
Quantum Markov chains (QMCs) on graphs and trees were investigated in connection with many important models arising from quantum statistical mechanics and quantum information. These quantum states generate many important properties such as quantum phase transition and clustering properties. In the present paper, we propose a construction of QMCs associated with an XX-Ising model over the comb graph N⊳0Z. Mainly, we prove that the QMC associated with the disordered phase, enjoys a clustering property.
Citation: Abdessatar Souissi, El Gheteb Soueidy, Mohamed Rhaima. Clustering property for quantum Markov chains on the comb graph[J]. AIMS Mathematics, 2023, 8(4): 7865-7880. doi: 10.3934/math.2023396
[1] | Luigi Accardi, Amenallah Andolsi, Farrukh Mukhamedov, Mohamed Rhaima, Abdessatar Souissi . Clustering quantum Markov chains on trees associated with open quantum random walks. AIMS Mathematics, 2023, 8(10): 23003-23015. doi: 10.3934/math.20231170 |
[2] | Luigi Accardi, El Gheted Soueidi, Abdessatar Souissi, Mohamed Rhaima, Farrukh Mukhamedov, Farzona Mukhamedova . Structure of backward quantum Markov chains. AIMS Mathematics, 2024, 9(10): 28044-28057. doi: 10.3934/math.20241360 |
[3] | Sakander Hayat, Hafiz Muhammad Afzal Siddiqui, Muhammad Imran, Hafiz Muhammad Ikhlaq, Jinde Cao . On the zero forcing number and propagation time of oriented graphs. AIMS Mathematics, 2021, 6(2): 1833-1850. doi: 10.3934/math.2021111 |
[4] | Bader Alshamary, Milica Anđelić, Edin Dolićanin, Zoran Stanić . Controllable multi-agent systems modeled by graphs with exactly one repeated degree. AIMS Mathematics, 2024, 9(9): 25689-25704. doi: 10.3934/math.20241255 |
[5] | Chaofeng Guan, Ruihu Li, Hao Song, Liangdong Lu, Husheng Li . Ternary quantum codes constructed from extremal self-dual codes and self-orthogonal codes. AIMS Mathematics, 2022, 7(4): 6516-6534. doi: 10.3934/math.2022363 |
[6] | Ridho Alfarisi, Liliek Susilowati, Dafik . Local multiset dimension of comb product of tree graphs. AIMS Mathematics, 2023, 8(4): 8349-8364. doi: 10.3934/math.2023421 |
[7] | Lin Xu, Linlin Wang, Hao Wang, Liming Zhang . Optimal investment game for two regulated players with regime switching. AIMS Mathematics, 2024, 9(12): 34674-34704. doi: 10.3934/math.20241651 |
[8] | Ahmed Ghezal, Mohamed balegh, Imane Zemmouri . Markov-switching threshold stochastic volatility models with regime changes. AIMS Mathematics, 2024, 9(2): 3895-3910. doi: 10.3934/math.2024192 |
[9] | Wanchun Fan, Yan Jiang, Songyang Huang, Weiguo Liu . Research and prediction of opioid crisis based on BP neural network and Markov chain. AIMS Mathematics, 2019, 4(5): 1357-1368. doi: 10.3934/math.2019.5.1357 |
[10] | Sergio Iglesias-Perez, Alberto Partida, Regino Criado . The advantages of k-visibility: A comparative analysis of several time series clustering algorithms. AIMS Mathematics, 2024, 9(12): 35551-35569. doi: 10.3934/math.20241687 |
Quantum Markov chains (QMCs) on graphs and trees were investigated in connection with many important models arising from quantum statistical mechanics and quantum information. These quantum states generate many important properties such as quantum phase transition and clustering properties. In the present paper, we propose a construction of QMCs associated with an XX-Ising model over the comb graph N⊳0Z. Mainly, we prove that the QMC associated with the disordered phase, enjoys a clustering property.
Over the past few decades, quantum Markov chains (QMCs) (see [1,2,7,8,18,21]) have undergone great development through the vast volume of relevant scientific literature in different research areas such as computational physics [42], interacting particle systems [10], quantum spin models [19,20,24], quantum information [15,41], quantum cryptography [21].
In view of the absence of a satisfactory theory of quantum Markovian fields on general graphs, it is quite natural to restrict the study to special class of graphs for which the considered problem is more solvable. Thanks to their hierarchical simplified structure consist a natural candidate. Quantum Markov chains on graphs [3,9,23,44] are multi-dimensional extensions of 1D QMCs. Namely, QMCs on the Cayley trees (CT) have been investigated in connection with quantum phase transitions for Pauli models [4,5,6,25,28,29,30]. This consists a quantum extension of an increasing number of works on classical Gibbs measures [31,32]. Types of von Neumann factors associated with QMCs based on Ising and XY type models have been investigated [27,33]. Moreover, clustering property have been showed for QMCs on the Cayley trees [26]. The structure of quantum Markov states (QMS) on CT have been studied in details [34,35,36,37]. However, an extension of Fermi Markov states [17] to trees is still missing.
Recently, the notions of stopping rules and recurrence for QMCs on trees have been introduced [45]. In [13], a bridge between recurrence and phase transition for QMCs on CT have been established. In [39,40,46] QMCs on CT were associated with open quantum random walks [11,14] extending the 1D case [16].
QMCs on Trees have been investigated in connection with interesting phenomena such as quantum phase transition [28], quantum walks [39] and clustering properties [38]. Namely, the clustering analysis [22] play a key role in several areas such as data science [47], image segmentation [43].
In this paper, we show a clustering property for QMCs associated with an XX-Ising model on the comb graph N⊳oZ. Unlike the CT the underlying comb graph is non-homogeneous. Therefore, we construct QMCs based on two different types of hamiltonian. We prove the uniqueness of QMCs associated with the considered model. Notice that the present work extends the clustering property for QMCs on CT [26] to non-homogeneous trees. Moreover, it generalizes a previous work [38] that deals with QMCs on the comb N⊳oN.
