On Topological Indices for New Classes of Benes Network

Department of Mathematics, King Abdulaziz University, P.O. Box 80203, Jeddah 21589, Saudi Arabia Department of Mathematics, COMSATS University, Attock Campus, Islamabad, Pakistan Department of Mathematics, COMSATS University, Lahore Campus, Islamabad, Pakistan Department of Natural Sciences and Humanities, University of Engineering and Technology, Lahore (RCET), Pakistan Department of Mathematics, COMSATS University, Islamabad Vehari Campus, Vehari 61110, Pakistan


Introduction
e processing nodes in an interconnection network are the multiprocessor used to build a network of homogeneously same processor memory pairs. Programs are compiled and executed through message sending. Considerable importance to the architecture and utilization of multiprocessor interconnection network is due to low cost, more efficient microprocessors, and chips [1]. Interconnection networks resembled the communication pattern of the natural scenario, which make it more valuable and important. Most of the networks are interconnected and dependent on each other which need to be reviewed for the future work.
Graphs are used to design interconnected networks in a very natural way, in which the processors or components represent vertices and edges represent the communication links, e.g., fiber optic cables. e way in which all these components work will be carried out by incidence functions. Graphs show the topological properties of the networks. erefore, the graph and networks are basically the same in a sense; that is, when we are considering networks, components, and links, we actually speak of graphs, vertices, and edges. In Fourier transform networks, butterfly graphs are the basic graphs that can accomplish FFT very proficiently.
ere are series of interconnection patterns and switching stages in a butterfly network that permits n inputs to be linked to n outputs. e Benes network consists of butterflies connected back to back. e Benes network is known for permutation routing [2]. ey are significant multistage interconnection networks, which enjoy striking topologies for communication networks [3]. e use of Benes network is in parallel computing systems such as SP1/SP2, IBM, NEC Cenju-3, and MIT Transit Project. It is also used in the internal structures of optical couplers [4,5]. ere are 2r + 1 levels in an r-dimensional Benes network where every level has 2r nodes. An r-dimensional butterfly is formed between the level 0 to level r nodes. e back to back butterflies share the middle level to form a Benes network. We denote an r-dimensional Benes by B(r). Figure 1 shows the graph 3dimensional Benes network B (3).
From now onward, we denote by G a connected, simple graph with vertex and edge set by V(G) and E(G). Moreover, for a ∈ V(G), d a , N(a), and S a present its degree, neighbors, and sum of degrees. For more details on undefined terminology, we refer the readers to [6][7][8].
Several invariants assigned to graphs (or molecular structure) establish correlations between its structure and chemical/physical aspects. ese graph invariants are named as topological indices. Several degree-dependent TIs have been formulated so far and are found to be useful in predicting the physical/chemical properties of an underlying molecular structure. For instance, there is a virtuous correlation among the Randic � index (RI) and numerous physicochemical features of alkanes including chromatographic retention times, boiling points, and surface areas. e Zagreb indices (ZIs) and its different variants contribute to understand the complexity of (molecular) structures [9][10][11][12][13], chirality [14], ZE isomerism [15], and heterosystems [16], whilst the overall ZIs show a latent applicability to derive multilinear regression models. e ABC index offers a good model for the strain energy of cycloalkanes and stability of linear and branched alkanes [17,18]. In this respect, computing different forms of TIs has provided the indicators of such medicinal conduct of numerous compounds and drugs. e mathematical forms of degree-based topological descriptors, which we discuss in this article, are depicted in Table 1. For more details on computation of TIs of different chemical structures, the readers can see [29][30][31][32][33][34][35][36].
In this paper, we constructed two classes of Benes network, i.e., horizontal cylindrical Benes network HCB(r) and vertical cylindrical Benes network VCB(r) obtained by identification of vertices of first rows with last row and first column with last column of Benes network, respectively. Also, we embed the Benes network on torus denoted by TB(r). Figures 2 and 3 show the graph of horizontal and vertical Benes networks, respectively. e graph of embedding of Benes network on torus is shown in Figure 4. We compute R α , M α , χ α , M 1 , M 2 , PM 1 , PM 2 , HM, ABC, and GA indices of these networks. Also, the explicit formula of ABC 4 and GA 5 indices is computed for these networks.

Results for Cylindrical Representation of Benes Network
In this section, we construct two types of cylindrical Benes network by identifying its vertices. First, we identify the vertices of top row to the vertices of bottom row of Benes network B(r). is will result in a cylindrical network, and we call it a horizontal cylindrical Benes network and denote it by HCB(r). Following the construction, a 3-dimensional horizontal cylindrical Benes network HCB(3) is presented in Figure 2, where the same label to the vertices represents identification of vertices. Clearly, |V(HCB(r))| � (2r + 1)(2 r − 1) and |E(HCB(r))| � 2r(2 r+1 − 1), respectively. Now, we compute the R α , M α , and χ α of horizontal cylindrical Benes network HCB(r).

