1 Introduction

More on energy and Randić energy of specific graphs

Saeid Alikhani111Corresponding author. E-mail: alikhani@yazd.ac.irand Nima Ghanbari222E-mail: n.ghanbari@stu.yazd.ac.ir

Department of Mathematics, Yazd University

89195-741, Yazd, Iran

Abstract

Let be a simple graph of order . The energy of the graph is the sum of the absolute values of the eigenvalues of . The Randić matrix of , denoted by , is defined as the matrix whose -entry is if and are adjacent and for another cases. The Randić energy of is the sum of absolute values of the eigenvalues of . In this paper we compute the energy and Randić energy for certain graphs. Also we propose a conjecture on Randić energy.

Mathematics Subject Classification: 15A18.
Keywords: 3-regular graphs; energy; Randić energy; characteristic polynomial; Petersen graph.

## 1 Introduction

In this paper we are concerned with simple finite graphs, without directed, multiple, or weighted edges, and without self-loops. Let be adjacency matrix of and its eigenvalues. These are said to be the eigenvalues of the graph and to form its spectrum [4]. The energy of the graph is defined as the sum of the absolute values of its eigenvalues

 E(G)=n∑i=1|λi|.

Details and more information on graph energy can be found in [5, 6, 8, 12].

The Randić matrix is defined as [2, 3, 9]

 rij=⎧⎪⎨⎪⎩1√didjif vi∼vj0otherwise.

Denote the eigenvalues of the Randić matrix by and label them in non-increasing order. The Randić energy [2, 3, 9] of is defined as

 E(G)=n∑i=1|ρi|.

Two graphs and are said to be Randić energy equivalent, or simply -equivalent, written , if . It is evident that the relation of being -equivalence is an equivalence relation on the family of graphs, and thus is partitioned into equivalence classes, called the -equivalence classes. Given , let

 [G]={H∈G:H∼G}.

We call the equivalence class determined by . A graph is said to be Randić energy unique, or simply -unique, if .

Similarly, we can define -equivalence for energy and -unique for a graph.

A graph is called -regular if all vertices have the same degree . One of the famous graphs is the Petersen graph which is a symmetric non-planar 3-regular graph. In the study of energy and Randić energy, it is interesting to investigate the characteristic polynomial and energy of this graph. We denote the Petersen graph by .

In this paper, we study the energy and Randić energy of specific graphs. In the next section, we study energy and Randić energy of -regular and -regular graphs. We study cubic graphs of order 10 and list all characteristic polynomial, energy and Randić energy of them. As a result, we show that Petersen graph is not -unique (-unique) but can be determined by its Randić energy (energy) and its eigenvalues. In the last section we consider some another families of graphs and study their Randić characteristic polynomials.

## 2 Energy of 2-regular and 3-regular graphs

The energy and Randić energy of regular graphs have not been widely studied. In this section we consider -regular and -regular graphs. The following theorem gives a relationship between the Randić energy and energy of -regular graphs.

###### Lemma 1

[10] If the graph is -regular then .

Also we have the following easy lemma:

###### Lemma 2

. Let . Then

1. .

2. .

Randić characteristic polynomial of the cycle graph can be determined by the following theorem:

###### Lemma 3

[1] For , the Randić characteristic polynomial of the cycle graph is

 RP(Cn,λ)=λΛn−1−12Λn−2−(12)n−1,

where for every , with and .

By Lemma 3, we can find all the eigenvalues of Randić matrix of cycle graphs. So we can compute the Randić energy of cycles. Also every cycle is -regular. By Lemma 1, we have . Hence we can compute energy of cycle graphs too. Every -regular graph is a disjoint union of cycles. Therefore by Lemma 2, we can find energy and Randić energy of 2-regular graphs.

Let to consider the characteristic polynomial of -regular graphs of order . Also we shall compute energy and Randić energy of this class of graphs. There are exactly cubic graphs of order given in Figure 1 (see [11]).

