Abstract
Calculating the number of spanning trees of a graph is one of the widely studied graph problems since the Pioneer Gustav Kirchhoff (1847). In this work, using knowledge of difference equations we drive the explicit formulas for the number of spanning trees in the sequence of some Nonahedral (nine faced polyhedral) graphs generated by triangle using electrically equivalent transformations and rules of the weighted generating function. Finally, we evaluate the entropy of graphs in this manuscript with different studied graphs with an average degree being 4, 5 and 6.
Keywords: Mathematics, Number of spanning trees, Entropy, Electrically equivalent transformations
Mathematics; Number of spanning trees; Entropy; Electrically equivalent transformations.
1. Introduction
The trouble of counting spanning trees turns to be essential and more importantly, interesting. For instance, it has been shown that, if the graph represents an electrical community with each edge a unit resistor, the effective resistance of an edge is equal to the proportion of spanning trees that the edge is in. Also, the wide variety of spanning trees is used as an invariant for computing the entropy of certain networks related to physical processes. In addition, there are various applications of the wide variety of spanning trees within mathematics as well [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12].
A spanning tree of a connected graph is a subgraph that is a tree reaching all vertices.
There exist numerous strategies for finding the number of spanning trees of a graph .
A classic technique called the matrix tree theorem, also called Kirchhoff''s matrix-tree theorem [13] which states that the number of nonidentical spanning trees of a graph is same to any cofactor of its Laplacian matrix , in which is the degree matrix and is the adjacency matrix of the graph .
Another method to count this number is using Laplacian eigenvalues. Kelmans and Chelnoknov [14] derived the following formula:
(1.1) |
in which is a connected graph with vertices and are the eigenvalues of the Laplacian matrix .
One popular technique for finding the number of spanning tress, , is the deletion-contraction method. This technique is a reliable method which lets into enumerate the number of spanning trees of a multigraph . This method makes use of the fact that
(1.2) |
where denotes the graph obtained by deleting an arbitrary edge , and denotes the graph obtained by contracting an arbitrary edge [15, 16]. For more results, see [17, 18, 19, 20].
2. Electrically equivalent transformations
An electrical network is an interconnection of electrical components (e.g. inductors, capacitors, batteries, resistors, switches, etc).
Kirchhoff's motivation was studied of electrical networks: an edge-weighted graph can be regarded as an electrical network, where weights are the conductance of the respective edges. The effect conductance between two specific nodes can be written as the quotient of (weighted) number of spanning trees and the (weighted) number of so called thickets, i.e., spanning forests with exactly two components and property that each of the components contains precisely one of the nodes [21, 22]. Next, we list the effect of some simple transformations on the number of spanning trees, suppose that is an edge weighted graph, is the corresponding electrically equivalent graph and denotes the weighted number of spanning trees of .
-
i.
Parallel edges: when two parallel edges with conductances and in are merged into a single edge with conductances in , then
-
ii.
Serial edges: when two serial edges with conductances and in are merged into a single edge with conductance in , then .
-
iii
transformation: when a triangle with conductances and in is changed into an electrically equivalent star graph with conductances and in , then
-
iv
transformation: when a star graph with conductances and in is changed into an electrically equivalent triangle with conductances and in , then
3. Main results
In mathematics, one constantly attempts to get new structure from given once. This also applies to the realm of graphs, where one can generate many new graphs from given one.
In the following we take into account six nonahedral graphs generated by triangle and we derive the explicit formulas for the number of spanning trees of sequences of those graphs.
Consider the sequence of graphs formed as illustrated in Figure 1.
Figure 1.
Some sequences of the graph Gn.
According to this formation, the number of total vertices and edges are , It is obvious that the average degree is convergently for a large .
Theorem 1
The number of spanning trees in the sequence of the graph , where , is given by
Proof: Let us be using the electrically equivalent transformation to transform to . Figure 2 clarifies the transformation process from to
Figure 2.
The transformations from G2 to G1.
By utilizing the properties that are given in section 2, the following the transformations are given:
Merging these seven transformations, we get
(3.1) |
Moreover,
(3.2) |
where
Its characteristic equation is which has two roots and . Subtracting these two roots into both sides of , we have
(3.3) |
(3.4) |
Let . Then by Eqs. (3.3) and (3.4), we get and
Thus Therefore,
(3.5) |
Utilizing the expression and indicating the coefficients of and as and , we obtain
(3.6) |
(3.7) |
Thus, we get
(3.8) |
where a0 = 0, b0 = 1 and a1 = 9, b1 = 8. By the expression and Eqs. (3.6) and (3.7), we obtain
(3.9) |
The characteristic equation of Eq. (3.9) is which has two roots and . The general solution of Eq. (3.9) are .
Utilizing the initial conditions and , yields
(3.10) |
If , yields has no any electrically equivalent transformation. Substituting Eq. (3.10) into Eq. (3.8), we get
(3.11) |
When , which verifies Eq. (3.11). Thus, the number of spanning trees in the sequence of the graph is given by
(3.12) |
Where
(3.13) |
Putting Eq. (3.13) into Eq. (3.12), the result is obtained.
Consider the sequence of graphs formed as illustrated in Figure 3.
Figure 3.
Some sequences of the graph Hn.
According to this formation, the number of total vertices and edges are , It is obvious that the average degree is convergently for a large .
Theorem 2
The number of spanning trees in the sequence of the graph , where , is given as
Proof: The electrically equivalent transformation to transform to is using. Figure 4 clarifies the transformation process from to
Figure 4.
The transformations from H2 to H1.
By utilizing the properties that are given in section 2, the following the transformations are given:
Merging these nine transformations, we get
(3.14) |
Moreover,
(3.15) |
where
Its characteristic equation is which has two roots and . Subtracting these two roots into both sides of , we get
(3.16) |
(3.17) |
Let . Then by Eqs. (3.16) and (317), we get and
Thus Therefore,
(3.18) |
Utilizing the expression and indicating the coefficients of and as and , we obtain
(3.19) |
(3.20) |
Thus, we obtain
(3.21) |
where and . By the expression and Eqs. (3.19) and (3.20), we have
(3.22) |
The characteristic equation of Eq. (3.22) is which has two roots and . The general solution of Eq. (3.22) are .
Utilizing the initial conditions and , yields
(3.23) |
If , yields has no any electrically equivalent transformation. Substituting Eq. (3.23) into Eq. (3.21), we get
(3.24) |
When , which verifies Eq. (3.24). Thus, the number of spanning trees in the sequence of the graph is given by
(3.25) |
where
(3.26) |
Putting Eq. (3.26) into Eq. (3.25), we obtain the result.
Consider the sequence of graphs formed as illustrated in Figure 5.
Figure 5.
Some sequences of the graph Tn.
According to this formation, the number of total vertices and edges are , It is obvious that the average degree is convergently for a large .
Theorem 3
The number of spanning trees in the sequence of the graph , where , is given by
Proof: The electrically equivalent transformation to transform to is using. Figure 6 clarifies the transformation process from to
Figure 6.
The transformations from T2 to T1.
By utilizing the properties that are given in section 2, the following the transformations are given:
Merging these nine transformations, we get
(3.27) |
Moreover,
(3.28) |
where
Its characteristic equation is which has two roots and . Subtracting these two roots into both sides of , we have
(3.29) |
(3.30) |
Let . Then by Eqs. (3.29) and (3.30), we have and
Thus Therefore,
(3.31) |
Utilizing the expression and indicating the coefficients of and as and , we obtain
(3.32) |
(3.33) |
Thus, we get
(3.34) |
where and . By the expression and Eqs. (3.32) and (3.33), we obtain
(3.35) |
The characteristic equation of Eq. (3.35) is which has two roots and . The general solution of Eq. (3.35) are .
Utilizing the initial conditions and , yields
(3.36) |
If , yields has no any electrically equivalent transformation. Substituting Eq. (3.36) into Eq. (3.34), we get
(3.37) |
When , which verifies Eq. (3.37). Thus, the number of spanning trees in the sequence of the graph is given by
(3.38) |
where
(3.39) |
Putting Eq. (3.39) into Eq. (3.38), then the result is obtained.
Consider the sequence of graphs formed as illustrated in Figure 7.
Figure 7.
Some sequences of the graph Xn.
According to this formation, the number of total vertices and edges are , It is obvious that the average degree is convergently for a large .
Theorem 4
The number of spanning trees in the sequence of the graph , where , is given by
Proof: The electrically equivalent transformation to transform to is using. Figure 8 clarifies the transformation process from to
Figure 8.
The transformations from X2 to X1.
By utilizing the properties that are given in section 2, the following the transformations are given:
Merging these seven transformations, we get
(3.40) |
Moreover,
(3.41) |
where
Its characteristic equation is which has two roots and . Subtracting these two roots into both sides of , we have
(3.42) |
(3.43) |
Let . Then by Eqs. (3.42) and (3.43), we get and
Thus Therefore,
(3.44) |
Utilizing the expression and indicating the coefficients of and as and , we obtain
(3.45) |
(3.46) |
Thus, we obtain
(3.47) |
where and . By the expression and Eqs. (3.45) and (3.46), we get
(3.48) |
The characteristic equation of Eq. (3.48) is which has two roots and . The general solution of Eq. (3.48) are .
Utilizing the initial conditions and , yields
(3.49) |
If , yields has no any electrically equivalent transformation. Substituting Eq. (3.49) into Eq. (3.47), we get
(3.50) |
When , which verifies Eq. (3.50). Thus, the number of spanning trees in the sequence of the graph is given by
(3.51) |
where
(3.52) |
Putting Eq. (3.52) into Eq. (3.51), the result is obtained.
Consider the sequence of graphs formed as illustrated in Figure 9.
Figure 9.
Some sequences of the graph Yn.
According to this formation, the number of total vertices and edges are It is obvious that the average degree is convergently for a large .
Theorem 5
The number of spanning trees in the sequence of the graph , where , is given by
Proof: The electrically equivalent transformation to transform to is using. Figure 8 clarifies the transformation process from to (see Figure 10).
Figure 10.
The transformations from Y2 to Y1.
By utilizing the properties that are given in section 2, the following the transformations are given:
Merging these nine transformations, we obtain
(3.53) |
Moreover,
(3.54) |
where
Its characteristic equation is which has two roots and . Subtracting these two roots into both sides of , we have
(3.55) |
(3.56) |
Let . Then by Eqs. (3.55) and (3.56), we have t and
Thus Therefore,
(3.57) |
Utilizing the expression and indicating the coefficients of and as and , we obtain
(3.58) |
(3.59) |
Thus, we have
(3.60) |
where and . By the expression and Eqs. (3.58) and (3.59), we have
(3.61) |
The characteristic equation of Eq. (3.61) is which has two roots and . The general solution of Eq. (3.61) are .
Utilizing the initial conditions and , yields
(3.62) |
If , yields has no any electrically equivalent transformation. Substituting Eq. (3.62) into Eq. (3.60), we get
(3.63) |
When , which verifies Eq. (3.63). Thus, the number of spanning trees in the sequence of the graph is given by
(3.64) |
where
(3.65) |
Putting Eq. (3.65) into Eq. (3.64), then the result is obtained.
Consider the sequence of graphs formed as illustrated in Figure 11.
Figure 11.
Some sequences of the graph Zn.
According to this formation, the number of total vertices and edges are , It is obvious that the average degree is convergently for a large .
Theorem 6
The number of spanning trees in the sequence of the graph , where , is given by , where , is given by
Proof: Let us be using the electrically equivalent transformation to transform to . Figure 2 clarifies the transformation process from to
Proof: The electrically equivalent transformation to transform to is using. Figure 12 clarifies the transformation process from to
Figure 12.
The transformations from Z2 to Z1.
By utilizing the properties that are given in section 2, the following the transformations are given:
Merging these nine transformations, we obtain
(3.66) |
Moreover,
(3.67) |
where
Its characteristic equation is which has two roots and . Subtracting these two roots into both sides of , we have
(3.68) |
(3.69) |
Let . Then by Eqs. (3.68) and (3.69), we have and
Thus Therefore,
(3.70) |
Utilizing the expression and indicating the coefficients of and as and , we obtain
(3.71) |
(3.72) |
Thus, we get
(3.73) |
where and . By the expression and Eqs. (3.71) and (3.72), we have
(3.74) |
The characteristic equation of Eq. (3.74) is which has two roots and . The general solution of Eq. (3.74) are .
Utilizing the initial conditions and , yields
(3.75) |
If , yields Zn has no any electrically equivalent transformation. Substituting Eq. (3.75) into Eq. (3.73), we get
(3.76) |
When , which verifies Eq. (3.76). Thus, the number of spanning trees in the sequence of the graph Zn is given by
(3.77) |
where
(3.78) |
Putting Eq. (3.77) into Eq. (3.78), hence the result is obtained.
4. Numerical results
Next tables illustrate some the values of the number of spanning trees in the graphs and Zn.
3 | 3 | 3 | |
2 | 8664 | 1176 | 1734 |
3 | 22852800 | 596748 | 881292 |
4 | 60019201536 | 303141984 | 447690264 |
5 | 157597728780288 | 153993738288 | 227423130672 |
6 | 413814073710182400 | 78227606477184 | 115529159623776 |
n | |||
---|---|---|---|
1 | 3 | 3 | 3 |
2 | 31104 | 8214 | 36300 |
3 | 188940288 | 18960588 | 320498688 |
4 | 1136394240000 | 43761009624 | 2829362006208 |
5 | 6833482751803392 | 101000334380592 | 24977602663502592 |
6 | 41091617468631220224 | 233108596706389344 | 220502231043611492352 |
5. Spanning tree entropy
After having explicit formulas for the number of spanning trees of the sequence of the six graphs and Zn, we can calculate its spanning tree entropy Z which is a finite number and a very interesting quantity characterizing the network structure, defined in [23, 24] : for a graph ,
(5.1) |
By comparing the value of entropy in our graphs with other graphs, we notice that the entropy of the graph is larger than the entropy of the graph of the same average degree 5. Also, the entropy of the graph Zn is larger than the entropy of the graph of the same average degree 6. In addition the entropy of the graphs of the same average degree 4 are equal and smaller than the entropy of the fractal scale-free lattice [25] which has the entropy , two dimensional Sierpinski gasket [26] which has the entropy and the 3-prism graph [27] which has the entropy .
6. Conclusions
In this paper, we have calculated the number of spanning trees in the sequences of some nonahedral (polyhedral graphs having nine vertices) graphs generated by triangle using electrically equivalent transformations. The feature of this technique lies in the parry of strenuous computation of Laplacian spectra that is prerequisite for a generic method for determining spanning trees. Also, our results have been shown that entropy is related to the average degree of the graphs.
Declarations
Author contribution statement
S. Daoud, W. Saleh: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.
Funding statement
Authors were supported by Taibah University.
Competing interest statement
The authors declare no conflict of interest.
Additional information
No additional information is available for this paper.
Acknowledgements
The authors are grateful to the anonymous reviewers for their helpful comments and suggestions for improving the original version of the paper.
