This paper investigates how graft transformations affect the distance signless Laplacian spectral radius of graphs, providing characterizations of extremal trees with respect to this spectral radius.
Contribution
It introduces graft transformations that influence the spectral radius and characterizes extremal trees among non-starlike and non-caterpillar trees.
Findings
01
Identifies transformations that minimize or maximize the spectral radius.
02
Characterizes graphs with extremal spectral radius among certain tree classes.
03
Provides theoretical bounds for the spectral radius based on graph structure.
Abstract
Suppose that the vertex set of a connected graph G is V(G)={v1,⋯,vn}. Then we denote by TrG(vi) the sum of distances between vi and all other vertices of G. Let Tr(G) be the n×n diagonal matrix with its (i,i)-entry equal to TrG(vi) and D(G) be the distance matrix of G. Then QD(G)=Tr(G)+D(G) is the distance signless Laplacian matrix of G. The largest eigenvalues of QD(G) is called distance signless Laplacian spectral radius of G. In this paper we give some graft transformations on distance signless Laplacian spectral radius of the graphs and use them to characterize the graphs with the minimum and maximal distance signless Laplacian spectral radius among non-starlike and non-caterpillar trees.
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsGraph theory and applications · Synthesis and Properties of Aromatic Compounds · Complex Network Analysis Techniques
Full text
The effect of a graft transformation on distance signless Laplacian spectral radius of the graphs ††thanks: This work is supported by NSFC (No. 11461071).
a College of Mathematical and Physical Sciences, Xinjiang Agricultural University,
Urumqi 830052, Xinjiang, P. R. China.
b School of Mathematical Sciences, Xinjiang Normal University,
Urumqi 830054, Xinjiang, P. R. China.
Abstract. Suppose that the vertex set of a connected graph G is V(G)={v1,⋯,vn}.
Then we denote by TrG(vi) the sum of distances between vi and all other vertices of G.
Let Tr(G) be the n×n diagonal matrix with its (i,i)-entry equal to TrG(vi)
and D(G) be the distance matrix of G. Then QD(G)=Tr(G)+D(G) is the distance signless Laplacian matrix of G.
The largest eigenvalues of QD(G) is called distance signless Laplacian spectral radius of G.
In this paper we give some graft transformations on distance signless Laplacian spectral radius of the graphs
and use them to characterize the graphs with the minimum and maximal distance signless Laplacian spectral radius
among non-starlike and non-caterpillar trees.
Let G be a simple connected graph with the vertex set V(G)={v1,…,vn}.
The distancedG(u,v) between the vertices u and v is the length of shortest path between u and v in G.
For u∈V(G), the transmissionTrG(u) of u is the sum of distances between u and all other vertices of G.
Let Tr(G) be the n×n diagnonal matrix with its (i,i)-entry equal to TrG(vi) and D(G) be the distance matrix of G.
Then QD(G)=Tr(G)+D(G) is the distance signless Laplacian matrix of G.
The largest eigenvalues ρQ(G) of QD(G) is the distance signless Laplacian spectral radius of G.
The distance spectral radius of a connected graph has been studied extensively.
S. Bose, M. Nath and S. Paul [4] determined the unique graph with maximal distance
spectral radius in the class of graphs without a pendant vertex.
G. Yu, Y. Wu, Y. Zhang and J. Shu [17] obtained respectively the extremal graph and unicyclic graph
with the maximum and minimum distance spectral radius.
W. Ning, L. Ouyang and M. Lu [10] characterized the graph with minimum distance spectral radius
among trees with given number of pendant vertices.
G. Yu, H. Jia, H. Zhang and J. Shu [16] characterized the unique graphs
with the minimum and maximum distance spectral radius among trees and graphs with given number of pendant vertices.
For more about the distance spectra of graphs see the surveys [3, 12] as well as the references therein.
M. Aouchiche and P. Hansen introduced in [1] the distance Laplacian and distance signless Laplacian spectra of graphs,
and proved in [2] that the star Sn attains minimum distance
Laplacian spectral radius among all trees of order n.
R. Xing and B. Zhou [13] gave the unique graph with minimum distance and
distance signless Laplacian spectral radii among bicyclic graphs.
R. Xing, B. Zhou and J. Li [14] determined the graphs with minimum
distance signless Laplacian spectral radius among the trees, unicyclic graphs,
bipartite graphs, the connected graphs with fixed pendant vertices and fixed connectivity, respectively.