We show that Petersen graph is not -unique (-unique) but can be determined by its Randić energy (energy) and its eigenvalues. There are just two non-connected cubic graphs of order . The following theorem gives us characteristic polynomial of -regular graphs of order . We denote the characteristic polynomial of the graph by .

Using Maple we computed the characteristic polynomials of -regular graphs of order in Table 1.

Table 1. Characteristic polynomial , for .

By computing the roots of characteristic polynomial of cubic graphs of order , we can have the energy of these graphs. We compute them to four decimal places. So we have table 2:

15.1231 5.0410 15.1231 5.0410 14.7943 4.9314
14.8596 4.9532 15.3164 5.1054 14.0000 4.6666
14.8212 4.9404 14.4721 4.8240 16.0000 5.3333
13.5143 4.5047 14.7020 4.9006 13.5569 4.5189
14.2925 4.7641 16.0000 5.3333 15.5791 5.1930
14.9443 4.9814 14.3780 4.7926 14.0000 4.6666
15.0777 5.0259 15.0895 5.0298 12.0000 4.0000

Table 2. Energy and Randić energy of cubic graphs of order .

###### Theorem 1

. Six cubic graphs of order are not -unique (-unique). If two cubic graphs of order have equal energy (Randić energy), then their eigenvalues are different in exactly values.

Proof. Using Table 2, we see that , and . Now, it suffices to find the eigenvalues of , , , , and . By Table 1 we have:

 P(G1,λ) =λ10−15λ8−8λ7+71λ6+64λ5−101λ4−104λ3+44λ2+48λ =λ(λ−3)(λ+2)2(λ−1)2(λ+1)2(λ−1−√172)(λ−1+√172),
 P(G8,λ) =λ10−15λ8+71λ6−16λ5−133λ4+64λ3+76λ2−48λ =λ(λ−3)(λ+2)2(λ−1)3(λ+1)(λ−−1+√172)(λ−−1−√172).

Also

 P(G12,λ) =λ10−15λ8−4λ7+75λ6+24λ5−157λ4−36λ3+144λ2+16λ−48 =(λ−3)(λ−2)(λ+2)3(λ−1)3(λ+1)2,
 P(G17,λ) =λ10−15λ8+75λ6−24λ5−165λ4+120λ3+120λ2−160λ+48 =(λ−3)(λ+2)4(λ−1)5.

And

 P(G16,λ) =λ10−15λ8+63λ6−85λ4+36λ2 =λ2(λ−3)(λ+3)(λ−2)(λ+2)(λ−1)2(λ+1)2,
 P(G20,λ) =λ10−15λ8−12λ7+63λ6+96λ5−13λ4−84λ3−36λ2 =λ2(λ−3)2(λ+2)2(λ−1)(λ+1)3.

So we have the result.

Now we consider Petersen graph . We have shown this graph in Figure 2.

###### Theorem 2

. Let be the family of -regular graphs of order . For the Petersen graph , we have the following properties:

• is not -unique (-unique) in .

• has the maximum energy (Randić energy) in .

• can be identify by its energy (Randić energy) and its eigenvalues in .

Proof.

• The adjacency matrix of is

 A(P)=⎛⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜⎝0100110000101000100001010001000010100010100100000110000001100100000011001001000100010110000000101100⎞⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟⎠.

So Therefore we have:

 λ1=3  ,  λ2=λ3=λ4=λ5=−2  ,  λ6=λ7=λ8=λ9=λ10=1,

and so we have . By Table 2, we have . Hence is not -unique (and -unique) in .

• It follows from Part (i) and Table 2.

• It follows from Part (i) and Theorem 1. So is the Petersen graph.

The following result gives a relationship between energy and permanent of adjacency matrix of two connected graphs in the family of cubic graphs of order whose have the same -equivalence class.

###### Theorem 3

. If two connected cubic graphs of order have the same energy, then their adjacency matrices have the same permanent.

Proof. By Table 2, it suffices to find , , and . For graph , we have

 A(G1)=⎛⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜⎝0100010001101100000001011000000110100000001101000010001010000000010110000000101100000011011000000110⎞⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟⎠.

By Ryser’s method, we have . Similarly we have:

 per(A(G8))=72  ,  per(A(G12))=60  ,  per(A(G17))=60.

So we have the result.

Remark 1. The converse of Theorem 4 is not true. Because , but as we see in Table 2, .

###### Corollary 1

. If two connected cubic graphs of order have the same Randić energy, then their adjacency matrices have the same permanent.

Proof. It follows from Lemma 1, Table 2 and Theorem 3.

## 3 Randić characteristic polynomial of a kind of Dutch-Windmill graphs

We recall that a complex number is called an algebraic number (resp. an algebraic integer) if it is a root of some monic polynomial with rational (resp. integer) coefficients (see [13]). Since the Randić characteristic polynomial is a monic polynomial in with integer coefficients, its roots are, by definition, algebraic integers. This naturally raises the question: Which algebraic integers can occur as zeros of Randić characteristic polynomials? And which real numbers can occur as Randić energy of graphs? We are interested to numbers which are occur as Randić energy. Clearly those lying in are forbidden set, because we know that if graph possesses at least one edge, then . We think that the Randić energy of graphs are dense in . In this section we would like to study some further results of this kind.

Let be any positive integer and be Dutch Windmill graph with vertices and edges. In other words, the graph is a graph that can be constructed by coalescence copies of the cycle graph of length with a common vertex. We recall that is friendship graphs. Figure 3 shows some examples of this kind of Dutch Windmill graphs. In this section we shall investigate the Randić characteristic polynomial of Dutch Windmill graphs.

By Lemma 3, we know that

 RP(Cm,λ)=λΛm−1−12Λm−2−(12)m−1,

where for every , with and . We show that the Randić characteristic polynomial of Dutch Windmill graphs can compute by the cycle which constructed it.

###### Theorem 4

. For , the Randić characteristic polynomial of the Dutch Windmill graph is

 RP(Dnm,λ)=Λn−1m−1.RP(Cm,λ),

where for every , with and .

Proof. For every , consider

 Bk:=⎛⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜⎝λ−1200…000−12λ−120…0000−12λ−12…00000−12λ…000⋮⋮⋮⋮⋱⋮⋮⋮0000…λ−1200000…−12λ−120000…0−12λ⎞⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟⎠k×k,

and let . It is easy to see that .

Suppose that . We have

 RP((Dnm,λ)=det⎛⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜⎝λAA…AAtBm−10…0At0Bm−1…0⋮⋮⋮⋱⋮At00…Bm−1⎞⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟⎠,

where . So

 det(λI−R((Dnm))=λΛnm−1+(−14Λm−2+2((−1)m+1(−12)m)+(−1)2m+1(14)Λm−2)Λn−1m−1.

Therefore

 det(λI−R((Dnm))=λΛnm−1+(−12Λm−2−(12)m−1)Λn−1m−1.

Hence

 det(λI−R((Dnm))=Λn−1m−1(λΛm−1−12Λm−2−(12)m−1)=Λn−1m−1RP(Cm,λ).\vbox{\hrule width 100% height 1px\hbox% {\vrule width 1px\kern 0.0pt\vbox{\kern 0.0pt\vbox to 5.690551pt{\hfill}\kern 0% .0pt}\kern 0.0pt\vrule width 1px}\hrule width 100% height 1px}\vskip12.0ptplus4.0ptminus4.0pt

In [1] we have presented two families of graphs such that their Randić energy are and . Here we recall the following results:

###### Theorem 5

. [1]

1. The Randić energy of friendship graph is

2. The Randić energy of Dutch-Windmill graph is

3. For every , the Randić energy of is

We can use Theorem 4 to obtain . Here using the definition of Randić characteristic polynomial, we prove the following result:

###### Theorem 6

. The Randić energy of is

 RE(Dn5)=1+n√5.