References
- 1.Applegate D.L., Bixby R.E.V., Chvátal, Cook W.J. Princeton University Press; 2006. The Traveling Salesman Problem: A Computational Study. [Google Scholar]
- 2.Cvetkoviĕ D., Doob M., Sachs H. third ed. Johann Ambrosius Barth; Heidelberg: 1995. Spectra of Graphs: Theory and Applications. [Google Scholar]
- 3.Kirby E.C., Klein D.J., Mallion R.B., Pollak P., Sachs H. A theorem for counting spanning trees in general chemical graphs and its particular application to toroidal fullerenes. Croat. Chem. Acta. 2004;77:263–278. [Google Scholar]
- 4.Boesch F.T., Salyanarayana A., Suffel C.L. A survey of some network reliability analysis and synthesis results. Networks. 2009;54:99–107. [Google Scholar]
- 5.Boesch F.T. On unreliability polynomials and graph connectivity in reliable network synthesis. J. Graph Theor. 1986;10:339–352. [Google Scholar]
- 6.Wu F.Y. Number of spanning trees on a Lattice. J. Phys. 1977;10:113–115. [Google Scholar]
- 7.Zhang F., Yong X. Asymptotic enumeration theorems for the number of spanning trees and Eulerian trail in circulant digraphs & graphs. Sci. China, Ser. A. 1999;43:264–271. [Google Scholar]
- 8.Chen G., Wu B., Zhang Z. Properties and applications of Laplacian spectra for Koch networks. J. Phys. Math. Theor. 2012;45 [Google Scholar]
- 9.Atajan T., Inaba H. Network reliability analysis by counting the number of spanning trees, ISCIT 2004. IEEE Int. Symposium Communication Info. Technol. 2004;1:601–604. [Google Scholar]
- 10.Brown T.J.N., Mallion R.B., Pollak P., Roth A. Some methods for counting the spanning trees in labelled molecular graphs, examined in relation to certain fullerenes. Discrete Appl. Math. 1996;67:51–66. [Google Scholar]
- 11.Myrvold W., H Cheung K., Page L.B., Perry J.E. Uniformly-most reliable networks do not always exist. Networks. 1991;21:417–419. [Google Scholar]
- 12.Petingi L., Boesch F., Suffel C. On the characterization of graphs with maximum number of spanning trees, Discrete Appl. Math. 1998;179:155–166. [Google Scholar]
- 13.Kirchhoff G.G. Über die Auflösung der Gleichungen auf welche man bei der Untersucher der linearen Verteilung galuanischer Strome gefhrt wird. Ann. Phg. Chem. 1847;72:497–508. [Google Scholar]
- 14.Kelmans A.K., Chelnokov V.M. A certain polynomial of a graph and graphs with an extremal number of trees. J. Combin. Theor. B. 1974;16:197–214. [Google Scholar]
- 15.Biggs N.L. second ed. Cambridge Univ. Press; Cambridge: 1993. Algebraic Graph Theory; p. 205. [Google Scholar]
- 16.Daoud S.N. The deletion- contraction method for counting the number of spanning trees of graphs. Eur. J. Plant Pathol. 2015;130(10):1–14. [Google Scholar]
- 17.Daoud S.N. Complexity of graphs generated by wheel graph and their asymptotic limits. J. Egypt. Math. Soci. 2017;25(4):424–433. [Google Scholar]
- 18.Daoud S.N. Number of spanning trees in different products of complete and complete tripartite graphs. Ars Cmbinatoria. 2018;139:85–103. [Google Scholar]
- 19.Daoud S.N. Number of Spanning trees of cartesian and composition products of graphs and Chebyshev polynomials. IEEE AccessVol. 2019;7:71142–71157. [Google Scholar]
- 20.Liu Jia-Bao, Daoud S.N. Complexity of some of pyramid graphs created from a gear graph. Symmetry. 2018;10:689. [Google Scholar]
- 21.Teufl E., Wagner S. Determinant identities for Laplace matrices. Linear Algebra Appl. 2010;432:441–457. [Google Scholar]
- 22.Liu Jia-Bao, Daoud S.N. Vol. 2019. Hindawi Publ. Corp.; 2019. p. 22. (Number of Spanning Trees in the Sequence of Some Graphs, Complecity, Complexity). Article ID 4271783. [Google Scholar]
- 23.Wu F.Y. Number of spanning trees on a lattice. J. Phys. Math. Gen. 1977;10:113–115. [Google Scholar]
- 24.Lyons R. Asymptotic enumeration of spanning trees. Combinator. Probab. Comput. 2005;14:491–522. [Google Scholar]
- 25.Zhang Z., Liu H., Wu B., Zou T. Spanning trees in a fractal scale –free lattice. Phys. Rev. E. 2011;83 doi: 10.1103/PhysRevE.83.016116. [DOI] [PubMed] [Google Scholar]
- 26.Chang S., Chen L., Yang W. Spanning trees on the Sierpinski gasket. J. Stat. Phys. 2007;126:649–667. [Google Scholar]
- 27.Sun W., Wang S., Zhang J. Counting spanning trees in prism and anti-prism Graphs. J. Appl. Anal. And Comp. 2016;6:65–75. [Google Scholar]