A. Niu, D. Fan and G. Wang [11] determined the extremal graph with the minimum distance Laplacian spectral radius
among bipartite graphs with the fixed connectivity and matching numbers.
H. Lin and B. Zhou [7] characterized the unique graphs with the minimum distance Laplacian spectral radius with fixed number of pendant
vertices and edge connectivity among graphs, the unique tree with the minimum distance Laplacian spectral radius among trees with fixed bipartition.
A tree T is starlike if it contains at most one vertex of degree at least 3, and otherwise it is non-starlike.
A tree is caterpillar if the deletion of all pendant vertices yields a path, and otherwise it is non-caterpillar. Set
[TABLE]
[TABLE]
R. Xing, B. Zhou and F. Dong [15] determined the graphs with the minimum and maximal distance spectral radius in T(n) and ℜ(n) respectively.
In this paper we give some graft transformations on distance signless Laplacian spectral radius of the graphs
and then use them to characterize the graphs with the minimum and maximal distance signless Laplacian spectral radius in T(n) and ℜ(n) respectively.
2. Trees with minimum distance Signless Laplacian spectral radius in ℜ(n) and T(n)
Suppose that G is a connected graph with V(G)={v1,v2,…,vn}.
Let x=(xv1,xv2,…,xvn)T, where xvi corresponds to vi, i.e. x(vi)=xvi for i=1,2,…,n.
Then
[TABLE]
which shows that QD(G) is positive definite for n≥3.
Suppose that x is an eigenvector of QD(G) corresponding to the eigenvalue λ. Then for each v∈V(G),
λx(v)=u∈V(G)∑dG(u,v)(x(u)+x(v)).
By the Perron-Frobenius theorem, we know that there is the unique unit vector x>0 satisfies QD(G)x=ρQ(G)x, and we call x the Perron vector of G corresponding to ρQ(G).
Suppose that G is a graph. If a vertex of G is of degree one then it is a pendant vertex.
An edge and a path are respectively pendant edge and pendant path if one of their end vertices are pendant vertices.
Suppose that uv is a cut edge of connected graph G.
Then we denote by Guv the graph obtained from G by contracting uv to a vertex u and
attaching a pendant edge at u, and call such a process C-transformation of G for the edge uv.
Lemma 2.1.Let uv be a cut edge of connected graph G.
If uv is not a pendant edge and there exists a pendant edge uz, then ρQ(G)>ρQ(Guv).
Proof. Let x be the Perron vector of Guv corresponding to ρQ(Guv).
By symmetry, we have xv=xz.
Suppose that G\{uv}=G1∪G2 where u∈G1 and v∈G2.
It can be easily observed that dG(u,w)=dGuv(u,w) for w∈V(G1) and that dG(w,vj)=dGuv(w,vj)+1 for w∈V(G1) and vj∈V(G2)\{v}.
Thus,
[TABLE]
Suppose on the contrary that ρQ(Guv)=ρQ(G). Then xTQD(G)x=ρQ(Guv).
Therefore, (ρQ(G)−ρQ(Guv))xu=vj∈V(G2)\{v}∑(xu+xvj)>0.
This contradiction shows that ρQ(G)>ρQ(Guv).
□
Lemma 2.2.Suppose that n=n0+n1+n2+n3+4 with max{ni∣i=1,2,3}>1.
If G and G′ are shown in Fig. 1, then ρQ(G)>ρQ(G′).
Proof. Let x be the Perron vector of G′ corresponding to ρQ(G′).
By symmetry, we can set x=(a,a,a,b,b,b,n−7d,…,d,e)T labeled in Fig. 1.
Without lost of generality, we assume that n3≥n2≥n1.
Then n3≥2, and hence
[TABLE]
Let Trmax(G) be the maximum transmission of vertices of G. Then we have the following
Lemma 2.3 [8]. Let G be a connected graph. Then ρQ(G)>Trmax(G).
Lemma 2.4.Suppose that n=n0+n1+n2+n3+4 with n1+n2+n3>4.
If G and G∗ the trees as in Fig. 2, then ρQ(G)>ρQ(G∗).
Proof. Let x be the Perron vector of G∗ corresponding to ρQ(G∗).
By symmetry, we can set x=(a,a,b,b,c,c,d,f,n−8e,…,e)T labeled in Fig. 2.
Without lost of generality we can assume that n3≥n2≥n1.
From QD(G∗)x=ρQ(G∗)x, we have
[TABLE]
By Lemma 2.3, ρQ(G∗)>Trmax(G∗)=3n−3 and n>8, and so we easily observe from the above equations that a>b, c>d, a>c and 2c>a.
We also have (ρQ(G∗)−(2n−5))(b−d)=2b+4c−2a, from which we obtain b>d.
Thus we know that 2(e+a)2+2(e+b)2+(e+f)2+(n−n3−6)(e+e)2−2(e+c)2−(e+d)2>0.
It follows from n1+n2+n3>4 that either n1≥2 and n2≥2 or n3≥3.
Therefore,
[TABLE]
Let B(n;n0,n1,…,nr) be the graph obtained from the star S1,r by attaching n0 edges to its center and ni edges to the i-th
pendant vertex for each i, where n0,n1,…,nr are integers
satisfying n0≥0, nr≥nr−1≥…≥n1≥1 and i=0∑rni=n−r−1.
Theorem 2.5.If T∈T(n) with n≥7, then ρQ(T)≥ρQ(B(n;n−7,1,1,1)), where the equality holds if and only if T≅B(n;n−7,1,1,1).
Proof. Suppose that T′ is a tree with minimal distance signless Laplacian spectral radius in T(n).
Let d be the diameter of T′.
Now we verify that the following three claims are true.
Claim 1.d=4.
Since T′∈T(n), neither d=2 nor d=3.
Suppose by a way of contradiction that d≥5.
Let P=v1v2…vd+1 be a path of T′.
Making a C-transformation of T′ for the edge v2v3 and vd−1vd
we obtain the resulting graph Tv2v3′ and Tvd−1vd′, respectively.
Obviously, at least one of Tv2v3′ and Tvd−1vd′ falls into T(n).
By Lemma 2.1, ρQ(T′)>ρQ(Tv2v3′) and ρQ(T′)>ρQ(Tvd−1vd′).
This contradicts minimality of T′, and so d=4.
The Claim 1 shows that T′≅B(n;n0,n1,n2,…,nr), where r≥3.
Claim 2.r=3.
Suppose on the contrary that r≥4. Let T′′=B(n;n0+nr+1,n1,n2,…,nr−1).
Then T′′∈T(n), and by Lemma 2.1, ρQ(T′)>ρQ(T′′).
This contradiction shows r=3, and so T′≅B(n;n0,n1,n2,n3).
Claim 3.T′≅B(n;n−7,1,1,1).
If some one of n1, n2 and n3 is greater than 1,
then, by Lemma 2.2, ρQ(B(n;n0,n1,n2,n3))>ρQ(B(n;n−7,1,1,1)),
and so T′≅B(n;n−7,1,1,1). □
Theorem 2.6.If T∈T(n)∩ℜ(n) and n≥8, then ρQ(T)≥ρQ(B(n;n−8,1,1,2)),
where the equality holds if and only if T≅B(n;n−8,1,1,2).
Proof. Suppose that T′∈T(n)∩ℜ(n) is a graph with minimal distance signless Laplacian spectral radius, where n≥8.
Let d be the diameter of T′. Now we verify the following three claims.
Claim 1.d=4.
Suppose on the contrary that d≥5. Let P=v1v2…vd+1 be a path.
Making a C-transformation of T′ for the edge v2v3 and vd−1vd
we obtain the resulting graph Tv2v3′ and Tvd−1vd′, respectively.
Obviously, at least one of Tv2v3′ and Tvd−1vd′ falls into T(n)∩ℜ(n).
By Lemma 2.1, ρQ(T′)>ρQ(Tv2v3′) and ρQ(T′)>ρQ(Tvd−1vd′).
This contradicts minimality of T′, and so d=4.
The Claim 1 shows that T′≅B(n;n0,n1,n2,…,nr), where r≥3.
Claim 2.r=3.
Suppose on the contrary that r≥4. Let T′′=B(n;n0+n1+1,n2,n3,…,nr).
Thus T′′∈T(n)∩ℜ(n), and by Lemma 2.1, ρQ(T′)>ρQ(T′′).
This contradiction shows r=3, and so T′≅B(n;n0,n1,n2,n3).
Claim 3.T′≅B(n;n−8,1,1,2).
If either one of n1 and n2 is not less than 2 or n3≥3, then by Lemma 2.4, ρQ(B(n;n0,n1,n2,n3))>ρQ(B(n;n−8,1,1,2)), and so T′≅B(n;n−8,1,1,2). □
3. The tree with maximal distance Signless Laplacian spectral radius in ℜ(n) or T(n)
Lemma 3.1.Let G=G1∪G2∪G3, where V(Gi)∩V(Gj)=v0 for 1≤i=j≤3 and V(Gi)≥2 for i=1,2,3.
Suppose that u∈V(G2)\{v0}, and
let G be the graph obtained from G by moving G3 from v0 to u.
Let x=(xv1,xv2,⋯,xvn)T be a Perron vector corresponding to ρQ(G).
If vi∈V(G3)∖{v0}∑vj∈V(G1)∑(xvi+xvj)2≥vi∈V(G3)∖{v0}∑vj∈V(G2)∑(xvi+xvj)2,
then ρQ(G)>ρQ(G).
Proof. It can be easily observed that dG(u,v0)=dG(u,v0) and that for vi∈V(G3)\{v0}, vj∈V(G1), dG(vi,vj)=dG(vi,vj)+dG(u,v0);
and for vi∈V(G3)\{v0}, vj∈V(G2),
dG(vi,vj)≤dG(vi,vj)+dG(u,v0).
Suppose on the contrary that ρQ(G)=ρQ(G). Then xTQD(G)x=ρQ(G).
Therefore, (ρQ(G)−ρQ(G))xv0=vk∈V(G3)\{v0}∑dG(u,v0)(xv0+xvk)>0.
This contradiction shows that ρQ(G)>ρQ(G). □
Lemma 3.2 [6]. Let Gp,q be the connected graph obtained from a graph G of order at least two by attaching two pendant paths of length p and q at a vertex u of G.
If q≥p≥1, then ρQ(Gp−1,q+1)>ρQ(Gp,q).
Denote by T(n,k;t1,t2) the graph obtained from a path Pℓ of order ℓ (ℓ≥2) by attaching t1≥1 edges at an end of Pℓ and t2≥1 edges at other end
of Pℓ, where t1+t2=k and ℓ+k=n. Clearly, Pn≅T(n,2;1,1).
Let
[TABLE]
Theorem 3.3.The graph with the maximum distance signless Laplacian spectral radius in ℜ(n) lies to D(n,k).
Proof. Suppose that T0 is a tree with the maximum distance signless Laplacian spectral radius in ℜ(n).
Let k be the number of pendant vertices of T0. Then k≥4.
Let nT03 be the number of vertices which are of degree at least three in T0. Now we verify that the following two claims are true.
Claim 1.nT03=2.
It is clear that nT03≥2. Now suppose by a way of a contradiction that nT03≥3.
Let A={w1,w2,…,wnT03} be the set of all vertices in T0 which are of degree at least three,
where dT0(w1,w2)=max{dT0(wp,wq)∣1≤p=q≤nT03}.
Let T0=T1∪T2∪T3,
where V(Ti)∩V(Tj)=w3(1≤i=j≤3) and w1∈T1, w2∈T2.
If vi∈V(T3)∖{w3}∑vj∈V(T1)∑(xvi+xvj)2≥vi∈V(T3)∖{w3}∑vj∈V(T2)∑(xvi+xvj)2,
then let T0′=T0−vj∈NT3(w3)∑w3vj+vj∈NT3(w3)∑w2vj,
and otherwise T0′=T0−vj∈NT3(w3)∑w3vj+vj∈NT3(w3)∑w1vj.
Thus T0′∈ℜ(n), and by Lemma 3.1, ρQ(T0′)>ρQ(T0).
This contradicts maximality of T0, and so nT03=2.
The Claim 1 shows that T0 consists of a path with its two end vertices w1 and w2 attached some pendant paths respectively.
Claim 2.T0≅T(n,k;t1,t2).
Suppose on the contrary that T2=w1u1…us is a pendant path of length s≥2.
Let T0=T1∪T2∪T3 be shown in Fig. 3,
where V(Ti)∩V(Tj)=w1 for 1≤i=j≤3, NT0(w1)\{u1,v1}⊂V(T1) and w2∈V(T3).
If vi∈V(T1)∑vj∈V(T3)∑(xvi+xvj)2≥vi∈V(T1)∑vj∈V(T2)∑(xvi+xvj)2,
then let T0=T0−vj∈V(T1)∑w1vj+vj∈V(T1)∑us−1vj,
and otherwise T0=T0−vj∈V(T1)∑w1vj+vj∈V(T1)∑w2vj.
By Lemma 3.1, ρQ(T0)>ρQ(T0).
This contradicts maximality of T0, and so T0≅T(n,k;t1,t2).
Let n1,n2,…,nr be positive integers satisfying n1≤n2≤…≤nr and i=1∑rni=n−1.
Then we denote by S(n;n1,n2,…,nr) the tree obtained by adding an edge between a fixed vertex u and an end vertex of the path Pni for each 1≤i≤r.
Theorem 3.4.If T∈T(n), then ρQ(S(n;2,2,n−5))≥ρQ(T), where the equality holds
if and only if T≅S(n;2,2,n−5).
Proof. It is clear that there is such a vertex v in T that N(v) contains at least three non-pendant vertices.
Thus, if n1,n2,…,nr are respectively orders of r connected components of T−v then at least three of them are not less than two.
By Lemma 3.2, ρQ(S(n;n1,n2,…,nr))≥ρQ(T).
We can further see by Lemma 3.2 that ρQ(S(n;2,2,n−5))≥ρQ(S(n;n1,n2,…,nr)).□
Let B(n,k)={T∣T\mboxisatreeofordern\mboxwithk\mboxpendantvertices,\mboxwhere2≤k≤n−2}.
Let P(n;i,j) be the tree obtained from the path P=v1v2…vn−3 by attaching a pendant edge at vi and a pendant
path of length 2 at vj.
Lemma 3.5.Let T∈T(n)∩ℜ(n)∩B(n,4). Then we have
[TABLE]
where the equality holds if and only if T≅P(n;2,3) or T≅P(n;2,n−5).
Proof. Suppose that T0∈T(n)∩ℜ(n)∩B(n,4) is a tree with the maximum distance signless Laplacian spectral radius.
Let d be the diameter of T0, and P=u1u2…ud+1 be a path in T0.
Then there exist one pendant path P1 of length p≥1 at some ui and another one pendant path P2 of length q≥2 at another vertex uj(=ui).
We assume without loss of generality that q≥p. Now we verify the following three claims.
Claim 1.p=1 and q=2.
Set P1=uiz1z2…zp and P2=ujw1w2…wq.
Suppose on the contrary that p≥2.
Note that p≤i−1. Hence we let T0′=T0−zp−1zp+u1zp.
Clearly, T0′∈T(n)∩ℜ(n)∩B(n,4), and by Lemma 3.2, ρQ(T0′)>ρQ(T0).
This contradiction shows that p=1.
Similarly, we can prove that q=2.
The Claim 1 shows that T0≅P(n;i,j), where 2≤i<j≤n−5 and P=u1u2…un−3.
Claim 2.i=2.
Suppose that un−2 and un−1 are respectively the neighbors of ui and uj in T0.
Let x be a Perron vector corresponding to ρQ(T0).
Suppose on the contrary that i≥3. Let U1={u1,u2,…,ui−1}, U2={ui+1,ui+2,…,un−3,un−1,un} and U3={ui,un−2}.
If vi∈U3∖{ui}∑vj∈U1∑(xvi+xvj)2≥vi∈U3∖{ui}∑vj∈U2∑(xvi+xvj)2,
then let T0′=T0−uiun−2+un−1un−2, and otherwise T0′=T0−uiun−2+ui−1un−2.
Thus T0′∈T(n)∩ℜ(n)∩B(n,4), and by Lemma 3.1, ρQ(T0′)>ρQ(T0).
This contradicts maximality of T0, and so i=2.
The Claim 2 shows that T0≅P(n;2,j).
Claim 3.T0≅P(n;2,3) or T0≅P(n;2,n−5).
Suppose on the contrary that T0≆P(n;2,3) and T0≆P(n;2,n−5). Then 4≤j≤n−6.
Let U1={u1,u2,…,uj−1,un−2}, U2={uj+1,uj+2,…,un−3} and U3={uj,un−1,un}.
If vi∈U3∖{uj}∑vj∈U1∑(xvi+xvj)2≥vi∈U3)∖{uj}∑vj∈U2∑(xvi+xvj)2,
then let T0′′=T0−ujun−1+uj+1un−1, and otherwise T0′′=T0−ujun−1+uj−1un−1.
Thus T0′′∈T(n)∩ℜ(n)∩B(n,4), and by Lemma 3.1, ρQ(T0′′)>ρQ(T0).
This contradicts maximality of T0, and so T0≅P(n;2,3) or T0≅P(n;2,n−5). □
Suppose that P=v1v2…vs is a path of a graph G and u∈V(G). Then we let dG(u,P)=min{dG(u,vi)∣1≤i≤s}.
Theorem 3.6.Let T∈T(n)∩ℜ(n). Then we have
[TABLE]
where the equality holds if and only if T≅P(n;2,3) or T≅P(n;2,n−5).
Proof. Suppose that T0∈T(n)∩ℜ(n) is a tree with the maximum distance signless Laplacian spectral radius.
Let A={w1,w2,…,wnT03} be the set of all vertices in T0 which are of degree at least three.
It is obvious that nT03≥2.
Let P=u1u2…ud+1 be a path, where d is the diameter of T0. Next we distinguish two cases.
Case 1.∣A∩V(P)∣≥2.
Suppose for a contradiction that A⊈V(P).
Without loss of generality we assume that wℓ∈A∖V(P) such that dT0(wℓ,P)=max{dT0(wi,P)∣1≤i≤nT03}.
Then there are at least two pendant paths at wℓ, say, P1=wℓz1…zp and P2=wℓv1…vq (p≥q≥1).
Let T0′=T0−vq−1vq+zpvq. Since ∣A∩V(P)∣≥2, we have ∣A\{wℓ}∣≥2, and so T0′∈T(n)∩ℜ(n).
By Lemma 3.2, ρQ(T0′)>ρQ(T0).
This contradicts maximality of T0, and so A⊆V(P).
Suppose by a way of contradiction that nT03≥3.
Let dT0(w1,w2)=max{dT0(wp,wq)∣1≤p=q≤nT03}.
Then there are two pendant paths at w1, say, P3=w1y1…ys and P4=w1f1…ft (s≥t≥1).
Let T0′=T0−ft−1ft+ysft.
Thus T0′∈T(n)∩ℜ(n),
and by Lemma 3.2, ρQ(T0′)>ρQ(T0).
This contradicts maximality of T0, and so nT03=2.
Suppose for a contradiction that dT0(w1)≥4.
Then there exist three pendant paths at w1 in T0, say P3=w1y1…ys, P4=w1f1…ft and P5=w1h1…hm (s≥t≥m≥1).
Let T0′=T0−hm−1hm+fthm.
Thus T0′∈T(n)∩ℜ(n), and by Lemma 3.2, ρQ(T0′)>ρQ(T0).
This contradicts maximality of T0, and so dT0(w1)=3.
Similarly, we can prove dT0(w2)=3.
Therefore, T0∈T(n)∩ℜ(n)∩B(n,4), and
from Lemma 3.5 we know that the assertion is true.
Case 2.∣A∩V(P)∣=1.
Let A∩V(P)={w1}.
Since nT03≥2, A\V(P)=∅.
Suppose on the contrary that nT03≥3.
Let w2∈A\{w1} such that dT0(w1,w2)=max{dT0(w1,wj)∣2≤j≤nT03}.
Then there are two pendant paths at w2, say, P6=w2a1…ar and P7=w2b1…bj (r≥j≥1).
Let T0′′=T0−bj−1bj+arbj.
Since ∣A\{w2}∣≥2, T0′′∈T(n)∩ℜ(n), and by Lemma 3.2, ρQ(T0′′)>ρQ(T0).
This contradicts maximality of T0, and so nT03=2.
By similar argument as Case 1, we can verify that dT0(w1)=dT0(w2)=3,
and so T∈T(n)∩ℜ(n)∩B(n,4).
By Lemma 3.5, we know that the assertion is true. □
Bibliography17
The reference list from the paper itself. Each links out to its DOI / PubMed record.
1[1] M. Aouchiche, P. Hansen, Two Laplacians for the distance matrix of a graph, Linear Algebra Appl. 439 (2013) 21-33.
2[2] M. Aouchiche, P. Hansen, Some properties of the distance Laplacian eigenvalues of a graph, Les Cahiers du GERAD, G-2013-28.
3[3] M. Aouchiche, P. Hansen, The distance spectra of a graphs: A survey, Linear Algebra Appl. 458 (2014) 301-386.
4[4] S. Bose, M. Nath, S. Paul, On the maximal distance spectral radius of graphs without a pendant vertex, Linear Algebra Appl. 438 (2013) 4260-4278.
5[5] A. Ilić, Distance spetral radius of trees with given matching number, Discrete Appl. Math. 158 (2010) 1799-1806.
6[6] H. Liu, M. Lu, Bounds on the distance signless Laplacian spectral radius in terms of clique number, Linear and Multilinaer Algebra 63 (2015) 1750-1759.
7[7] H. Liu, B. Zhou, On the distance Laplacian spectral radius of graphs, Linear Algebra Appl. 475 (2015) 265-275.
8[8] H. Minc, Nonnegative matrices. New York; 1988.