Skip to main content
Elsevier - PMC COVID-19 Collection logoLink to Elsevier - PMC COVID-19 Collection
. 2020 Sep 6;1223:129210. doi: 10.1016/j.molstruc.2020.129210

On neighborhood Zagreb index of product graphs

Sourav Mondal a, Nilanjan De b,, Anita Pal a
PMCID: PMC7474663  PMID: 32921807

Abstract

The properties and activities of chemicals are strongly related to their molecular structures. Topological indices defined on these molecular structures are capable to predict those properties and activities. In this article, a new topological index named as neighborhood Zagreb index (MN) is presented. Here the chemical importance of the MN index is investigated and it is shown that the newly introduced index is useful in predicting physico-chemical properties with high accuracy compared to some well-established and often used indices. The isomer-discrimination ability of MN is also examined. To demonstrate how the computational formula of the novel index for chemical compounds is simple and convenient, the chemical structures of favipiravir and hydroxychloroquine are used. In addition, some explicit results for this index of different product graphs such as Cartesian, tensor and wreath product are derived. Some of these results are applied to obtain the MN index of some special structures.

Keywords: Molecular graph, Molecular descriptor, Neighborhood Zagreb index, Cartesian product, Tensor product, Wreath product

1. Introduction

A molecular graph [12], [32] is a connected graph where loops and parallel connections are not allowed and in which nodes and edges are supposed to be atoms and chemical bonds of compound respectively. Throughout this work, we use only molecular graphs. For the node and edge sets of a graph G, we consider the notations V(G) and E(G), respectively. The degree (valency) of a node u, written as degG(u), is the total count of edges associated with u. The set of neighbors of a node u is written as NG(u). For molecular graph, |NG(u)| = degG(u).

In mathematical chemistry, molecular descriptors play a leading role specifically in the field of quantitative structure property relationship/quantitative structure activity relationship modeling. Amongst them, an outstanding area is preserved for the well-known topological indices or graph invariants. A real valued mapping considering graph as an argument is called a graph invariant if it gives the same value to isomorphic graphs. The order(total count of nodes) and size(total count of edges) of a graph are examples of two graph invariants. In mathematical chemistry, the graph invariants are named as topological indices. Some familiar topological indices are Wiener index, Randić index, connectivity indices, Zagreb indices etc. The idea of topological indices was initiated when the eminent chemist Harold Wiener found the first topological index, known as Wiener index [6], in 1947 for searching boiling points of alkanes. Amidst the topological indices invented on initial stage, the Zagreb indices are associated with the most popular molecular descriptors. It was firstly presented by Gutman and Trinajestić [14], where they investigated how the total energy of π-electron depends on the structure of molecules and it was recognized on [13]. The first (M 1) and second (M 2) Zagreb indices are as follows:

M1(G)=uV(G)degG(u)2, (1)
M2(G)=uvE(G)degG(u)degG(v). (2)

For more discussion regarding the Zagreb indices, see the articles [4], [5], [11], [22]. In addition to the Zagreb indices, there are some other well-established and most used degree based topological indices such as forgotten topological index (F) [8], [14], Randić index (R) [30], sum connectivity index (SCI) [35] and symmetric division degree index (SDD) [33] to model different structure-property/structure-activity relationships, which are defined as follows.

F(G)=uV(G)degG(u)3, (3)
R(G)=uvE(G)1degG(u)degG(v), (4)
SCI(G)=uvE(G)1degG(u)+degG(v), (5)
SDD(G)=uvE(G)[degG(u)degG(v)+degG(v)degG(u)]. (6)

Let the degree sum of all nodes connected to u in G be denoted by δG(u), i.e.

δG(u)=vNG(u)degG(v). (7)

Following the construction of first Zagreb index as described in Eq. (1), we present here a novel index known as the neighborhood Zagreb index(MN) which is defined below.

MN(G)=uV(G)δG(u)2. (8)

In mathematical chemistry, graph operations are very significant since certain graphs of chemical interest can be evaluated by various graph operations of different simple graphs. H.Yousefi Azari and co-authors [3] derived some exact formulae of PI index for Cartesian product of bipartite graphs. P. Paulraja and V.S. Agnes [28] evaluated some explicit expressions of the degree distance for the Cartesian and wreath products. De et al. [25] found explicit expressions of the F-index under several graph operations. For further illustration on this area, interested readers are suggested some articles [1], [2], [9], [10], [15], [19], [24], [26], [29]. We continue this progress for MN index. The objective of this work is to determine the usefulness of the newly designed index defined in Eq. (8) and compute some exact results for the index under different product graphs. Also we intend to apply that results to some special graphs and nano-materials.

