中国科学院数学与系统科学研究院期刊网

15 April 1978, Volume 21 Issue 2
    

  • Select all
    |
    论文
  • Acta Mathematica Sinica, Chinese Series. 1978, 21(2): 97-118. https://doi.org/10.12386/A1978sxxb0012
    Abstract ( )   Knowledge map   Save
    1.Consider an analytic function (s) defined by a Dirichlet series whose axis of convergence is σ = 0, where 0 ≤λ_n< + ∞ and lim (log n/λ_n) = 0. We define the Ritt order and proximate order of f(s) in σ > 0 and have obtained some relations between the growth of f(s) and the coefficients, which extend some of G. Valiron's results.2. Let (Ω, P) be a probability space. Consider a random Dirichlet series whose abscissa of convergence is σ(ω), where a_n(ω) are random variables in (Ω,P) and λ_n satisfy the same conditions as in 1(n = 0,1,2,…). To σ(ω) we extend L. Arnold's results on the radius of a.s. convergence of random Taylor series and have solved some of his problems.In the case σ(ω) = 0 a.s. the results mentioned in 1 is applied to studying the a.s. growth of f(s, ω) in σ> 0. As a special case we find that if |a_n(ω)| are independent and have the same non-degenerate distribution function F(x) and if the radius of a.s. convergence of g(s, ω)is 1, then the a.s. growth of g(s, ω |z|<1 ean be determined by the convergence or divergence of(logx)~k dF(x) (k≥1).3. Let the probability space in 2 be such that Ω = [0, 1], is composed of all Lebesgue measurable sets E on Ω and P[E] is the Lebesgue measure of E. Let a_n(ω)= b_nε_n(ω) or b_nγ_n(ω), where {ε_n(ω)} or {γ_n(ω)}is a Rademaeher or Steinhaus sequenee and b_n are constants such that the abscissa of convergence of f(s, ω) is zero. Then, under certain conditions every point on σ = 0 is a Pieard or Borel (R) point of f(s, ω) a.S.
  • Acta Mathematica Sinica, Chinese Series. 1978, 21(2): 119-129. https://doi.org/10.12386/A1978sxxb0013
    Abstract ( )   Knowledge map   Save
    The theory of linear symmetric positive systems of PDE is extended to the quasilinear ease. An existence theorem for a large elass of boundary value problems is established provided the systems are sufficiently positive. Using this theorem we prove the existences of solutions of some boundary problems for a certain class of 2nd order quasilinear mixed PDE with n independent variables. The equations can not be linealized by any known transformation and the elliptie regions and hyperbolie regions can not be identified before the solutions are determined.
  • Acta Mathematica Sinica, Chinese Series. 1978, 21(2): 130-134. https://doi.org/10.12386/A1978sxxb0014
    Abstract ( )   Knowledge map   Save
    Certain initial-boundary value problems for quasilinear hyperbolie symmetrie systems of PDE are eonsidered. The eoeffieients of the systems are independent of the time variable. It is proved that if the system is sufficiently positive a global solution of the problem exists and that the solution tends to the solution of the eorresponding stationary boundary value problem as t→∞.
  • Acta Mathematica Sinica, Chinese Series. 1978, 21(2): 135-150. https://doi.org/10.12386/A1978sxxb0015
    Abstract ( )   Knowledge map   Save
    The discounted Markovian decision programming of our concern consists of statespace and action-set eorresponding to every state, both of whieh are denumerable infinite sets, substochastie transition law family and boundedreward function We have given a succassive approximation method of accelerative eonvergence for (ε-) optimal stationary policy, This algorithm converges to the optimal solution more qurckly than White's successive approximation method. It has also been furnished with a testing criterion for non-optimal policy, making the algorithm more efficient.Let β be the discounting factor. Generally the β(or (ε,β))- optimal stationary policy is often not unique, and even has as many policies as contained in the stationary policy class, It is natural to hope that a policy with homogeneously (ε-) minimized variance (to the initial states) be found in the β(or (ε, β))- optimal stationary policies. We have proved that a poliey of this kind does exist, and have given an algorithm for this poliey.
  • Acta Mathematica Sinica, Chinese Series. 1978, 21(2): 151-160. https://doi.org/10.12386/A1978sxxb0016
    Abstract ( )   Knowledge map   Save
    For a class of systems K, R. J. Diperna has proved, by using Glimm scheme, the existence of the solution of (1.1) with prescribed initial data having finite total variation. In §§2-3 of the present paper, we construct the solution of (1.1) with prescribed step initial data by using the method of charaeterstic parameter. In §4, we extend the existence theorem of §§2-3 to the initial problem studied in [1].
  • Acta Mathematica Sinica, Chinese Series. 1978, 21(2): 161-170. https://doi.org/10.12386/A1978sxxb0017
    Abstract ( )   Knowledge map   Save
    In this paper, we introduce three topologies in the perfect matrix algebras ∑(λ). The neighborhood systems. V,(θ,N,M,ε), V_k(θ,N,M,ε), V_ω(θ,N,M,ε) of the point θ ∈ ∑(λ) are respectively called strong, k-and weak topologies, where ε > 0 and N, M are respectively the bounded, weak compact and finite sets of λ, λ. Obviously, ∑(λ) are locally convex algebras relative to the above three topologies.Our principal results are as follows:Theorem 1. The following propositions are equivalent:1°In ∑(λ), multiplication is eontinuous relative to the strong topology;°In λ there exists the bounded set N_o, whieh absorbs any bounded set N of λ, i.e. N a N_o for some a > 0;3°λ is a Banaeh space relative to the strong topology;4°∑(λ) is a Banach algebra relative to the strong topology;5°∑(λ) is a m-convex algebra relative to the strong topotogy, i.e. in ∑(λ) there exists a base for the strong neighborhood system {V_s} of the point θ∈∑(λ) satisfying V_s.V_s V_s.Corollary 1. If ∑() is not a Banach algebra relative to the strong topology, then it is also not a B_o-algebra.Corollary 2. If λ is perfect and convergence-free, then ∑(λ) is multiplicatively discontinuous and it is not a B_o-algebra relative to the strong topology.Corollary 3. If "gestufen" space λ is generated by the enumerable, non-negative and increasing "stufen" system {B~((n))}, then ∑(λ) is multiplieatively discontinuous and it is not a B_o-algebra relative to the strong topology.Theorem 2. The following propositions are equivalent:1°In ∑(λ), multiplication is continuous relative to the k-topology;2°λ and λ are sequentially separable, and in λ there exists the bounded set N', which absorbs any bounded set N of λ;3°In ∑(λ), multiplication is continuous relative to the strong topology which is equivalent to the k-topology.Theorem 3. In ∑(λ), multiplication is impossible to be eontinous relative to the weak topology.
  • Acta Mathematica Sinica, Chinese Series. 1978, 21(2): 171-175. https://doi.org/10.12386/A1978sxxb0018
    Abstract ( )   Knowledge map   Save
    In this paper, a necessary and sufficient condition for produciblity of CW-complexes is shown as follows.Theorem. Let K and L be CW-complexes. Suppose that either (i) K or L is locally finite, or (ii) K and L are locally countable, then product complex K × L is a CW-complex. Furthmore, under the Cantor's continuum hypothesis the converse is also true.Thus we solve a problem from J. H. C. Whitehead [4; p. 563].
  • Acta Mathematica Sinica, Chinese Series. 1978, 21(2): 176-186. https://doi.org/10.12386/A1978sxxb0019
    Abstract ( )   Knowledge map   Save
    For a linear system of differential equations with constant coefficients. dx/dt=Ax(1)It. is well-known thai, if all eigenvalues of A have negative real parts, then the trival solution of (1) is asymptotieally stable.However, the problem is much more difficult for a linear time-varying system dx/dt=A(t)x in the relation between the signs of Re(λ_i(t)) of the eigenvalues of A(t) and the stability of the dynamical systems. Examples of all kinds of the combinations are given in this paper to show its eomplexity. But if the rate of change of the elements a_(ij)(t) of A(t) is sufficiently slow, by using the method of Lyapunov funetion, the problem ean be solved thoroghly. In addition, an explieit bound for the a_(ij)'(t) is given, no matter what form matrix A(t) has.