Proof. The Randić matrix of is

 R(Dn5)=⎛⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜ ⎜⎝012√n12√n00⋯12√n12√n0012√n00120…000012√n00012…00000120012…00000012120…0000⋮⋮⋮⋮⋮⋱⋮⋮⋮⋮12√n0000…0012012√n0000…0001200000…12001200000…012120⎞⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟⎠(4n+1)×(4n+1).

Let and . Then

 det(λI−R((Dn5))=λdet(A)n+√ndet(C)det(A)n−1.

So

 det(λI−R((Dn5))=det(A)n−1(λ−1)(λ−(√54−14))2(λ+(√54+14))2.

Hence

 RE(Dn5)=1+n√5.\vbox{\hrule width 100% % height 1px\hbox{\vrule width 1px\kern 0.0pt\vbox{\kern 0.0pt\vbox to 5.690551% pt{\hfill}\kern 0.0pt}\kern 0.0pt\vrule width 1px}\hrule width 100% height 1px% }\vskip12.0ptplus4.0ptminus4.0pt

Part (iii) of Theorem 5 implies that the Randić energy of graphs are dense in . Motivated by this notation, Theorems 5 and 6, we think that the Randić energy of graphs are dense in . We close this paper by the following conjecture:

###### Conjecture 1

. Randić energy of graphs are dense in .

## References

• [1] S. Alikhani, N. Ghanbari, Randić energy of specific graphs, submitted. Availabel at http://arxiv.org/abs/1411.2544.
• [2] Ş. B. Bozkurt, A. D. Güngör, I. Gutman, A. S. Çevik, Randić matrix and Randić energy, MATCH Commum. Math. Comput. Chem. 64 (2010) 239–250.
• [3] Ş. B. Bozkurt, A. D. Güngör, I. Gutman, Randić spectral radius and Randić energy, MATCH Commum. Math. Comput. Chem. 64 (2010) 321–334.
• [4] D. Cvetković, M. Doob, H. sachs, Spectra of graphs - Theory and Aplication, Academic Press, New York, 1980.
• [5] I. Gutman, The energy of a graph: Old and new results, in: A. Betten, A.Kohnert, R. Laue, A. Wassermannn (Eds.), Algebraic Combinatorics and Applications, Springer-Verlag, Berlin, 2001, pp. 196–211.
• [6] I. Gutman, Topology and stability of conjugated hydrocarbons. The dependence of total -electron energy on molecular topology, J. Serb. Chem. Soc. 70 (2005) 441–456.
• [7] I. Gutman, B. Furtula, Ş. B. Bozkurt, On Randić energy, Linear Algebra Appl., 442 (2014) 50–57.
• [8] I. Gutman, X. Li, J. Zhang, Graph energy, in: M. Dehmer, F. Emmert-Streib (Eds.), Analysis of Complex Networks. From Biology to Linguistics, Wiley-VCH, Weinheim, 2009, pp. 145–174.
• [9] I. Gutman, B. Furtula, Ş. B. Bozkurt, On Randić energy, Linear Algebra Appl., 442 (2014) 50–57.
• [10] R. Gu, F. Huang, X. Li, General randić matrix and general randic energy, Transactions on Combinatorics., Vol. 3 No. 3 (2014), pp. 21-33.
• [11] G. B. Khosrovshahi, Ch. Maysoori, Tayfeh-Rezaie, A Note on 3-Factorizations of , J. Combin. Designs 9 (2001), 379-383.
• [12] S. Majstorović, A. KlobuÂ·car, I. Gutman, Selected topics from the theory of graph energy: hypoenergetic graphs, in: D. Cvetković, I. Gutman (Eds.), Applications of Graph Spectra, Math. Inst., Belgrade, 2009, pp. 65–105.
• [13] I. Stewart, D. Tall, Algebraic Number Theory, 2nd ed, Chapman and Hall, London- New York, 1987.
You are adding the first comment!
How to quickly get a good reply:
• Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
• Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
• Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters