
Previously, boundary control problems for parabolic type equations were considered. A portion of the thin rod boundary has a temperature-controlled heater. Its mode of operation should be found so that the average temperature in some region reaches a certain value. In this article, we consider the boundary control problem for the pseudo-parabolic equation. The value of the solution with the control parameter is given in the boundary of the interval. Control constraints are given such that the average value of the solution in considered domain takes a given value. The auxiliary problem is solved by the method of separation of variables, and the problem under consideration is reduced to the Volterra integral equation. The existence theorem of admissible control is proved by the Laplace transform method.
Citation: Farrukh Dekhkonov. On a boundary control problem for a pseudo-parabolic equation[J]. Communications in Analysis and Mechanics, 2023, 15(2): 289-299. doi: 10.3934/cam.2023015
[1] | Zhi-Yu Shi, Jia-Bao Liu . Topological indices of linear crossed phenylenes with respect to their Laplacian and normalized Laplacian spectrum. AIMS Mathematics, 2024, 9(3): 5431-5450. doi: 10.3934/math.2024262 |
[2] | Jahfar T K, Chithra A V . Central vertex join and central edge join of two graphs. AIMS Mathematics, 2020, 5(6): 7214-7233. doi: 10.3934/math.2020461 |
[3] | Zenan Du, Lihua You, Hechao Liu, Yufei Huang . The Sombor index and coindex of two-trees. AIMS Mathematics, 2023, 8(8): 18982-18994. doi: 10.3934/math.2023967 |
[4] | Jia-Bao Liu, Kang Wang . The multiplicative degree-Kirchhoff index and complexity of a class of linear networks. AIMS Mathematics, 2024, 9(3): 7111-7130. doi: 10.3934/math.2024347 |
[5] | Akbar Ali, Sadia Noureen, Akhlaq A. Bhatti, Abeer M. Albalahi . On optimal molecular trees with respect to Sombor indices. AIMS Mathematics, 2023, 8(3): 5369-5390. doi: 10.3934/math.2023270 |
[6] | Kun Wang, Wenjie Ning, Yuheng Song . Extremal values of the modified Sombor index in trees. AIMS Mathematics, 2025, 10(5): 12092-12103. doi: 10.3934/math.2025548 |
[7] | Zhen Lin, Ting Zhou, Xiaojing Wang, Lianying Miao . The general Albertson irregularity index of graphs. AIMS Mathematics, 2022, 7(1): 25-38. doi: 10.3934/math.2022002 |
[8] | Zhibin Du, Ayu Ameliatul Shahilah Ahmad Jamri, Roslan Hasni, Doost Ali Mojdeh . Maximal first Zagreb index of trees with given Roman domination number. AIMS Mathematics, 2022, 7(7): 11801-11812. doi: 10.3934/math.2022658 |
[9] | Tariq A. Alraqad, Igor Ž. Milovanović, Hicham Saber, Akbar Ali, Jaya P. Mazorodze, Adel A. Attiya . Minimum atom-bond sum-connectivity index of trees with a fixed order and/or number of pendent vertices. AIMS Mathematics, 2024, 9(2): 3707-3721. doi: 10.3934/math.2024182 |
[10] | Zhenhua Su, Zikai Tang, Hanyuan Deng . Higher-order Randić index and isomorphism of double starlike trees. AIMS Mathematics, 2023, 8(12): 31186-31197. doi: 10.3934/math.20231596 |
Previously, boundary control problems for parabolic type equations were considered. A portion of the thin rod boundary has a temperature-controlled heater. Its mode of operation should be found so that the average temperature in some region reaches a certain value. In this article, we consider the boundary control problem for the pseudo-parabolic equation. The value of the solution with the control parameter is given in the boundary of the interval. Control constraints are given such that the average value of the solution in considered domain takes a given value. The auxiliary problem is solved by the method of separation of variables, and the problem under consideration is reduced to the Volterra integral equation. The existence theorem of admissible control is proved by the Laplace transform method.
Throughout this article, we only consider simple, undirected, and finite graphs and assume that all graphs are connected. Suppose G is a graph with the vertex set V(G)={v1,v2,⋅⋅⋅,vn} and the edge set E(G)={e1,e2,⋅⋅⋅,em}. The adjacency matrix A(G) is a 0−1n×n matrix indexed by the vertices of G and defined by aij=1 if and only if vsvt∈EG. For more notation, one refer to [1].
The Laplacian matrix of graph G is defined as L(G)=D(G)−A(G), and assume that the eigenvalues of L(G) are labeled 0≤μ1<μ2<⋅⋅⋅<μn.
(L(G))st={1,s=t;−1,s≠t,vs∼vt;0,otherwise. | (1.1) |
The normalized Laplacian matrix is given by
(L(G))st={1,s=t;−1√dsdt,s≠t,vs∼vt;0,otherwise. | (1.2) |
The distance, dij=dG(vs,vt), between vertices us and ut of G is the length of the shortest us, ut-path in G. The Wiener index [2,3] is the sum of the distances of any two vertices in the graph G, that is
W(G)=∑s<tdst. |
In 1947, the distance-based invariant first appeared in chemistry [2,3] and began to the applied to mathematics 30 years later [5]. Nowadays, the Wiener index is widely used in mathematics [6,7,8] and chemistry [9,10,11].
