This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

Invariants of Bipartite Kneser B type-k graphs

Jayakumar C Department of Mathematics, University College, Thiruvananthapuram, India. [email protected] Sreekumar K.G Department of Mathematics, University of Kerala, Thiruvananthapuram, India. [email protected] Manilal K Department of Mathematics, University College, Thiruvananthapuram, India. [email protected]  and  Ismail Naci Cangul Department of Mathematics, Bursa Uludag University, Bursa 16059 Bursa-Turkey. [email protected]
Abstract.

Let n={±x1,±x2,±x3,,±xn1,xn}\mathscr{B}_{n}=\{\pm x_{1},\pm x_{2},\pm x_{3},\cdots,\pm x_{n-1},x_{n}\} where n>1n>1 is fixed, xi+x_{i}\in\mathbb{R}^{+}, i=1,2,3,,ni=1,2,3,\cdots,n and x1<x2<x3<<xnx_{1}<x_{2}<x_{3}<\cdots<x_{n}. Let ϕ(n)\phi(\mathscr{B}_{n}) be the set of all non-empty subsets S={u1,u2,,ut}S=\{u_{1},u_{2},\cdots,u_{t}\} of n\mathscr{B}_{n} such that |u1|<|u2|<<|ut1|<ut|u_{1}|<|u_{2}|<\cdots<|u_{t-1}|<u_{t} where ut+u_{t}\in\mathbb{R}^{+}. Let n+={x1,x2,x3,,xn1,xn}\mathscr{B}_{n}^{+}=\{x_{1},x_{2},x_{3},\cdots,x_{n-1},x_{n}\}. For a fixed kk, let V1V_{1} be the set of kk-element subsets of n+\mathscr{B}_{n}^{+}, 1k<n1\leq k<n. V2=ϕ(n)V1V_{2}=\phi(\mathscr{B}_{n})-V_{1}. For any AV2A\in V_{2}, let A={|x|:xA}A^{\dagger}=\{\lvert x\rvert:x\in A\}. Define a bipartite graph with parts V1V_{1} and V2V_{2} and having adjacency as XV1X\in V_{1} is adjacent to YV2Y\in V_{2} if and only if XYX\subset Y^{\dagger} or YXY^{\dagger}\subset X. A graph of this type is called a bipartite Kneser B type-kk graph and denoted by HB(n,k)H_{B}(n,k). In this paper, we calculated various graph invariants of HB(n,k)H_{B}(n,k).

Key words and phrases:
Bipartite Kneser graphs, Bipartite Kneser B type-k graph, Degree sequence, cyclomatic number
2010 Mathematics Subject Classification:
05C07, 05C12, 05C69

1. Introduction

Named after the German mathematician Martin Kneser, Kneser graphs are an interesting family of combinatorial structures in the field of graph theory. These graphs have applications in several fields, such as algebraic geometry, combinatorics, and topology.

Numerous fields, such as combinatorics, topology, coding theory, and combinatorial optimisation have Kneser graph applications. These are fundamental building blocks of combinatorial theory and can lead to interesting problems and conjectures. In this subject, questions about their chromatic number [10] and other graph-theoretic properties continue to be crucial to research. Kneser graphs are related to topological problems, for example, by helping to understand the homotopy type of some spaces[6]. Kneser graphs are used in coding theory[2] to design codes with efficient error-correcting features.

The Kneser graph K(n,k)K(n,k) has the kk-subsets of [n][n] as vertices. If the kk-subsets are disjoint, then two vertices are adjacent. For integers k1k\geq 1 and n2k+1n\geq 2k+1, any vertex in the bipartite Kneser graph H(n,k)H(n,k) is either a kk-element subset or an nkn-k element subset of [n][n]. Here, the sets AA and BB are adjacent if ABA\subseteq B.

The algebraic structures of different varieties of bipartite Kneser graphs, [1], [7], were then built and investigated.

Here is a modified version of the bipartite Kneser graph : Let 𝒮n={1,2,3,,n}\mathscr{S}_{n}=\{1,2,3,\cdots,n\} for a fixed integer n>1n>1. Let ϕ(𝒮n)\phi(\mathscr{S}_{n}) be the set of all non-empty subsets of 𝒮n\mathscr{S}_{n}. Let V1V_{1} be the set of 1-element subsets of 𝒮n\mathscr{S}_{n} and V2=ϕ(𝒮n)V1V_{2}=\phi(\mathscr{S}_{n})-V_{1}. Define a bipartite graph with an adjacency of vertices as described: A vertex AV1A\in V_{1} is adjacent to a vertex BV2B\in V_{2} if and only if ABA\subset B. This graph is called a bipartite Kneser type-1 graph[8], and is denoted by HT(n,1)H_{T}(n,1). Sreekumar K. G. et al., [9], defined a bipartite Kneser B type-kk graph, G=HB(n,k)G=H_{B}(n,k), which are more general bipartite graphs.

This paper determines the following invariants of the bipartite Kneser B type-kk graph HB(n,k)H_{B}(n,k): Order, size, independence number, covering number, domination number, vertex connectivity, edge connectivity, girth, circuit rank, distance between two vertices, eccentricity, periphery, centre, median, and degree sequence.

2. Preliminaries

The greatest distance between any two vertices in a graph is known as its diameter. The eccentricity of a vertex, e(v)e(v), is the largest possible distance between it and any other vertex. The maximum eccentricity obtained by the vertices of a connected simple graph GG is the diameter, diam(G)diam(G). The least eccentricity among all vertices of GG is the radius, rad(G)rad(G). The centre C(G)C(G) of a graph GG is the subgraph induced by the set of vertices with the lowest eccentricity. The periphery of GG is P(G)={vV:e(v)=diam(G)}P(G)=\{v\in V:e(v)=diam(G)\}. The length of the shortest cycle in a graph is its girth.For any vertex vv of a connected graph GG, the status of vv denoted as s(v)s(v) is the sum of the distances from vv to other vertices of GG. That is, s(v)=uV(G)d(v,u)s(v)=\sum\limits_{u\in V(G)}d(v,u). The set of vertices with minimal status is the median M(G)M(G) of the graph.

The girth of a graph is the length of the shortest cycle in it.

3. Basic definitions and examples

Definition 3.1.

Let n={±x1,±x2,±x3,,±xn1,xn}\mathscr{B}_{n}=\{\pm x_{1},\pm x_{2},\pm x_{3},\dots,\pm x_{n-1},x_{n}\} where n>1n>1 is fixed, xi+x_{i}\in\mathbb{R}^{+}, i=1,2,3,,ni=1,2,3,\dots,n and x1<x2<x3<<xnx_{1}<x_{2}<x_{3}<\dots<x_{n}. Let ϕ(n)\phi(\mathscr{B}_{n}) be the set of all non-empty subsets S={u1,u2,,ut}S=\{u_{1},u_{2},\dotsc,u_{t}\} of n\mathscr{B}_{n} such that |u1|<|u2|<<|ut1|<ut|u_{1}|<|u_{2}|<\dotsc<|u_{t-1}|<u_{t} where ut+u_{t}\in\mathbb{R}^{+}. Let n+={x1,x2,x3,,xn1,xn}\mathscr{B}_{n}^{+}=\{x_{1},x_{2},x_{3},\dots,x_{n-1},x_{n}\}. For a fixed kk, let V1V_{1} be the set of kk-element subsets of n+\mathscr{B}_{n}^{+}, 1k<n1\leq k<n. V2=ϕ(n)V1V_{2}=\phi(\mathscr{B}_{n})-V_{1}. For any AV2A\in V_{2}, let A={|x|:xA}A^{\dagger}=\{\lvert x\rvert:x\in A\}. Define a bipartite graph with parts V1V_{1} and V2V_{2} and having adjacency as XV1X\in V_{1} is adjacent to YV2Y\in V_{2} if and only if XYX\subset Y^{\dagger} or YXY^{\dagger}\subset X. A graph of this type is called the bipartite Kneser B type-kk graph [9] and is denoted by HB(n,k)H_{B}(n,k).

Definition 3.2.

An rr-vertex in HB(n,k)H_{B}(n,k) is an element in ϕ(n)\phi(\mathscr{B}_{n}) containing rr elements, where 1rn1\leq r\leq n. Members of ϕ(n)\phi(\mathscr{B}_{n}) are called rr-vertices.

Example 3.3.

An example of a bipartite Kneser B type-1 graph, for n=2n=2,namely HB(2,1)H_{B}(2,1) is shown in figure 1.

Refer to caption
Figure 1. HB(2,1)H_{B}(2,1)
Example 3.4.

HB(n,k)H_{B}(n,k) for n=3,k=2n=3,\ k=2 is illustrated in figure 2.

The 22-vertex {1,2}V1\{1,2\}\in V_{1} is adjacent to the 11-vertex {1}\{1\} as {1}{1,2}={|1|,|2|}={1,2}\{1\}\subset\{1,2\}^{\dagger}=\{|1|,|2|\}=\{1,2\}. By the same argument, {1,2}\{1,2\} is adjacent to the 11-vertex {2}\{2\}. Also, {1,2}\{1,2\} is adjacent to {1,2}\{-1,2\} as {1,2}{1,2}={|1|,|2|}={1,2}\{1,2\}\subset\{-1,2\}^{\dagger}=\{|-1|,|2|\}=\{1,2\}. Similar arguments explain the adjacency shown in figure 2.

Refer to caption
Figure 2. HB(3,2)H_{B}(3,2)
Example 3.5.

Given the large size of HB(4,2)H_{B}(4,2), we will present its bipartition here. According to the definition of HB(n,k)H_{B}(n,k), a vertex AA in V1V_{1} is adjacent to a vertex BB in V2V_{2} if and only if ABA\subset B^{\dagger} or BAB^{\dagger}\subset A.

4={±1,±2,±3, 4}\mathscr{B}_{4}=\{\pm 1,\;\pm 2,\;\pm 3,\;4\},

V1={{1,2},{1,3},{1,4},{2,3},{2,4},{3,4}}V_{1}=\big{\{}\{1,2\},\{1,3\},\{1,4\},\{2,3\},\{2,4\},\{3,4\}\big{\}},

V2\displaystyle V_{2} ={{1},{2},{3},{4},{1,2},{1,3},{1,4}\displaystyle=\big{\{}\{1\},\{2\},\{3\},\{4\},\{-1,2\},\{-1,3\},\{-1,4\}
{2,3},{2,4},{3,4},{1,2,3},{1,2,3},{1,2,3},{1,2,3},\displaystyle\quad\{-2,3\},\{-2,4\},\{-3,4\},\{1,2,3\},\{-1,2,3\},\{1,-2,3\},\{-1,-2,3\},
{1,2,4},{1,2,4},{1,2,4},{1,2,4},{2,3,4},{2,3,4},\displaystyle\quad\{1,2,4\},\{-1,2,4\},\{1,-2,4\},\{-1,-2,4\},\{2,3,4\},\{-2,3,4\},
{2,3,4},{2,3,4},{1,3,4},{1,3,4},{1,3,4},{1,3,4}\displaystyle\quad\{2,-3,4\},\{-2,-3,4\},\{1,3,4\},\{-1,3,4\},\{1,-3,4\},\{-1,-3,4\}
{1,2,3,4},{1,2,3,4},{1,2,3,4},{1,2,3,4},{1,2,3,4}}\displaystyle\quad\{1,2,3,4\},\{-1,2,3,4\},\{1,-2,3,4\},\{1,2,-3,4\},\{-1,-2,3,4\}\big{\}}
{1,2,3,4},{1,2,3,4},{1,2,3,4}}.\displaystyle\quad\{1,-2,-3,4\},\{-1,2,-3,4\},\{-1,-2,-3,4\}\big{\}}.

While V1V_{1} contains six 22-vertices, V2V_{2} contains four 11-vertices, six 22-vertices, sixteen 33-vertices, and eight 44-vertices.

4. Some parameters of bipartite Kneser B type-kk graphs

Theorem 4.1.

The order of G=HB(n,k)G=H_{B}(n,k), |V(G)|=3n12|V(G)|=\dfrac{3^{n}-1}{2}.

Proof.

Every vertex of HB(n,k)H_{B}(n,k) is a set formed using the elements of n\mathscr{B}_{n}. Here n={±a1,±a2,±a3,,±an1,an},where ,a1<a2<anandann.\mathscr{B}_{n}=\{\pm a_{1},\pm a_{2},\pm a_{3},\dots,\pm a_{n-1},a_{n}\},\text{where },a_{1}<a_{2}\dots<a_{n}\;\text{and}\;-a_{n}\notin\mathscr{B}_{n}. HB(n,k)H_{B}(n,k) has vertices as subsets of cardinality 11 to nn. |V(G)||V(G)| is the total number of sets of cardinality 11 to nn.
Let NiN_{i} be the number of subsets of n\mathscr{B}_{n} of cardinality ii, where 1in1\leq i\leq n.
N1=1(n1)=20(n1),N2=2(n2),N3=22(n3),,Nn=2n1(nn)N_{1}=1\binom{n}{1}=2^{0}\binom{n}{1},N_{2}=2\binom{n}{2},N_{3}=2^{2}\binom{n}{3},\dotsc,N_{n}=2^{n-1}\binom{n}{n}.
Thus, |V(G)|=i=1nNi=20(n1)+2(n2)+22(n3)+2n1(nn)=3n12|V(G)|=\sum\limits_{i=1}^{n}N_{i}=2^{0}\binom{n}{1}+2\binom{n}{2}+2^{2}\binom{n}{3}+\dotsb 2^{n-1}\binom{n}{n}=\dfrac{3^{n}-1}{2}, by binomial theorem. ∎

Definition 4.2.

A set SV(G)S\subseteq V(G) is an independent set of GG if no two vertices of SS are adjacent in GG. The independence number of GG, denoted by α\alpha, is the number of vertices in a maximum independent set of GG. SS is maximum if it has the maximum cardinality among all independent subsets of GG.

Definition 4.3.

A set SS is a dominating set [5] if for every vertex uVSu\in V-S, there exists vSv\in S such that the edge uvEuv\in E. The domination number of GG, denoted by γ(G)\gamma(G), is the minimum cardinality of a dominating set of GG.

Theorem 4.4.

For G=HB(n,k)G=H_{B}(n,k), the vertex independence number,

α(G)=i=1n2i1(ni)(nk)\alpha(G)=\sum\limits_{i=1}^{n}2^{i-1}\binom{n}{i}-\binom{n}{k}, for n2n\geq 2.

Proof.

By the construction of HB(n,k)=V1V2H_{B}(n,k)=V_{1}\cup V_{2}, |V2||V_{2}| forms a maximum independent set and |V2|=|ϕ(n)|(nk)|V_{2}|=|\phi(\mathscr{B}_{n})|-\binom{n}{k}.
Therefore, α(G)=i=1n2i1(ni)(nk)\alpha(G)=\sum\limits_{i=1}^{n}2^{i-1}\binom{n}{i}-\binom{n}{k}. ∎

Corollary 4.5.

For G=HB(n,k)G=H_{B}(n,k), the vertex covering number, β(G)=(nk)\beta(G)=\binom{n}{k}.

Proof.