Theorem 1. Let H be the graph of horizontal cylindrical
Benes network (HCB(r)), then Proof. In graph H, we have (2r + 1)(2 r − 1) vertices among which 2(2 r − 2) vertices are of degree 2, 2 vertices are of degree 3, (2r − 1)(2 r − 2) vertices are of degree 4, and the remaining (2r − 1) vertices are of degree 6. Now, by using the formula of general Zagreb index, we get To compute the R α and χ α indices, we need the partition of the edges of G based on the degree of the end vertices. is partition is given in Table 1. Using the values from Table 1, we can compute the R α and χ α indices as follows: e values of first Zagreb, second Zagreb, and hyper-Zagreb indices can be computed from the eorem 6 by taking the value of α � (−1/2), 1, 2.

Corollary 1. Let H be the graph of horizontal cylindrical Benes network, then
, Next, we compute the PM 1 , PM 2 , ABC, and GA indices of horizontal cylindrical Benes network.
Proof. e result follows by using the values of the edge partition given in Table 2 to the formulas of the PM 1 , PM 2 , ABC, and GA indices presented in Table 1.

Journal of Mathematics
Now, we compute the values ABC 4 and GA 5 indices of HCB(r), where r ≥ 4. is requires the partition of its edges based on the sum of degree of neighboring vertices of end vertices of each edge. is partition is given in Table 3.

Theorem 3. Let H be the graph of Horizontal cylindrical
Benes network HCB(r), r ≥ 4, then , Proof. Let m i,j denotes the number of edges of H with i � S a and j � S b . en, by using Table 3, ABC 4 and GA 5 indices can be computed as follows: is completes the proof. Next, we consider the vertical identification of the vertices. We identify the vertices on the left column with the corresponding vertices on the right column to form a cylindrical network. We name this network as the vertical cylindrical Benes network and denote it by VCB(r). e graph of VCB(3) is shown in Figure 3, where the arcs from vertical sides represent vertical sides identification. e rdimensional vertical cylindrical Benes network has r2 r+1 vertices and r2 r+2 edges. One can see that VCB(r) is a 4regular graph. erefore, the computation of TIs for VCB(r) is straightforward and is presented in the following theorem.

Results for Toroidal Representation of Benes Network TB(r)
If we identify the vertices in the top row of VCB(r) to the corresponding vertices of the bottom row, we get a toroidal network. We call this network Toroidal Benes network and denote it by TB(r). Figure 4 shows the graph of toroidal Benes network TB(3). Clearly, |V(TB(r))| � (2r + 1)(2 r − 1) and |E(TB(r))| � 2r(2 r+1 − 1), respectively. Now, we compute the R α , χ α , and M α of toroidal Benes network TB(r).

Theorem 5. Let L be the graph of toroidal cylindrical Benes network, then
Proof. In graph L, we have (2r + 1)(2 r − 1) vertices among which 2r(2 r − 2) vertices are of degree 4 and the remaining 2r vertices are of degree 6. By using the formula of general Zagreb index, we get To compute the general Randic and general sum connectivity indices, we need the partition of the edges of L based on the degree of the end vertices. is partition is given in Table 4. Using the values from Table 4, we can compute the Randic and general sum connectivity indices as follows: 4 Journal of Mathematics � 2r 2 r+1 − 6 (4 + 4) α + 8r ×(4 + 6) α + 2r(6 + 6) α � 2r 2 r+1 − 6 × 8 α + 8r × 10 α + 2r × 12 α .
e values of first Zagreb, second Zagreb, and hyper-Zagreb indices can be computed from the eorem 6 by taking value of α � (−1/2), 1 and 2.

Corollary 2. Let L be the graph of toroidal Benes network, then
Proof. e result follows by using the values of the edge partition given in Table 4 to the formulas of the PM 1 , PM 2 , ABC, and GA indices presented in Table 1. Now, we compute the values of ABC 4 and GA 5 indices of TB(r), where r ≥ 4.
is requires the partition of its edges based on the sum of degree of neighboring vertices of end vertices of each edge.
is partition is given in Table 5.
Theorem 7. Let L be the graph of toroidal Benes network TB(r), r ≥ 4, then , Proof. Let m i,j denote the number of edges of G with i � S a and j � S b . en, by using Table 5, ABC 4 and GA 5 indices can be computed as follows: is completes the proof.

Conclusion and General Remarks
Designing new network structure always attract and open ways for the researchers in the networking and other structural sciences. In this paper, we constructed new classes of Benes networks embedded on the surface of cylinder and torus called horizontal cylindrical Benes network HCB(r), vertical cylindrical Benes network VCB(r), and toroidal Benes network TB(r). en, some degree-based TIs are studied for the abovementioned networks. In future, it will help to those who are interested in problems related to interconnected networks and will be able to deal with the complex networks and their topological properties.

Data Availability
No additional data set has been used to support the study.

Disclosure
is research was carried out as a part of the employment of the authors.

Conflicts of Interest
e authors declare that there are no conflicts of interest.