In a simple graph G, the degree, dij=dG(vi), of a vertex vi is the number of edges at vi. The Gutman index [12] of the simple graph G is expressed by
Gut(G)=∑s<tdsdtdst. |
Klein and Randic initially outlined the concepts associated with the resistance distance of the graph. Assume that each edge is replaced by a unit resistor, and we use rst to denote the resistance distance between two vertices s and t. Similar to the Wiener index, the Kirchhoff index [13,14] of graph G is expressed as the sum of the resistance distances between each two vertices, that is
Kf(G)=∑s<trst. |
In 2007, Chen and Zhang [15] defined the multiplicative degree-Kirchhoff index [16,17], that is
Kf∗(G)=∑s<tdsdtrst. |
Phenyl is a conjugated hydrocarbon, and L6,4,4n denote a linear chain, containing n hexagons and 2n−1 squares, as shown in Figure 1.
With the rapid changes of the times, organic chemistry has also developed rapidly, which has led to a growing interest in polycyclic aromatic compounds.
In 1985, the computational method and procedure of the matrix decomposition theorem were proposed by Yang [18]. This led to the solution of some problems in graph networks and allowed the unprecedented development of self-homogeneous linear hydrocarbon chains. For example, in 2021, X.L. Ma [20] got the normalized Laplacian spectrum of linear phenylene, and the linear phenylene containing it has n hexagons and n−1 squares. L. Lan [21] explored linear phenylene with n hexagons and n squares. Umar Ali [22] analyzed the strong prism of a graph G, which is the strong product of the complete graph of order 2, and X.Y. Geng [23] obtained the Laplacian spectrum of L6,4,4n, which contains n hexagons and 2n−1 squares. J.B. Liu [24] derived the Kirchhoff index and complexity of On, which denoting linear octagonal-quadrilateral networks. C. Liu [25] got the Laplacian spectrum and Kirchhoff index of Ln, and Ln has t hexagons and 3t+1 quadrangles.
Inspired by these recent works, we try to investigate the Laplacians and the normalized Laplaceians for graph transformations on phenyl dicyclobutadieno derivatives.
The various sections of this article are as follows: In Section 2, we proposed some concepts and lemmas and used them in subsequent articles. In Section 3 and Section 4, we acquired the Laplacian matrix and the normalized Laplacian matrix, then we made sure the Kirchoff index, the multiplicative degree-Kirchoff index, and the complexity of Ln. In Section 5, we obtained conclusions based on the calculations in this paper.
In this article, graph Ln and graph L6,4,4n are portrayed in Figure 1. Define the characteristic polynomial of matrix U of order n as PU(x)=det(xI−U).
It is easy to understand that π=(1,1′)(2,2′)⋯((4n),(4n)′) is an automorphism. Let V1={u1,u2,⋯,u4n+1,v1,⋯,v4n},V2={u′1,u′2,⋯,u′4n,v′1,⋯,v′4n+1}, |V(Ln)|=8n and |E(Ln)|=19n−4. Thus, the (normalized) Laplacians matrix can be expressed in the form of a block matrix, that is
L(Ln)=(LV0V0LV0V1LV0V2LV1V0LV1V1LV1V2LV2V0LV2V1LV2V2), |
where LVsVt and LVsVt are a submatrix consisting of rows corresponding to the vertices in Vs and columns corresponding to the vertices in Vt, s,t=0,1,2. Let
Q=(It0001√2I4n1√2I4n01√2I4n−1√2I4n), |
then
QL(LG)Q′=(LA(G)00LS(G)),QL(LG)Q′=(LA(G)00LS(G)), |
and Q′ is the transposition of Q.
LA=LV1V1+LV1V2, LS=LV1V1−LV1V2, LA=LV1V1+LV1V2, LS=LV1V1−LV1V2
Lemma 2.1. [20] If G is a graph and suppose that LA(G),LS(G),LA(G), and LS(G) are determined as above, then
ϑL(Ln)(y)=θLA(G)(y)θLS(G)(y),ϑL(Ln)(y)=θLA(G)(y)θLS(G)(y). |
Lemma 2.2. [26] With the extensive study of the Kirchhoff index, Gutman and Mohar proposed an algorithm based on the relation between the Kirchhoff index and the Laplacian eigenvalues, namely
Kf(G)=nn∑t=21ξt, |
and the eigenvalues of L(G) are 0=ξ1<ξ2≥⋯≥ξn(n≥2).