As α(G)+β(G)=|G|\alpha(G)+\beta(G)=|G|, we get β(G)=(nk)\beta(G)=\binom{n}{k}. ∎

Theorem 4.6.

The domination number, γ(HB(n,k))=(nk)\gamma(H_{B}(n,k))=\binom{n}{k}, for all n2n\geq 2 .

Proof.

In HB(n,k)H_{B}(n,k), no two vertices in V1V_{1} are adjacent, and every vertex in it is adjacent to some other vertex in V2V_{2}. Thus, V1V_{1} forms a dominating set for HB(n,k)H_{B}(n,k).
V1V_{1}, being the smallest dominating set,we get γ(HB(n,k))=(nk)\gamma(H_{B}(n,k))=\binom{n}{k}, for all n2n\geq 2. ∎

Theorem 4.7.

The size of G=HB(n,k)G=H_{B}(n,k) is |E(G)|=(nk)(3k32+2k1(3nk1))|E(G)|=\binom{n}{k}\left(\dfrac{3^{k}-3}{2}+2^{k-1}(3^{n-k}-1)\right).

Proof.

Let uu be a kk-vertex in V1V_{1}.
The number of 11-vertices in V2V_{2} adjacent to uu is 20(k1)2^{0}\binom{k}{1}.
The number of 22-vertices in V2V_{2} adjacent to uu is 21(k2)2^{1}\binom{k}{2}.
The number of 33-vertices in V2V_{2} adjacent to uu is 22(k3)2^{2}\binom{k}{3}.

The number of kk-vertices in V2V_{2} adjacent to uu is 2k1(kk)12^{k-1}\binom{k}{k}-1.
The number of k+1k+1-vertices in V2V_{2} adjacent to uu is 2k(nk1)2^{k}\binom{n-k}{1}.
The number of k+2k+2-vertices in V2V_{2} adjacent to uu is 2k+1(nk2)2^{k+1}\binom{n-k}{2}.

The number of nn-vertices in V2V_{2} adjacent to uu is 2n1(nknk)2^{n-1}\binom{n-k}{n-k}.
The degree of uu,

d(u)\displaystyle d(u) =20(k1)+21(k2)++2k1(kk)1+2k(nk1)\displaystyle=2^{0}\binom{k}{1}+2^{1}\binom{k}{2}+\dotsb+2^{k-1}\binom{k}{k}-1+2^{k}\binom{n-k}{1}
+2k+1(nk2)++2n1(nknk).\displaystyle\quad+2^{k+1}\binom{n-k}{2}+\dotsb+2^{n-1}\binom{n-k}{n-k}.
=(3k32+2k1(3nk1)).\displaystyle=\left(\dfrac{3^{k}-3}{2}+2^{k-1}(3^{n-k}-1)\right).

Since every vertex in V1V_{1} is of the same degree and d(u)d(u) is the maximum, we get, |E(G)|=(nk)(3k32+2k1(3nk1))|E(G)|=\binom{n}{k}\left(\dfrac{3^{k}-3}{2}+2^{k-1}(3^{n-k}-1)\right).

Theorem 4.8.

The vertex connectivity of G=HB(n,k)G=H_{B}(n,k) when k>1k>1 is κ(G)=1.\kappa(G)=1.

Proof.

As every kk-vertex, vv in V2V_{2}, is of degree 11, when a kk-vertex uu in V1V_{1} adjacent to vv is removed, vv becomes isolated. Accordingly, we get κ(G)=1\kappa(G)=1. ∎

Theorem 4.9.

The edge connectivity of G=HB(n,k)G=H_{B}(n,k) when k>1k>1is λ(G)=1\lambda(G)=1.

Proof.

Let vv be any kk-vertex in V2V_{2}. Since d(v)=1d(v)=1, the graph becomes disconnected when the only edge adjacent to it is removed. Consequently, we get λ(G)=1\lambda(G)=1. ∎

Theorem 4.10.

The Girth of G=HB(n,k)G=H_{B}(n,k) when k>1k>1 is Girth(G)=4.

Proof.

Let uu be a kk-vertex in V1V_{1}. Then, uu is adjacent to a 11-vertex xx in V2V_{2}, and xx is adjacent to another kk-vertex yy in V1V_{1}. The vertex yy is adjacent to an nn-vertex vv in V2V_{2} and vv is adjacent to uu. Thus, there is a cycle uxyvuu-x-y-v-u of length 44. Since any cycle in a bipartite graph is of even length, eventually, we get Girth(G)=4.

Theorem 4.11.

The circuit rank of G=HB(n,k)G=H_{B}(n,k) is
(nk)(20(k1)+21(k2)++2k1(kk)1+2k(nk1)+2k+1(nk2)++2n1(nknk))i=1n2i1(ni)+1\binom{n}{k}\Big{(}2^{0}\binom{k}{1}+2^{1}\binom{k}{2}+\dotsb+2^{k-1}\binom{k}{k}-1+2^{k}\binom{n-k}{1}+2^{k+1}\binom{n-k}{2}+\dotsb+2^{n-1}\binom{n-k}{n-k}\Big{)}\\ -\sum\limits_{i=1}^{n}2^{i-1}\binom{n}{i}+1.

Proof.

The circuit rank, which is also called the cyclomatic number of a graph, is nm+cn-m+c, where n, m, and c are the order, size, and the number of connected components, respectively. As c=1c=1 and n and m are already determined for HB(n,k)H_{B}(n,k), the result follows. ∎

Remark 4.12.

Cyclomatic  number is related to a recently defined graph invariant called the Omega invariant, which allows some combinatorial and graph theoretical properties to be calculated. The Omega invariant is an additive number defined for a given degree sequence with

(4.1) Ω(G)=i=1Δai(di2).\Omega(G)=\sum\limits_{i=1}^{\Delta}a_{i}(d_{i}-2).

It is shown that Ω(G)=2(mn)\Omega(G)=2(m-n), and therefore it is always an even number. For further properties of the Omega invariant, see [3].

Theorem 4.13.

The degree of every vertex in G=HB(n,k)G=H_{B}(n,k) and the number of vertices having a specific degree are determined. The degree sequence is obtained by arranging the sequence {dV2(1)NV2(1),dV2(2)NV2(2),dV2(k1)NV2(k1),dV1(k)NV1(k),dV2(k)NV2(k)\Big{\{}d_{V_{2}}(1)^{N_{V_{2}}(1)},d_{V_{2}}(2)^{N_{V_{2}}(2)}\dots,d_{V_{2}}(k-1)^{N_{V_{2}}(k-1)},d_{V_{1}}(k)^{N_{V_{1}}(k)},\\ d_{V_{2}}(k)^{N_{V_{2}}(k)}, dV2(k+1)NV2(k+1),dV2(n)NV2(n)}d_{V_{2}}(k+1)^{N_{V_{2}}(k+1)},\dots d_{V_{2}}(n)^{N_{V_{2}}(n)}\Big{\}}of degrees with corresponding multiplicities as a monotonic non-increasing sequence.

Proof.

Let dV2(r)d_{V_{2}}(r), where r=1,2,3,,k1,k+1,,nr=1,2,3,\dotsc,k-1,k+1,\dotsc,n, denote the degrees of rr-vertices in V2V_{2}, and dV1(k)d_{V_{1}}(k) denote the degree of any kk-vertex in V1V_{1}. Let the multiplicities of degrees of any kk-vertex in V1V_{1}, and rr-vertices in V2V_{2} be denoted by NV1(k)N_{V_{1}}(k) and  NV2(r)N_{V_{2}}(r) respectively.

We have, dV2(1)=(n1k1),NV2(1)=20(n1)d_{V_{2}}(1)=\binom{n-1}{k-1},\;N_{V_{2}}(1)=2^{0}\binom{n}{1}, dV2(2)=(n2k2),NV2(2)=21(n2)d_{V_{2}}(2)=\binom{n-2}{k-2},\;N_{V_{2}}(2)=2^{1}\binom{n}{2}, dV2(k1)=(n(k1)k(k1)),NV2(k1)=2k2(nk1)d_{V_{2}}(k-1)=\binom{n-(k-1)}{k-(k-1)},\;N_{V_{2}}(k-1)=2^{k-2}\binom{n}{k-1}, dV2(k+1)=(k+1k),NV2(k+1)=2k(nk+1)d_{V_{2}}(k+1)=\binom{k+1}{k},\;N_{V_{2}}(k+1)=2^{k}\binom{n}{k+1}, \cdots, dV2(n)=(nk)d_{V_{2}}(n)=\binom{n}{k}, and NV2(n)=2n1(nn)N_{V_{2}}(n)=2^{n-1}\binom{n}{n}.

Let uu be any kk-vertex in V1V_{1}. For s=1,2,3,,k1s=1,2,3,\dotsc,k-1, the number of ss-vertices adjacent to uu is 2s1(ks)2^{s-1}\binom{k}{s}. The number of kk-vertices adjacent to uu is 2k1(kk)12^{k-1}\binom{k}{k}-1. For s=k+1,k+2,,ns=k+1,k+2,\dotsc,n, the number of ss-vertices adjacent to uu is 2s1(nkt)2^{s-1}\binom{n-k}{t}. Here, t=1,2,,nkt=1,2,\dotsc,n-k. Hence, the degree of any kk-vertex,which is denoted as uu, in V1V_{1} is dV1(k)=20(k1)+21(k2)++2k1(kk)1+2k(nk1)+2k+1(nk2)++2n1(nknk)d_{V_{1}}(k)=2^{0}\binom{k}{1}+2^{1}\binom{k}{2}+\dotsb+2^{k-1}\binom{k}{k}-1+2^{k}\binom{n-k}{1}+2^{k+1}\binom{n-k}{2}+\dotsb+2^{n-1}\binom{n-k}{n-k}.

The number of kk-vertices in V1V_{1} is NV1(k)=(nk)N_{V_{1}}(k)=\binom{n}{k}. Every kk-vertex in V2V_{2} has degree dV2(k)=1d_{V_{2}}(k)=1. Then, the number of kk-vertices in V2V_{2} is NV2(k)=(2k11)(nk)N_{V_{2}}(k)=(2^{k-1}-1)\binom{n}{k}. Thus, the degree of every vertex in HB(n,k)H_{B}(n,k) and the number of vertices having a specific degree are determined. The degree sequence is obtained by arranging the sequence, {dV2(1)NV2(1),dV2(2)NV2(2),dV2(k1)NV2(k1),dV1(k)NV1(k),dV2(k)NV2(k)\Big{\{}d_{V_{2}}(1)^{N_{V_{2}}(1)},d_{V_{2}}(2)^{N_{V_{2}}(2)}\dots,d_{V_{2}}(k-1)^{N_{V_{2}}(k-1)},d_{V_{1}}(k)^{N_{V_{1}}(k)},d_{V_{2}}(k)^{N_{V_{2}}(k)}, dV2(k+1)NV2(k+1),dV2(n)NV2(n)}d_{V_{2}}(k+1)^{N_{V_{2}}(k+1)},\dots d_{V_{2}}(n)^{N_{V_{2}}(n)}\Big{\}} of degrees with corresponding multiplicities as a monotonic non-increasing sequence. ∎

Example 4.14.

The degree sequence for HB(4,2)H_{B}(4,2) is obtained by arranging the sequence, {dV2(1)NV2(1),dV1(2)NV1(2),dV2(2)NV2(2),dV2(3)NV2(3),dV2(4)NV2(4)}={34,196,16,316,68}\left\{d_{V_{2}}(1)^{N_{V_{2}}(1)},d_{V_{1}}(2)^{N_{V_{1}}(2)},d_{V_{2}}(2)^{N_{V_{2}}(2)},d_{V_{2}}(3)^{N_{V_{2}}(3)},d_{V_{2}}(4)^{N_{V_{2}}(4)}\right\}=\\ \{3^{4},19^{6},1^{6},3^{16},6^{8}\} of degrees with corresponding multiplicities as a monotonic, non-increasing sequence. That is, the degree sequence is {196,68,316,34,16}\{19^{6},6^{8},3^{16},3^{4},1^{6}\}.

The following result gives the Omega invariant of G=HB(n,k)G=H_{B}(n,k):

Theorem 4.15.

The Omega invariant of G=HB(n,k)G=H_{B}(n,k) is

Ω(HB(n,k))=(nk)i=1n(ki)2i1i=1n(ni)2i.\Omega(H_{B}(n,k))=\binom{n}{k}\sum\limits_{i=1}^{n}\binom{k}{i}2^{i-1}-\sum\limits_{i=1}^{n}\binom{n}{i}2^{i}.
Proof.

By the definition of Omega invariant, we have

Ω(HB(n,k))\displaystyle\Omega(H_{B}(n,k)) =i=1n(di2)Ni\displaystyle=\sum_{i=1}^{n}(d_{i}-2)N_{i}
=i=1n((niki)2)2i1(ni)\displaystyle=\sum_{i=1}^{n}\left(\binom{n-i}{k-i}-2\right)2^{i-1}\binom{n}{i}
=(nk)i=1n(ki)2i1i=1n(ni)2i.\displaystyle=\binom{n}{k}\sum_{i=1}^{n}\binom{k}{i}2^{i-1}-\sum_{i=1}^{n}\binom{n}{i}2^{i}.

In [3, 4], it was shown that the number of closed regions of a graph is given by

(4.2) r(G)=Ω(G)2+c(G).r(G)=\frac{\Omega(G)}{2}+c(G).

Here c(G)c(G) is the number of components of GG. Hence the number of faces of the graph HB(n,k)H_{B}(n,k) is given by the following result:

Theorem 4.16.

The number of faces of the graph HB(n,k)H_{B}(n,k) is

r(HB(n,k))=(nk)i=1n(ki)2i2i=1n(ni)2i1+c.r(H_{B}(n,k))=\binom{n}{k}\sum_{i=1}^{n}\binom{k}{i}2^{i-2}-\sum_{i=1}^{n}\binom{n}{i}2^{i-1}+c.
Proof.

It follows by the formula of rr given in Eqn.(4.2). ∎

Theorem 4.17.

Consider the graph G=HB(n,k),n>2, 1<k<nwithV(G)=V1V2G=H_{B}(n,k),\>n>2,\;1<k<n\;\;\text{with}\;V(G)=V_{1}\cup V_{2} and u,vV(G)u,v\in V(G), then

d(u,v)={1if uV1 and vV2 are adjacent,2if u and v are in V1,3if uV1 and vV2 are not adjacent,2,4if u and v are in V2.d(u,v)=\begin{cases}1&\text{if $u\in V_{1}$ and $v\in V_{2}$ are adjacent},\\ 2&\text{if $u$ and $v$ are in $V_{1}$},\\ 3&\text{if $u\in V_{1}$ and $v\in V_{2}$ are not adjacent},\\ 2,4&\text{if $u$ and $v$ are in $V_{2}$}.\end{cases}
Proof.

Let uV1u\in V_{1} and vV2v\in V_{2}. If uu and vv are adjacent, then d(u,v)=1d(u,v)=1. Suppose that vv is not adjacent to uu. Since the degree of vv is at least 11, it must be adjacent to some vertex wV1w\in V_{1}. Let xx be an nn-vertex in V2V_{2}. As xx is a common neighbour of uu and ww, uxwvu-x-w-v is the shortest path from uu to vv, and hence d(u,v)=3d(u,v)=3. Let u,vV1u,v\in V_{1}. As any nn-vertex xx is a common neighbour of uu and vv, uxvu-x-v is the shortest path from uu to vv, and hence d(u,v)=2d(u,v)=2. Let u,vV2u,v\in V_{2}. Then, d(u,v)=2d(u,v)=2 in one of the following three cases.

Case 1:

There exists xV1x\in V_{1} such that xx is a superset of both uu^{\dagger} and vv^{\dagger}.

Case 2:

There exists yV1y\in V_{1} such that yy is a subset of both uu^{\dagger} and vv^{\dagger} .

Case 3:

There exists wV1w\in V_{1} such that ww is a subset of uu^{\dagger} and a
          super set of vv^{\dagger} or ww is a subset of vv^{\dagger} and a superset of uu^{\dagger} .

In other words, d(u,v)=2d(u,v)=2 if either |uv|k|u^{\dagger}\cup v^{\dagger}|\leq k or |uv|k|u^{\dagger}\cap v^{\dagger}|\geq k or |uv|=|u|k|u^{\dagger}\cap v^{\dagger}|=|u^{\dagger}|\leq k. If none of these three conditions are satisfied, then d(u,v)2d(u,v)\neq 2. Choose x,yV1x,y\in V_{1} such that d(x,u)=1d(x,u)=1 and d(y,v)=1d(y,v)=1. Let ww be any nn-vertex in V2V_{2}. Then ww is a common neighbour of xx and yy. Thus, we get the shortest path uxwyvu-x-w-y-v of length 44 and hence d(u,v)=4d(u,v)=4. ∎

The following proposition and corollary are immediate consequences of the lemma.

Corollary 4.18.

If n>2n>2 and 1<k<n1<k<n, for the graph HB(n,k)H_{B}(n,k), the eccentricity is given by

e(v)={3if vV1,2if vV2 is an n-vertex ,4if vV2 is an r-vertex ,1r<ne(v)=\begin{cases}3&\text{if $v\in V_{1}$,}\\ 2&\text{if $v\in V_{2}$ is an n-vertex },\\ 4&\text{if $v\in V_{2}$ is an r-vertex ,$1\leq r<n$. }\end{cases}
Proof.

Let vV1v\in V_{1}. For any uV1u\in V_{1}, d(v,u)=2d(v,u)=2. If wV2w\in V_{2} such that it is adjacent to vv, then d(v,w)=2d(v,w)=2. If ww is not adjacent to vv, then d(v,w)=3d(v,w)=3. Thus, e(v)=3e(v)=3. Let vV2v\in V_{2} be an nn-vertex. Since vv is adjacent to all vertices in V1V_{1} and d(v,u)=2d(v,u)=2 for any uV2u\in V_{2}, we get e(v)=2e(v)=2. Let vV2v\in V_{2} be an rr-vertex, 1r<n1\leq r<n. The distance from vv to any vertex in V1V_{1} is either 11 or 33, and the distance from vv to any vertex in V2V_{2} is either 22 or 33.

Corollary 4.19.

The diameter of G=HB(n,k),n>2,1<k<nG=H_{B}(n,k),n>2,1<k<n is diam(G)=4\emph{diam}(G)=4 and radius is rad(G)=2\emph{rad}(G)=2.

Proof.

diam(G)=max{e(v):vV}=max{2,3,4}=4\text{diam}(G)=\text{max}\{e(v):v\in V\}=\text{max}\{2,3,4\}=4.

rad(G)=min{e(v):vV}=min{2,3,4}=2\text{rad}(G)=\text{min}\{e(v):v\in V\}=\text{min}\{2,3,4\}=2. ∎

Corollary 4.20.

For G=HB(n,k)),n>2, 1<k<nG=H_{B}(n,k)),\>n>2,\;1<k<n,

  1. (1)

    Periphery,P(G)={vV2|vis an r-vertex,1r<n}.\emph{Periphery},P(G)=\{v\in V_{2}|v\;\text{is an r-vertex},1\leq r<n\}.

  2. (2)

    Center, C(G)={vV2| v is an n-vertex}.C(G)=\{v\in V_{2}|\text{ v is an n-vertex}\}.

  3. (3)

    Median, M(G)= C(G).

Proof.

As the eccentricity of any rr-vertex vv, where 1r<n1\leq r<n is e(v)=4=diam(G)e(v)=4=\text{diam}(G), we get, Periphery,P(G)={vV2| vis an r-vertex,1r<n}\text{Periphery},P(G)=\{v\in V_{2}| v\;\text{is an r-vertex},1\leq r<n\} As the eccentricity of any nn-vertex vv is e(v)=2=rad(G)e(v)=2=\text{rad}(G), we get, Center,C(G)={vV2| vis an n-vertex}\text{Center},C(G)=\{v\in V_{2}| v\;\text{is an n-vertex}\}.

For finding the median(GG), we find the status of vertices in VV. We have the status of any vertex vGv\in G,  s(v)=uV(G)d(v,u)s(v)=\sum\limits_{u\in V(G)}d(v,u). First, we find the status of any nn-vertex vv in V2V_{2}. The sum of the distances from vv to (nk)\binom{n}{k} vertices in V1V_{1} is (nk)\binom{n}{k}.  The sum of the distance from vv to 3n12(nk)2n1\frac{3^{n}-1}{2}-\binom{n}{k}-2^{n-1}, rr-vertices, where 1r<n1\leq r<n in V2V_{2}  is  2(3n12(nk)2n1)2\left(\frac{3^{n}-1}{2}-\binom{n}{k}-2^{n-1}\right). Therefore, status of vv is 2(3n12(nk)2n1)+(nk)2\left(\frac{3^{n}-1}{2}-\binom{n}{k}-2^{n-1}\right)+\binom{n}{k}. There are vertices at distances 1,2,1,2, and 33 from vertices in V1V_{1}. There are vertices at distances 1,2,31,2,3, and 44 from  rr-vertices in V2V_{2}. This leads to the conclusion that the status of any nn-vertex is minimum compared to other vertices in VV. Thus, Median, M(G)M(G)=C(G)={vV2| v is an n-vertex}.C(G)=\{v\in V_{2}|\text{ v is an n-vertex}\}.

Remark 4.21.

Let G=HB(n,k)G=H_{B}(n,k) with bipartition, V(G)=V1V2V(G)=V_{1}\cup V_{2}. We denote the set of all pairs (unordered) of vertices of V(G)V(G) by S={{u,v}u,vV(G)}S=\{\{u,v\}\mid u,v\in V(G)\}. Then, SS contains vertex pairs at distances 1,2,3,1,2,3, and 44. The subsets of SS are of the form S(V(G),h)={{u,v} d(u,v)=h,  1h4}S\big{(}V(G),h\big{)}=\{\{u,v\}\mid d(u,v)=h,\;\;1\leq h\leq 4\}. Then, S=h=14S(V(G),h)S=\bigcup_{h=1}^{4}S\big{(}V(G),h\big{)}. Let d(h)(u,v)d^{(h)}(u,v) be the cardinality of S(V(G),h)S\big{(}V(G),h\big{)} for h=1,2,3,h=1,2,3, and 44. We denote by S(V1,2)S(V_{1},2) and S(V2,2)S(V_{2},2), respectively, the sets of vertex pairs of V1V_{1} and V2V_{2} at distance 22. We have, S(V(G),2)=S(V1,2)S(V2,2)S\big{(}V(G),2\big{)}=S(V_{1},2)\cup S(V_{2},2). Let dV1(2)(u,v)d^{(2)}_{V_{1}}(u,v) and dV2(2)(u,v)d^{(2)}_{V_{2}}(u,v), respectively, denote the cardinalities of S(V1,2)S(V_{1},2) and S(V2,2)S(V_{2},2). Consequently, we get, d(2)(u,v)=dV1(2)(u,v)+dV2(2)(u,v)d^{(2)}(u,v)=d^{(2)}_{V_{1}}(u,v)+d^{(2)}_{V_{2}}(u,v). As vertex pairs at distance 44 exists only in V2V_{2}, we denote the set containing them as S(V2,4)S(V_{2},4). We have, S(V(G),4)=S(V2,4)S\big{(}V(G),4\big{)}=S(V_{2},4). Let dV2(4)(u,v)d^{(4)}_{V_{2}}(u,v) be the cardinality of S(V2,4)S(V_{2},4). Then, d(4)(u,v)=dV2(4)(u,v)d^{(4)}(u,v)=d^{(4)}_{V_{2}}(u,v).