2. Materials and methods

Our main outcomes are organized in two parts. In the first part, the chemical applicability of the newly designed index is investigated. We consider the benchmark data set of octane isomers for such testing and corresponding experimental values of physico-chemical properties are collected from www.moleculardescriptors.eu/dataset/dataset.htm. Different topological indices of octanes are obtained using Dev-C++ software. All the properties are correlated with the index by MATLAB. After that a regression analysis for well correlated properties is performed using MATLAB and Excel data analysis tools. Linear fittings of the obtained models are plotted by MATLAB basic fitting tools. The degeneracy of the indices are checked using ”unique” command in MATLAB. In the second part, some explicit expressions of the novel index for different product graphs are computed. We consider combinatorial computing, graph theoretic tools and mathematical induction to obtain the results. Different composite graphs are drawn using Latex tikzpicture environment.

3. Results and discussion

Laboratory testing of chemicals to understand their different properties is very expensive. To overcome this, lots of topological indices have been presented in the theoretical chemistry. To introduce a topological index, one should check two aspects: On the one hand, it should correlate well with at least one physico-chemical properties for a benchmark data set, while on the other hand its formulation should be simple and give some theoretical insight. We split this section into two subsections. Firstly, we establish the applicability of MN index for octane isomer. We study the following linear regression model

P=I(±2E)+S(±2E)T, (9)

where P, I, E, S, and T are properties, intercept, standard error of coefficients, slope, and topological index respectively. The results are interpreted graphically using MATLAB software. Later, we study the index for some product graphs. Throughout this section, for the graph Gi, we use Vi and Ei for the node and the edge sets, respectively. Also for path, cycle and complete graphs with n nodes, we use Pn, Cn and Kn, respectively. From the definition (8), it is clear that MN(Pn)=16n38(n4), MN(Cn)=16n(n3), and MN(Kn)=n(n1)4(n1). Since various significant graphs can be obtained from different product of Pn, Cn and Kn, the MN index of them are also obtained in the second subsection.

3.1. Chemical significance of the neighborhood Zagreb index (MN)

According to the instruction of the International Academy of Mathematical Chemistry (IAMC), to investigate the effectiveness of a topological index to model physico-chemical attributes, we use regression analysis. Usually octane isomers are helpful for such investigation, since they represent a sufficiently large and structurally diverse group of alkanes for the preliminary testing of indices [16], [31]. Furtula et al. [8] derived that the correlation coefficient of both M 1 and F for octane isomers is greater than 0.95 with acentric factor and entropy. They also enhanced the skill of prediction of these indices by devising a linear model (M1+λF), where λ was varied from -20 to 20.

In this article, we find the correlation of entropy (S) and acentric factor with the neighborhood Zagreb index for octane isomers. The data related to octanes are listed in Table 1 . Here we have computed that the correlation coefficient (r) between acentric factor and MN is -0.99456 and between entropy (S) and MN is -0.95261. Thus MN can help to predict the entropy (r2=0.90746) and acentric factor (r2=0.98915) with powerful accuracy. These results confirm the suitability of the indices in QSPR analysis. The Eq. (9) yields the following regression models for the MN index.

Acentricfactor=0.51918(±0.00977)0.00137(±7.16964×105)MN,r2=0.98915,Se=0.00381,F=1457.77859,SF=3.8081×1017, (10)
S=127.80036(±3.63605)0.16707(±0.02667)MN,r2=0.90746,Se=1.41645,F=156.92469,SF=1.1012×109, (11)

where Se, F and SF are the statistical parameters: standard error of model, F-test and significance F, respectively. The linear fittings of the models are depicted in Figs. 1 and 2 . In both the figures, the solid circles represent the data point (x, y), where x, y denote the MN value and the physico-chemical property for octane isomers, respectively and the blue line represents the regression line. The Fig. 1 reveals the strength of structure property relationship between MN and acentric factor and the Fig. 2 shows that between MN and acentric factor. If we round the r 2 values to two digits, then it is clear that 99% and 91% of our observations fit the models (10) and (11) respectively and are shown visually in Figs. 1 and 2 respectively. The data points in Fig. 1 are more closed to the best fitting line compared to the Fig. 2. It confirms that the linear fitting of the model (10) is more accurate that the model (11). The smaller the Se values, the more confident we are regarding the regression equation. The Se values of both the equations are significantly low. The average distance of the data points to the regression line is also very low in Figs. 1 and 2. In fact, Fig. 1 yields a lower average distance than Fig. 2. The consistency of the model improves as the F-value increases. In each model, F-value is considerably high. When the SF value is less than 0.05, then the model is considered to be statistically reliable. In each case, SF value is far less than 0.05. Correlation of some well-established and most used degree based indices like first (M 1) and second (M 2) Zagreb indices [14], forgotten topological index (F) [8], [14], connectivity index (R) [30], sum connectivity index (SCI)[35] and symmetric division degree index (SDD) [33] with acentric factor and S is shown in Table 2 . It reveals the supremacy of MN compared to the indices in Table 2 in modelling acentric factor. Sometimes the novel index shows better predictive capability than the existing indices for S.