Lemma 2.3. [14] Suppose that the eigenvalues of L(G) are ξ1≤ξ2≤⋯≤ξn, then its multiplicative degree-Kirchhoff index can be denoted by
Kf∗(G)=2mn∑t=21ξt. |
Lemma 2.4. [1] The number of spanning trees in G can also be called the complexity of G. If G is a graph with |VG|=n and |EG|=m. Let λi(i=2,3,⋯,n) be the eigenvalues of L(G). Then the complexity of G is
2mτ(G)=n∏i=1di⋅n∏i=2λi. |
In this section, the main objective is to find out the Kirchhoff index of Ln. Then, combining the definition of the Laplacian matrix and Eq (1.1), we can write these block matrices as follows:
LV1V1=(3−1−14−1−15−1−15−1−15−1−14−1⋱−15−1−14−1−15−1−13)(4n)×(4n), |
and
LV1V2=(−1−1−10−1−1−1−1−1−1−1−1−1−1−10−1⋱−1−1−1−10−1−1−1−1−1−1)(4n)×(4n). |
Hence,
LA=(2−2−24−2−24−2−24−2−24−2−24−2⋱−24−2−24−2−24−2−22)(4n)×(4n), |
and
LS=diag(4,4,6,6,6,4,⋅⋅⋅,6,4,6,4). |
Assume that 0≤α1<α2≤α3≤⋯≤α4n are the roots of PLA(x)=0, and 0≤β1≤β2≤β3≤⋯≤β4n are the roots of PLS(x)=0. By Lemma 2.2, we immediately have
Kf(L2n)=2(4n)(4n∑i=21αi+4n∑j=11βj). | (3.1) |
Obviously, ∑4nj=11βj can be obtained according to LS.
4n∑j=11βj=16×(3n−2)+14×(n+2)=9n+212. | (3.2) |
Next, we focus on computing ∑4ni=11αi. Let
PLA(x)=det(xI−LA)=x(x4n−1+a1x4n−2+⋯+a4n−2x+a4n−1),a4n−1≠0. |
Based on Vieta's theorem of PLA(x), we can exactly get the following equation:
4n∑i=21αi=(−1)4n−2a4n−2(−1)4n−1a4n−1. | (3.3) |
For the sake of convenience, let Ms be used to express the s−th order principal minors of matrix A, and ms=detMs is recorded. We can get m1=2,m2=4,m3=8.
And
ms=4ms−1−4ms−2,4≤s≤4n, |
by further induction, we have
ms=2s. |
In this way, we can get two theorems.
Theorem 3.1. (−1)4n−1a4n−1=(4n)24n−1.
Proof. Due to the sum of all the principal minors of order 4n−1 of LA is (−1)4n−1a4n−1,
(−1)4n−1a4n−1=4n∑s=1detLA[s]=4n∑s=1det(Ms−100U4n−s)=4n∑s=1detMs−1⋅detU4n−s, |
where
Ms−1=(l11−2⋯0−2l22⋯0⋮⋮⋱⋮00⋯ls−1,s−1)(s−1)×(s−1), |
U4n−s=(ls+1,s+1⋯⋯0⋮⋱⋮⋮⋮⋯l4n−1,4n−1⋮0⋯−2l4n,4n)(4n−s)×(4n−s). |
Let m0=1,detU0=1, because of the symmetry of matrix LA, then detU4n−s=detM4n−s. Hence
(−1)4n−1a4n−1=4n∑s=1detms−1⋅detm4n−s=(4n)24n−1, |
as desired.
Theorem 3.2. (−1)4n−2a4n−2=(4n−1)(4n)(4n+1)24n−13.
Proof. Since the (−1)4n−2a4n−2 is the tatal of all the principal minors of order 4n−2 of LA, we have
(−1)4n−1a4n−1=∑1≤s<t≤4ndetLA[s,t]=∑1≤s<t≤4ndetMs−1detNt−s−1detU4n−t, |
where
LA[s,t]=(Mp−1000Nt−s−1000U4n−t),1≤s<t≤4n |
and
Nt−s−1=|4−2−24−2−24−2⋱−24−2−24−2−24|(t−s−1)=(t−s)2t−s−1. |
Therefore, we can have
(−1)4n−2a4n−2=∑1≤s<t≤4ndetMs−1⋅detNt−s−1⋅detU4n−t=∑1≤s<t≤4n(t−s)2t−s−1⋅detms−1⋅m4n−t=(4n−1)(4n)(4n+1)24n−13. |
The proof is over.
From the results of Theorem 3.1 and Theorem 3.2, we can get
4n∑i=21αi=(−1)4n−2a4n−2(−1)4n−1a4n−1=16n2−112, | (3.4) |
where the eigenvalues of LA are 0≤α1<α2≤α3≤⋯≤α4n.
Lemma 3.3. Suppose L6,4,4n be the dicyclobutadieno derivative of phenylenes and the graph Ln be obtained from the transformation of the graph L6,4,4n.
Kf(Ln)=32n3+18n2+2n3. |
Proof. Substituting Eqs (3.5) and (3.6) into (3.4), the Kirchhoff index of Ln can be expressed
Kf(Ln)=2(4n)(4n∑i=21αi+4n∑j=11βj)=(8n)(9n+212+(4n+1)(4n−1)12)=32n3+18n2+2n3. |
The result is as desired.
The Kirchhoff index of Ln from L1 to L12 is shown in Table 1.
