1),3r2i+8for j=2r+4i+2 and 1i12(r1),r2i+2for j=2r+4i+4 and 1i12(r1).

Case 2. For any even r2, set

f(v1)=r+1,f(v2)=r+5,f(v3)=3r+9,and
f(wj)={ifor j=2i1 and 1i12(r+2),2r+2i+5for j=4i2 and 1ir2,2r+2i+6for j=4i and 1ir,3r+2i+9for j=2r+4i2 and 1ir2,2r+2i2for j=4r+2i5 and 1i2,r+3for j=4r+1,2r+6for j=n=4r+2.

To label the rest of vertices, we consider two subcases.

  • Subcase 2.1. For r2(mod4), set
    f(wj)={12(3r2)for j=r+3,12(r+2i+2)for j=r+2i+3 and 1i12(r4),r+2i+5for j=2r+4i3 and 1i14(r2),r+2i2for j=2r+4i1 and 1i14(r2),12(3r+4i2)for j=3r+4i5 and 1i14(r+2),12(3r+4i+8)for j=3r+4i3 and 1i14(r2).
  • Subcase 2.2. For r0(mod4), set
    f(wj)={12(r+2i+4)for j=r+2i+1 and 1i12(r4),r+2ifor j=2r+4i5 and 1ir4,r+2i+5for j=2r+4i3 and 1i14(r4),12(r+4)for j=3r3,12(3r+4i+6)for j=3r+4i5 and 1ir4,12(3r+4i)for j=3r+4i3 and 1ir4.

It can be checked that under the labeling f, the set of all edge-sums is S={2r+6,2r+7,,6r+13}. By Lemma 1.1, f extends to a SEM labeling of G with magic constant k=10r+23. Thus, μs(2C3Cn)=1 for n2(mod4).

To show that μs(2C3C5)2, label {u1,u2,u3}{v1,v2,v3}{w1,w2,w3,w4,w5}{2K1} with {3,7,9}{5,6,12}{1,8,11,2,13}{4,10}, respectively. Finally, μs(2C3Cn)=+ for n0(mod4) is a direct consequence of Theorem 3.1. □

In 2013, Ichishima et al. [9] introduced the concept of a pseudo SEM graph. A graph G with isolated vertices is called pseudo SEM if there exists a bijection f:V(G){1,2,...,|V(G)|} such that the set {f(u)+f(v):uvE(G)}{2f(u):deg(u)=0} consists of |E(G)|+|{uV(G):deg(u)=0}| consecutive integers. Such a function is called a pseudo SEM labeling. They also proved the following results.

Theorem 3.4

[9] Let m be an odd integer. If G=(i=1k1Cni)k2K1 is any pseudo SEM 2-regular graph then H=(i=1k1(m,ni)C[m,ni])k2Cm is a SEM 2-regular graph, where (m,ni) and [m,ni] are the greatest common divisor and the least common multiple of integers m and ni, respectively.

Corollary 3.5

[9] Let m be an odd integer such that (m,ni)=1 and (m,ni)=ni, 1ik1. If G=(i=1k1Cni)k2K1 is any pseudo super edge-magic graph then H=(i=1k1Cmni)k2Cm is a SEM graph.

By applying Theorem 3.4 and Corollary 3.5 to a part of Theorem 3.3, we have the following Lemma.

Lemma 3.6

For any odd m and n2(mod4), the following SEMD of some graphs are hold.

(i)μs(2C3m(m,n)C[m,n]Cm)=0.(iii)μs(2C3mC5m2Cm)=0.(ii)μs((n+6)PmCm)=0.(iv)μs(12PmCm)=0.

Proof

It is easy to verify that the labeling f in the proof of Theorem 3.3 is a pseudo SEM labeling of 2C3CnK1 for n2(mod4) and 2C3C52K1. By applying Theorem 3.4 and Corollary 3.5 to these 2-regular pseudo SEM graphs, we obtain (i) and (iii). By removing n+6 edges of the SEM 2-regular graph 2C3m(m,n)C[m,n]Cm with the smallest edge-sums {a,a+1,,a+n+5}, where a=12[(n+10)+(m1)(n+7)], we get (ii). In a similar way, we obtain the result (iv) by removing 12 edges of the SEM 2-regular graph 2C3mC5m2Cm with the smallest edge-sums {b,b+1,,b+11}, where b=12(13m+3). □

As an illustration of the proof of Lemma 3.6 (i) and (ii), see Fig. 4.

Figure 4.