The paper is organized as follows: Section 2, is devoted to some preliminaries. In Section 3, we provide a construction of QMCs on the comb graph. In Section 4, we investigate QMCs associated with XX-Ising type model on the comb graph. Moreover, we prove the uniqueness of QMCs associated with the considered model. Section 5 is devoted to the main result which concerns the clustering property of the considered QMC.
Let G=(L,E) be a connected, locally finite, infinite graph, here L denotes the set of vertices and E denotes the vertex set. Each edge l∈E is identified to non ordered pair of vertices (its endpoints) l=<u,v>=<v,u> and E is then identifiable to a part of L×L.
E⊂{{u,v}:u,v∈L}. |
Let us recall some basic notions on graph theory:
(i) We call nearest-neighbors vertices u and v, and we denote by u∼v, if l=<u,v>∈E.
(ii) A path on the graph is a finite list of vertices u∼u1∼⋯∼ud−1∼v.
(iii) The distance d(u,v),u,v∈L, on the graph, is defined to be the length of the shortest path joining u to v.
Let G(1)=(L(1),E(1)) and G(2)=(L(2),E(2)) be two graphs where L(1)=N and L(2)=Z with distinguished vertex 0∈L(2). Let
L=N×Z, |
and
E={{(u,v),(u′,v′)},(u,v),(u′,v′)∈L,(u,v)∼(u′,v′)}, |
where (u,v)∼(u′,v′) if and only if one of the following assertions is satisfied:
(i) u∼u′ and v=v′=0.
(ii) u=u′ and v∼v′.
The graph G=(L,E) is the comb product of G(1) and G(2) with a distinguished vertex o∈Z denoted by N⊳0Z.
Let us fix a root o=(0,0). Define the sets
Wm={u∈L:d(u,o)=m},Λm=m⋃k=0Wk. | (2.1) |
For u∈L, define the set of its direct successors
S(u)={v∈Λcm:u∼v}. | (2.2) |
One can see that elements of Wm are on the form
u=(k,l),k+|l|=m, |
where k={0,…,m}andl={−m,…,m}. It follows the enumeration
u(−m)Wm=(0,−m),…,u(0)Wm=(m,0),…,u(m)Wm=(0,m). | (2.3) |
The graph under consideration (see Figure 1) is a tree. There are two types of vertices according to the number of nearest-neighbors (or also the number of direct successors). We distinguish vertices with three direct successors and others with only one.
Define
L1={u1∈L:|S(u1)|=1}, | (2.4) |
and
L3={u3∈L:|S(u3)|=3}. | (2.5) |
It is clear that L=L1∪L3. Each element u3∈L3 has the form u3=(k,0) with k∈N. Its set of direct successors is
S(u3)={u3+e1,u3±e2}, | (2.6) |
whereas, elements of L1 have the form v1=(k,l) where l≥1 and
S(v1)={{v1+e2},ifl>0,{v1−e2},ifl<0, | (2.7) |
where e1=(1,0) and e2=(0,1).
Let us define the restriction of the usual addition of the commutative group Z2 on comb+ as follows: for any two elements u=(k,l) and v=(k′,l′)
u∘v=(k+k′,l+l′). | (2.8) |
For these notations, one has
u∘u(0)=u(0)∘u=u. |
Operation (2.8) induces on comb+ a semi-group of translation with unit u(0). With the above structure of semi-group, we define the translations τg:comb+→comb+, g∈L3 as follows
τg(u)=g+u, | (2.9) |
and τ(0)=id. Let
Λm={u∈L:d(u,u(0))=m}. |
Thanks to the tree structure, one has
Λm+1=⋃u∈ΛmS(u)andS(u)∩S(v)=∅,∀u≠v. | (2.10) |
Let C⊆B⊆A be three unitary C∗–algebras. A completely positive identity preserving linear map E: A→B satisfies
E(ca)=cE(a),a∈A,c∈C, | (3.1) |
is called Quasi-conditional expectation (QCE).
Remark 3.1. Any K∈A, satisfies
E0(K∗K)=I | (3.2) |
is called E0-conditional amplitude. where E0: A→B is a conditional expectation [1]. For any B⊆A sub–∗–algebra, denote by
B′:={x∈A : xy=yx , ∀y∈B}, |
the commutant of B in A, If K∈C′ then E0(K∗( ⋅)K):A→B is a (normalized) QCE w.r.t. the triplet C⊆B⊆A.
Remark 3.2. Every NQCE w.r.t. C⊆B⊆A satisfies
E(ac)=E(a)c,a∈A,c∈C, | (3.3) |
E(C′∩A)⊆C′∩B. | (3.4) |
To each vertex u∈V an algebra of observable Bu≡B(Hu) is associated, where Hu is a finite dimensional Hilbert space. Consider the quasi-local algebra
BL:=⨂u∈LBu, |
which is obtained as inductive limit of the net
BΛ:=⨂u∈ΛBu⊗IΛc,Λ⊂L,|Λ|<∞. |
where for each Λ′⊂L, we denote IΛ′ the identity of BΛ′. See [12] for a systematic study of quasi-local algebras.
Remark 3.3. Starting from any QCE EΛm]: BΛm+1]→BΛm] with respect to BΛn−1]⊆BΛn]⊆BΛn+1] one can derive a transition expectation (TE) from BΛ[n,n+1] into BΛn by the next restriction
EΛ[m,m+1]:BΛm:=EΛm]|B[m,m+1]. |
Conversely, every TE EΛ[m,m+1]: BΛ[m,m+1]→BΛm is extendable to a QCE EΛn] w.r.t. the given triplet in the following way
EΛm]:=idBm−1]⊗EΛ[m,m+1]. | (3.5) |
The reader is referred to [7] for further details about the extendability of transition expectations in a generalized framework, including both the tensor and the Fermi cases.
Definition 3.4. A backward quantum Markov chain (b-QMC) on the algebra BL is a triplet (ρ0,(E[m,m+1],),(hm)) of positive linear functional ρ0 on B(u0), a sequence of TE E[m,m+1]:BΛ[m,m+1]→BΛm and a sequence hm∈B+Λm such that for each a∈BL, the limit
φ(a):=limm→+∞ρ0(EΛ0](EΛ1](⋯(EΛm](a⊗hm+1))⋯))) | (3.6) |
exists for the weak-∗-topology and it defines a state φ on the full algebra BL. In this case the limit state φ is also called QMC. The sequence (hm) is called sequence of boundary conditions of the QMC.
Thanks to (3.5), one can immediately check that, φ evaluated on the element a=a0⊗a1⊗⋯⊗am∈BΛm], aj∈BΛj, is provited by the correlations
φ(a)=ρ0(EΛ[0,1](aΛ0⊗EΛ[1,2](a1⊗⋯⊗EΛ[m,m+1](am⊗hm+1)⋯))), |
which highlight the quantum Markov structure.
We are going to construct a state on BΛm] with initial state ω0∈B(0),+ and boundary conditions {hu∈Bu,+: u∈L}.
For every n∈N, denote
Ku∨S(u):=∏v∈S(u)K<u,v>, | (3.7) |
K[m,m+1]:=∏u∈→ΛmKu∪S(u), 1≤m≤n, | (3.8) |
h1/2n:=∏u∈→Λnh1/2u, hn=h1/2n(h1/2n)∗, | (3.9) |
Kn:=ω1/20n−1∏m=1K[m,m+1]h1/2n, | (3.10) |
Wn]:=K∗nKn. | (3.11) |
One can see that Wn] is positives.
In the sequel, Trn]: BL→BΛn denotes the (normalized) partial trace i.e. Trn](IΛn)=IΛn, here IΛn=⨂u∈ΛnIu), for any finite part Λn.
Let's set a positive functional φ(n,b)w0,h on BΛn by
φ(n,b)w0,h(a)=Tr(Wn+1](a⊗IWn+1)), | (3.12) |
for each a∈BΛn. Note that, the trace Tr is normalized (i.e. Tr(Iu)=1).
To obtain a state φ(b) on BL satisfying
φ(b)⌈BΛn=φ(n,b)w0,h, |
we must impose some constrains on the boundary conditions {w0,h} so that the positive functionals {φ(n,b)w0,h} satisfy the following compatibility condition, i.e.
φ(n+1,b)w0,h⌈BΛn=φ(n,b)w0,h. | (3.13) |
Theorem 3.5. Let w0∈B(0),+ and h={hu∈Bu,+}u∈L. If
Tr(ω0h0)=1, | (3.14) |
Tru](K∗u∨S(u)I(u)⊗hS(u)Ku∨S(u))=h(u),∀u∈L. | (3.15) |
Then the sequebce {φ(m,b)w0,h} satisfy condition (3.13). Moreover, there exists a unique b-QMC φ(b)w0,h on BL such that
φ(b)w0,h=w−limm→∞φ(m,b)w0,h. |
Remark 3.6. Theorem 3.5 extends results of [25,26,28,29,34] where only considered Bethe lattice or Cayley tree. The first attempt to investigate QMCs on the comb graph was done in [38] by considering N⊳0N.
In this section, we define the model and study the b-QMC φ associated to the XX-Ising model on the Comb graph N⊳Z. Let Bu=M2(C), for all u∈L. The Pauli spin operators σx, σy, σz are given by
I=(1001),σx=(0110),σy=(0−ii0),σz=(100−1). |
The shift of an element a∈M2(C) to the uth component of the infinite tensor product BL=⨂x∈LBx will be denoted by
a(u):=τu(a). |
Define the nearest neighbors interactions: for each u1∈L1, v∈S(u1),
K<u1,v>:=cos(β)I(u1)⊗IS(u1)−isin(β)σ(u1)x⊗σS(u1)x, β>0, | (4.1) |
and for u3∈L3,w∈S(u3),
K<u3,w>=exp{βH<u3,w>},β>0, | (4.2) |
where
H<u3,w>=12(I(u3)⊗I(w)+σ(u3)z⊗σ(w)z). | (4.3) |
A simple calculation leads to
K<u3,w>=K0I(u3)⊗I(w)+K3σ(u3)z⊗σ(w)z, |
where
K0=exp(J0β)+12, K3=exp(J0β)−12, J0>0. |
One finds: for u1∈L1 and v∈S(u1)
Ku1∨S(u1)=K<u1,v>=cos(β)I(u1)⊗I(v)−isin(β)σ(u1)x⊗σ(v)x, β>0. | (4.4) |
And for v∈L3 (its successors S(v)={v+e1,v±e2}) one finds,
Kv∨S(v)=K<v,v+e1>K<v,v+e2>K<v,v−e2>=K30I(v)⊗I(v+e1)⊗I(v+e2)⊗I(v−e2)+K20K3σ(v)z⊗I(v+e1)⊗I(v+e2)⊗σ(v−e2)z+K20K3σ(v)z⊗I(v+e1)⊗σ(v+e2)z⊗I(v−e2)+K0K23I(v)⊗I(v+e1)⊗σ(v+e2)z⊗σ(v−e2)z+K3K20σ(v)z⊗σ(v+e1)z⊗I(v+e2)⊗I(v−e2)+K0K23I(v)⊗σ(v+e1)z⊗I(v+e2)⊗σ(v−e2)z+K0K23I(v)⊗σ(v+e1)z⊗σ(v+e2)z⊗I(v−e2)+K33σ(v)z⊗σ(v+e1)z⊗σ(v+e2)z⊗σ(v−e2)z. |
Recall that, a net {hu} is translation-invariant if
h(u)=hτg(u),∀u,g∈L. |
This means that
h(u)=h(v),∀u,v∈L. | (4.5) |
In what follows, we consider only translation-invariant solutions of (3.14), (3.15). Put h(u)=h for all u∈L, where
h=(h11h12h21h22). |
Theorem 4.1. For the XX-Ising model (4.1), (4.2) there exists a unique b-QMC φα with translation-invariant boundary condition hα satisfying (3.14). Moreover, for each a∈BΛn] one has
φ(b)α(a)=α2n+1Tr(n∏j=0∏u∈ΛjK∗{u}∨S(u)an∏j=0∏u∈ΛjK{u}∨S(u)). | (4.6) |
Proof. Let u1∈L1 and v its unique successor (S(u1)={v}), then (3.15) is reduced to
h(u1)=Tru1](K∗u1∨S(u1)I(u1)⊗h(v)Ku1∨S(u1))=Tru1](cos2(β)I(u1)⊗h(v)+sin2(β)I(u1)⊗σxh(v)σx+isinβcos(β)(σ(u1)x⊗h(v)σx−σ(u1)x⊗σxh(v))). |
One can check that
Tr(σxh(v)σx)=Tr(h(v))andTr(h(v)σx)=Tr(σxh(v)). |
Then, we find that
h(u1)=Tr(h(v))I(u1). | (4.7) |
Now for u3∈L3 according to the above computation (3.15) becomes
h(u3)=Tru3](K∗{u3∨S(u3)}I(u3)⊗hS(u3)K{u3∨S(u3)}). |
Since the boundary condition satisfy (4.5), according to (4.7) we have
h(u3)=h(u1)=h=Tr(h(v))I=αI, |
for some α>0, then (3.15) is reduced to
h=Tru3](K∗{u3∨S(u3)}I(u3)⊗hS(u3)K{u3∨S(u3)})=α3Tru3](K∗{u3∨S(u3)}K{u3∨S(u3)})=α3(K20+K23)3I. |
Therefore, α=α3(K20+K23)3 and this is equivalent to
α=1(K20+K23)3/2=23/2(e2J0β+1)3/2. |
Hence,
h=αI | (4.8) |
is the unique commune solution of (3.15). The initial state ω0 can be chosen ω0=1αI.
From (3.12) one has
φ(b)α(a)=Tr(Wn]a)=Tr(ω0K∗n]ahnKn])=α|Λn|−1Tr(n∏j=0∏u∈ΛjK∗{u}∨S(u)an∏j=0∏u∈ΛjK{u}∨S(u)). |
Since |Λn|=2n+1 one gets (4.6). This finishes the proof.
Remark 4.2. The QMC φα given in Theorem 4.1 is the state associated with the disordered phase of the underlying quantum. This state always exists for some fixed point reasons. Therefore, the existence of phase transition requires a t least one additional state satisfying conditions of under some conditions (see [28,39]. Notice that, if the boundary condition is non-homogeneous, phenomena of phase transitions may appear even for the above considered model.
A state φ on BL is said to enjoy the clustering property if for every a,b∈BL one has
lim|g|→∞φ(aτg(b))=φ(a)φ(b). | (5.1) |
From Theorem 4.1, there is a unique b-QMC φ(b)α with translation-invariant boundary condition hα satisfying (3.14). Now, let establish clustering property for this B-QMC φ(b)α.
First, Let denote for each n∈N∗
Λn={u(−n)Λn,⋯,u(−1)Λn,u(0)Λn,u(1)Λn,⋯,u(n)Λn}:φ(b)α⌈BΛn=:φ(n,b)α, |
such that u(0)Λn∈L3 and u(i)Λn∈L1, i=±1,±2,⋯±n.
Moreover, to prove Theorem 5.2, we need the following:
Lemma 5.1. Let um0∈BΛm0, for a certain integer m0, and fn∈BΛn of the form
fn=f⊗IΛn∖{uΛn(0)}, |
where f=f(uΛn(0)). Then one has
limn→∞φ(b)α(um0⊗fn)=φ(b)α(um0)φ(b)α(f). | (5.2) |
Proof. For n≥m0, one has
φ(b)α(um0⊗f)=φ(b,n)α(um0⊗f)=Tr(ω0E0∘E1∘...∘EM0(um0⊗Em0+1(IΛm0+1⊗⋯⊗En−1(IΛn−1⊗ˆEn(f⊗IΛn+1))⋯)), |
here, as before, {w0=1α,h0=αI} is the fixed point of the system with α=23/2(e2J0β+1)3/2. One finds
ˆEn(f⊗IΛn+1)=Trn](K[n,n+1]h1/2n+1f⊗IΛn+1h1/2n+1K∗[n,n+1])=Trn](⨂u∈ΛnKu∨S(u)h1/2n+1f⊗IΛn+1h1/2n+1⨂u∈ΛnK∗u∨S(u))=TruΛn(0)](KuΛn(0)∨S(uΛn(0))f⊗h(u)⊗h(u)⊗h(u)K∗uΛn(0)∨S(uΛn(0)))⊗⨂u∈Λn∖{uΛn(0)}Tru](Ku∨S(u)I⊗h(u)K∗u∨S(u))=α3((K60+3K20K43)f+(K63+3K40K23)σzfσz)⊗⨂x∈Λn∖{xΛn(0)}h(x)=α3((K60+3K20K43)f+(K63+3K40K23)σzfσz)⊗⨂u∈Λn∖{uΛn(0)}h(u)=α3guΛn(0)⊗⨂u∈Λn∖{uΛn(0)}h(u), |
where,
guΛn(0)=(K60+3K20K43)f+(K63+3K40K23)σzfσz. |
Hence,
En−1(IΛn−1⊗ˆEn(f⊗IΛn+1))=α3TruΛn−1(0)](KuΛn−1(0)∨S(uΛn−1(0))IuΛn−1(0)⊗guΛn(0)⊗h(u)⊗h(u)K∗uΛn−1(0)∨S(uΛn(0)))⊗⨂u∈Λn−1∖{uΛn−1(0)}Tru](Ku∨S(u)I⊗h(u)K∗u∨S(u))=α3(Tr(g)I+2α2/3K0K3Tr(σzg)σz)⊗⨂u∈Λn−1∖{uΛn−1(0)}h(u)=α3Tr(g)IuΛn−1(0)⊗⨂u∈Λn−1∖{uΛn−1(0)}h(u)+2α3α2/3K0K3Tr(σzg)σ(uΛn−1(0))z⊗⨂u∈Λn−1∖{uΛn−1(0)}h(u). |
Therefore,
En−2(IΛn−2⊗En−1(IΛn−1⊗ˆEn(f⊗IΛn+1)))=α3Tr(g)IuΛn−2(0)⊗⨂u∈Λn−1∖{uΛn−2(0)}h(u)+22α3α2/3+2/3K20K23Tr(σzg)σ(uΛn−2(0))z⊗⨂u∈Λn−1∖{uΛn−2(0)}h(u). |
Now iterating n−m0−1 times, we find
Em0+1(Em0+2⋯En−1(IΛn−1⊗ˆEn(f⊗IΛn+1))⋯))=α32n−m0−1Tr(gσz)α2(n−m0−1)3Kn−m0−10Kn−m0−13σ(u(0)Λm0+1)z⊗⨂u∈Λn−1∖{u(0)Λm0+1}h(u)+Tr(g)α3Iu(0)Λm0+1⊗⨂u∈Λn−1∖{u(0)Λm0+1}h(u). |
Hence, one get
φ(b)α(um0⊗f)=Tr(ω0E0∘E1∘⋯∘Em0(um0⊗I))Tr(g)α2α|Λm0+1|+Tr(ω0E0∘E1∘⋯∘Em0(um0⊗σ(x(1)Λm0+1)z))Tr(gσz)α2α2(n−m0−1)3α|Λm0+1|2n−m0−1Kn−m00Kn−m0−13=Tr(ω0E0∘E1∘⋯∘ˆEm0(um0⊗I))Tr(f)+Tr(ω0E0∘E1∘⋯∘Em0(um0⊗σ(x(1)Λm0+1)z))Tr(gσz)α(2+2(n−m0−1)3+|Λm0+1|)2n−m0−1Kn−m00Kn−m0−13. |
One can see that,
2n(K0K3)nα2n3=(e2J0β−1e2J0β+1)n. |
Therefore, by taking the limit n→∞, we obtain,
limn→∞φ(b)α(um0⊗fn)=Tr(ω0E0∘E1∘⋯∘ˆEm0(am0⊗I))Tr(f)=φ(b)α(um0)φ(b)α(f). |
Thus, this completes the proof.
Now we are ready to satate the main result of this paper.
Theorem 5.2. Let φ(b)α be the b-QMC associated with the XX-Ising model on the comb graph N⊳0Z. Then for each g∈G+
lim|g|→+∞φ(b)α(aτg(f))=φ(b)α(a)φ(b)α(f), | (5.3) |
for all a,f∈BL.
Proof. Let a,f∈BL,loc, then a,f∈BΛ[0,l0] for a certain integer l0.
Then, let denote
ul0:=a∈BΛ[0,l0]andfl0:=f∈BΛ[0,l0]. |
fl0 can be rewritten in the following form
fl0=⨂u∈Λ[0,l0]fu=l0⨂k=0fΛk,withfΛk=⨂u∈Λkfu∈BΛk. |
Furthermore, one can see that
τgm(fl0)=⨂u∈Λ[0,l0]f(u+me1)u=⨂v∈Λ[m,m+l0]˜fv∈BΛ[m,m+l0], |
where
˜fv={fv−me1,if v−me1∈Λ[0,l0],1,otherwise. |
For k∈[0,l0], g∈G+ and b∈BΛ[k,k+1] we denote
E(τg)[k,k+1](τg(b)):=⨂v∈τg(Λk)Trv](K{v}∨S(v)τg(b)K∗{v}∨S(v)), | (5.4) |
the τg-shift of the transition expectation E[k,k+1], in fact one can check that
E(τg)[k,k+1](τg(bl0))=τg(E[k,k+1](bl0)). |
In light of (5.4), one finds
ˆE[m+l0,m+l0+1](˜fΛm+l0)))=(⨂v∈τgm(Λl0)Trv](K{v}∨S(v)fvhS(v)K∗{v}∨S(v)))⊗(⨂v∈Λm+l0∖τgm(Λl0)hv)=α|Λl0+1|E(τgm)[l0,l0+1](τgm(fΛl0))⊗(⨂v∈Λm+l0∖τgn(Λm0)hv). |
The comb graph Z⊳0N satisfies
τgm(Λk+1)=⋃v∈τgm(Λk)S(v). |
Therefore,
E[m+l0−1,m+l0](˜fΛm+l0−1⊗ˆE[m+l0,m+l0+1](˜fΛm+l0))))=α|Λl0+1|⨂u∈τgm(Λl0−1)Tru](K{u}∨S(u)(˜fu⊗E(τgm)[l0−1,l0](τgm(fΛl0)))K{u}∨S(u))⊗⨂w∈Λm+l0−1∖τgm(Λl0−1)hw,=α|Λl0+1|E(τgm)[l0−1,l0](τgm(fΛl0−1)⊗E(τgm)[l0−1,l0](τgm(fΛl0)))⊗⨂w∈Λm+l0−1∖τgm(Λl0−1)hw. |
An iterative process leads to
E[m,m+1](˜fΛm⊗⋯E[m+l0−1,m+l0](˜fΛm+l0−1⊗ˆE[m+l0,m+l0](˜fm+l0)))=α|Λl0+1|E(τgm)[0,1](τgm(fΛ0)⊗⋯E(τgm)[l0−1,l0](τgm(fΛl0−1)⊗E(τgm)[l0−1,l0](τgm(fΛl0))))⊗⨂w∈Λm∖τgm(Λ0)hw. |
Let denote
ˆfo:=E[0,1](fΛ0⊗⋯E[l0−1,N0](fΛl0−1⊗E[l0−1,l0](fΛl0⊗hl0+1)))∈Bo. |
Since τgm(Λ0)={u(0)Λm} and hw=αI, for each w∈L, one gets
E[m,m+1](˜fΛm⊗⋯E[m+N0−1,m+N0](˜fΛm+N0−1⊗ˆE[m+N0,m+N0](˜fm+N0)))=ˆf(u(0)Λm)o⊗⨂w∈Λm∖{u(0)Λm}hw. |
This leads to
φ(b)α(ul0⊗τgm(fl0)=ρ0(E[0,1](uΛ0⋯E[l0,l0+1](uΛl0⊗E[l0,l0+1](IΛl0+1⊗⋯E[m,m+1](˜fΛm⊗⋯E[m+l0−1,m+l0](˜fΛm+l0−1⊗ˆE[m+l0,m+l0](˜fm+l0)))))))=ρ0(E[0,1](uΛ0⋯E[l0,l0+1](uΛl0⊗E[l0,l0+1](IΛl0+1⊗⋯E[m,m+1](IΛm−1⊗(ˆf(u(0)Λm)o⊗hm+1))))). |
Therefore, Lemma 5.1 implies that
limm→+∞φ(b)α(aτgm(f))=limm→+∞φ(b)α(ul0τgm(fl0))=φ(b)α(ul0)φ(b)α(ˆfo)=φ(b)α(a)φ(b)α(f), |
and this concludes the proof.
We investigate an XX-Ising model on the comb graph N⊳0Z. Namely, we show the uniqueness of QMC with homogeneous boundary condition associated with the model. Indeed, the considered quantum Markov chain is the one associated with the disordered phase of the system. Our main result concerns a clustering property for this QMC. Notice that, further relevant open problems can be investigated such as the recurrence problem for QMCs on the comb graph, the existence of phase transitions and the QMCs associated with open quantum random walks on the comb graph.
The authors extend their appreciation to the Deputyship for Research & Innovation, Ministry of Education, Saudi Arabia for funding this research work through the project number (QU-IF-04-01-27970). The authors also thank to Qassim University for technical support.
All authors declare no conflicts of interest in this paper.
[1] | L. Accardi, Noncommutative Markov chains, Proc. Int. Sch. Math. Phys., 1974,268–295. |
[2] | L. Accardi, A. Frigerio, Markovian cocycles, Proc. R. Ir. Acad., 83 (1983), 251–263. |
[3] |
L. Accardi, F. Mukhamedov, A. Souissi, Construction of a new class of quantum Markov fields, Adv. Oper. Theory, 1 (2016), 206–218. https://doi.org/10.22034/aot.1610.1031 doi: 10.22034/aot.1610.1031
![]() |
[4] |
L. Accardi, F. Mukhamedov, M. Saburov, On quantum Markov chains on Cayley tree Ⅰ: niqueness of the associated chain with XY-model on the Cayley tree of order two, Inf. Dimens. Anal., 14 (2011), 443–463. https://doi.org/10.1142/S021902571100447X doi: 10.1142/S021902571100447X
![]() |
[5] |
L. Accardi, F. Mukhamedov, M. Saburov, On quantum Markov chains on Cayley tree Ⅱ: phase transitions for the associated chain with XY-model on the Cayley tree of order three, Ann. Henri Poincaré, 12 (2011), 1109–1144. https://doi.org/10.1007/s00023-011-0107-2 doi: 10.1007/s00023-011-0107-2
![]() |
[6] |
L. Accardi, F. Mukhamedov, M. Saburov, On quantum Markov chains on Cayley tree Ⅲ: Ising model, J. Stat. Phys., 157 (2014), 303–329. https://doi.org/10.1007/s10955-014-1083-y doi: 10.1007/s10955-014-1083-y
![]() |
[7] |
L. Accardi, A. Souissi, E. G. Soueidy, Quantum Markov chains: a unification approach, Inf. Dimens. Anal., 23 (2020), 2050016. https://doi.org/10.1142/S0219025720500162 doi: 10.1142/S0219025720500162
![]() |
[8] |
L. Accardi, Y. G. Lu, A. Souissi, A Markov–Dobrushin inequality for quantum channels, Open Sys. Inf. Dyn., 28 (2021), 2150018. https://doi.org/10.1142/S1230161221500189 doi: 10.1142/S1230161221500189
![]() |
[9] |
L. Accardi, H. Ohno, F. Mukhamedov, Quantum Markov fields on graphs, Inf. Dimens. Anal., 13 (2010), 165–189. https://doi.org/10.1142/S0219025710004000 doi: 10.1142/S0219025710004000
![]() |
[10] | L. Accardi, F. Fagnola, Quantum interacting particle systems, World Scientific, 2002. |
[11] |
S. Attal, F. Petruccione, C. Sabot, I. Sinayskiy, Open quantum random walks, J. Stat. Phys., 147 (2012), 832–852. https://doi.org/10.1007/s10955-012-0491-0 doi: 10.1007/s10955-012-0491-0
![]() |
[12] | O. Bratteli, D. W. Robinson, Operator algebras and quantum statistical mechanics I, Springer Verlag, 1987. |
[13] |
A. Barhoumi, A. Souissi, Recurrence of a class of quantum Markov chains on trees, Chaos Solitons Fract., 164 (2022), 112644. https://doi.org/10.1016/j.chaos.2022.112644 doi: 10.1016/j.chaos.2022.112644
![]() |
[14] | R. Carbone, Y. Pautrat, Open quantum random walks: reducibility, period, ergodic properties, Ann. Henri Poincaré, 17 (2016), 99–135. |
[15] |
J. I. Cirac, D. Perez-Garcia, N. Schuch, F. Verstraete, Matrix product unitaries, structure, symmetries, and topological invariants, J. Stat. Mech. Theory Exp., 2017 (2017), 083105. https://doi.org/10.1088/1742-5468/aa7e55 doi: 10.1088/1742-5468/aa7e55
![]() |
[16] |
A. Dhahri, F. Mukhamedov, Open quantum random walks, quantum Markov chains and recurrence, Rev. Math. Phys., 31 (2019), 1950020. https://doi.org/10.1142/S0129055X1950020X doi: 10.1142/S0129055X1950020X
![]() |
[17] | F. Fidaleo, Fermi Markov states, J. Oper. Theory, 66 (2011), 385–414. |
[18] |
F. Fidaleo, F. Mukhamedov, Diagonalizability of non homogeneous quantum Markov states and associated von Neumann algebras, arXiv, 24 (2004), 401–418. https://doi.org/10.48550/arXiv.math/0411200 doi: 10.48550/arXiv.math/0411200
![]() |
[19] |
M. Fannes, B. Nachtergaele, R. F. Werner, Finitely correlated states on quantum spin chains, Commun. Math. Phys., 144 (1992), 443–490. https://doi.org/10.1007/BF02099178 doi: 10.1007/BF02099178
![]() |
[20] |
M. Fannes, B. Nachtergaele, R. F. Werner, Ground states of VBS models on Cayley trees, J. Stat. Phys., 66 (1992), 939–973. https://doi.org/10.1007/BF01055710 doi: 10.1007/BF01055710
![]() |
[21] |
Y. Feng, N. Yu, M. Ying, Model checking quantum Markov chains, J. Comput. Sys. Sci., 79 (2013), 1181–1198. https://doi.org/10.1016/j.jcss.2013.04.002 doi: 10.1016/j.jcss.2013.04.002
![]() |
[22] |
J. A. Hartigan, Statistical theory in clustering, J. Classif., 2 (1985), 63–76. https://doi.org/10.1007/BF01908064 doi: 10.1007/BF01908064
![]() |
[23] |
V. Liebscher, Markovianity of quantum random fields, Quantum Probab. White Noise Anal., 15 (2003), 151–159. https://doi.org/10.1142/9789812704290-0011 doi: 10.1142/9789812704290-0011
![]() |
[24] |
A. Mohari, Spontaneous SU2(C) symmetry breaking in the ground states of quantum spin chain, J. Math. Phys., 59 (2018), 111701. https://doi.org/10.1063/1.5078597 doi: 10.1063/1.5078597
![]() |
[25] |
F. Mukhamedov, S. El Gheteb, Uniqueness of quantum Markov chain associated with XY-Ising model on the Cayley tree of order two, Open Syst. Inf. Dyn., 24 (2017), 175010. https://doi.org/10.1142/S123016121750010X doi: 10.1142/S123016121750010X
![]() |
[26] |
F. Mukhamedov, S. El Gheteb, Clustering property of quantum Markov chain associated to XY-model with competing Ising interactions on the Cayley tree of order two, Math. Phys. Anal. Geom., 22 (2019), 10. https://doi.org/10.1007/s11040-019-9308-6 doi: 10.1007/s11040-019-9308-6
![]() |
[27] |
F. Mukhamedov, S. El Gheteb, Factors generated by XY-model with competing Ising interactions on the Cayley tree, Ann. Henri Poincaré, 21 (2020), 241–253. https://doi.org/10.1007/s00023-019-00853-9 doi: 10.1007/s00023-019-00853-9
![]() |
[28] |
F. Mukhamedov, A. Barhoumi, A. Souissi, Phase transitions for quantum Markov chains associated with Ising type models on a Cayley tree, J. Stat. Phys., 163 (2016), 544–567. https://doi.org/10.1007/s10955-016-1495-y doi: 10.1007/s10955-016-1495-y
![]() |
[29] |
F. Mukhamedov, A. Barhoumi, A. Souissi, On an algebraic property of the disordered phase of the Ising model with competing interactions on a Cayley tree, Math. Phys. Anal. Geom., 19 (2016), 21. https://doi.org/10.1007/s11040-016-9225-x doi: 10.1007/s11040-016-9225-x
![]() |
[30] |
F. Mukhamedov, A. Barhoumi, A. Souissi, S. El Gheteb, A quantum Markov chain approach to phase transitions for quantum Ising model with competing XY-interactions on a Cayley tree, J. Math. Phys., 61 (2020), 093505. https://doi.org/10.1063/5.0004889 doi: 10.1063/5.0004889
![]() |
[31] |
F. Mukhamedov, U. Rozikov, On Gibbs measures of models with competing ternary and binary interactions on a Cayley tree and corresponding von Neumann algebras, J. Stat. Phys., 114 (2004), 825–848. https://doi.org/10.1023/B:JOSS.0000012509.10642.83 doi: 10.1023/B:JOSS.0000012509.10642.83
![]() |
[32] |
F. Mukhamedov, U. A. Rozikov, On Gibbs measures of models with competing ternary and binary interactions on a Cayley tree and corresponding von Neumann algebras Ⅱ, J. Stat. Phys., 119 (2005), 427–446. https://doi.org/10.1007/s10955-004-2056-3 doi: 10.1007/s10955-004-2056-3
![]() |
[33] |
F. Mukhamedov, A. Souissi, Types of factors generated by quantum Markov states of Ising model with competing interactions on the Cayley tree, Inf. Dimens. Anal., 23 (2020), 2050019. https://doi.org/10.1142/S0219025720500198 doi: 10.1142/S0219025720500198
![]() |
[34] |
F. Mukhamedov, A. Souissi, Quantum Markov states on Cayley trees, J. Math. Anal. Appl., 473 (2019), 313–333. https://doi.org/10.1016/j.jmaa.2018.12.050 doi: 10.1016/j.jmaa.2018.12.050
![]() |
[35] |
F. Mukhamedov, A. Souissi, Diagonalizability of quantum Markov states on trees, J. Stat. Phys., 182, (2021), 9. https://doi.org/10.1007/s10955-020-02674-1 doi: 10.1007/s10955-020-02674-1
![]() |
[36] |
F. Mukhamedov, A. Souissi, Refinement of quantum Markov states on trees, J. Stat. Mech., 2021 (2021), 083103. https://doi.org/10.1088/1742-5468/ac150b doi: 10.1088/1742-5468/ac150b
![]() |
[37] |
F. Mukhamedov, A. Souissi, Entropy for quantum Markov states on trees, J. Stat. Mech., 2022 (2022), 093101. https://doi.org/10.1088/1742-5468/ac8740 doi: 10.1088/1742-5468/ac8740
![]() |
[38] |
F. Mukhamedov, A. Souissi, T. Hamdi, Quantum Markov chains on comb graphs: Ising model, Proc. Steklov Inst. Math., 313 (2021), 178–192. https://doi.org/10.1134/S0081543821020176 doi: 10.1134/S0081543821020176
![]() |
[39] |
F. Mukhamedov, A. Souissi, T. Hamdi, Open quantum random walks and quantum Markov chains on trees Ⅰ: phase transitions, Open Syst. Inf. Dyn., 29 (2022), 2250003. https://doi.org/10.1142/S1230161222500032 doi: 10.1142/S1230161222500032
![]() |
[40] |
F. Mukhamedov, A. Souissi, T. Hamdi, A. A. Andolsi, Open quantum random walks and quantum Markov chains on trees Ⅱ: the recurrence, arXiv, 2022. https://doi.org/10.48550/arXiv.2208.04320 doi: 10.48550/arXiv.2208.04320
![]() |
[41] |
R. Orús, A practical introduction of tensor networks: matrix product states and projected entangled pair states, Ann. Phys., 349 (2014), 117–158. https://doi.org/10.1016/j.aop.2014.06.013 doi: 10.1016/j.aop.2014.06.013
![]() |
[42] |
S. Rommer, S. Ostlund, A class of ansatz wave functions for 1D spin systems and their relation to DMRG, Phys. Rev., 55 (1997), 2164. https://doi.org/10.1103/PhysRevB.55.2164 doi: 10.1103/PhysRevB.55.2164
![]() |
[43] |
P. Singh, S. S. Bose, A quantum-clustering optimization method for COVID-19 CT scan image segmentation, Expert Syst. Appl., 185 (2021), 115637. https://doi.org/10.1016/j.eswa.2021.115637 doi: 10.1016/j.eswa.2021.115637
![]() |
[44] |
A. Souissi, A class of quantum Markov fields on tree-like graphs: Ising-type model on a Husimi tree, Open Syst. Inf. Dyn., 28 (2021), 2150004. https://doi.org/10.1142/S1230161221500049 doi: 10.1142/S1230161221500049
![]() |
[45] |
A. Souissi, On stopping rules for tree-indexed quantum Markov chains, Inf. Dim. Anal., (2022). https://doi.org/10.1142/S0219025722500308 doi: 10.1142/S0219025722500308
![]() |
[46] |
A. Souissi, M. Mukhamedov, A. Barhoumi, Tree-homogeneous quantum Markov chains, Int. J. Theor. Phys., 62 (2023), 19. https://doi.org/10.1007/s10773-023-05276-1 doi: 10.1007/s10773-023-05276-1
![]() |
[47] |
O. R. Zaïane, A. Foss, C. H. Lee, W. Wang, On data clustering analysis: scalability, constraints, and validation, Adv. Knowl. Discovery Data Min., 28 (2022), 2030. https://doi.org/10.1007/3-540-47887-6-4 doi: 10.1007/3-540-47887-6-4
![]() |
1. | Abdessatar Souissi, El Gheteb Soueidy, Abdessatar Barhoumi, On a ψ-Mixing property for Entangled Markov Chains, 2023, 613, 03784371, 128533, 10.1016/j.physa.2023.128533 | |
2. | Dewang Li, Meilan Qiu, Shuiping Yang, Chao Wang, Zhongliang Luo, An optimal fractional-order accumulative Grey Markov model with variable parameters and its application in total energy consumption, 2023, 8, 2473-6988, 26425, 10.3934/math.20231349 | |
3. | Dewang Li, Meilan Qiu, Zhongliang Luo, Huizhou resident population, Guangdong resident population and elderly population forecast based on the NAR neural network Markov model, 2024, 9, 2473-6988, 3235, 10.3934/math.2024157 | |
4. | Luigi Accardi, Amenallah Andolsi, Farrukh Mukhamedov, Mohamed Rhaima, Abdessatar Souissi, Clustering quantum Markov chains on trees associated with open quantum random walks, 2023, 8, 2473-6988, 23003, 10.3934/math.20231170 | |
5. | Hasan Akın, Phase transition analysis of the Potts-SOS model with spin set {−1,0,+1} on the Cayley tree, 2024, 99, 0031-8949, 125204, 10.1088/1402-4896/ad88b5 | |
6. | Dewang Li, Chingfei Luo, Meilan Qiu, Optimal Weighted Markov Model and Markov Optimal Weighted Combination Model with Their Application in Hunan’s Gross Domestic Product, 2025, 13, 2227-7390, 533, 10.3390/math13030533 |