G | Kf(G) | G | Kf(G) | G | Kf(G) |
L1 | 17.3 | L5 | 1486.7 | L9 | 8268.0 |
L2 | 110.7 | L6 | 2524.0 | L10 | 24457.3 |
L3 | 344.0 | L7 | 3957.3 | L11 | 30454.7 |
L4 | 781.3 | L8 | 5850.7 | L12 | 37360.0 |
Next, we will further consider the Wiener index of Ln.
Lemma 3.4. Let L6,4,4n be the dicyclobutadieno derivative of [n]phenylenes, and the graph Ln be obtained from the transformation of the graph L6,4,4n, then
limn→∞Kf(Ln)W(Ln)=14. |
Proof. Consider dst for all vertices. For the sake of convenience, we divide the vertices of the graph into the following five categories:
Case 1. Vertex 1 of Ln:
g1(i)=1+2(4n−1∑k=1k). |
Case 2. Vertex 4j−3(j=1,2,⋯,n) of Ln, i=4j−3:
g2(i)=1+2(4n−1∑k=1k+4n−i∑k=1k). |
Case 3. Vertex 4j−2(j=1,2,⋯,n) of Ln, i=4j−2:
g3(i)=1+2(4n−1∑k=1k+4n−i∑k=1k). |
Case 4. Vertex 4j−1(j=1,2,⋯,n) of Ln, i=4j−1:
g4(i)=1+2(4n−1∑k=1k+4n−i∑k=1k). |
Case 5. Vertex 4j(j=1,2,⋯,n) of Ln, i=4j:
g5(i)=1+2(4n−1∑k=1k+4n−i∑k=1k). |
Hence, we have
W(Ln)=4g1(i)+2∑i=4j−3g2(i)+2∑i=4j−2g3(i)+2∑i=4j−1g4(i)+2∑i=4jg5(i)2=4(1+2∑4n−1k=1k)+2∑nj=1[1+2(∑4j−kk=1k+∑4n−4j+2k=1k)]2+2∑nj=1[2+2(∑4j−3k=1k+∑4n−4j+2k=1k)]+2∑nj=1[1+2(∑4j−2k=1k+∑4n−4j+1k=1k)]2+2∑n−1j=1[1+2(∑4j−1k=1k+∑4n−4jk=1k)]2=128n3+48n2−5n+33. |
Consider the above results of the Kirchhoff index and the Wiener index. We can get following equation when n tends to infinity:
limn→∞Kf(Ln)W(Ln)=14. |
The result is as desired.
In this section, we use the eigenvalues of the normalized Laplacian matrix to determine the multiplicative degree-Kirchhoff index of Ln. Besides, we calculate the complexity of Ln. Then
LV1V1=(1−1√12−1√121−1√20−1√201−15−151−15−151−1√20−1√201−1√20⋱−151−1√20−1√201−1√20−1√201−1√15−1√151)(4n)×(4n), |
and
LV1V2=(−13−1√12−1√120−1√20−1√20−15−15−15−15−15−15−15−1√20−1√200−1√20⋱−15−15−1√20−1√200−1√20−1√20−15−1√15−1√15−13)(4n)×(4n). |
Therefore,
LA=(23−1√3−1√31−1√5−1√545−25−2545−25−2545−1√5−1√51−1√5⋱−2545−1√5−1√51−1√5−1√545−2√15−2√1523)(4n)×(4n), |
and
LS=diag(43,1,65,65,65,⋅⋅⋅,65,1,65,43). |
Assume that the roots of PLA(x)=0 are 0≤ξ1<ξ2≤ξ3≤⋯≤α4n, and 0≤γ1≤γ2≤γ3≤⋯≤γ4n are the roots of PLS(x)=0. By Lemma 2.3, we can get
Kf∗(L2n)=2(19n−4)(4n∑i=21ξi+4n∑j=11γj). |
Since Ls is a diagonal matrix. Obviously, its diagonal elements 1,43 and 65 correspond to the eigenvalues of Ls, respectively. Then, it can be clearly obtained as
4n∑i=11γi=21n−16. | (4.1) |
Let
PLA(x)=det(xI−LA)=x4n+b1x4n−1+⋯+b4n−1x,b4n−1≠0, |
1ξ2,1ξ3,⋅⋅⋅,1ξ4n+1are the roots of the following equation
b4n−1x4n−1+b4n−2x4n−2+⋅⋅⋅+b1x+1=0. |
Based on the Vieta' s theorem of PLA(x), we can get
4n∑i=21ξi=(−1)4n−2b4n−2(−1)4n−1b4n−1. |
Similarly, we can get another two interesting facts.
Theorem 4.1. (−1)4n−1b4n−1=259(38n−8)(4125)n.
Proof. Let sp=detFp, then we have s1=23,s2=13,s3=215,s4=475, and
{s4p=45s4p−1−425s4p−2;s4p+1=45s4p−425s4p−1;s4p+2=s4p+1−15s4p;s4p+3=45s4p+2−15s4p+1. |
After further simplification, the transformation form of the above formula is obtained.
{s4p=53⋅(4125)p,1≤p≤n;s4p+1=23⋅(4125)i,0≤p≤n−1;s4p+2=13⋅(4125)i,0≤p≤n−1;s4p+3=115⋅(4125)i,0≤p≤n−1. |
Similarly, we have t1=23,t2=415,t3=215,t4=475, and
{t4p=25t4p−1−25t4p−2;t4p+1=45t4p−425t4p−1;t4p+2=45t4p+1−425t4p;t4p+3=t4p+2−15t4p+1. |
Therefore, the transformation form of the above formula is obtained.
{t4p−4=53⋅(4125)p,1≤p≤n;t4p−3=23⋅(4125)p,0≤p≤n−1;t4p−2=415⋅(4125)p,0≤p≤n−1;t4p−1=215⋅(4125)p,0≤p≤n−1. |
Since the (−1)4n−1b4n−1 is the total of all the principal minors of order 4n−1 of LA, we have
(−1)4n−1b4n−1=4n∑i=2detNLA[i]+s4n+t4n=145(38n−8)(4125)n. |
The proof of Theorem 4.1 completed.
Theorem 4.2. (−1)4n−2b4n−2=13240(14520n3+4599n2−1496n+3)(4125)n.
Proof. We observe that the sum of all the principal minors of order 4n in LA is (−1)4n−2b4n−2, then
(−1)4n−2b4n−2=∑1≤s<t≤4ndetLA[s,t]⋅fs−1⋅f′4n−t. | (4.2) |
By Eq (4.8), we know that the result of detLA[s,t] will change with the values of s and t. Then we can get the following twenty cases:
Case 1. i=4s, j=4t, 1≤s<t≤n,
detφ=|45−1√5−2√51−15−1545−2√5⋱−2545−25−25−45−1√5−151−1√5−25−45|(4t−4s−1)=10(t−s)(4125)t−s. |
Case 2. i=4s, j=4t+1, 1≤s≤t≤n−1,
detφ=|45−1√5−1√51−15−1545−2√5⋱−2545−15−151−1√5−1√545−2√5−25−45|(4t−4s)=(4t−4s+1)(4125)t−s. |
Case 3. i=4s, j=4t+2, 1≤s≤t≤n−1,
detφ=|45−1√5−1√51−15−1545−2√5⋱−1√51−1√5−1√545−25−2545−25−25−45|(4t−4s+1)=45(2t−2s+1)(4125)t−s. |
Case 4. i=4s, j=4t+3, 1≤s≤t≤n−1,
detφ=|45−1√5−1√51−15−1545−2√5⋱−1√545−25−2545−25−2545−1√5−1√51|(4t−4s+2)=15(4t−4s+3)(4125)t−s. |
Case 5. i≡0, j=4n, 1≤s≤t,
detφ=|45−1√5−2√51−15−1√545−25⋱−2545−25−2545−1√5−1√51−1√5−2√545|(4n−4s−1)=10(n−s)(4125)n−s. |
Case 6. i=4s+1, j=4t, 1≤s<t≤n,
detφ=|1−1√5−1√545−2√5−2√545−2√5⋱−2545−25−2545−1√5−1√51−1√5−1√545|(4t−4s−2)=254(4t−4s−1)(4125)t−s. |
Case 7. i=4s+1, j=4t+1, 1≤s<t≤n−1,
detφ=|1−1√5−1√545−2√5−2√545−2√5⋱−2545−1√5−1√51−1√5−1√545−25−2545|(4t−4s−1)=10(t−s)(4125)t−s. |
Case 8. i=4s+1, j=4t+2, 1≤s<t≤n−1,
detφ=|1−1√5−1√545−2√5−2√545−2√5⋱−251−1√5−1√545−25−2545−25−2545|(4t−4s)=(4t−4s+1)(4125)t−s. |
Case 9. i=4s+1, j=4t+3, 0≤s≤t≤n−1,
detφ=|1−1√5−1√545−2√5−2√545−2√5⋱−1√545−25−2545−25−2545−1√5−1√51|(4t−4s+1)=(2t−2s+1)(4125)t−s. |
Case 10. i≡1, j=4n+1, 0≤s≤n,
detφ=|1−1√5−1√545−2√5−2√545−2√5⋱−2545−25−2545−1√5−1√51−1√5−1√545|(4n−4s−2)=254(4n−4s−1)(4125)n−s. |
Case 11. i=4s+2, j=4t, 0≤s<t≤n,
detφ=|45−25−2545−25−2545−25⋱−2545−25−2545−1√5−1√51−1√5−1√545|(4t−4s−3)=25(2t−2s−1)(4125)t−s. |
Case 12. i=4s+2, j=4t+1, 0≤s<t≤n−1,
detφ=|45−25−2545−25−2545−25⋱−1√51−1√5−1√545−25−2545−25−2545|(4t−4s−2)=5(4t−4s−1)(4125)t−s. |
Case 13. i=4s+2, j=4t+2, 0≤s<t≤n−1,
detφ=|45−25−2545−25−2545−25⋱−1√51−1√5−1√545−25−2545−25−2545|(4t−4s−1)=8(t−s)(4125)t−s. |
Case 14. i=4s+2, j=4t+3, 0≤s≤t≤n−1,
detφ=|45−25−2545−25−2545−25⋱−1√545−25−2545−25−2545−1√5−1√51|(4t−4s)=(4t−4s+1)(4125)t−s. |
Case 15. i≡2, j=4t+3, 0≤s≤n−1,
detφ=|45−25−2545−25−2545−25⋱−2545−25−2545−1√5−1√51−1√5−1√545|(4n−4s−3)=25(2n−2s−1)(4125)n−s. |
Case 16. i=4s+3, j=4t, 0≤s<t≤n,
detφ=|45−25−2545−1√5−1√51−1√5⋱−2545−25−2545−1√5−1√51−1√5−1√545|(4t−4s−4)=1254(4t−4s−3)(4125)t−s. |
Case 17. i=4s+3, j=4t+1, 0≤s<t≤n−1,
detφ=|45−25−2545−1√5−1√51−1√5⋱−2545−1√5−1√51−1√5−1√545−1√5−1√545|(4t−4s−3)=25(2t−2s−1)(4125)t−s. |
Case 18. i=4s+3, j=4t+2, 0≤s<t≤n−1,
detφ=|45−25−2545−1√5−1√51−1√5⋱−1√51−1√5−1√545−25−2545−25−2545|(4t−4s−3)=253(4t−4s−1)(4125)t−s. |
Case 19. i=4s+3, j=4t+3, 0≤s<t≤n−1,
detφ=|45−25−2545−1√5−1√51−1√5⋱−1√545−25−2545−25−2545−1√5−1√51|(4t−4s−1)=10(t−s)(4125)t−s. |
Case 20. i≡3, j=4t, 0≤s≤n−1,
detφ=|45−25−2545−1√5−1√51−1√5⋱−2545−25−251−1√5−1√51−1√5−1√545|(4n−4s−4)=1254(4n−4s−3)(4125)n−s. |
Therefore, we can get
(−1)4n−2b4n−2=E1+E2+E3+E4, |
where
E1=118(227n3+347n2−574n+4)(4125)n−1. |
E2=172(908n3+3431n2+523n)(4125)n. |
E3=145(454n3+1375n2−1079n)(4125)n. |
E4=181(92n3+561n2−611n)(4125)n−1. |
Hence
(−1)4n−2b4n−2=E1+E2+E3+E4=13240(14520n3+4599n2−1496n+4)(4125)n. |
The proof of Theorem 4.2 completed.
Let 0≤ξ1≤ξ2≤ξ3≤⋯≤ξ3n+2 are the eigenvalues of LA. We can get the following exact equation:
4n∑i=21ξi=(−1)4n−2b4n−2(−1)4n−1b4n−1=172(14520n3+4599n2−1496n+838n−8). |
Theorem 4.3. Set L6,4,4n be the derivative [n]pheylenes, and the expression of the multiplicative degree-Kirchhoff index is
Kf∗(Ln)=(29040n3+8996n2−3198n+8144). |
Proof. Together with Eq.(4.7) and Theorems 4.1 and 4.2, one can get
Kf∗(Ln)=2(19n−4)(4n∑i=21ξi+4n∑i=11γi)=2(19n−4)[172(14520n3+4599n2−1496n+838n−8)+21n−16]=(29040n3+8996n2−3198n+8144). |
The result as desired.
The multiplicative degree-Kirchhoff indices of Ln from L1 to L12, see Table 2.
G | Kf∗(G) | G | Kf∗(G) | G | Kf∗(G) |
L1 | 241.98 | L5 | 26659.15 | L9 | 151875.4 |
L2 | 1818.86 | L6 | 45675.81 | L10 | 207691.9 |
L3 | 5940.68 | L7 | 72077.4 | L11 | 275733.2 |
L4 | 13817.44 | L8 | 107073.9 | L12 | 357209.6 |
Then we want to calculate the Gutman index of Ln.
Theorem 4.4. Suppose that L6,4,4n is the dicyclobutadieno derivative of [n]phenylenes and the graph Ln is obtained from the transformation of the graph L6,4,4n, then
limn→∞Kf∗(Ln)Gut(Ln)=14. |
Proof. Consider dij for all vertices. We divide the vertices of Ln into the following four categories.
Case 1. Vertex 4i−2(i=1,2,⋯,n) of Ln:
f4i−2=103n(56n2−24n+37). |
Case 2. Vertex 4i−1(i=1,2,⋯,n) of Ln:
f4i−1=103n(152n2−48n+29). |
Case 3. Vertex 4i(i=1,2,⋯,n) of Ln:
f4i=103n(140n2−48n+43). |
Case 4. Vertex 4i−3(i=1,2,⋯,n) of Ln:
f4i−3=103n(136n2−6n+71). |
According to Eq.(1.3), the Gutman index of Ln is
Gut(Ln)=f4i+f4i−1+f4i−2+f4i−32=10n3(242n2−63n+61). |
Therefore, combining with Kf∗(Ln) and Gut(Ln), we have
limn→∞Kf∗(Ln)Gut(Ln)=14. |
The result as desired.
Finally, we want to know the complexity of Ln.
Theorem 4.5. For the graph Ln, we have
τ(Ln)=23n+2⋅33n−2. |
Proof. Based on Lemma 2.4, we can get
8n∏i=1di4n∏i=2αi4n∏j=1βi=2(19n−4)⋅τ(Ln). |
Note that
8n∏i=1di=34⋅42n⋅56n−4. |
4n∏i=2αi=259⋅(38n−8)⋅(4125)n. |
4n∏j=1βi=(43)2⋅(65)3n−2. |
Hence,
τ(Ln)=23n+2⋅33n−2. |
The proof is over.
Thus, we can get the complexity of Ln from W1 to W8 which are listed in Table 3.
G | τ(G) | G | τ(G) |
W1 | 96 | W5 | 208971104256 |
W2 | 20736 | W6 | 45137758519296 |
W3 | 4478976 | W7 | 9749755840167936 |
W4 | 967458816 | W8 | 2105947261476274176 |
In this paper, the linear chain network with n hexagons and 2n−1 squares is considered. We have devoted ourselves to calculating the (multiplicative degree) Kirchhoff index, Wiener index, Gutman index, and complexity. In the meantime, we deduced that the ratio of the (multiplicative degree) Kirchhoff index to the (Gutman) Wiener index is nearly a quarter when n tends to infinity. These rules also apply to some other graphs.
The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.
This research was funded by the Anhui Provincial Natural Science Research Major Project (No. 2022AH040317) and the Anhui Provincial 2023 Action Project for Cultivating Young and Middle Aged Teachers in Universities (No. DTR2023095).
No potential conflicts of interest were reported by the authors.
[1] |
B. D. Coleman, W. Noll, An Approximation Theorem for Functionals, with Applications in Continuum Mechanics, Arch. Rational Mech. Anal., 6 (1960), 355–370. https://doi.org/10.1007/BF00276168 doi: 10.1007/BF00276168
![]() |
[2] |
B. D. Coleman, R. J. Duffin, V. J. Mizel, Instability, uniqueness, and nonexistence theorems for the equation on a strip, Arch. Rational Mech. Anal., 19 (1965), 100–116. https://doi.org/10.1007/BF00282277 doi: 10.1007/BF00282277
![]() |
[3] |
L. W. White, Point control approximations of parabolic problems and pseudo-parabolic problems, Appl. Anal., 12 (1981), 251–263. https://doi.org/10.1080/00036818108839365 doi: 10.1080/00036818108839365
![]() |
[4] |
I. E. Egorov, E. S. Efimova, Boundary value problem for third order equation not presented with respect to the highest derivative, Mat. Zamet., 24 (2017), 28–36. https://doi.org/10.25587/SVFU.2018.4.11314 doi: 10.25587/SVFU.2018.4.11314
![]() |
[5] | A. I. Kozhanov, The existence of regular solutions of the first boundary value Problem for one class of Sobolev type equations with alternating direction, Mat. Zamet. YaGU, 2 (1997), 39–48. |
[6] | S. I. Lyashko, On the solvability of pseudo-parabolic equations, Mat., 9 (1985), 71–72. |
[7] |
S. Albeverio, Sh. A. Alimov, On one time-optimal control problem associated with the heat exchange process, Appl. Math. Opt., 57 (2008), 58–68. https://doi.org/10.1007/s00245-007-9008-7 doi: 10.1007/s00245-007-9008-7
![]() |
[8] | Sh. A. Alimov, On a control problem associated with the heat transfer process, Eurasian. Math. J., 1 (2010), 17–30. |
[9] |
Sh. A. Alimov, F. N. Dekhkonov, On the time-optimal control of the heat exchange process, Uzbek Mathematical Journal, 2 (2019), 4–17. https://doi.org/10.29229/uzmj.2019-2-1 doi: 10.29229/uzmj.2019-2-1
![]() |
[10] | Yu. V. Egorov, Optimal control in Banach spaces, Dokl. Akad. Nauk SSSR, 150 (1967), 241–244. |
[11] |
Z. K. Fayazova, Boundary control of the heat transfer process in the space, Izv. Vyssh. Uchebe. Zaved. Mat., 12 (2019), 82–90. https://doi.org/10.26907/0021-3446-2019-12-82-90 doi: 10.26907/0021-3446-2019-12-82-90
![]() |
[12] |
Z. K. Fayazova, Boundary control for a Psevdo-Parabolic equation, Mathematical Notes of NEFU, 12 (2018), 40–45. https://doi.org/10.25587/SVFU.2019.20.57.008 doi: 10.25587/SVFU.2019.20.57.008
![]() |
[13] |
F. N. Dekhkonov, E.I. Kuchkorov, On the time-optimal control problem associated with the heating process of a thin rod, Lobachevskii. J. Math., 44 (2023), 1126–1136. https://doi.org/10.1134/S1995080223030101 doi: 10.1134/S1995080223030101
![]() |
[14] |
F. N. Dekhkonov, On a time-optimal control of thermal processes in a boundary value problem, Lobachevskii. J. Math., 43 (2022), 192–198. https://doi.org/10.1134/S1995080222040096 doi: 10.1134/S1995080222040096
![]() |
[15] |
F. N. Dekhkonov, On the control problem associated with the heating process, Mathematical notes of NEFU, 29 (2022), 62–71. https://doi.org/10.25587/SVFU.2023.82.41.005 doi: 10.25587/SVFU.2023.82.41.005
![]() |
[16] | H. O. Fattorini, Time-optimal control of solutions of operational differential equations, DSIAM J. Control, 2 (1965), 49–65. |
[17] |
H. O. Fattorini, Time and norm optimal controls: a survey of recent results and open problems, Acta Math. Sci. Ser. B Engl. Ed., 31 (2011), 2203–2218. http://dx.doi.org/10.1016/S0252-9602(11)60394-9 doi: 10.1016/S0252-9602(11)60394-9
![]() |
[18] | A. V. Fursikov, Optimal Control of Distributed Systems, Theory and Applications, Translations of Math. Monographs, Amer. Math. Soc., Providence, Rhode Island, (2000). https://doi.org/10.1090/mmono/187D |
[19] |
A. Friedman, Optimal control for parabolic equations, J. Math. Anal. Appl., 18 (1967), 479–491. https://doi.org/10.1016/0022-247X(67)90040-6 doi: 10.1016/0022-247X(67)90040-6
![]() |
[20] | J. L. Lions, P. Lelong, Contróle optimal de systèmes gouvernés par des équations aux dérivées partielles, Dunod Gauthier-Villars, Paris, (1968). |
[21] |
A. Altmüller, L. Grüne, Distributed and boundary model predictive control for the heat equation, Technical report, University of Bayreuth, Department of Mathematics, (2012). https://doi.org/10.1002/gamm.201210010 doi: 10.1002/gamm.201210010
![]() |
[22] |
S. Dubljevic, P. D. Christofides, Predictive control of parabolic PDEs with boundary control actuation, Chem. Eng. Sci., 61 (2006), 6239–6248. https://doi.org/10.1016/j.ces.2006.05.041 doi: 10.1016/j.ces.2006.05.041
![]() |
[23] | A. N. Tikhonov, A. A. Samarsky, Equations of Mathematical Physics, Nauka, Moscow, (1966). |
[24] | F. G. Tricomi, Differential equations, Moscow, (1962). |
[25] | V. S. Vladimirov, Equations of Mathematical Physics, Marcel Dekker, New York, (1971). https://doi.org/10.1063/1.3022385 |
[26] |
V. E. Vladykina, Spectral characteristics of the Sturm-Liouville operator under minimal restrictions on smoothness of coefficients, Vestnik Moskov. Univ. Ser. 1. Mat. Mekh., 6 (2019), 23–28. https://doi.org/10.3103/S0027132219060044 doi: 10.3103/S0027132219060044
![]() |
G | Kf(G) | G | Kf(G) | G | Kf(G) |
L1 | 17.3 | L5 | 1486.7 | L9 | 8268.0 |
L2 | 110.7 | L6 | 2524.0 | L10 | 24457.3 |
L3 | 344.0 | L7 | 3957.3 | L11 | 30454.7 |
L4 | 781.3 | L8 | 5850.7 | L12 | 37360.0 |
G | Kf∗(G) | G | Kf∗(G) | G | Kf∗(G) |
L1 | 241.98 | L5 | 26659.15 | L9 | 151875.4 |
L2 | 1818.86 | L6 | 45675.81 | L10 | 207691.9 |
L3 | 5940.68 | L7 | 72077.4 | L11 | 275733.2 |
L4 | 13817.44 | L8 | 107073.9 | L12 | 357209.6 |
G | τ(G) | G | τ(G) |
W1 | 96 | W5 | 208971104256 |
W2 | 20736 | W6 | 45137758519296 |
W3 | 4478976 | W7 | 9749755840167936 |
W4 | 967458816 | W8 | 2105947261476274176 |
G | Kf(G) | G | Kf(G) | G | Kf(G) |
L1 | 17.3 | L5 | 1486.7 | L9 | 8268.0 |
L2 | 110.7 | L6 | 2524.0 | L10 | 24457.3 |
L3 | 344.0 | L7 | 3957.3 | L11 | 30454.7 |
L4 | 781.3 | L8 | 5850.7 | L12 | 37360.0 |
G | Kf∗(G) | G | Kf∗(G) | G | Kf∗(G) |
L1 | 241.98 | L5 | 26659.15 | L9 | 151875.4 |
L2 | 1818.86 | L6 | 45675.81 | L10 | 207691.9 |
L3 | 5940.68 | L7 | 72077.4 | L11 | 275733.2 |
L4 | 13817.44 | L8 | 107073.9 | L12 | 357209.6 |
G | τ(G) | G | τ(G) |
W1 | 96 | W5 | 208971104256 |
W2 | 20736 | W6 | 45137758519296 |
W3 | 4478976 | W7 | 9749755840167936 |
W4 | 967458816 | W8 | 2105947261476274176 |