Figure 4

(a) The pseudo SEM labeling of 2C3 ∪ C10 ∪ K1; (b) The SEM labeling of 2C9 ∪ C30 ∪ C3, which is obtained by applying Corollary 3.5 to Fig. 4(a) for m = 3; (c) The SEM labeling of 16P3 ∪ C3, which is obtained by removing 16 edges with the smallest edge-sums {27,28,⋯,42} of the graph in Fig. 4(b).

Cichacz et al. [10] in 2017 introduced a new method to expand the classes of known (strong) vertex-magic labeling of 2-regular graphs as the following theorem.

Theorem 3.7

[10] For 1it, let ni3 be an integer and G=i=1tCni. If G is (strong) vertex-magic then H=i=1tCni(2r+1) is (strong) vertex-magic for every integer r.

Ngurah [19] discovered that the method introduced in the proof of Theorem 3.7 is also valid for pseudo SEM graphs. By applying this fact to our results, we have the following lemma.

Lemma 3.8

For any odd m and n2(mod4), the following SEMD of some graphs are satisfied.

(i)μs(2C3mCnmCm)=0.(v)μs(2C3mC5mCm)m.(ii)μs(mP3C3mCnmCm)=0.(vi)μs(2C3mmP5Cm)m.(iii)μs(mP3C3mmPnCm)=0.(vii)μs(mP3C3mmP5Cm)m.(iv)μs(2C3mC5m2Cm)=0.(viii)μs(2mP3mP5Cm)m.

Proof

The result (i) and (iv) are obtained by applying Theorem 3.7 to the fact that 2C3CnK1 for n2(mod4) and 2C3C52K1 are pseudo SEM, respectively. We gain (ii) and (iii) from (i) by removing edges with the edge-sums {a,a+1,,a+m1} and {a,a+1,,a+2m1}, where a=12(nm+7m+3), respectively. Moreover, we obtain the result (v), (vi), (vii), and (viii) from (iv) by deleting edges with the edge-sums {b,b+1,,b+m1}, {b,b+1,,b+2m1}, {b,b+1,,b+3m1}, and {b,b+1,,b+4m1}, where b=12(13m+3), respectively. □

Fig. 5 shows the construction given in the proof of Lemma 3.8 (i), (ii), and (iii).

Figure 5.

Figure 5

(a) The SEM labeling of 2C9 ∪ C30 ∪ C3, which is obtained by applying Theorem 3.7 to Fig. 4(a) for r = 1; (b) The SEM labeling of 3P3 ∪ C9 ∪ C30 ∪ C3, which is obtained by removing 3 edges with the smallest edge-sums {27,28,29} in Fig. 5(a); (c) The SEM labeling of 3P3 ∪ C9 ∪ 3P10 ∪ C3, which is obtained by removing 3 edges with the smallest edge-sums {30,31,32} in Fig. 5(b).

4. The SEMD of join product of union of a star and a path with an isolated vertex

To present our results, we need the concept of dual labeling. A dual labeling f of a SEM labeling f is defined as f(x)=p+1f(x) for all xV(G) and f(xy)=2p+q+1f(xy) for all xyE(G). Baskoro et al. [20] proved that the dual of a SEM labeling is also a SEM labeling.

Ngurah and Simanjuntak [21] studied the SEMD of join products of a path, a star, and a cycle, respectively, with isolated vertices. Generally, they showed that the join product of a SEM graph with isolated vertices has finite SEMD. In [22], the same authors investigated the SEMD of join product of a graph G which has certain properties with an isolated vertex. They gave a necessary condition for G+K1 to have zero SEMD as the following lemma.

Lemma 4.1

[22] Let G be a graph with no cycle and minimum degree one. If μs(G+K1)=0 then G is a tree or a forest.

They showed that Lemma 4.1 is attainable. In particular, they proved that the join product of some forests with an isolated vertex has zero SEMD, such as μs([PnP2]+K1)=0 if and only if 3n5 and μs([K1,nP2]+K1)=0 if and only if n=2. In this section, we study the SEMD of join product graph (K1,nPm)+K1 for any integer n1 and m3.

For any integer n1 and m3, let Gn,m=(K1,nPm)+K1 be a graph having vertex and edge sets:

V(Gn,m)={c,z}{xi:1in}{yj:1jm}and E(Gn,m)={cz}{cxi,zxi:1in}{zyj:1jm}{yjyj+1:1lm1}.

Thus, the graph Gn,m has n+m+2 vertices and 2n+2m edges.

Now, we give necessary and sufficient conditions of Gn,m to have zero SEMD for m=3,4,5.

