中国科学院数学与系统科学研究院期刊网

15 July 1955, Volume 5 Issue 3
    

  • Select all
    |
    论文
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 285-294. https://doi.org/10.12386/A1955sxxb0020
    Abstract ( )   Knowledge map   Save
    As we know, the Riemann ζ function has many interesting ways of generalization. Here I present still a different one.Let n be an even integer, then I define for σ>kv, where "′" denotes that x_1,…,x_n do not vanish simultaneously. It is easy to generalize a little bit more so as to include the Epstein's Z function as a special case. However, I prefer to deal with the more special case for it is more interesting.Let B(m) denote the number of integer solutions of the equation then, for σ>kv, Let and then ω(x) is bounded and furnishes the analytic continuation of Z_(n,k)(s) to the left of σ=kv-v, except at the sole singular point s = kv, which is a simple pole with the residue 2~kΓ~k(1+v)/Γ(kv).To show that Z_(n,k)(s) can be continued to the whole plane we have to make use of the representation which is true for σ>kv By Poisson's formula we find that Since h can be taken as large as we please, (5) shows that Z_(n,k)(s) can be continued analytically to the half plane σ>kv. For h=0, we must intet the 0-th derivative as the function itself. So (5)can be written in a more symmetrical form as follow: where denotes that the sum is taken to be Cauchy's principal value. It is worth while to note that for n=2, (6) gives the functional equation for the Epstein Z function Z_(2,k)(s).
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 295-311. https://doi.org/10.12386/A1955sxxb0021
    Abstract ( )   Knowledge map   Save
    1. We generalize one of Mandelbrojt's theorems Considering Valiron's product we can replace Mandelbrojt's condition C by the following: C~n. Suppose n runs over all the positive integers, and suppose and we can study the representation of F(s) by instead of that of F(s) by In the present case, Mandelbrojt's theorem can be easily modified, and the inequality in the conclusion takes a more convenient form: where2. The Cauchy-Hadamard formula on the radius of convergence of power series can be extended to certain generalized Dirichlet series with complex exponents. Our results are also extensions of those of Hille and,and our form is more convenient. As a consequence of our general theorem, we have, for example, the following result:Theorem If {λ_n}satisfies the condition C′and the following conditions: then the series converges in a half-plane bounded by and it diverges in a half-plane bounded by3. According to Mandelbrojt's method, the preceding results can be used to detect the singularities of a nction represented by some generalized Dirichlet series From our general theorem, we deduce, for example, the following:Theorem Suppose {λ_n}°satisfies C′and (3) and the following conditions:C_1′. For x is sufficiently large, the number of λ_n, whose modulii lie between x and x + 1, is less than a fixed positive number K;C_2~′. Given any δ>0, we have, for n sufficiently large, then each point on the boundry of the half-plane of convergence of is a sigularity of the function represented by the series.4. Applying the inequality (1), we can extend our previous results on the growth of entire functions defined by the Dirichlet series to the case of generalized Dirichlet series whose exponents satisfy the conditions C′, C_1~′ and C_2~′
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 313-324. https://doi.org/10.12386/A1955sxxb0022
    Abstract ( )   Knowledge map   Save
    Let d_k(n)be the numbers of expressions of n as k factors, and let D_k(x) R_k(x) = (a_(ko) + a_(k1) In x +…+a_(k,k-1) In~(k-1) x) x,(x>0) be the residue of ζ~k(s)x~s/s at s=1. Define The following two theorems are proved in this paper.Theorem 1. There exist two positive constants c_k and C_k, such that for every x≥1,and every t lying in the interval the equation △_k(y)=t has, at least, one solution lying in the interval [x, x+C_k x~(1-1/k)].Theorem 2. There exists a positive constant C_k, such that for every λ≥1 and every x≥1, the following inequality holds:Theorem 1 is a little better than Hardy's resultsThe particular case λ= 2 of theorem 2 is a little better than the result of Titchmarsh:The proof of these two theorems is based on an identity, the generalized s identity: given by the author.
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 325-332. https://doi.org/10.12386/A1955sxxb0023
    Abstract ( )   Knowledge map   Save
    Norgen has defined a pair of affine connections G_(ik)~k, Γ_(ik)~i as conjugate connections with respect to a symmetric tensor b_(ij), when they satisfy the condition In a recent paper the author has generalized this notion to the case of a symmetric tensor of order m, and has solved the problem of determing the mth connection from the given m-1 connections. We have system of equations and for the consistency and uniqueness of the solution of the system of equations we obtained some restrictions which must be added to the tensor or the m-1 connections.In the present note we consider n conjugate connections with respect to an antisymmetric tensor ei_1 i_2 ... i_n and prove the following theorem:If n connections are conjugate with respect to an antisymmetric tensor ei_1 i_2 ... i_n, then(i)one of them for example, can be choosen arbitrarily,(ii)the remaining connections are common-pseudo with(hence mutually common-pseudo),(iii)the common-pseudo vectors must satisfy the following condition:As a geometrical interpretation of these connections we have: If there exist a volume in an affinely connected space formed by n vectors such that it remains unaltered when these vectors undergo parallel displacements respectively with regard to the connections then these connections must satisfy the conditions (i), (ii), (iii).If G_(ik)~i is such an affine connection in our space, then we can always find n connections common-pseudo with G_(ik)~i such that the volume formed by certain n vectors remains unaltered when the forming vectors undergo parallel displacements with regard to respectively.If the connection G_(ik)~i is equi-affine and any one of the n connections is commonpseudo with it, then the directions of the vectors remain the same when they are parallelly displaced under respectively.Especially, when the n connections are all symmetric, then they become one and the same.
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 333-346. https://doi.org/10.12386/A1955sxxb0024
    Abstract ( )   Knowledge map   Save
    In this paper scalar quantities are complex numbers. The transformations studied in this paper are those of the form A→B = PAP~(-1), carrying a square matrix A into a square matrix B, where P is any non-singular matrix and P is formed of the conjugate imagineries of the elements of P. Any such transformation will be called a transformation. If two square matrices A and B can be transformed into each other by a transformation, they are said to be similar. We give an enumeration of the results arrived at in this paper. 1°Any square matrix A can be transformed by means of transformations into a certain canonical form which is uniquely determined by A.2°The similarity of the matrices is a necessary and sufficient condition for the similarity of two square matrices A and B; the same may be said of the similarity of the matrices3°Necessary and sufficient conditions (in terms of elementary divisors) are found for a square matrix to be(i) expressible in the form AA,(ii) similar to a matrix of the form(ii) similar to a matrix of the form 4°Any square matrix is similar to a real matrix.5°Any square matrix can be expressed as a product HS (or SH), where H is a Hermitian matrix and S is a symmetric matrix, a pre-assigned one of which is non-singular.We have also studied the natural extension of transformations, namely transformations of the following form: A_1 → B_1 = PA_1Q, A_2 → B_2 = PA_2, (P and Q non-singular) carrying a matric pair (A_1, A_2 ) into another matric pair (B_1, B_2). The results will be published in another paper.
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 347-368. https://doi.org/10.12386/A1955sxxb0025
    Abstract ( )   Knowledge map   Save
    This paper contains various distributional results about the ratio F_n(x) :F(x), where F(x) is a distribution function, and F_n(x) is the corresponding empirical distribution function formed of a sample of size n. The main results are:(A) Concerning stability. It is proved that, as n→∞, where a_n is any sequence of positive numbers tending to infinity.(B) Concerning exact distributions. We have derived exact formulae for the following distribution functions: where 0
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 369-381. https://doi.org/10.12386/A1955sxxb0026
    Abstract ( )   Knowledge map   Save
    ~~
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 383-392. https://doi.org/10.12386/A1955sxxb0027
    Abstract ( )   Knowledge map   Save
    ~~
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 393-399. https://doi.org/10.12386/A1955sxxb0028
    Abstract ( )   Knowledge map   Save
    En considerant la question de l'unicite de solution quasi-reguliere du probleme de Tricomi pour l'equation de Chaplygin nous avons constate: Si F(y)=2(K/K′)′+1 a un zero-1/ad′ordre impair, et que l'integrale ulors K a necessairement un point singulier audessous du point-1/a.Pour la question de l'unicite, nous avons demontre le theoreme: Si F(y) prend des valeurs negatives dans le domaine hyperbolique, en designant et que le support dans le domaine elliptique du probleme de Tricomi se trouve compris entre les deux droites X = a, X = a+2π/β, a constante, alors l'unicite de solution quasi-reguliere sera assuree.Quand le sommet du triangle caracteristique dans le domaine hyperbolique tend vers un point de la droite y= -1/aou-1/a est le plus grand zero de F(y), β tend vers zero, on retrouve le resultat de F. I. Frankl.
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 401-410. https://doi.org/10.12386/A1955sxxb0029
    Abstract ( )   Knowledge map   Save
    This paper gives the explicit formula for classes introduced in Ⅱ of this series of papers as follows. Let M be an oriented closed differentiable manifold of dimension m. For any odd prime p, let S_p~(2i(p-1)), T_p~(2i(p-1)) ∈H~(2i)(p-1))(M, I_p) be defined by S_p~(2i)(p-1)) U X~(m-2i(p-1))=X~(m-2i(p-1)),X~(m-2i(p-1)) ∈H~(m-2i(p-1)) (M, I_p), and in which are Steenrod powers in the new notations of J. P. Serre. Then the classes are given by In particular, for p=3, we deduce for the mod 3 Pontrjagin classes P_3~(4j) of the oriented manifold M: P_3~(4j)=T_3~(4j).As a corollary, we get the following result: In an oriented closed differentiable manifold M of dimension 2(p-l) or 2p-1, Q_p~(2(p-1) = 0. In particular, we have p_3~4 = 0 in an oriented closed differentiable manifold of dimension 4 or 5.
  • P.TURN
    Acta Mathematica Sinica, Chinese Series. 1955, 5(3): 417-423. https://doi.org/10.12386/A1955sxxb0031
    Abstract ( )   Knowledge map   Save