Table 1.

Experimental values of the acentric factor, entropy(S) and the corresponding values of different topological indices for octane isomers.

Octane isomers Acentric factor S MN M1 M2 F SCI R SDD
n-octane 0.397898 111.67 90 26 24 50 3.6547 3.9142 15
2-methyl heptane 0.377916 109.84 104 28 26 62 3.5246 3.7701 17.3333
3-methyl heptane 0.371002 111.26 108 28 27 62 3.5491 3.8081 16.6667
4-methyl heptane 0.371504 109.32 110 28 27 62 3.5491 3.8081 16.6667
3-ethyl hexane 0.362472 109.43 114 28 28 62 3.5737 3.8461 16
2,2-dimethyl hexane 0.339426 103.42 138 32 30 92 3.3272 3.5607 21.75
2,3-dimethyl hexane 0.348247 108.02 126 30 30 74 3.4328 3.6807 18.6667
2,4-dimethyl hexane 0.344223 106.98 124 30 29 74 3.419 3.6639 19
2,5-dimethyl hexane 0.35683 105.72 118 30 28 74 3.3944 3.6259 19.6667
3,3-dimethyl hexane 0.322596 104.74 146 32 32 92 3.3656 3.6213 20.5
3,4-dimethyl hexane 0.340345 106.59 130 30 31 74 3.4574 3.7187 18
2-methyl-3-ethyl pentane 0.332433 106.06 132 30 31 74 3.4574 3.7187 18
3-methyl-3-ethyl pentane 0.306899 101.48 152 32 34 92 3.404 3.682 19.25
2,2,3-trimethyl pentane 0.300816 101.31 162 34 35 104 3.2442 3.4814 22.8333
2,2,4-trimethyl pentane 0.30537 104.09 156 34 32 104 3.1971 3.4165 24.0833
2,3,3-trimethyl pentane 0.293177 102.06 164 34 36 104 3.258 3.504 22.25
2,3,4-trimethyl pentane 0.317422 102.39 144 32 33 86 3.3165 3.5535 20.6667
2,2,3,3-tetramethyl butane 0.255294 93.06 194 38 40 134 3.0368 3.25 27.5

Fig. 1.

Fig. 1

Linear fitting of MN with acentric factor for octane isomers.

Fig. 2.

Fig. 2

Linear fitting of MN with S for octane isomers.

Table 2.

The square of correlation coefficient of different topological indices with acentric factor and entropy.

M1 M2 F SCI R SDD
Acentric factor 0.9468 0.973 0.9313 0.8647 0.8176 0.8118
S 0.9107 0.8868 0.9077 0.8518 0.8205 0.8276

In addition to their application to different structure-property and structure-activity correlations, topological indices are also used for discrimination against isomers. The discrimination ability of an index has remarkable importance for the coding and the computer processing of chemical structures. Most of the indices have a flaw that more than one isomers occupy the same index which is known as degeneracy. But this novel index is exceptional for octane isomers. Konstantinova [21] proposed the sensitivity, the measure of degeneracy, formulated as

ST=NNTN, (12)

where N and NT are the total number of isomers and the count of isomers that cannot be discriminated by the descriptor T, respectively. The isomer discrimination ability of an index is directly proportional to ST. Clearly, its maximum value is 1. Therefore, ST plays a major role in the discriminating power of an index. The indices having good discrimination ability captures more structural information. For octane isomers, MN index exhibits better response (ST=1) compared to some well established and most utilized degree based indices that are reported in Table 3 .

Table 3.

Sensitivity of different indices for octane isomers.

Indices Sensitivity (ST)
M1 0.333
M2 0.722
F 0.389
R 0.889
SCI 0.889
SDD 0.889
Neighborhood Zagreb index (MN) 1.000

Correlation of MN with some existing indices are shown in Table 4 .

Table 4.

The square of correlation coefficient of MN with some existing indices.

M1 M2 F SCI R SDD
MN 0.9716 0.9646 0.9657 0.891 0.8471 0.8539

Apart from chemical importance, an effective topological index should have a convenient and straightforward computational formula. To show how the computation of MN index for chemical compound is clear and easy, we consider chemical structures favipiravir and hydroxychloroquine in Fig. 3 . Favipiravir has been researched for the treatment of life-threatening pathogens such as Ebola, Lassa, and now COVID-19. Hydroxychloroquine is an antimalarial drug. It is one of the antiviral agents that is being investigated currently to prevent COVID-19. The hydrogen suppressed molecular graphs of the aforesaid compounds are shown in Fig. 4 . Let G 1 and G 2 be the hydrogen deleted molecular graphs of favipiravir and hydroxychloroquine, respectively. Then, we have

MN(G1)=uV(G1)δG(u)2=4(3)2+4(5)2+2(6)2+(8)2=272,
MN(G2)=uV(G2)δG(u)2=2(2)2+2(3)2+5(4)2+8(5)2+2(6)2+2(7)2+(8)2=540,

Fig. 3.

Fig. 3

Chemical structures of favipiravir and hydroxychloroquine from left to right.

Fig. 4.

Fig. 4

Hydrogen deleted molecular graphs of favipiravir and hydroxychloroquine from left to right.

3.2. MN Index of some product graphs

Product graphs are applicable in a number of areas, including automata theory, communication networks, information theory, computer architecture, algebraic structures and chemistry. They help to construct many network topologies for interconnection networks. In this section, we evaluate the newly introduced index for different product graphs such as Cartesian, wreath and tensor product of graphs. We proceed with the following lemma directly followed from definitions.

Lemma 3.1

If G be a graph, then we have

  • (i)

    uV(G)δG(u)=M1(G),

  • (ii)

    uV(G)degG(u)δG(u)=2M2(G),

where M1(G), M2(G) are formulated in Eqs.(1),(2)and δG(u) is defined in Eq.(7).

3.2.1. Cartesian product

Definition 3.2

The Cartesian product of G 1, G 2, written as G 1 G 2, containing node set V 1  ×  V 2 and (u 1, v 1) is connected to (u 2, v 2) iff [u 1 is connected with u 2 in G 1 and v1=v2] or [v 1 is connected with v 2 in G 2 and u1=u2]. We consider the symbol  ×  for the Cartesian product of two sets.

Clearly the above definition yield the lemma stated below.

Lemma 3.3

For graphs G1and G2, we have

  • (i)

    δG1G2(u,v)=δG1(u)+δG2(v)+2degG1(u)degG2(v),

  • (ii)

    |E(G1G2)|=|V2||E1|+|V1||E2|.

In [18], [20] different topological descriptors were studied for Cartesian product. Here we intend to go forward for the MN index.

Proposition 3.4

The MN index of Cartesian product of G1and G2is given by

MN(G1G2)=6M1(G1)M1G2)+|V2|MN(G1)+|V1|MN(G2)+16[|E2|M2(G1)+|E1|M2(G2)]. (13)
Proof

From definition of neighborhood Zagreb index and applying Lemma 3.3 and Lemma 3.1, we get

MN(G1G2)=(u1,u2)V1×V2δG1G22(u1,u2)=u1V1u2V2[δG1(u1)+δG2(u2)+2degG1(u1)degG2(u2)]2=6M1(G1)M1G2)+|V2|MN(G1)+|V1|MN(G2)+16[|E2|M2(G1)+|E1|M2(G2)].

Hence the result. Using the Eq. (13), we have the following results. □

Example 3.5

The Cartesian product of P 2 and Pn+1 produces the ladder graph Ln (Fig. 5 ). By the above proposition, we derive the following result.

MN(Ln)=162n130,n3. (14)
Fig. 5.

Fig. 5

The ladder graph Ln.

Carbon nanotube is the most popular nanomaterial having low weight, high strength, and very well thermal and electric conductivity. It has diverse usage in electromagnetic devices, Coatings and films, water and air filtration, bio-medical industry etc. The MN index for C4nanotorus and C4nanotube are obtained in Eqs. (15) and (16), respectively.

Example 3.6

For a C4nanotorus TC4(m,n)=CmCn, the MN index is given by

MN(TC4(m,n))=256mn. (15)
Example 3.7

The Cartesian product of Pm and Cn yields a C4nanotube TUC4(m,n)=PmCn. Its MN index is as follows:

MN(TUC4(m,n))=256mn374n,m4. (16)
Example 3.8

The MN index of the grid (PnPm) (Fig. 6 ) is given by

MN(PnPm)=256mn374m374n+472,m,n4. (17)
Example 3.9

For a n-prism (K2Cn) (Fig. 7 ), the neighborhood Zagreb index is given below.

MN(K2Cn)=162n. (18)
Example 3.10

The Cartesian product of Kn and Km yields the rook’s graph (Fig. 8 ). All legal move of a rook on a chessboard can be represented by a rook’s graph. Its each node correspond to a square of the chessboard and edges correspond to legal moves from one square to another. Applying the Proposition 3.4, we have computed the MN index of rook’s graph as follows.

MN(KmKn)=mn[2(m1)(n1)(2m2+2n2+3mn7m7n+7)+(m1)4+(n1)4]. (19)
Fig. 6.

Fig. 6

The grid graph P5P4.

Fig. 7.

Fig. 7

The example of n-Prism graph (n=6).

Fig. 8.

Fig. 8

The rook’s graph K6K6.

Now we generalize the Proposition 3.4. We begin with the following lemma.

Lemma 3.11

If G1, G2, ...., Gn be n graphs andV=V(p=1nGp),E=E(p=1nGp),then we have

  • (i)

    |E(p=1nGp)|=|V|p=1n|Ep||Vp|,

  • (ii)

    M1(p=1nGp)=|V|p=1nM1(Gp)|Vp|+4|V|pq,p,q=1n1|Ep||Eq||Vp||Vq|,

  • (iii)

    M2(p=1nGp)=|V|[p=1nM2(Gp)|Vp|+4pqr,p,q,r=1n|EpEqEr||VpVqVr|]+3p=1nM1(Gp)(|E||Vp||VEp||Vp|2).

Proof

Applying Lemma 3.3(ii) and an inductive argument, (i) is clear. In order to proof (ii) and (iii), we refer to Khalifeh et al. [19]. □

Proposition 3.12

If G1, G2, ..., Gn be n graphs, then we have

MN(p=1nGp)=|V|[p=1nMN(Gp)|Vp|+3pq,p,q=1nM1(Gp)M1(Gq)|Vp||Vq|+24pqr,p,q,r=1nM1(Gp)|Eq||Er||Vp||Vq||Vr|+16pq,p,q=1nM2(Gp)|Eq||Vp||Vq|+16pqrs,p,q,r,s=1n|Ep||Eq||Er||Es||Vp||Vq||Vr||Vs|]. (20)
Proof

We derive the formula by mathematical induction. Evidently the result holds for n=2. Let us take the proposition to be true for (n1) graphs. Then we obtain

MN(p=1nGp)=MN(p=1n1GpGn)=6M1(p=1n1Gp)M1(Gn)+|Vn|MN(p=1n1Gp)+|V(p=1n1Gp)|MN(Gn)+16[M2(p=1n1Gp)|En|+M2(Gn)|E(p=1n1Gp)|]. (21)

Using Lemma 3.11 in Eq. (21), we get

MN(p=1nGp)=6|V|M1(Gn)|Vn|[p=1n1M1(Gp)|Vp|+4pq,p,q=1n1|Ep||Eq||Vp||Vq|]+|V|[p=1n1MN(Gp)|Vp|+3pq,p,q=1n1M1(Gp)M1(Gq)|Vp||Vq|+24pqr,p,q,r=1n1M1(Gp)|Eq||Er||Vp||Vq||Vr|+16pq,p,q=1n1M2(Gp)|Eq||Vp||Vq|+16pqrs,p,q,r,s=1n1|Ep||Eq||Er||Es||Vp||Vq||Vr||Vs|]+|V|MN(Gn)|Vn|+16|V||Vn|[|En|{p=1n1M2(Gp)|Vp|+3p=1n1M1(Gp)|Vp|(q=1n1|Eq||Vq||Ep||Vp|)+4pqr,p,q,r=1n1|Ep||Eq||Er||Vp||Vq||Vr|}+M2(Gn)p=1n1|Ep||Vp|].

After simplification, we have

MN(p=1nGp)=|V|[p=1nMN(Gp)|Vp|+3{pq,p,q=1n1M1(Gp)M1(Gq)|Vp||Vq|+2M1(Gn)|Vn|p=1n1M1(Gi)|Vp|}+24{pqr,p,q,r=1n1M1(Gp)|Eq||Er||Vp||Vq||Vr|+pq,p,q=1n1M1(Gn)|Ep||Eq||Vn||Vp||Vq|+2(p,q=1n1M1(Gp)|Eq||En||Vp||Vq||Vn|p=1n1M1(Gp)|Ep||En||Vp||Vp||Vn|)}+16{pq,p,q=1n1M2(Gp)|Eq||Vp||Vq|+p=1n1M2(Gn)|Ep||Vn||Vp|+p=1n1M2(Gp)|En||Vp||Vn|}]+16|V|pqrs,p,q,r,s=1n|Ep||Eq||Er||Es||Vp||Vq||Vr||Vs|. (22)

Thus, the result (20) can be obtained easily from the Eq. (22). □

Definition 3.13

Consider the graph G containing m-tuples b1,b2,,bm with bp {0,1,,np1},np2, as vertices and let whenever the difference of two tuples is exactly one place, the corresponding two vertices are adjacent. This graph is known as Hamming graph. The necessary and sufficient criteria for a graph G to be a Hamming graph is that G=p=1mKnp and that is why such a graph G is naturally written as Hn1,n2,,nm.