Theorem 4.2

μs(Gn,3)=0 if and only if n=1,2,3; and for m=4,5, μs(Gn,m)=0 if and only if n=1,2.

Proof

Firstly, we show that for n=1,2,3, μs(Gn,3)=0. For n=1,2,3, label (c,z),(x1,,xn),(y1,y2,y3) with (3,2),(1),(5,4,6); (2,3),(1,4),(6,5,7); and (1,4),(2,3,7),(5,8,6), respectively. These vertex labelings can be extended to a SEM labeling of Gn,3 for n=1,2,3.

Next, we prove that μs(Gn,3)>0 for every n4. Suppose that μs(Gn,3)=0 for every n4. Then by Lemma 1.1, there exists a bijection f:V(Gn,3){1,2,,n+5} such that S={f(x)+f(y):xyE(Gn,3)} is a set of 2n+6 consecutive integers. Since Gn,3 has n+5 vertices and 2n+6 edges, there are two possibilities of S, namely {3,4,,2n+8} or {4,5,,2n+9} and they are dual to each other. Thus, we can consider S={3,4,,2n+8}. Based on the degree of all vertices of Gn,3, the sum of all elements in S contains n+4 times of label of z, n+1 times of label of c, three times of label of y2 and two times of label of the remaining vertices. Hence,

(n+4)f(z)+(n+1)f(c)+3f(y2)+2[f(y1)+f(y3)+i=1nf(xi)]=sSs=(n+4)(2n+9)3

or

(n+2)f(z)+(n1)f(c)+f(y2)=n2+6n+3.

Furthermore, to get edge-sums 3, 4, and 5 in S, then the vertices of labels 1, 2 and 3 are adjacent to each other or the vertex of label 1 is adjacent to the vertices of labels 2, 3, and 4. By using the above equation and this fact, we have two following cases.

Case 1. There is a triangle with the vertices of labels 1, 2 and 3. Since every triangle in Gn,3 share a common vertex z, hence f(z){1,2,3} and f(c),f(y2){1,2,3}{f(z)}. We can check that the equation has no solution for f(z){1,2}. If f(z)=3 then (n1)f(c)+f(y2)=n2+3n3 and its solution is f(y2)=1 and f(c)=n+4. However, the solution does not lead to a SEM labeling of Gn,3 for any n4.

Case 2. The vertex of label 1 is adjacent to the vertices of label 2, 3, and 4. Since the vertex of label 1 must have minimum degree 3, the possibilities of its vertex are z, c or y2.

  • (i)

    If f(z)=1 then (n1)f(c)+f(y2)=n2+5n+3, where f(c),f(y2){2,3,,n+5}. It is easy to verify that this equation has no solution.

  • (ii)

    If f(c)=1 then (n+2)f(z)+f(y2)=n2+5n+4. It can be checked that this equation has no solution if f(z){2,3,,n+1}{n+3,n+4,n+5}. If f(z)=n+2 then f(y2)=n. However, this solution does not lead to a SEM labeling of Gn,3 for any n4.

  • (iii)

    If f(y2)=1 then (n+2)f(z)+(n1)f(c)=n2+6n+2. Since y2 has degree three, which is adjacent to z, thus f(z){2,3,4}. If f(z)=2 then f(c)>n+5=p. If f(z)=3 then either f(z)+f(y1)=f(y2)+f(y3)=5 or f(z)+f(y3)=f(y1)+f(y2)=5. If f(z)=4 then f(c)=n+33n1, thus n=4. However, this solution cannot be extended to a SEM labeling of G4,3.

Hence, μs(Gn,3)>0 for every n4. Therefore, μs(Gn,3)=0 if and only if n=1,2,3.

Now, we show that for m=4,5, μs(Gn,m)=0 if and only if n=1,2. For n=1,2 and m=4, label (c,z),(x1,,xn),(y1,y2,y3,y4) with (3,2),(1),(4,6,5,7) and (2,3),(1,4),(5,7,6,8), respectively. Moreover, for n=1,2 and m=5, label (c,z),(x1,,xn),(y1,y2,y3,y4,y5) with (2,1),(3),(4,7,5,8,6) and (3,2),(1,4),(5,8,6,9,7), respectively. It can be checked that these labelings extend to a SEM labeling of Gn,4 and Gn,5 for n=1,2. Conversely, suppose that μs(Gn,4)=μs(Gn,5)=0 for n3. Then for m=4, we have

(n+3)f(z)+(n1)f(c)+f(y2)+f(y3)=n2+8n+10

and for m=5,

(n+4)f(z)+(n1)f(c)+f(y2)+f(y3)+f(y4)=n2+10n+19.

By a similar argument as in the proof of Gn,3, we can show that μs(Gn,4)>0 and μs(Gn,5)>0 for every n3. Hence, for m=4,5, μs(Gn,m)=0 if and only if n=1,2. □

Since Gn,3, Gn,4, and Gn,5 do not SEM for almost all of n, thus we try to find its SEMD. Our result is as follows.

Theorem 4.3

For any integer m,n3, the SEMD of Gn,m is given by:

  • (i)

    μs(Gn,m)=1 if m=3 and n=4,5; m=4 and n=3,4,5,6; and m=5 and n=3,4,5.

  • (ii)

    μs(Gn,m)n1 if m3 and n=rm12, where r is any positive integer.

Proof

  • (i)
    As a consequence of Theorem 4.2, we obtain that μs(Gn,3)1 for n4, μs(Gn,4)1 for n3, and μs(Gn,5)1 for n3. Hence, the remaining case is to show that μs(Gn,3)1 for n=4,5; μs(Gn,4)1 for n=3,4,5,6; and μs(Gn,5)1 for n=3,4,5. Thus, let us consider the graph Gn,mK1 and suppose w is the vertex of K1.
    • For n=4,5 and m=3, label (c,z),(x1,x2,,xn),(y1,y2,y3),(w) with (1,5),(2,3,4,9),(7,6,10),(8) and (1,6),(2,3,4,5,11),(7,8,10),(9), respectively.
    • For n=3,4,5,6 and m=4, label (c,z),(x1,x2,,xn),(y1,y2,y3,y4),(w) with (1,4),(2,3,7),(6,10,5,8),(9); (1,5),(2,3,4,9),(6,11,7,8),(10); (1,6),(2,3,4,5,11),(8,12,7,9),(10); and (1,7),(2,3,4,5,6,13),(8,10,12,9),(11), respectively.
    • For n=3,4,5 and m=5, label (c,z),(x1,x2,,xn),(y1,y2,,y5),(w) with (1,4),(2,3,7),(5,8,10,6,11),(9); (1,5),(2,3,4,9),(6,10,8,12,7),(11); and (1,6),(2,3,4,5,11),(7,9,13,8,12),(10), respectively.
    It is not hard to check that these labelings extend to a SEM labeling of Gn,3K1 for n=4,5; Gn,4K1 for n=3,4,5,6; and Gn,5K1 for n=3,4,5.
  • (ii)
    Let H=Gn,m(n1)K1, where m3 and n=rm12. Define a labeling f:V(H){1,2,,2n+m+1} as follows.
    f(c)=m+1,f(z)=3m+12,
    f(xi)={(2l+1)m2l+52+i((l1)(m12)+1), if mis odd,(2l+1)m4l+62+i((l1)(m22)+1), if m is even,
    where (l1)m12+1ilm12, 1lr. For 1jm,
    f(yj)={lforj=2l1and1lm+12,m+12+lforj=2land1lm2.
    The remaining labels are used to label isolated vertices of H. It can be verified that the set of all edge-sums is a consecutive integer S={a,a+1,,a+2n+2m1}, where a=m+52. By Lemma 1.1, f extends to a SEM labeling of H with magic constant k=4n+3m+a+1. Therefore, for any m3 and n=rm12, rZ+, μs(Gn,m)n1. □

5. Conclusion

In this paper, we study the (consecutively) SEMD of some graphs. We find the exact value or upper bound of the (consecutively) SEMD of forests with two components. We also find the exact value of the SEMD of a 2-regular graph 2C3Cn for almost all of n. By using this and previous known results, we obtain the exact value or upper bound of the SEMD of some 2-regular graphs and union of cycles and paths. Moreover, we provide the necessary and sufficient conditions of (K1,nPm)+K1 to gain zero SEMD and the upper bound of SEMD of this graph for the remaining cases.

Declarations

Author contribution statement

V.H. Krisnawati: Conceived and designed the analysis; Analyzed and interpreted the data; Wrote the paper.

A.A.G. Ngurah: Analyzed and interpreted the data; Contributed analysis tools or data; Wrote the paper.

N. Hidayat, A.R. Alghofari: Conceived and designed the analysis; Contributed analysis tools or data.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of interests statement

The authors declare no conflict of interest.

Additional information

No additional information is available for this paper.

Acknowledgements

The authors would like to thank to the reviewers for their valuable comments and suggestions.

Contributor Information

Vira Hari Krisnawati, Email: virahari@ub.ac.id.

Anak Agung Gede Ngurah, Email: aag.ngurah@unmer.ac.id.

References

  • 1.Kotzig A., Rosa A. Magic valuation of finite graphs. Can. Math. Bull. 1970;13:451–461. [Google Scholar]
  • 2.Enomoto H., Llado A., Nakamigawa T., Ringel G. Super edge-magic graphs. SUT J. Math. 1998;34:105–109. [Google Scholar]
  • 3.Muntaner-Batle F.A. Special super edge-magic labelings of bipartite graphs. J. Comb. Math. Comb. Comput. 2001;39:107–120. [Google Scholar]
  • 4.Oshima A. Consecutively super edge-magic tree with diameter 4. Rev. Bull. Calcutta Math. Soc. 2007;15:87–90. [Google Scholar]
  • 5.Figueroa-Centeno R., Ichishima R., Muntaner-Batle F.A. The place of super edge-magic labelings among other classes of labelings. Discrete Math. 2001;231:153–168. [Google Scholar]
  • 6.Figueroa-Centeno R., Ichishima R., Muntaner-Batle F.A. On the super edge-magic deficiency of graphs. Electron. Notes Discrete Math. 2002;11:299–314. [Google Scholar]
  • 7.Ichishima R., Muntaner-Batle F.A., Oshima A. The consecutively super edge-magic deficiency of graphs and related concepts. Electron. J. Graph Theory Appl. 2020;8:71–92. [Google Scholar]
  • 8.Gallian J.A. A dynamic survey of graph labeling. Electron. J. Comb. 2019;DS6 [Google Scholar]
  • 9.Ichishima R., Muntaner-Batle F.A., Oshima A. Enlarging the classes of super edge-magic 2-regular graphs. AKCE Int. J. Graphs Comb. 2013;340:129–146. [Google Scholar]
  • 10.Cichacz S., Froncek D., Singgih I. Vertex magic total labelings of 2-regular graphs. Discrete Math. 2017;340:3117–3124. [Google Scholar]
  • 11.Figueroa-Centeno R., Ichishima R., Muntaner-Batle F.A. Some new results on the super edge-magic deficiency of graphs. J. Comb. Math. Comb. Comput. 2005;55:17–31. [Google Scholar]
  • 12.Baig A.Q., Ahmad A., Baskoro E.T., Simanjuntak R. On the super edge-magic deficiency of forests. Util. Math. 2011;86:147–159. [Google Scholar]
  • 13.Javed S., Riasat A., Kanwal S. On super edge-magicness and deficiencies of forests. Util. Math. 2015;98:149–169. [Google Scholar]
  • 14.Imran M., Mukhtar A. On super edge-magic total labeling of forests consisting of stars and subdivided stars. Int. J. Math. Soft Comput. 2017;7:1–14. [Google Scholar]
  • 15.Krisnawati V.H., Ngurah A.A.G., Hidayat N., Alghofari A.R. On the super edge-magic deficiency of forests. AIP Conf. Proc. 2019;2192 doi: 10.1016/j.heliyon.2020.e05561. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Holden J., McQuillan D., McQuillan J.M. A conjecture on strong magic labelings of 2-regular graphs. Discrete Math. 2009;309:4130–4136. [Google Scholar]
  • 17.Figueroa-Centeno R., Ichishima R., Muntaner-Batle F.A., Oshima A. A magical approach to some labeling conjectures. Discuss. Math., Graph Theory. 2011;31:79–113. [Google Scholar]
  • 18.Ichishima R., Oshima A. On the super edge-magic deficiency of 2-regular graphs with two components. Ars Comb. 2016;129:437–447. [Google Scholar]
  • 19.Ngurah A.A.G. On (super) edge-magic deficiency of some classes of graphs. Electron. J. Graph Theory Appl. 2019 submitted for publication. [Google Scholar]
  • 20.Baskoro E.T., Sudarsana I.W., Cholily Y.M. How to construct new super edge-magic graphs from some old ones. J. Indones. Math. Soc. 2005;11:155–162. [Google Scholar]
  • 21.Ngurah A.A.G., Simanjuntak R. Super edge-magic deficiency of join product graphs. Util. Math. 2017;105:279–289. [Google Scholar]
  • 22.Ngurah A.A.G., Simanjuntak R. On the super edge-magic deficiency of join product and chain graphs. Electron. J. Graph Theory Appl. 2019;7:157–167. [Google Scholar]

Articles from Heliyon are provided here courtesy of Elsevier

RESOURCES