The cardinalities of S(V(G),1)S\big{(}V(G),1\big{)} and S(V(G),3)S\big{(}V(G),3\big{)}, the total number of vertex pairs at distance 22 from V2V_{2} denoted by dV2(2)(u,v)d^{(2)}_{V_{2}}(u,v), and the total number of vertex pairs {u,v}\{u,v\} at distances of 22 and 44, where uu and vv are from V2V_{2}, are determined in the next theorem

Theorem 4.22.

Consider the bipartite Kneser B type-k graph G=HB(n,k)),n>2, 1<k<nG=H_{B}(n,k)),\>n>2,\;1<k<n. For u,vVu,v\in V,

  1. (1)

    d(1)(u,v)=(nk)(3k32+2k1(3nk1))d^{(1)}(u,v)=\binom{n}{k}\left(\dfrac{3^{k}-3}{2}+2^{k-1}(3^{n-k}-1)\right).

  2. (2)

    d(3)(u,v)=(nk)(3n12(nk)(3k32)2k1(3nk1))d^{(3)}(u,v)=\binom{n}{k}\left(\frac{3^{n}-1}{2}-\binom{n}{k}-\left(\frac{3^{k}-3}{2}\right)-2^{k-1}(3^{n-k}-1)\right).

  3. (3)

    dV1(2)(u,v)=((nk)2)d^{(2)}_{V_{1}}(u,v)=\binom{\binom{n}{k}}{2}.

  4. (4)

    dV2(2)(u,v)+dV2(4)(u,v)=(3n12(nk)2)d_{V_{2}}^{(2)}(u,v)+d_{V_{2}}^{(4)}(u,v)=\dbinom{\frac{3^{n}-1}{2}-\tbinom{n}{k}}{2}.

Proof.

d(u,v)=1d(u,v)=1 when uV1u\in V_{1} and vV2v\in V_{2} are adjacent. Thus, d(1)(u,v)=|E(G)|=(nk)(3k32+2k1(3nk1))d^{(1)}(u,v)=|E(G)|=\binom{n}{k}\left(\dfrac{3^{k}-3}{2}+2^{k-1}(3^{n-k}-1)\right). d(u,v)=2d(u,v)=2 when uu and vv are in V1V_{1}. As there are (nk)\binom{n}{k} kk-vertices in V1V_{1}, dV1(2)(u,v)=((nk)2)d^{(2)}_{V_{1}}(u,v)=\binom{\binom{n}{k}}{2}. d(u,v)=3d(u,v)=3 when uV1u\in V_{1} and vV2v\in V_{2} are not adjacent.

The total number of unordered pairs of vertices such that one vertex is from V1V_{1} and the other from V2V_{2} is (nk)(|V(G)|(nk))\binom{n}{k}(|V(G)|-\binom{n}{k}). Thus, d(3)(u,v)=(nk)(|V(G)|(nk))|E(G)|=(nk)(3n12(nk)(3k32)2k1(3nk1)).d^{(3)}(u,v)=\binom{n}{k}(|V(G)|-\binom{n}{k})-|E(G)|=\binom{n}{k}\left(\frac{3^{n}-1}{2}-\binom{n}{k}-\left(\frac{3^{k}-3}{2}\right)-2^{k-1}(3^{n-k}-1)\right).

Given that dV2(2)(u,v)d_{V_{2}}^{(2)}(u,v) and dV2(4)(u,v)d_{V_{2}}^{(4)}(u,v) are the counts of pairs at distances 2 and 4 respectively, it follows from theorem 4.17 that

dV2(2)(u,v)+dV2(4)(u,v)\displaystyle d_{V_{2}}^{(2)}(u,v)+d_{V_{2}}^{(4)}(u,v) =(|V(G)|2)(|E(G)|+(nk)(|V(G)|(nk))|E(G)|)\displaystyle=\binom{|V(G)|}{2}-\left(|E(G)|+\binom{n}{k}\left(|V(G)|-\binom{n}{k}\right)-|E(G)|\right)
=(3n12(nk)2)\displaystyle=\dbinom{\frac{3^{n}-1}{2}-\tbinom{n}{k}}{2}

.

dV2(4)(u,v)d_{V_{2}}^{(4)}(u,v) is computed in the following theorem. As any ii-vertex uu and jj-vertex vv in V2V_{2} can have both positive and negative components, whenever we say common elements in an ii vertex and a jj-vertex, we mean |uv||u^{\dagger}\cap v^{\dagger}|.

Theorem 4.23.

Let Pi,jP_{i,j} be the number of unordered pairs of ii-vertices and jj-vertices of V2V_{2} that are at distance 44.

For u,vV2u,v\in V_{2} in HB(n,k)H_{B}(n,k), dV2(4)(u,v)=i,j{1,2,n1}ijk+1i+jn+k1Pi,jd_{V_{2}}^{(4)}(u,v)=\sum\limits_{\begin{subarray}{c}i,j\in\{1,2,\dotsc n-1\}\\ i\leq j\\ k+1\leq i+j\leq n+k-1\end{subarray}}P_{i,j}.

Here, Pi,j={t(ni)2i12j1(nijt)(it)for ik and jkt(nk)(2k11)2j1(nkjt)(kt)fori=kandjkt(ni)2i1(2k11)(nikt)(it)forikandj=k12(t(nk)(2k11)2(nkkt)(kt))fori=j=k12(t(ni)22(i1)(niit)(it))fori=j and jkP_{i,j}=\begin{cases}\sum\limits_{t}\binom{n}{i}2^{i-1}2^{j-1}\binom{n-i}{j-t}\binom{i}{t}\quad\text{for $i\neq k$ and $j\neq k$}\\ \sum\limits_{t}\binom{n}{k}(2^{k-1}-1)2^{j-1}\binom{n-k}{j-t}\binom{k}{t}\quad\text{for}\ i=k\;\text{and}\;j\neq k\\ \sum\limits_{t}\binom{n}{i}2^{i-1}(2^{k-1}-1)\binom{n-i}{k-t}\binom{i}{t}\quad\text{for}\ i\neq k\;\text{and}\;j=k\\ \frac{1}{2}{}\left(\sum\limits_{t}\binom{n}{k}(2^{k-1}-1)^{2}\binom{n-k}{k-t}\binom{k}{t}\right)\quad\text{for}\ i=j=k\\ \frac{1}{2}\left(\sum\limits_{t}\binom{n}{i}2^{2(i-1)}\binom{n-i}{i-t}\binom{i}{t}\right)for\;i=j\;\text{ and }j\neq k\par\end{cases}

Here, |uv|=t|u^{\dagger}\cap v^{\dagger}|=t and tt is a non-negative integer such that i+jnt<i+jki+j-n\leq t<i+j-k and t<min{i,k}t<\text{min}\{i,k\}

Proof.

Choose an ii-vertex u={x1,x2,,xi}u=\{x_{1},x_{2},\dotsc,x_{i}\} and a jj-vertex v={y1,y2,,yj}v=\{y_{1},y_{2},\dotsc,y_{j}\} in V2V_{2} such that d(u,v)=4d(u,v)=4. Here i,j{1,2,n1},ij,andk+1i+jn+k1i,j\in\{1,2,\dotsc n-1\},i\leq j,\text{and}\>k+1\leq i+j\leq n+k-1.By the construction of HB(n,k)H_{B}(n,k), the vertices at distance 44 satisfy the conditions: |uv|=t|u^{\dagger}\cap v^{\dagger}|=t, t0t\geq 0, i+jnt<i+jki+j-n\leq t<i+j-k, and t<min{i,k}t<\text{min}\{i,k\}. Corresponding to an ii-element subset of n+={a1, a2, a3,, an1,an}\mathscr{B}_{n}^{+}=\{a_{1}, a_{2}, a_{3},\dots, a_{n-1},a_{n}\}, 2i12^{i-1}, ii-element subsets or ii-vertices are seen in ϕ(n)\phi(\mathscr{B}_{n}). Similarly, corresponding to a jj-element subset of n+\mathscr{B}_{n}^{+}, 2j12^{j-1}, jj-vertices are there in ϕ(n)\phi(\mathscr{B}_{n}). Pi,jP_{i,j} is calculated in various cases.

Case 1:

For i,ji,j such that i<ji<j and iki\neq k and jkj\neq k.

There are 2j12^{j-1}, jj-vertices at distance 44 to uu. As there are 2i12^{i-1} vertices corresponding to uu, the total number of 44 pairs between uu and vv and their corresponding vertices is 2i12j12^{i-1}2^{j-1}. As |uv|=t|u^{\dagger}\cap v^{\dagger}|=t, the remaining jtj-t elements in any other jj-vertex at distance 44 can be selected from nin-i elements of n+\mathscr{B}_{n}^{+} in (nijt)\binom{n-i}{j-t} ways. Also, tt elements can be selected from ii-vertex in (it)\binom{i}{t}ways.   The total number of ii element subsets of n+\mathscr{B}_{n}^{+} is (ni)\binom{n}{i}. Using the restrictions on t,iandjt,i\;\;\text{and}\;\;j, we get Pi,j=t(ni)2i12j1(nijt)(it)P_{i,j}=\sum\limits_{t}\binom{n}{i}2^{i-1}2^{j-1}\binom{n-i}{j-t}\binom{i}{t}.

Case 2:

For i,ji,j such that i<ji<j, i=ki=k and jkj\neq k.

Of the 2k12^{k-1}, kk-vertices corresponding to {x1,x2,,xk}\{x_{1},x_{2},\dotsc,x_{k}\} in ϕ(n)\phi(\mathscr{B}_{n}), (2k11)(2^{k-1}-1)are in V2V_{2} and 11 in V1V_{1}. Therefore, Pi,j=t(ni)(2k11)2j1(nkjt)(kt)P_{i,j}=\sum\limits_{t}\binom{n}{i}(2^{k-1}-1)2^{j-1}\binom{n-k}{j-t}\binom{k}{t}.

Case 3:

For i,ji,j such that i<ji<j, iki\neq k and j=kj=k.

Of the 2k12^{k-1}, kk-vertices corresponding to {y1,y2,,yk}\{y_{1},y_{2},\dotsc,y_{k}\} in ϕ(n)\phi(\mathscr{B}_{n})(2k11)(2^{k-1}-1)are in V2V_{2} and 11 in V1V_{1}. Therefore, Pi,j=t(ni)2i1(2k11)(nikt)(it)P_{i,j}=\sum\limits_{t}\binom{n}{i}2^{i-1}(2^{k-1}-1)\binom{n-i}{k-t}\binom{i}{t}.

case 4:

For i,ji,j such that i=j=ki=j=k.

Using similar arguments, we conclude that Pi,j=12(t(nk)(2k11)2(nkkt)(kt))P_{i,j}=\frac{1}{2}\left(\sum\limits_{t}\binom{n}{k}(2^{k-1}-1)^{2}\binom{n-k}{k-t}\binom{k}{t}\right)

Case 5:

For i,ji,j such that i=ji=j and j kj\neq k.

Here,  Pi,j=12(t(ni)22(i1)(niit)(it)).P_{i,j}=\frac{1}{2}\left(\sum\limits_{t}\binom{n}{i}2^{2(i-1)}\binom{n-i}{i-t}\binom{i}{t}\right).

For i,j{1,2,n1},ij,andk+1i+jn+k1i,j\in\{1,2,\dotsc n-1\},i\leq j,\text{and}\>k+1\leq i+j\leq n+k-1, the total number of unordered pairs of vertices from  V2V_{2} such that d(u,v)=4d(u,v)=4 is

dV2(4)(u,v)=\displaystyle d^{(4)}_{V_{2}}(u,v)= Pi,j\displaystyle\sum P_{i,j}
=\displaystyle= P1,k+P1,k+1+P1,n1+\displaystyle P_{1,k}+P_{1,k+1}+\dotsb P_{1,n-1}+
P2,k1+P2,k+P2,n1+\displaystyle P_{2,k-1}+P_{2,k}+\dotsb P_{2,n-1}+
+\displaystyle\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb+
Pk,k+Pk,n1+Pk+1,k+1+Pk+1,n2+\displaystyle P_{k,k}+\dotsb P_{k,n-1}+P_{k+1,k+1}+\dotsb P_{k+1,n-2}+
+\displaystyle\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb+
Pn+k12,n+k12.(Whenn+k1is even)\displaystyle P_{\frac{n+k-1}{2},\frac{n+k-1}{2}}.\qquad(\text{When}\;\;n+k-1\;\;\text{is even})
dV2(4)(u,v)=\displaystyle d^{(4)}_{V_{2}}(u,v)= Pi,j\displaystyle\sum P_{i,j}
=\displaystyle= P1,k+P1,k+1+P1,n1+\displaystyle P_{1,k}+P_{1,k+1}+\dotsb P_{1,n-1}+
P2,k1+P2,k+P2,n1+\displaystyle P_{2,k-1}+P_{2,k}+\dotsb P_{2,n-1}+
+\displaystyle\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb+
Pk,k+Pk,n1+Pk+1,k+1+Pk+1,n2+\displaystyle P_{k,k}+\dotsb P_{k,n-1}+P_{k+1,k+1}+\dotsb P_{k+1,n-2}+
+\displaystyle\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb\dotsb+
Pn+k22,n+k22+Pn+k22,n+k2.(Whenn+k1is odd)\displaystyle P_{\frac{n+k-2}{2},\frac{n+k-2}{2}}+P_{\frac{n+k-2}{2},\frac{n+k}{2}}.\qquad(\text{When}\;\;n+k-1\;\;\text{is odd})

Remark 4.24.

From theorem 4.22, we have got the cardinalities of S(V(G),1)S\big{(}V(G),1\big{)} and S(V(G),3)S\big{(}V(G),3\big{)}. Using theorems 4.22 and 4.23, the cardinalites of S(V(G),2)S\big{(}V(G),2\big{)} and S(V(G),4)S\big{(}V(G),4\big{)} are obtained as d(2)(u,v)=dV1(2)(u,v)+dV2(2)(u,v)d^{(2)}(u,v)=d^{(2)}_{V_{1}}(u,v)+d^{(2)}_{V_{2}}(u,v) and d(4)(u,v)=dV2(4)(u,v)d^{(4)}(u,v)=d^{(4)}_{V_{2}}(u,v).

Example 4.25.

Consider the partite sets V1V_{1} and V2V_{2} of HB(4,2)H_{B}(4,2) as given in example 3.5.

d(1)(u,v)\displaystyle d^{(1)}(u,v) =(nk)(3k32+2k1(3nk1)).\displaystyle=\binom{n}{k}\left(\dfrac{3^{k}-3}{2}+2^{k-1}(3^{n-k}-1)\right).
=(42)(3232+221(3421)).\displaystyle=\binom{4}{2}\left(\dfrac{3^{2}-3}{2}+2^{2-1}(3^{4-2}-1)\right).
=114.\displaystyle=114.

dV1(2)(u,v)=((nk)2)=((42)2)=15.d^{(2)}_{V_{1}}(u,v)\;=\binom{\binom{n}{k}}{2}=\binom{\binom{4}{2}}{2}=15.

d(3)(u,v)\displaystyle d^{(3)}(u,v) =(nk)(3n12(nk)(3k32)2k1(3nk1)).\displaystyle=\binom{n}{k}\left(\frac{3^{n}-1}{2}-\binom{n}{k}-\left(\frac{3^{k}-3}{2}\right)-2^{k-1}(3^{n-k}-1)\right).
=(42)(3412(42)(3232)221(3421))\displaystyle=\binom{4}{2}\left(\frac{3^{4}-1}{2}-\binom{4}{2}-\left(\frac{3^{2}-3}{2}\right)-2^{2-1}(3^{4-2}-1)\right)
=90.\displaystyle=90.

Then, d(4)(u,v)=dV2(4)(u,v)=i,j{1,2,3}ij3i+j5Pi,j=P1,2+P1,3+P2,2+P2,3d^{(4)}(u,v)=d^{(4)}_{V_{2}}(u,v)=\sum\limits_{\begin{subarray}{c}i,j\in\{1,2,3\}\\ i\leq j\\ 3\leq i+j\leq 5\end{subarray}}P_{i,j}=P_{1,2}+P_{1,3}+P_{2,2}+P_{2,3}.

For t=0t=0, P1,2=(41)211(2211)(4120)(10)=12P_{1,2}=\binom{4}{1}2^{1-1}(2^{2-1}-1)\binom{4-1}{2-0}\binom{1}{0}=12.

For t=0t=0, P1,3=(41)211231(4130)(10)=16P_{1,3}=\binom{4}{1}2^{1-1}2^{3-1}\binom{4-1}{3-0}\binom{1}{0}=16.

For t=0,1t=0,1, P2,2=12((42)(2211)2)(4220)(20)+(42)(2211)2(4221)(21))=15P_{2,2}=\frac{1}{2}\left(\binom{4}{2}(2^{2-1}-1)^{2})\binom{4-2}{2-0}\binom{2}{0}+\binom{4}{2}(2^{2-1}-1)^{2}\binom{4-2}{2-1}\binom{2}{1}\right)=15.

For t=1t=1, P2,3=(42)(2211)231(4231)(21)=48P_{2,3}=\binom{4}{2}(2^{2-1}-1)2^{3-1}\binom{4-2}{3-1}\binom{2}{1}=48.

Therefore, d(4)(u,v)=dV2(4)(u,v)=12+16+15+48=91d^{(4)}(u,v)=d^{(4)}_{V_{2}}(u,v)=12+16+15+48=91.

Using the identity (4) in theorem (4.22), we get

dV2(2)(u,v)+dV2(4)(u,v)=(3n12(nk)2)d_{V_{2}}^{(2)}(u,v)+d_{V_{2}}^{(4)}(u,v)=\binom{\frac{3^{n}-1}{2}-\binom{n}{k}}{2}.

Therefore, dV2(2)(u,v)=(3n12(nk)2)dV2(4)(u,v)=56191=470d_{V_{2}}^{(2)}(u,v)=\binom{\frac{3^{n}-1}{2}-\binom{n}{k}}{2}-d_{V_{2}}^{(4)}(u,v)=561-91=470

d(2)(u,v)=dV1(2)(u,v)+dV2(2)(u,v)=15+470=485d^{(2)}(u,v)=d^{(2)}_{V_{1}}(u,v)+d^{(2)}_{V_{2}}(u,v)=15+470=485.

Table 1. A table showing the values of d(h)(u,v)d^{(h)}(u,v), 1h41\leq h\leq 4 and HB(n,k)H_{B}(n,k) for some values of nn and kk.
d(1)(u,v)d^{(1)}(u,v) d(2)(u,v)d^{(2)}(u,v) d(3)(u,v)d^{(3)}(u,v) d(4)(u,v)d^{(4)}(u,v)
HB(4,2)H_{B}(4,2) 114 485 90 91
HB(4,3)H_{B}(4,3) 80 486 64 150
HB(5,2)H_{B}(5,2) 550 5275 560 875
HB(5,3)H_{B}(5,3) 440 4125 670 2025
HB(5,4)H_{B}(5,4) 275 4715 305 1965
HB(6,2)H_{B}(6,2) 2445 54050 2790 6781

5. Conclusion

In this paper, we have determined various invariants of HB(n,k)H_{B}(n,k). There is scope for applications in various disciplines. As the degree sequence of HB(n,k)H_{B}(n,k) and the distance between any pair of vertices in HB(n,k)H_{B}(n,k) are determined, hundreds of molecular descriptors can be calculated. Distance properties also lead to the determination of various types of metric dimensions. Centrality measures such as degree centrality, Closeness centrality, betweenness centrality and eigen vector centrality can also be obtained.

Conflict of Interest

The authors hereby declare that there is no potential conflict of interest.

Acknowledgement

The first author is a doctoral fellow in mathematics at University College, Thiruvananthapuram. This research has been promoted and supported by  the University of Kerala.

References

  • [1] Louis Anthony Agong, Carmen Amarra, John S Caughman, Ari J Herman, and Taiyo S Terada, On the girth and diameter of generalized johnson graphs, Discrete Mathematics 341 (2018), no. 1, 138–142.
  • [2] Dean Crnković, Daniel R. Hawtin, Nina Mostarac, and Andrea Švob, Neighbour-transitive codes in kneser graphs, Journal of Combinatorial Theory, Series A 204 (2024), 105850.
  • [3] Sadik Delen and Ismail Naci Cangul, A new graph invariant, Turkish Journal of Analysis and Number Theory 6 (2018), 30–33.
  • [4] by same author, Extremal problems on components and loops in graphs, Acta Mathematica Sinica, English Series 35 (2019), no. 2, 161–171.
  • [5] Teresa W. Haynes, Stephen T. Hedetniemi, and Michael A. Henning, Domination in graphs: Core concepts, Springer Monographs in Mathematics, pp. 1–644, Springer Science and Business Media Deutschland GmbH, Germany, 2023 (English), Publisher Copyright: © Springer Nature Switzerland AG 2023.
  • [6] L. LOVASZ, Kneser’s conjecture, chromatic number, and homotopy, Journal of Combinatorial Theory, series A 25 25 (1978), 319 – 324.
  • [7] S Morteza Mirafzal, The automorphism group of the bipartite kneser graph, Proceedings-Mathematical Sciences 129 (2019), no. 3, 1–8.
  • [8] K.G Sreekumar, P Ramesh Kumar, and K Manilal, Automorphism groups of bipartite kneser type-k graphs, Asian-European Journal of Mathematics (2022), 2350047.
  • [9] K.G Sreekumar, G Suresh Singh, and C. S. Preenu, Elliptic curves obtained from bipartite kneser B type-k graphs, TWMS J. App. and Eng. Math. Special Issue -1(2023) (2023), no. 01, 102–111.
  • [10] Dmitriy Zakharov, Chromatic numbers of kneser-type graphs, Journal of Combinatorial Theory, Series A 172 (2020), 105188.