Hamming graph is very useful in coding theory specially in error correcting codes. Also such type of graph is effective in association schemes. Applying the result (20), we have the corollary stated below.

Corollary 3.14

The neighborhood Zagreb index of Hamming graph is obtained as follows:

MN(G)=p=1mnp[p=1m(np1)4+3pq,p,q=1m(np1)2(nq1)2+6pqr,p,q,r=1m(np1)2(nq1)(nr1)+4pq,p,q=1m(ni1)3(nj1)+pqrs,p,q,r,s=1m(np1)(nq1)(nr1)(ns1)].
Example 3.15

When n 1, n 2, ....,nm are all equal to 2, the graph Hn1,n2,,nm is known as a hyper cube (Fig. 9 ) with dimension m and written as Qm. We compute the following.

MN(Qm)=2mm4. (23)
Fig. 9.

Fig. 9

Example of Hypercube.

3.2.2. Tensor product

Definition 3.16

The tensor product of G 1, G 2, written as G 1G 2, contains the node set V 1 × V 2 and (u 1, v 1) is connected to (u 2, v 2) iff u 1 u 2 ∈ E 1 and v 1 v 2 ∈ E 2.

Clearly the definition gives the lemma as follows:

Lemma 3.17

For graphs G1and G2, we have

δG1G2(u,v)=δG1(u)δG2(v). (24)

The tensor product was thoroughly studied in terms of graph coloring, graph identification and decomposition, graph embedding, matching theory and graph stability in [17]. Z. Yarahmadi studied about degree based indices for tensor product in [34]. Also in [23], [27] various topological descriptors of tensor product graphs are calculated. Here we continue this journey for the MN index.

Proposition 3.18

The MN index of tensor product for G1, G2is given by

MN(G1G2)=MN(G1)MN(G2). (25)
Proof

By the definition of the MN index and applying Eq. (24), we get

MN(G1G2)=(u1,u2)V1×V2δG1G22(u1,u2)=u1V1u2V2[δG1(u1)δG2(u2)]2=MN(G1)MN(G2).

Which is the required result. □

Example 3.19

Using the result (25), we have the following computations.

  • (i)

    MN (PnPm) = (16n38)(16m38),   m, n ≥ 4,

  • (ii)

    MN(CnCm) = 256mn,

  • (iii)

    MN(KnKm)= mn(m1)4(n1)4,

  • (iv)

    MN(PnCm)= 16m(16n38),   n ≥ 4,

  • (v)

    MN(PnKm)= m(m1)4(16n38),   n ≥ 4,

  • (vi)

    MN(CnKm) = 16mn(m1)4.

3.2.3. Wreath product

Definition 3.20

The wreath product (also known as composition) of G 1 and G 2 having V 1 and V 2 as vertex sets with no common vertex and edge sets E 1 and E 2 is the graph G 1[G 2] containing node set V 1 × V 2 and (u 1, v 1) is connected to (u 2, v 2) iff (u 1 u 2 ∈ E 1) or (u1=u2 and v 1 v 2 ∈ E 2).

From the definition, we have the following obvious lemma.

Lemma 3.21

For graphs G1and G2, we have

δG1[G2](u,v)=|V2|2δG1(u)+δG2(v)+2|E2|degG1(u)+|V2|degG1(u)degG2(v). (26)

In [7], [25] different topological indices for wreath product of graphs are derived. Here we proceed for the MN index of wreath product.

Proposition 3.22

The MN index of wreath product for G1, G2is obtained as follows:

MN(G1[G2])=|V2|5MN(G1)+|V1|MN(G2)+12|V2||E2|2M1(G1)+8|E1||E2|M1(G2)+16|V2|3|E2|M2(G1)+8|V2||E1|M2(G2)+3|V2|2M1(G1)M1(G2). (27)
Proof

From definition of neighborhood Zagreb index and using Eq. (26), we have

MN(G1[G2])=(u,v)V1×V2δG1[G2]2(u,v)=uV1vV2[|V2|2δG1(u)+δG2(v)+2|E2|degG1(u)+|V2|degG1(u)degG2(v)]2=uV1vV2[|V2|4δG1(u)2+|V2|2degG1(u)2degG2(v)2+4|E2|2degG1(u)2+δG2(v)2+2|V2|3δG1(u)degG1(u)degG2(v)+4|V2|2|E2|δG1(u)degG1(u)+2|V2|2δG1(u)δG2(v)+4|V2||E2|degG1(u)2degG2(v)+2|V2|δG2(v)degG1(u)degG2(v)+4degG1(u)|E2|δG2(v)].

Applying Lemma 3.1, we have

MN(G1[G2])=|V2|5MN(G1)+|V1|MN(G2)+12|V2||E2|2M1(G1)+8|E1||E2|M1(G2)+16|V2|3|E2|M2(G1)+8|V2||E1|M2(G2)+3|V2|2M1(G1)M1(G2),

which is the desired result. □

Example 3.23

The wreath product of the path graphs Pn and P 2 yield the fence graph (Fig. 10 ), whereas the wreath product of the cycle Cn and the path P 2 gives the closed fence graph (Fig. 10). Thus from the result (27), we compute the followings.

  • (i)

    MN(Pn[P2])=1250n2560,n4,

  • (ii)

    MN(Cn[P2])=1250n,n3.

Fig. 10.

Fig. 10

Fence graph (Pn[P2]) and closed fence graph (Cn[P2]).

4. Conclusion

In this article, we introduced the MN index, examined its chemical applicability, computed some exact formulae for MN of some product graphs and applied the results to some special graphs. As a future work, we derive the results for some other graph operations and compute some bounds of this index. Also some exact expressions of it for different networks can also be derived. As the pharmacological activity of a compound depends on its physico-chemical properties and the correlations of MN index with some of these properties are attractive, there is nothing to be surprised that MN index can be used in designing new drugs.

CRediT authorship contribution statement

Sourav Mondal: Conceptualization, Writing - original draft, Software. Nilanjan De: Conceptualization, Investigation, Writing - review & editing. Anita Pal: Supervision, Visualization, Validation.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The first author is very obliged to the Department of Science and Technology (DST), Government of India for the Inspire Fellowship [IF170148].

Footnotes

Supplementary material associated with this article can be found, in the online version, at 10.1016/j.molstruc.2020.129210

Appendix A. Supplementary materials

Supplementary Data S1

Supplementary Raw Research Data. This is open data under the CC BY license http://creativecommons.org/licenses/by/4.0/

mmc1.xml (271B, xml)

References

  • 1.Akhter S., Imran M. Computing the forgotten topological index of four operations on graphs. AKCE Int. J. Graphs Comb. 2017;14:70–79. [Google Scholar]
  • 2.Alameri A., Al-Rumaima M., Almazah M. Y-coindex of graph operations and its applications of molecular descriptors. J. Mol. Struct. 2020;1221:128754. [Google Scholar]
  • 3.Azari H.Y., Manoochehrian B., Ashrafi A.R. The PI index of product graphs. Appl. Math. Lett. 2008;21:624–627. [Google Scholar]
  • 4.Azari M., Iranmanesh A. Chemical graphs constructed from rooted product and their zagreb indices. MATCH Commun. Math. Comput. Chem. 2013;70:901–919. [Google Scholar]
  • 5.CMD F., Stevanovic D. Further properties of the second zagreb index. MATCH Commun. Math. Comput. Chem. 2014;72:655–668. [Google Scholar]
  • 6.Das K.C., Xu K., Gutman I. On zagreb and harary indices. MATCH Commun. Math. Comput. Chem. 2013;70:301–314. [Google Scholar]
  • 7.Donno A. Spectrum, distance spectrum, and wiener index of wreath products of complete graphs. Ars Math. Contemp. 2017;13:207–225. [Google Scholar]
  • 8.Furtula B., Gutman I. A forgotten topological index. J. Math. Chem. 2015;53:1184–1190. [Google Scholar]
  • 9.Ghorbani M., Hossainzadeh M.A. The third version of zagreb index. Discrete Math Algorithm Appl. 2013;5(4) [Google Scholar]; Article ID:1350039
  • 10.Gunes A.Y., Togan M., Celik F., Cangul I.N. Cut vertex and cut edge problem for topological graph indices. J. Tybah Univ. Sci. 2019;13:1175–1183. [Google Scholar]
  • 11.Gutman I., Das K.C. The first zagreb index 30 years after. MATCH Commun. Math. Comput. Chem. 2004;50:83–92. [Google Scholar]
  • 12.Gutman I., Polansky O.E. Springer; Berlin: 1986. Mathematical Concepts in Organic Chemistry. [Google Scholar]
  • 13.Gutman I., Rućić B., Trinajstić N., Wilcox C.F. Graph theory and molecular orbitals. XII. acyclic polyenes. J. Chem. Phys. 1975;62:3399–3405. [Google Scholar]
  • 14.Gutman I., Trinajstić N. Graph theory and molecular orbitals. total π-electron energy of alternate hydrocarbons. Chem. Phys. Lett. 1972;17:535–538. [Google Scholar]
  • 15.Hayat S., Khan S., Imran M., Liu J.B. Quality testing of distance-based molecular descriptors for benzenoid hydrocarbons. J. Mol. Struct. 2020;1222:128927. [Google Scholar]
  • 16.Hosoya H. Topological index, a newly proposed quantity characterizing the topological nature of structural isomers of saturated hydrocarbons. Bull. Chem. Soc. Jpn. 1971;44:2332–2339. [Google Scholar]
  • 17.Imrich W., Klavzar S. Wiley; New York: 2000. Product Graphs: Structure and Recognition. [Google Scholar]
  • 18.Khalifeh M.H., Azari H.Y., Ashrafi A.R. Vertex and edge PI indices of cartesian product graphs. Discrete Appl. Math. 2008;156:1780–1789. [Google Scholar]
  • 19.Khalifeh M.H., Azari H.Y., Ashrafi A.R. The first and second zagreb indices of some graph operations. Discrete Appl. Math. 2009;157:804–811. [Google Scholar]
  • 20.Klavzar S., Rajapakse A., Gutman I. The szeged and the wiener index of graphs. Appl. Math. Lett. 1996;9:45–49. [Google Scholar]
  • 21.Konstantinova E.V. The discrimination ability of some topological and information distance indices for graphs of unbranched hexagonal systems. J. Chem. Inf. Comput. Sci. 1996;36:54–57. [Google Scholar]
  • 22.Merrifield R.E., Simmons H.E. Wiley; New York: 1989. Topological Methods in Chemistry. [Google Scholar]
  • 23.MJN A., Khodashenas H. Distance-based topological indices of tensor product of graphs. Iranian J. Math. Chem. 2012;3:45–53. [Google Scholar]
  • 24.N. D., Pal A., SMA N. On some bounds and exact formulae for connective eccentric indices of graphs under some graph operations. Int J Comb. 2014;2014 [Google Scholar]; Article ID:579257
  • 25.N. D., SMA N., Pal A. F-index of some graph operations. Discrete Math Algorithm Appl. 2016;8(2) [Google Scholar]; Article ID:1650025
  • 26.Noureen S., Bhatti A.A., Ali A. Extremal trees for the modified first zagreb connection index with fixed number of segments or vertices of degree 2. J. Tybah Univ. Sci. 2020;14:31–37. [Google Scholar]
  • 27.Pattabiraman K., Paulraja P. On some topological indices of the tensor products of graphs. Discrete Appl. Math. 2012;160:267–279. [Google Scholar]
  • 28.Paulraja P., Agnes V.S. Degree distance of product graphs. Discrete Math Algorithm Appl. 2014;6(1) [Google Scholar]; Article ID:1450003
  • 29.Plavsić D., Nikolić S., Trinajstić N., Mihalić Z. On the harary index for the characterization of chemical graphs. J. Math. Chem. 1993;12:235–250. [Google Scholar]
  • 30.Randić M. Characterization of molecular branching. J. Am. Chem. Soc. 1975;97:6609–6615. [Google Scholar]
  • 31.Randić M., Trinajstić N. In search for graph invariants of chemical interest. J. Mol. Struct. 1993;300:551–571. [Google Scholar]
  • 32.Trinajstić N. CRC Press; Boca Raton: 1993. Chemical Graph Theory. [Google Scholar]
  • 33.Vukicevic D. Bond additive modeling 2 mathematical properties of max-min rodeg index. Croat. Chem. Acta. 2010;54:261–273. [Google Scholar]
  • 34.Yarahmadi Z. Computing some topological indices of tensor product of graphs. Iranian J. Math. Chem. 2011;2:109–118. [Google Scholar]
  • 35.Zhou B., Trinajstić N. On a novel connectivity index. J. Math. Chem. 2009;46:1252–1270. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data S1

Supplementary Raw Research Data. This is open data under the CC BY license http://creativecommons.org/licenses/by/4.0/

mmc1.xml (271B, xml)

Articles from Journal of Molecular Structure are provided here courtesy of Elsevier

RESOURCES