Research article

On symmetry of the product of two higher-order quasi-differential operators

  • Received: 29 December 2022 Accepted: 13 February 2023 Published: 20 February 2023
  • MSC : 34B20, 34B24, 47B25

  • The symmetric realizations of the product of two higher-order quasi-differential expressions in Hilbert space are investigated. By means of the construction theory of symmetric operators, we characterize symmetric domains determined by two-point boundary conditions for product of two symmetric differential expressions with regular or limit-circle singular endpoints. The presented result contains the characterization of self-adjoint domains as a special case. Several examples of singular symmetric product operators are given.

    Citation: Yanyu Xiang, Aiping Wang. On symmetry of the product of two higher-order quasi-differential operators[J]. AIMS Mathematics, 2023, 8(4): 9483-9505. doi: 10.3934/math.2023478

    Related Papers:

    [1] Xiulin Hu, Lei Wang . Positive solutions to integral boundary value problems for singular delay fractional differential equations. AIMS Mathematics, 2023, 8(11): 25550-25563. doi: 10.3934/math.20231304
    [2] Yanyan Cui, Chaojun Wang . Dirichlet and Neumann boundary value problems for bi-polyanalytic functions on the bicylinder. AIMS Mathematics, 2025, 10(3): 4792-4818. doi: 10.3934/math.2025220
    [3] Kun Li, Peng Wang . Properties for fourth order discontinuous differential operators with eigenparameter dependent boundary conditions. AIMS Mathematics, 2022, 7(6): 11487-11508. doi: 10.3934/math.2022640
    [4] Jianfang Qin, Kun Li, Zhaowen Zheng, Jinming Cai . Eigenvalues of fourth-order differential operators with eigenparameter dependent boundary conditions. AIMS Mathematics, 2022, 7(5): 9247-9260. doi: 10.3934/math.2022512
    [5] Mukhamed Aleroev, Hedi Aleroeva, Temirkhan Aleroev . Proof of the completeness of the system of eigenfunctions for one boundary-value problem for the fractional differential equation. AIMS Mathematics, 2019, 4(3): 714-720. doi: 10.3934/math.2019.3.714
    [6] Tao Wang, Jijun Ao, Anton Zettl . A class of dissipative differential operators of order three. AIMS Mathematics, 2021, 6(7): 7034-7043. doi: 10.3934/math.2021412
    [7] A. Kamal, M. Hamza. Eissa . A new product of weighted differentiation and superposition operators between Hardy and Zygmund Spaces. AIMS Mathematics, 2021, 6(7): 7749-7765. doi: 10.3934/math.2021451
    [8] Karmina K. Ali, Resat Yilmazer . Discrete fractional solutions to the effective mass Schrödinger equation by mean of nabla operator. AIMS Mathematics, 2020, 5(2): 894-903. doi: 10.3934/math.2020061
    [9] Rabha W. Ibrahim, Dumitru Baleanu . Fractional operators on the bounded symmetric domains of the Bergman spaces. AIMS Mathematics, 2024, 9(2): 3810-3835. doi: 10.3934/math.2024188
    [10] Yanyan Cui, Chaojun Wang . The inhomogeneous complex partial differential equations for bi-polyanalytic functions. AIMS Mathematics, 2024, 9(6): 16526-16543. doi: 10.3934/math.2024801
  • The symmetric realizations of the product of two higher-order quasi-differential expressions in Hilbert space are investigated. By means of the construction theory of symmetric operators, we characterize symmetric domains determined by two-point boundary conditions for product of two symmetric differential expressions with regular or limit-circle singular endpoints. The presented result contains the characterization of self-adjoint domains as a special case. Several examples of singular symmetric product operators are given.



    Consider the equation

    My=λwyonJ=(a,b),a<b, (1.1)

    where M is a symmetric differential expression of order n, and w is a positive weight function.

    In 1995, M¨oller-Zettl [12] first characterized the symmetric realizations of (1.1) for regular problems. Motivated by the method used by Sun [10] for self-adjoint operators and by the method of M¨oller-Zettl [12], Wang-Zettl [3,4] characterized the domains of the symmetric realizations of (1.1) for singular problems with any deficiency index.

    Based on the self-adjoint GKN theorem and von Neumann's formula for the adjoint of a symmetric operator in Hilbert space, for classical expression M, Sun [10] gave a decomposition of the maximal domain using certain solutions for non-real value of the spectral parameter λ, and then characterized all the self-adjoint realizations. Sun's work is an important contribution to the study of self-adjoint domains. Wang et al. [6] established a new representation in terms of certain solutions for real λ. This leads to a classification of solutions as limit-point(LP) or limit-circle(LC). The LC solutions contribute to the singular boundary conditions, but the LP solutions do not. In 2012, Hao et al. [18] extended this result to the case when both endpoints are singular. This real λ decomposition of the maximal domain and the construction of LC and LP solutions also play a critical role in the investigation of the spectrum of self-adjoint operators, the classification of self-adjoint boundary conditions, and the characterization of symmetric domains.

    In [1,4,5], Wang-Zettl characterized the symmetric operators in H=L2(J,w) and proved a symmetric GKN-Type theorem. The result contains the self-adjoint GKN theorem as a special case. It is well known that the self-adjoin GKN theorem is widely used to study self-adjoint operators, difference operators, Hamiltonian systems, multi-interval operators, etc. The symmetric GKN-Type theorem maybe will have similar extensions for symmetric problems.

    For products and powers of differential expressions, in [2,15,16], the deficiency indices of powers of classical expressions and of quasi-differential expressions were discussed. In [20], the self-adjointness of the product of two second-order differential operators was obtained. Based on [10,20], An-Sun [8] characterized the self-adjointness of product of two nth-order real classical differential expressions with two regular endpoints and extended to problems with one regular endpoint and one singular endpoint [9]. In recent years, there are some works on the self-adjointness of products of differential operators, see [7,11,13,14,17,19].

    In this paper, we study the symmetric domain characterization for product of two quasi-differential expressions of order n, even or odd, with complex coefficients. We consider the cases when each endpoint is either regular or LC singular. The self-adjoint characterization for product of two differential expressions is a special case.

    The organization of this paper is as follows. Following this introduction, the quasi-differential expressions, Lagrange identity, maximal and minimal operators, and powers of differential expressions are given in Section 2. Section 3 is devoted to the symmetric domain characterization of product of two differential operators. In subsection 3.1, we consider the case when one endpoint is regular and the other LC singular. In subsection 3.2, we consider the case when both endpoints are singular. Some examples are given in Section 4.

    We first repeat some definitions and basic properties of quasi-differential expressions. See the book [1] for more details.

    Definition 1. For n>1, let

    Zn(J):={Q=(qrs)nr,s=1Mn(Lloc(J)),qr,r+10a.e.onJ,q1r,r+1Lloc(J),1rn1,qrs=0a.e.onJ,2r+1<sn;qrsLloc(J),sr+1,1rn1}. (2.1)

    For QZn(J) we define

    V0:={y:JC,yismeasurable}

    and y[0]=y(yV0). Inductively, for r=1,,n, we define

    Vr={yVr1:y[r1](ACloc(J))},y[r]=q1r,r+1{y[r1]rs=1qrsy[s1]}(yVr), (2.2)

    where qn,n+1:=1. Finally we set

    My=MQy=iny[n](yVn). (2.3)

    The expression M=MQ is called the quasi-differential expression generated by Q. For Vn we also use the notations D(Q) and D(M).

    Definition 2. Let QZn(J),J=(a,b). The expression M=MQ is said to be regular at a or we say a is a regular endpoint, if for some c,a<c<b, we have

    q1r,r+1L(a,c),r=1,,n1,qrsL(a,c),1r,sn,sr+1.

    Similarly the endpoint b is regular if for some c,a<c<b, we have

    q1r,r+1L(c,b),r=1,,n1,qrsL(c,b),1r,sn,sr+1.

    Note that from the definition of QZn(J) it follows that if the above hold for some cJ, then they hold for any cJ. We say that M is regular on J, if M is regular at both endpoints. An endpoint is singular if it is not regular.

    Definition 3. Let N2={2,3,4,,}. For kN2, we define the matrix Ek as follows:

    Ek=((1)rδr,k+1s)kr,s=1, (2.4)

    where δi,j is the Kronecker δ. Note that

    Ek=E1k=(1)k+1Ek. (2.5)

    Lemma 1 (Lagrange Identity). Let QZn(J),P=E1QE where E=En is defined in Definition 3. Then PZn(J) and for any yD(MQ),zD(MP), we have

    ˉzMQyy¯MPz=[y,z], (2.6)

    where

    [y,z]=innr=0(1)n+1rˉz[nr1]Py[r]Q=inZEY. (2.7)

    Here we call [y,z] or just [,] a Lagrange bracket.

    Corollary 1. If My=λwy and Mz=ˉλwz, then [y,z] is constant on J. In particular, if λ is real and My=λwy, Mz=λwz, then [y,z] is constant on J.

    The above symplectic matrix Ek and the Lagrange Identity play an important role in the study of general symmetric differential expressions and the characterization of domains of symmetric and self-adjoint boundary conditions.

    Definition 4. Let QZn(J) and suppose that Q satisfies

    Q=E1QE,whereE=En=((1)rδr,n+1s)nr,s=1,i.e.qrs=(1)r+s1ˉqn+1s,n+1r,1r,sn.

    Then Q is called a Lagrange symmetric matrix and the expression M=MQ is called a Lagrange symmetric, or just a symmetric, differential expression.

    Definition 5. Let QZn(J), Q=E1QE, let M=MQ, H= L2(J,w), and let w be a weight function. The maximal operator Smax=Smax(Q,J) with domain Dmax=Dmax(Q,J) is defined by:

    Dmax={yH:yD(M),w1MyH},Smaxy=w1My,yDmax.

    Dmax(Q,J) is dense in H. The minimal operator Smin=Smin(Q,J) with domain Dmin=Dmin(Q,J) is defined as Smin=Smax. Smin is a closed symmetric operator in H with dense domain and Smin=Smax.

    The following result is immediate from the Lagrange Identity and integration.

    Corollary 2. For any y,zD(M),J=(a,b), the limits limtb[y,z](t),limta+[y,z](t) exist and are finite. And then we have:

    ba{ˉzMyy¯Mz}=[y,z](b)[y,z](a). (2.8)

    Lemma 2. Let a1<<akJ, where a1 and ak can also be regular endpoints. Let αjrC(j=1,,k;r=0,,n1). Then there is a yDmax such that

    y[r](aj)=αjr(j=1,,k;r=0,,n1).

    Lemma 3. Let QZn(J), Q=E1QE. Then

    D(Smin)={yDmax:[y,z](a)=0=[y,z](b),for allzDmax}. (2.9)

    Given a Lagrange symmetric matrix QZn(J) and its associated symmetric expression M=MQ the construction and properties of powers Ms (sN2) of quasi-differential expression M was given in [1,2].

    Lemma 4. Assume that QZn(J) is a Lagrange symmetric matrix. Let M=MQ and define M2 by M2y=M(My),,Msy=M(Ms1y). Let Q[1]=Q and for sN2, let Q[s] denote the block diagonal matrix

    Q[s]=[QQ], (2.10)

    where there are s matrices Q on the diagonal and all other entries in this sn×sn matrix are zero except for the entries in positions (n,n+1),(2n,2n+1),,((s1)n,(s1)n+1), these are all equal to 1. Then, for any positive integer s, the matrices Q[s] are in Zsn(J), and are Lagrange symmetric and the symmetric differential expression MS is given by

    Ms=MQ[s]. (2.11)

    Lemma 5. Let QZn(J) be Lagrange symmetric and let M=MQ be the associated symmetric expression. If all solutions of My=λwy are in L2(J,w) for some λC, then this is true for all solutions of Msy=λwy for every λC and every sN.

    Wang-Zettl [1] characterized the domains of symmetric realizations S of the equation

    My=λwyJ=(a,b),a<b+, (3.1)

    in the Hilbert space H=L2(J,w), where M=MQ is Lagrange symmetric, w is a positive weighted function. Based on the work of [1], we now study the symmetric realizations of product of two quasi-differential expressions in the case that each endpoint is either regular or limit-circle(LC) singular.

    In the following we always let QZn(J),J=(a,b),nN2, be a Lagrange symmetric matrix and M=MQ the corresponding symmetric differential expression of order n, even or odd, with real or complex coefficients.

    In this subsection, we consider the case when one endpoint of J=(a,b) is regular and the other LC singular. By the Patching Lemma 2 and the decomposition of the maximal domain given by Theorem 4.4.3 in [1], we have the following result.

    Lemma 6. Let the endpoint a be regular, b LC singular, and let a<c<b. Then the deficiency index of M is n and there exist n linearly independent solutions u1(t),,un(t) of My=0 in L2(J) such that

    u[j1]i(a)=δij(i,j=1,2,,n). (3.2)

    And then we have the decomposition of Dmax:

    Dmax=Dmin˙+span{z1,z2,,zn}˙+span{u1,u2,un}, (3.3)

    where ziDmax,i=1,n such that zi(t)=0 for tc and z[j1]i(a)= δij,i,j=1,,n, and δij is the Kronecker δ.

    By the Lagrange Identity and Corollary 1, we have

    [ui,uj]n(b)=[ui,uj]n(a),i,j=1,2,,n, (3.4)

    where [,]n denotes the Lagrange bracket of differential expression M.

    The following theorem can be found in Chapter 6 of [1].

    Theorem 1. Let M=MQ,QZn(J),J=(a,b),a<b+ be Lagrange symmetric, w a weight function, and let a be regular and b LC singular. Let the composed matrix U=(A:B) be a boundary condition matrix with rank(U)=l,0l2n, where Al,n,Bl,n are complex matrices. Define the operator S(U) in L2(J,w) by

    D(S(U))={yDmax:UYa,b=0},S(U)y=Smaxy  foryD(S(U)),

    where

    Ya,b=(YaYb),  Ya=(y[0](a)y[n1](a)),  Yb=([y,u1]n(b)[y,un]n(b)), (3.5)

    functions u1,,un are given by (3.2). Let C=AEnABEmb(b)B and r=rank(C), where En=((1)rδr,n+1s)nr,s=1. Then we have:

    (i) If l<n, then S(U) is not symmetric;

    (ii) If l=n, then S(U) is self-adjoint (and hence symmetric) if and only if r=0.

    (iii) Let l=n+s,0<sn. Then S(U) is symmetric and not self-adjoint if and only if r=2s.

    Proof. This follows directly from Theorem 6.3.3 of [1]. Note that here we choose the solutions u1(t),,un(t) of My=0. In terms of (3.4) and (3.2), for the case when a is regular and b is LC singular(then mb=n), the matrix Emb(b) given by Theorem 6.3.3 [1] can be written as

    Emb(b)=En(b)=[[u1,u1]n(b)[un,u1]n(b)[u1,un]n(b)[un,un]n(b)]=[[u1,u1]n(a)[un,u1]n(a)[u1,un]n(a)[un,un]n(a)]=inEn, (3.6)

    where i=1. This concludes the proof.

    For M2y=M(My), we obviously have Dmax(M2)Dmax(M) and Dmin(M2)Dmin(M).

    Lemma 7. For any y,zDmax(M2), we have

    [y,z]2n(t)=[My,z]n(t)+[y,Mz]n(t), (3.7)

    where [,]2n denotes the Lagrange bracket of differential expression M2.

    Proof. By the Lagrange Identity (2.6), we have

    [y,z]2n=¯zM2ydty¯M2zdt,
    [y,Mz]n=¯MzMydty¯M2zdt,
    [My,z]n=¯zM2ydtMy¯Mzdt=¯zM2ydt[y,Mz]ny¯M2zdt.

    Therefore

    [y,z]2n=[My,z]n+[y,Mz]n.

    Now we study the symmetric domain characterizations of product of two nth-order differential operators L1 and L2 which are generated by the same symmetric differential expression M (may be with same or different boundary conditions). Let

    Li(y):{Li(y)=My,yDi,Di={yDmax(M):UiYa,b=0},i=1,2, (3.8)

    where Ui=(Ai:Bi) is a composed matrix with rank(Ui)=n+s(0sn), Ai,Bi are (n+s)×n complex matrices, and Ya,b is given by (3.5).

    Set L(y)=(L2L1)(y)=L2(L1(y)), yD1,MyD2. By (3.8), we have

    L:{L(y)=M2y,U1Ya,b=0,U2(MY)a,b=0. (3.9)

    where

    (MY)a,b=((MY)a(MY)b),(MY)a=((My)[0](a)(My)[n1](a)),(MY)b=([My,u1]n(b)[My,un]n(b)), (3.10)

    and the functions u1,u2,,un are defined in Lemma 6.

    Theorem 2. The quasi-derivative (My)[m]=iny[n+m],0mn.

    Proof. From Lemma 4, we have

    Q[2]=[QF0Q], where F=(0010).

    When n=2, we have

    y[1]=q112(yq11y),y[2]=(y[1])q21yq22y[1],y[3]=q112(y[2]q11y[2]),y[4]=y[3]q21y[2]q22y[3].

    From My=i2y[2], it follows that

    (My)[1]=q112((My)q11My)=i2q112(y[2]q11y[2]).

    Therefore (My)[1]=i2y[3].

    When n>2, we have

    y[1]=q112(yq11y),y[2]=q123{(y[1])q21yq22y[1]},y[z]=q1z,z+1{y[z1]zh=1qzhy[h1]},

    where z=1,2,,2n. Since

    My=iny[n],y[n+1]=q112{y[n]q11y[n]},(My)[1]=q112((My)q11My)=inq112(y[n]q11y[n]),

    it follows that (My)[1]=iny[n+1]. Assume that for any given m=k1(2k<n), (My)[k1]=iny[n+k1] holds. Consider the case when m=k, then we have

    y[n+k]=q1n+k,n+k+1{y[n+k1]n+kh=1qn+k,hy[h1]}=q1k,k+1{y[n+k1]n+kh=n+1qn+k,hy[h1]}=q1k,k+1{y[n+k1]kh=1qkhy[n+h1]}

    and

    (My)[k]=q1k,k+1{My[k1]kh=1qkhMy[h1]}=inq1k,k+1{y[n+k1]kh=1qkhy[n+h1]}.

    Hence (My)[k]=iny[n+k] holds. This proof is completed by mathematical induction.

    Corollary 3.

    ((My)[0](t)(My)[n1](t))=in((y[n])[0](t)(y[n])[n1](t))=in(O:In)(y[0](t)y[2n1](t)), (3.11)

    where O is the n×n zero matrix, In is the n×n identity matrix, and (O:In) is a composed matrix.

    Proof. This result follows directly from Theorem 2.

    In the following, we will rewrite the operator defined in (3.9) in a clear form. We first give the maximal domain decomposition of the differential expression M2.

    It is obvious that u1,,un defined in Lemma 6 are linearly independent solutions of M2(y)=0. Let ψ1,ψ2,,ψn be n solutions of M2(y)=0 which satisfy

    ψ[j1]i(a)=0, ψ[n+j1]i(a)=δij,  i,j=1,2,,n. (3.12)

    Combining the conditions of (3.2), we can obtain that u1,,un,ψ1,,ψn are 2n linearly independent solutions of M2y=0. Similar to Lemma 6, we have the following decomposition.

    Corollary 4. The maximal domain Dmax(M2) can has the representation:

    Dmax(M2)=Dmin(M2)˙+span{z1,,zn,zn+1,,z2n}˙+span{u1,u2,un,ψ1,,ψn}, (3.13)

    where ziDmax(M2),i=1,2n such that zi(t)=0 for tc and z[j1]i(a)= δij,i,j=1,,2n.

    Proof. This proof can be completed by using the same method as the proof of Theorem 4.4.3 in [1].

    Since ψiDmax(M2) and Dmax(M2)Dmax(M), by Lemma 6, each ψi has a unique representation:

    ψi=yi0+nj=1dijzj+nj=1aijuj,  i=1,2,,n, (3.14)

    where yi0Dmin(M), dij,aijC. Set

    N=(a11a1nan1ann), (3.15)

    where the entries aij(i,j=1,2,,n) of N are given in (3.14).

    Theorem 3. The operator L defined in (3.9) can be rewritten as

    L:{L(y)=M2y,UYa,b=0, (3.16)

    where U=(A:B) is a composed matrix of matrices

    A=(A100inA2), (3.17)
    B=(inB1EnNTEninB1B20), (3.18)

    and matrices A1,A2,B1,B2 are given in (3.8).

    Proof. By (3.14) and Corollary 4, every yDmax(M2) can be uniquely written as

    y=y0+2ni=1dizi+ni=1ciui+ni=1ciψi=y0+2ni=1dizi+ni=1ciui+ni=1ci(yi0+nj=1dijzj+nj=1aijuj), (3.19)

    where y0Dmin(M2), di,ci,ciC. It follows from Lemma 6 and Lemma 7 that

    [ui,uj]2n(a)=[M(ui),uj]n(a)+[ui,M(uj)]n(a)=0,  i,j=1,2,,n.

    Let

    V(a)=(u1(a)un(a)ψ1(a)ψn(a)u[2n1]1(a)u[2n1]n(a)ψ[2n1]1(a)ψ[2n1]n(a))=(In0DIn), (3.20)
    W(t)=([u1,u1]2n[un,u1]2n[ψ1,u1]2n[ψn,u1]2n[u1,un]2n[un,un]2n[ψ1,un]2n[ψn,un]2n[u1,ψ1]2n[un,ψ1]2n[ψ1,ψ1]2n[ψn,ψ1]2n[u1,ψn]2n[un,ψn]2n[ψ1,ψn]2n[ψn,ψn]2n). (3.21)

    From (2.7), we have

    W(a)=i2nV(a)E2nV(a)=(1)n+1(InD0In)(0En(1)nEn0)(In0DIn)=(1)n+1((1)nDEn+EnDEn(1)nEn0). (3.22)

    Moreover

    W(a)=(1)n+1(0En(1)nEn0). (3.23)

    Therefore (1)nDEn+EnD=0. Note that uj,ψj,j=1,2,,n are solutions of M2y. By the Lagrange identity, we know W(t) is constant on J=(a,b). By (2.9), (3.2) and (3.19), we have

    ([y,u1]2n(b)[y,un]2n(b)[y,ψ1]2n(b)[y,ψn]2n(b))=W(b)(c1cnc1cn)=W(a)(c1cnc1cn). (3.24)

    Hence

    (c1cnc1cn)=(W(a))1([y,u1]2n(b)[y,un]2n(b)[y,ψ1]2n(b)[y,ψn]2n(b))=(1)n+1(0(1)nE1nE1n0)([y,u1]2n(b)[y,un]2n(b)[y,ψ1]2n(b)[y,ψn]2n(b)). (3.25)

    From the decomposition (3.19), we have

    ([y,u1]n(b)[y,un]n(b))=(inEninEnNT)(c1cnc1cn)=(i)n(EnEnNT)(0(1)nE1nE1n0)([y,u1]2n(b)[y,un]2n(b)[y,ψ1]2n(b)[y,ψn]2n(b))=(i)n(EnNTE1n(1)nIn)([y,u1]2n(b)[y,un]2n(b)[y,ψ1]2n(b)[y,ψn]2n(b)). (3.26)

    For any yDmax(M2), it follows from Lemma 7 and Mui=0 that

    [y,ui]2n=[My,ui]n+[y,Mui]n=[My,ui]n,  i=1,2,,n. (3.27)

    Namely

    [y,ui]2n=[My,ui]n,i=1,2,,n. (3.28)

    The proof of Theorem 3 is now immediate from the above discussion and (3.9). Furthermore, we have the following result.

    Theorem 4. The relationship ¯NEn+(1)nEnNT=0 holds.

    Proof. By Lemmas 7 and Lemma 3, we have

    [ψi,ψj]2n(b)=[Mψi,ψj]n(b)+[ψi,Mψj]n(b)=[Mψi,nk=1ajkuk]n(b)+[nk=1aikuk,Mψj]n(b)=nk=1¯ajk[Mψi,uk]n(b)+nk=1aik[uk,Mψj]n(b)=nk=1¯ajk[ψi,uk]2n(b)+nk=1aik[uk,ψj]2n(b)=nk=1¯ajk[ψi,uk]2n(a)+nk=1aik[uk,ψj]2n(a)=0.

    Let ¯NEn+(1)nEnNT=(bij)1i,jn. Then by (3.23), we have

    bij=nk=1(1)n+1¯ajk[ψi,uk]2n(a)+nk=1(1)n+1aik[uk,ψj]2n(a)=0.

    This completes the proof.

    Based on the above lemmas and theorems, we now obtain our main result: the symmetric characterization of product of two differential operators.

    Theorem 5. Let the hypothesis and notations of Theorem 3 hold. Then the product operator L=L2L1 is symmetric if and only if

    rank(A1EnA2B1EnB2)=2s, (3.29)

    where 0sn.

    Proof. Since rank(Ai:Bi)=n+s(i=1,2) and

    rank(U)=rank(A10inB1EnNTEninB10inA2B20)=rank(A1inB10inB1EnNTEn00inA2B2),

    we can obtain rank(U)=2n+2s. By computation, we have

    AE2nA=(0inA1EnA2inA2EnA10),
    BW(b)B=(1)n+1(inB1EnNTEninB1B20)(0En(1)nEn0)((1)n+1inEn¯NEnB1B2(i)nB10)=(1)n+1(B1(¯NEn+(1)nEnNT)B1(1)n+1inB1EnB2(1)n+1inB2EnB10).

    Now we will use the basic Theorem 1 to prove our result. Note that here the matrix Emb(b) given in Theorem 1 is W(b). Then, combining with Theorem 4, we have

    C=AEnABW(b)B=(0in(A1EnA2B1EnB2)in(A2EnA1B2EnB1)0).

    Therefore, by Theorem 1, the operator L is symmetric if and only if

    rank(C)=4s

    which is equal to

    rank(A1EnA2B1EnB2)=2s.

    Thus the proof is completed.

    Remark 1. For theorem 5, if s=0, then L is self-adjoint if and only if A1EnA2=B1EnB2.

    We will consider the symmetric characterization of product of two differential expressions for the case when both endpoints are LC singular.

    In the following we always let QZn(J),J=(a,b),nN2, be Lagrange symmetric, w a weight function, the two endpoints a and b LC singular, and let My=MQy=λwy be the corresponding symmetric differential equation.

    By Lemma 5, we have the next lemma.

    Lemma 8. If the expression M=MQ is LC singular at both endpoints of J=(a,b), then M2 is LC singular at both endpoints.

    We first reduce the decomposition Theorem 4.4.4 of [1] to the case when both endpoints are LC singular.

    Lemma 9. Let M be Lagrange symmetric, a and b LC singular, and let c(a,b). Then QZn((a,c)), QZn((c,b)), and the deficiency indices of My =λwy on (a,c),(c,b) and (a,b) are all n. Then

    (1) There exist n linearly independent solutions p1,,pn of My=0 on (a,c) such that p[j1]i(c)=δij (i,j=1,2,,n). For 1i,jn, we have

    [pi,pj]n(a)=[pi,pj]n(c). (3.30)

    The solutions p1,,pn can be extended to (a,b) such that the extended functions, also denoted by p1,,pn, are in Dmax(a,b) and are identically 0 near b.

    (2) There exist n linearly independent solutions v1,,vn of My=0 on (c,b) such that v[j1]i(c)=δij (i,j=1,2,,n). For 1i,jn, we have

    [vi,vj]n(b)=[vi,vj]n(c). (3.31)

    The solutions v1,,vn can be extended to (a,b) such that the extended functions, also denoted by v1,,vn, are in Dmax(a,b) and are identically 0 near a.

    (3) By the extended functions pi and vi, the maximal domain has the following decomposition:

    Dmax(a,b)=Dmin(a,b)˙+span{p1,,pn}˙+span{v1,,vn}. (3.32)

    Proof. This lemma can be directly obtained from Theorem 4.4.4 of [1]. Since pi and vi are respectively the solutions of My=0 on (a,c) and (c,b), together with the Lagrange identity, we can obtain (3.30) and (3.31) hold.

    Let

    ˆE(a)=([p1,p1]n(a)[pn,p1]n(a)[p1,pn]n(a)[pn,pn]n(a)),ˆE(b)=([v1,v1]n(b)[vn,v1]n(b)[v1,vn]n(b)[vn,vn]n(b)). (3.33)

    By (1) and (2) of Lemma 9, we have ˆE(a)=ˆE(b)=inEn.

    Let

    ˆYa,b=(ˆYaˆYb),ˆYa=([y,p1]n(a)[y,pn]n(a)),ˆYb=([y,v1]n(b)[y,vn]n(b)). (3.34)

    The following is a minor modification of Theorem 6.3.3 of [1].

    Lemma 10. Let the hypothesis and notations of Lemma 9 hold. Suppose U is a boundary matrix with rank(U)=l,0l2n. Let U=(A:B), where Al×n and Bl×n are complex matrices. Define the operator S(U) in L2(J,w) by

    D(S(U))={yDmax:UˆYa,b=0},S(U)y=Smaxy  foryD(S(U)).

    Let rank(AˆE(a)ABˆE(b)B)=r, where ˆE(a) and ˆE(b) are given in (3.33). Then we have:

    (i) If l<n, then S(U) is not symmetric;

    (ii) Let l=n+s,0sn. Then S(U) is symmetric if and only if r=2s.

    Let the operators ˆL1 and ˆL2 be generated by the same LC symmetric differential expression M :

    ˆLi:{ˆLi(y)=My,yˆDi,ˆDi={yDmax:UiˆYa,b=0},i=1,2, (3.35)

    where Ui=(Ai:Bi),rank(Ui)=n+s, 0sn, Ai,Bi are (n+s)×n matrices, and ˆYa,b is given in (3.34).

    Let ˆL=ˆL2ˆL1, yˆD1, MyˆD2, then we have:

    ˆL:{ˆL(y)=M2y,U1ˆYa,b=0,U2(^MY)a,b=0, (3.36)

    where

    (^MY)a,b=((^MY)a(^MY)b),  (^MY)a=([My,p1]n(a)[My,pn]n(a)),  (^MY)b=([My,v1]n(b)[My,vn]n(b)). (3.37)

    It is obvious that p1,,pn are solutions of M2(y)=0 on (a,c), and v1,,vn are solutions of M2(y)=0 on (c,b). Let θ1,θ2,,θn be solutions of M2(y)=0 on (a,c), and let β1,β2,,βn be solutions of M2(y)=0 on (c,b), which satisfy

    θ[j1]i(c)=0,  θ[n1+j]i(c)=δij,i,j=1,2,,n, (3.38)
    β[j1]i(c)=0,  β[n1+j]i(c)=δij,i,j=1,2,,n. (3.39)

    It is clear that p1,,pn,θ1,,θn are 2n linearly independent solutions of M2y=0 on (a,c) and v1,,vn,β1,,βn are 2n linearly independent solutions of M2y=0 on (c,b). By Naimark Patching Lemma 2, θ1,,θn can be extended to (a,b) such that the extended functions, also denoted by θ1,,θn, are in Dmax(a,b) and are identically 0 near b. Similarly, β1,,βn can be extended to (a,b) such that the extended functions, also denoted by β1,,βn, are in Dmax(a,b) and are identically 0 near a. Similar to Lemma 9, we have the following decomposition.

    Corollary 5. The maximal domain Dmax(M2)=Dmax(M2,(a,b)) has the representation:

    Dmax(M2)=Dmin(M2)˙+span{p1,,pn,θ1,,θn}˙+span{v1,v2,vn,β1,,βn}. (3.40)

    Since θi,βiDmax(M2) and Dmax(M2)Dmax(M), by (3.32), each θi and βi has a unique representation:

    θi=yi0+nj=1aijpj+nj=1bijvj,i=1,2,,n, (3.41)
    βi=y0i+nj=1cijpj+nj=1dijvj,i=1,2,,n, (3.42)

    where yi0,y0iDmin(M),aij,bij,cij,dijC.

    Set

    N1=(aij)n×n,N2=(dij)n×n, (3.43)

    where the entries aij,dij are given in (3.41) and (3.42).

    Theorem 6. The operator ˆL defined in (3.36) can be rewritten as

    ˆL:{ˆL(y)=M2y,UˆYa,b=0, (3.44)

    where U=(A:B),

    A=(inA1EnNT1EninA1A20), (3.45)
    B=(inB1EnNT2EninB1B20). (3.46)

    and matrices A1,A2,B1,B2 are given in (3.35).

    Proof. By Corollary 5, for every yDmax(M2), we have

    y=y0+ni=1aipi+ni=1ciθi+ni=1bivi+ni=1diβi, (3.47)

    where y0Dmin(M2),ai,bi,ci,diC.

    Let

    ˆU(c)=(p1(c)pn(c)θ1(c)θn(c)p[2n1]1(c)p[2n1]n(c)θ[2n1]1(c)θ[2n1]n(c))=(In0D1In), (3.48)
    ˆV(c)=(v1(c)vn(c)β1(c)βn(c)v[2n1]1(c)v[2n1]n(c)β[2n1]1(c)β[2n1]n(c))=(In0D2In), (3.49)
    ˆW(t)=([p1,p1]2n[pn,p1]2n[θ1,p1]2n[θn,p1]2n[p1,pn]2n[pn,pn]2n[θ1,pn]2n[θn,pn]2n[p1,θ1]2n[pn,θ1]2n[θ1,θ1]2n[θn,θ1]2n[p1,θn]2n[pn,θn]2n[θ1,θn]2n[θn,θn]2n), (3.50)
    ˆZ(t)=([v1,v1]2n[vn,v1]2n[β1,v1]2n[βn,v1]2n[v1,vn]2n[vn,vn]2n[β1,vn]2n[βn,vn]2n[v1,β1]2n[vn,β1]2n[β1,β1]2n[βn,β1]2n[v1,βn]2n[vn,βn]2n[β1,βn]2n[βn,βn]2n). (3.51)

    From (2.7), we have

    ˆW(c)=i2nˆU(c)E2nˆU(c)=(1)n+1(InD10In)(0En(1)nEn0)(In0D1In)=(1)n+1((1)nD1En+EnD1En(1)nEn0),
    ˆZ(c)=i2nˆV(c)E2nˆV(c)=(1)n+1(InD20In)(0En(1)nEn0)(In0D2In)=(1)n+1((1)nD2En+EnD2En(1)nEn0).

    By Lemma 7, we have

    [pi,pj]2n(c)=[Mpi,pj]n(c)+[pi,Mpj]n(c)=0,i,j=1,2,,n,
    [vi,vj]2n(c)=[Mvi,vj]n(c)+[vi,Mvj]n(c)=0,i,j=1,2,,n.

    Therefore

    ˆW(c)=ˆZ(c)=(1)n+1(0En(1)nEn0).

    At the same time, we obtain that (1)nDiEn+EnDi=0, i=1,2. Since pj,θj (j=1,2,,n) are solutions of M2y=0 on (a,c), vj,βj (j=1,2,,n) are solutions of M2y=0 on (c,b). It follows from the Lagrange identity that ˆW(t), ˆZ(t) are constant on (a,c) and (c,b), respectively. Note that pi,θi are identically 0 near b and vi,βi are identically 0 near a. Then, for i,j=1,2,n, we have

    [pi,vj]n(b)=[pi,βj]n(b)=[θi,vj]n(b)=[θi,βj]n(b)=0, (3.52)
    [vi,pj]n(a)=[vi,θj]n(a)=[βi,pjt]n(a)=[βi,θj]n(a)=0, (3.53)
    [pi,vj]2n(b)=[pi,βj]2n(b)=[θi,vj]2n(b)=[θi,βj]2n(b)=0, (3.54)
    [vi,pj]2n(a)=[vi,θj]2n(a)=[βi,pj]2n(a)=[βi,θj]2n(a)=0. (3.55)

    By (2.9), (3.47), (3.54) and (3.55), we have

    ([y,p1]2n(a)[y,pn]2n(a)[y,θ1]2n(a)[y,θn]2n(a))=ˆW(a)(a1anc1cn)=ˆW(c)(a1anc1cn), (3.56)
    ([y,v1]2n(b)[y,vn]2n(b)[y,β1]2n(b)[y,βn]2n(b))=ˆZ(b)(b1bnd1dn)=ˆZ(c)(b1bnd1dn). (3.57)

    Hence

    (a1anc1cn)=(ˆW(c))1([y,p1]2n(a)[y,pn]2n(a)[y,θ1]2n(a)[y,θn]2n(a))=(1)n+1(0(1)nE1nE1n0)([y,p1]2n(a)[y,pn]2n(a)[y,θ1]2n(a)[y,θn]2n(a)), (3.58)
    (b1bnd1dn)=(ˆZ(c))1([y,v1]2n(b)[y,vn]2n(b)[y,β1]2n(b)[y,βn]2n(b))=(1)n+1(0(1)nE1nE1n0)([y,v1]2n(b)[y,vn]2n(b)[y,β1]2n(b)[y,βn]2n(b)). (3.59)

    By (3.41), (3.42), (3.47), (3.52) and (3.53), we have

    ([y,p1]n(a)[y,pn]n(a))=(inEninEnNT1)(a1anc1cn)=(i)n(EnEnNT1)(0(1)nE1nE1n0)([y,p1]2n(a)[y,pn]2n(a)[y,θ1]2n(a)[y,θn]2n(a))=(i)n(EnNT1E1n(1)nIn)([y,p1]2n(a)[y,pn]2n(a)[y,θ1]2n(a)[y,θn]2n(a)),

    and

    ([y,v1]n(b)[y,vn]n(b))=(inEninEnNT2)(b1bnd1dn)=(i)n(EnEnNT2)(0(1)nE1nE1n0)([y,v1]2n(b)[y,vn]2n(b)[y,β1]2n(b)[y,βn]2n(b))=(i)n(EnNT2E1n(1)nIn)([y,v1]2n(b)[y,vn]2n(b)[y,β1]2n(b)[y,βn]2n(b)).

    For any yDmax(M2), it follows from Lemma 7, together with Mpi=0,Mvi=0, that

    [y,pi]2n=[My,pi]n+[y,Mpi]n=[My,pi]n,  i=1,2,,n; (3.60)
    [y,vi]2n=[My,vi]n+[y,Mvi]n=[My,vi]n,  i=1,2,,n. (3.61)

    Thus, in terms of above discussion, the proof for Theorem 6 is completed.

    Theorem 7. The relationships ¯N1En+(1)nEnNT1=0, ¯N2En+(1)nEnNT2=0 hold.

    Proof. Similar to the proof of Theorem 4, we have

    [θi,θj]2n(a)=[Mθi,θj]n(a)+[θi,Mθj]n(a)=nk=1¯ajk[θi,pk]2n(a)+nk=1aik[pk,θj]2n(a)=0

    and

    [βi,βj]2n(b)=[Mβi,βj]n(b)+[βi,Mβj]n(b)=nk=1¯djk[βi,vk]2n(b)+nk=1dik[vk,βj]2n(b)=0.

    Then we can easily obtain ¯N1En+(1)nEnNT1=0 and ¯N2En+(1)nEnNT2=0.

    Based on above discussion, we present the main result:

    Theorem 8. Let the hypothesis and notations of Theorem 6 hold. Then the product operator ˆL=ˆL2ˆL1 is symmetric if and only if

    rank(A1EnA2B1EnB2)=2s, (3.62)

    where 0sn.

    Proof. Since

    rank(U)=rank(inA1EnNT1EninA1inB1EnNT2EninB1A20B20)

    and rank(Ai:Bi)=n+s(i=1,2), we have rank(U)=2n+2s.

    By computation, we get

    AˆW(a)A=(1)n+1(inA1EnNT1EninA1A20)(0(1)nE1nE1n0)((1)n+1inEn¯N1EnA1A2(i)nA10)=(1)n+1(A1(¯N1En+(1)nEnNT1)A1(1)n+1inA1EnA2(1)n+1inA2EnA10)

    and

    BˆZ(b)B=(1)n+1(inB1(¯N2En+(1)nEnNT2)B1(1)n+1inB1EnB2(1)n+1inB2EnB10).

    To prove the furthermore part we note that here the matrices ˆE(a), ˆE(b) given in Lemma 10 are ˆW(a), ˆZ(b), respectively. Thus, by Lemma 10 and Theorem 7, we have

    AˆW(a)ABˆZ(b)B=(1)n+1(0(1)n+1in(A1EnA2B1EnB2)in(A1EnA2B1EnB2)0).

    It follows from Lemma 10 that L is symmetric if and only if

    rank(AˆW(a)ABˆZ(b)B)=4s

    which is equal to

    rank(A1EnA2B1EnB2)=2s.

    The next corollary is the self-adjoint special case of Theorem 8.

    Corollary 6. Let the hypothesis and notations of Theorem 6 hold. Then the product operator ˆL=ˆL2ˆL1 is self-adjoint if and only if

    A1EnA2B1EnB2=0.

    Proof. This is the special case s=0 of Theorem 8.

    Remark 2. Recall that the Lagrange brackets [y,z] are well defined at each singular endpoint. These brackets can be used to replace the quasi-derivatives. Our symmetric domain characterization Theorem 8 can be adapted to the maximal deficiency case which occurs when each endpoint is either regular or LC singular. Namely, our result can be used for the four cases for the endpoints: R/R, R/LC, LC/R, LC/LC.

    Example 1. Consider My=(py)+qy on J=(a,b),a<b+, where p1,qLloc(J,R). Let a and b be LC singular. Set

    A1=(100i00),  B1=(000110),  A2=(01i000),  B2=(000111).

    The operator L1y=My is determined by the boundary conditions:

    [y,p1](a)=0,  i[y,p2](a)+[y,v2](b)=0,  [y,v1](b)=0.

    The operator L2y=My is determined by the boundary conditions:

    [y,p2](a)=0,  i[y,p1](a)=[y,v2](b),  [y,v1](b)+[y,v2](b)=0.

    Since here s=1 and rank(A1E2A2B1E2B2)=2s=2, by Theorem 8, the product operator L=L2L1 is symmetric.

    Example 2. Consider My=(py)+qy on J=(a,b),a<b+, where p1,qLloc(J,R). Let a and b be LC singular. Set

    A1=(100100),  A2=(011000),  B1=(000010),  B2=(000001).

    The operator L1y=My is determined by the boundary conditions:

    [y,p1](a)=0,  [y,p2](a)=0,  [y,v1](b)=0.

    The operator L2y=My is determined by the boundary conditions:

    [y,p1](a)=0,  [y,p2](a)=0,  [y,v2](b)=0.

    Since here s=1 and rank(A1E2A2B1E2B2)=3, by Theorem 8, the product operator L=L2L1 is not symmetric.

    Remark 3. For the above examples, when the endpoint a is regular for this M, we can simply replace [y,p1](a),[y,p2](a) with y(a), y[1](a), respectively. Similarly for a regular endpoint b, we can replace [y,v1](b),[y,v2](b) with y(b), y[1](b), respectively.

    Example 3. Let QZn(J),nN2, J=(a,b),a<b+, M=MQ be the symmetric expression, and let Ns×n (0sn) be of full row rank, In the n×n identity matrix. Assume that a and b are LC singular. Choose

    A1=(Ns×n0n×n),B1=(0s×nIn),A2=(0s×nIn),B2=(Ns×n0n×n).

    Let

    U1=(A1:B1)=(N00In),U2=(A2:B2)=(0NIn0).

    We define the operators L1, L2 as (3.35). By computation, we have

    A1EnA2B1EnB2=(0NEnEnN0).

    Obviously, rank(U1)=rank(U2)=n+s, rank(A1EnA2B1EnB2)=2s. Then, by Theorem 8, we know L=L2L1 is a symmetric operator.

    The authors declare that there is no conflicts of interest in this paper.



    [1] A. Wang, A. Zettl, Ordinary Differential Operators, American Mathematical Soc., 2019.
    [2] A. Zettl, Powers of real symmetric differential expressions without smoothness assumptions, Quaest. Math., 1 (1976), 83–94. https://doi.org/10.1080/16073606.1976.9632515 doi: 10.1080/16073606.1976.9632515
    [3] A. Wang, A. Zettl, Characterization of domains of symmetric and self-adjoint ordinary differential operators, Electron. J. Differ. Eq., 15 (2018), 1–18.
    [4] A. Wang, A. Zettl, A symmetric GKN-type theorem and symmetric differential operators, J. Differ. Equations, 265 (2018), 5156–5176. https://doi.org/10.1016/j.jde.2018.06.028 doi: 10.1016/j.jde.2018.06.028
    [5] A. Wang, A. Zettl, A GKN theorem for symmetric operators, J. Math. Anal. Appl., 452 (2017), 780–791. https://doi.org/10.1016/j.jmaa.2017.03.004 doi: 10.1016/j.jmaa.2017.03.004
    [6] A. Wang, J. Sun, A. Zettl, Characterization of domains of self-adjoint ordinary differential operators, J. Differ. Equations, 246 (2009), 1600–1622. https://doi.org/10.1016/j.jde.2008.11.001 doi: 10.1016/j.jde.2008.11.001
    [7] C. Yang, Self-adjointness of products of the limit-point Sturm-Liouville operators, Journal of Systems Science and Mathematical Sciences (in Chinese), 26 (2006), 368–374.
    [8] J. An, J. Sun, On self-adjointness of the product of two n-order differential operators on [a,b], Annals Of Differential Equations, 18 (1998), 50–57.
    [9] J. An, J. Sun, the self-adjointness of the product operators of two mth-order differential operators on [0,+), Acta Math. Sin. (English Seris), 20 (2004), 793–802. https://doi.org/10.1007/s10114-004-0386-3 doi: 10.1007/s10114-004-0386-3
    [10] J. Sun, On the self-adjoint extensions of symmetric ordinary differential operators with middle deficiency indices, Acta Math. Sin., 2 (1986), 152–167. https://doi.org/10.1007/BF02564877 doi: 10.1007/BF02564877
    [11] J. Zheng, On the self-adjointness of product of two odd-order differential operators, Inner Mongolia University (in Chinese), 2009.
    [12] M. M¨oller, A. Zettl, Symmetrical differential operators and their friedrichs extension, J. Differ. Equations, 115 (1995), 50–69. https://doi.org/10.1006/jdeq.1995.1003 doi: 10.1006/jdeq.1995.1003
    [13] Q, Lin, Self-adjointness of class 4th-order and 6th-order differential operator products, Journal of Inner Mongolia Normal University (Natural Science Edition) (in Chinese), 32 (2019), 74–79.
    [14] Q. Qing, G. Wang, H. Kong, Self-adjointness of the products of high-order differential operators, Journal of Inner Mongolia Normal University(Natural Science Edition)(in Chinese), 41 (2012), 227–230.
    [15] R. M. Kauffman, T. T. Read, A. Zettl, The Deficiency Index Problem for Powers of Ordinary Differential Expressions, Lecture Notes in Mathematics, 1977.
    [16] W. D. Evans, A. Zettl, On the deficiency indices of powers of real 2nth-order symmetric differential expressions, J. Lond. Math. Soc., 13 (1976), 543–556. https://doi.org/10.1112/jlms/s2-13.3.543 doi: 10.1112/jlms/s2-13.3.543
    [17] X. Guo, The self-adjointness ofproduct of four second-orderdifferential operators, Inner Mongolia University, 2012.
    [18] X. Hao, J. Sun, A.Wang, A. Zettl, Characterization of domains of self-adjoint ordinary differential operators Ⅱ, Results Math., 61 (2012), 255–281. https://doi.org/10.1007/s00025-011-0096-y doi: 10.1007/s00025-011-0096-y
    [19] X. Zhang, W. Wang, Q. Yang, On self-adjointness of the product of three second-order differential operators, Journal of Inner Mongolia Normal University (Natural Science Edition) (in Chinese), 36 (2007), 35–42.
    [20] Z, Cao, J. Sun, D. E. Edmunds, On self-adjointness of the product of two second-order differential operators, Acta Math. Sin. (English Series), 15 (1999), 375–386. https://doi.org/10.1007/BF02650732 doi: 10.1007/BF02650732
  • Reader Comments
  • © 2023 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(1150) PDF downloads(39) Cited by(0)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog