中国科学院数学与系统科学研究院期刊网

15 September 1979, Volume 22 Issue 5
    

  • Select all
    |
    论文
  • Acta Mathematica Sinica, Chinese Series. 1979, 22(5): 515-529. https://doi.org/10.12386/A1979sxxb0048
    Abstract ( )   Knowledge map   Save
    In this note we establish the existence theorem of almost periodic solution for the almost periodic quasi-linear system dx/dt = A(t)x + f(t) + μF(X,t,μ)(1) where X is n-vector, ‖X‖ ≤ K, |t| < ∞,0 ≤μ≤μ, moreover suppose that (a) A(t) is an almost periodic and continuous n × n-matrix function, f(t) is an almost periodic and continuous vector function, and F(X, t,μ) is continuous in (X, t, μ), almost periodic in t, uniformly on ‖X‖≤K,0≤μ≤μ, and Lipshitzian in X; (b) A(t) is of class C~1, ‖A (t)‖≤K_o,‖dA(t)/dt‖≤K_1.The characteristic roots of A(t) are ρ1(t),ρ2(t),…,ρn(t), |Re (ρi(t) + ρi(t))|≥2δ_o>0,i,j = 1,2,3,…,n, and K_1≤1/2(1/K~o)~2(K_o/1+K_o)~((n-1)),K~o=2~(n-1)/2δ_o(1+K_o/2δ_o)~((n-1)(n+2))/2, where K_o,K_1, and δ_o are constants.At the same time we also give conditions for the existence of exponential dichotomies for linear differential equations. Now let us detail our results as follows:1. The system dX/dt=A(t)X(2)admits an exponential dichotomy, i. e. there exist positive constants α,β such that ‖Y_1(t)Z_1(τ)‖≤βexp(-α(t - τ)), t ≥τ, ‖Y_2(t)Z_2(τ)‖≤βexp(α(t - τ)),t ≤τ. where Y_1(t)+ Y_2(t)= Y(t) is the fundamental matrix of (1), and Z_1(t)+ Z_2(t)=Y~(-1)(t), iff there exists the quadric form V(X,t)=XG(t)X, where G(t) is bounded, symmetric and regular matrix, and total time derivative of V(X,t) with respect to the system (2) is positive definite (for the "if" part we require the matrix A(t) to be bounded).2. If A(t) satisfies the condition (b),then there is a quadric form V(X,t)=XG(t)X,such that (1) G(t) is a regular and symmetric matrix; (2) The total time derivative ofV(X,t) with respect to the system (2)is positive definite.3. The existence theorem: If (1) satisfies the condition (a) and (b), then there existsthe unique almost periodic solution of (1) for every μ in the interval 0≤μ≤μ_o for someu_o>0.
  • Acta Mathematica Sinica, Chinese Series. 1979, 22(5): 530-545. https://doi.org/10.12386/A1979sxxb0049
    Abstract ( )   Knowledge map   Save
    Let be a Banach space, T_(s, t) is a bounded linear operator from to family {T_(s,t),0≤s≤t} is said to be a semi-group for two parameters, of T_(s,r)=T_(s,t)·T(t,r), T_(s,s)=I,(0≤s≤t≤r) where I is an identical operator.Let be a measurable space, be all bounded measurable functions, its norm is defined by,then is a Banach space. To every Markov process P(s,t,x,A), let then {P_(s,t,)0≤s≤t} is a semi-group for two parameters.In this paper, the strong continuity and differentiability of the semi-group{T_(s.t,) 0≤s≤t}(particularly{P_(s,t),0≤s≤t}) are investigated.
  • Acta Mathematica Sinica, Chinese Series. 1979, 22(5): 546-555. https://doi.org/10.12386/A1979sxxb0050
    Abstract ( )   Knowledge map   Save
    Let △: a=x_0
  • Acta Mathematica Sinica, Chinese Series. 1979, 22(5): 556-568. https://doi.org/10.12386/A1979sxxb0051
    Abstract ( )   Knowledge map   Save
    Let R be a primitive ring, A be a faithful irreducible R-module. Suppose that F be the centralizer of R-module A, then it is clear that A Fu_i is a vecter space over F. Let Ω be the complete ring of linear transformations of A and let T_ν be the set of all linear transformations of A such that their ranks are all<.Then we can formulate the following definition.Definition: An ideal of R is said to be a v-socle if and only if satisfies the following conditions:(iv)let{u_i}r be a basis,{E_i}r be the set of Ω satisfying the conditions u_iE_j=δ_(ij)u_i,i,j∈Γand let l∈T_v,In this paper we have proved that the o-socle of R defined above is equivalent to the socle of R in the usual meaning.Furthermore, we introduce in this paper the concepts of v-topology,v-A'-summable dual vector space (A, A') and v-total A' too. Then we can obtain the following theorems:Theorem Ⅰ: Let (A,A') be a v-A'-summable dual vector space over F, A' be v-total and let (A) be the set of all continuous linear transformations (by the v-A'-topology), denote (A)=T_v(A). Suppose that R is isomorphic to a subring of (A), which includes (A), then R includes the v-socle.Theorem Ⅱ: Let R be a primitive ring with v-socle. Then there exist a v-A'-summable dual vector space (A,A') over F, A' is v-total and R is isomorphic to a subring of L(A), which includes T(A), where L(A) denotes the ring of all strong continuous linear transformations(by the v-A'-topology), T(A)=T_vL(A).If we set v=0, then it is clear that the o-topology introduced above is namely the usual finite topology and the theorems Ⅰ and Ⅱ are obviously consistent with the usual socalled "Structure Theorem".
  • Acta Mathematica Sinica, Chinese Series. 1979, 22(5): 569-578. https://doi.org/10.12386/A1979sxxb0052
    Abstract ( )   Knowledge map   Save
    The uniqueness in Cauchy problems for linear partial differential equations with C~∞ coefficients is of quite complication. In 1974, M. Strauss and F. Treves proved uniqueness in the Cauchy problem u_t+ib(x, t)u_x+c(x,t)u=0 and u(x, 0)=0, where t b(0,t) vanishes of finite order at t=0. In this paper, by using the technique of pseudo-differential operators, we give a general result on the uniqueness in Cauchy problems for systems of equations of order m≥1 with two independent variables.
  • Acta Mathematica Sinica, Chinese Series. 1979, 22(5): 579-583. https://doi.org/10.12386/A1979sxxb0053
    Abstract ( )   Knowledge map   Save
    In this paper, as another criterion for convex surface we prove the followingTheorem 2.In the 3-dimensional Euclidean space, a simply connected surface π with a closed curve γ as its boundary is a convex surface if and only if π is a local convex surface and through every point P∈π\γ there passes at least one local supporting plane S_p such that S_p~+γ=Φ.
  • Acta Mathematica Sinica, Chinese Series. 1979, 22(5): 584-595. https://doi.org/10.12386/A1979sxxb0054
    Abstract ( )   Knowledge map   Save
    In this paper we use the Leray-Schauder degree to investigate the number of positive solutions of Hammerstein integral equation where G is a bounded closed domain in Euclidean space R~N and f(x,u)=a_i>0(i=1,2,…, n).Theorem 1. Suppose that (i) the non-negative continuous kernel k(x,y) satisfies and among numbers a_i (i=1, 2,…, n) there exist a_(i_0)<1 and a_(i_1)>1 such that and Then equation (*) has at least two positive (i.e.≥0 and ) continuous solutions.Theorem 2. Let the hypothesis (i) of Theorem 1 be satisfied. Suppose that a_i(x) ≥0, a_i(x)∈L, a_i<1 (i=1,2,…, n) and among functions a_i(x) (i=1,2, …,n) there exists α_(i_0)(x) such that a_(i_0) Then equation (*) has at least one positive continuous solution.Theorem 3. Let the hypothesis (i) of Theorem 1 be satisfied. Suppose that α_i(x)≥0, α_i(x)∈L, α_i,>1 (i=1,2,…, n) and among functions α_i(x) (i=1,2,…,n) there exists α_(i_l)(x) such that a_(i_1) Then equation (*) has at least one positive continuous solution.If we assume, in addition, that there exists σ>0 such that,then equation (*) has exactly one positive continuous solution in the case of Theorem 2 and cannot have two comparable positive continuous solutions in the case of Theorem 3.The method mentioned above can also be used to establish the existence of positive solutions of the Dirichlet boundary value problem where L is a uniformly elliptic differential operator:Theorem 4. Let a_i(x)∈C~(o,λ)(Ω) (0 <λ<1),a_i(x)≥0 (i= 1,2,…, n) and among numbers a_i (i=1, 2,…,n) there exists a_(io)<1 such that a_(io)(x)>0 for all x∈ Ω. Suppose that where G(x,y) is the Green function for the Dirichlet problem on Ω of the operator L. Then the problem (**) has at least one solution u(x) such that u(x)∈C~(2,μ)(Ω) (μ=min{λ, α_1,…, α_n}) and 00 for all x∈Ω. Then the problem (**) has exactly one positive solution u(x)∈C~(2,μ)(Ω) (μ=min{λ,α_1,…,α_n}) and u(x)>0 for every x∈Ω.
  • Acta Mathematica Sinica, Chinese Series. 1979, 22(5): 596-619. https://doi.org/10.12386/A1979sxxb0055
    Abstract ( )   Knowledge map   Save
    This question was studied by von Neumann, J. Courant R. and Friedrichs K. in 1940's, but their results were experimental and incomplete. In this paper, we give a necessary supplement to their works. We prove rigorously some global geometrical properties of shock curves and rarefaction curves in the phase space (§2). Using these properties, we examine exhaustively all possible cases of interactions (§3,4). For most cases, by giving constructively the global solution in (x, t)-plane, the problem is solved thoroughly. For others, we discover by Glimm's scheme that its complexity is in the same way as general initial value problems.What are the main differences between their resalts and ours? For example, von Neumann discovered that a contact discontinuity may arise from the collision of two shock waves, but he didn't tell what kind of contact discontinuity it belongs to. We give the exact answer. For the overtaking of two shock waves, they only told that, besides a transmitted shock wave, there are a contact discontinuity and a reflected rarefaction wave or shock wave. We find the kind of the contact discontinuity and give the criterion that shows whether the reflected wave is a rarefaction wave or shock. For interaction involving rarefaction waves, they mainly discussed the circumstances after the penetration, while we analyse emphatically the process of penetration by Glimm's scheme.
  • Acta Mathematica Sinica, Chinese Series. 1979, 22(5): 620-632. https://doi.org/10.12386/A1979sxxb0056
    Abstract ( )   Knowledge map   Save
    Let x_1,…, x_m and y_1,…, y_n be random samples from populations with continuous distribution F and G respectively, Define the null hypothesis H_0 asH_0: There exist constants A and B, B>0, such that G(x)=F((x-A)/B),-∞