中国科学院数学与系统科学研究院期刊网

15 January 1955, Volume 5 Issue 1
    

  • Select all
    |
    论文
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(1): 1-25. https://doi.org/10.12386/A1955sxxb0001
    Abstract ( )   Knowledge map   Save
    Let n≥2 and z=(z_1,...,z_n) be an n-dimensional complex vector, denotes the space formed by vectors satisfying (1) and (2), where z’denotes the transposed column vector obtained from z. By means of spherical harmonics and the invariants of orthogonal group and its related techniques, we obtained an orthogonal system of. Moreover, the characteristic manifold of is defined by |zz′| = 1, zz′= 1.Cauchy formula has been obtained, which asserts that the function is uniquely determined in by its values on. Consequently, we proved that the maximal of the modulus of a function analytic in and on is taken on. From it we deduce several theorems which are refinements of Hadamard's three sphere theorem.
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(1): 27-36. https://doi.org/10.12386/A1955sxxb0002
    Abstract ( )   Knowledge map   Save
    Let a_1, a_2, …, a_n be n distinct finite points in the z-plane; G_1, G_2 G_n being a system of non-overlapping domains in the z-plane such that a_i ∈ G_i (i=1, 2…n), we denote the mapping radius of G_i with respect to a_i by R(a_i, G_i). Golousin invesfigated the extremal domains G_1,…, G_n, for which the product becomes maximum. He obtained a system of differential equations for the extremal functions corresponding to the extremal domains. Although Golousin's result has been simplified by the authort, the problem of determination of the extremal domains still leaves open. The first aim of present note is to establish the following:Theorem 1. Let P(z) be any polynomial such that (i) its degree is equal to or less than 2n-3, (ii) and (iii) for the roots z_1, z_2,…, z_m of P(z)=0, (The existence of P(z) is known.) Then the domains with the boundaries defined by the equation are extremal for the product .Further, the corresponding mapping functions are the inverse functions ofCorollary. If a_k=ae~(2πki/n)(k=1, 2,..., n), then In this case the ,boundary of the extremal domains are n rays ρexp{2π(ν+1/2)/n i+iarga} (0≤ρ≤∞, ν=1, 2,…, n) issuing from the origin.The proofs for the above results are based on the method of extremal length due to Ahlfors and Beurling.By the same method, we can investigate the problem proposed by Grotzsch and Lavrentieff. The problem is equivalent, as shown by Golousin to the statement: Given n points a_1,… a_n in the z-plane, let G be any region which contains z=0 but does not the points a_1,…,a_n and ∞. Determine the upper bound of the mapping radius of G with respect to z=0.The extremal domain G=G(0; a_1, a_2,…a_n) is given by the following:Theorem 2.Let (z) be any polynomial such that (i) the degree m is equal to or less than n-1, (ii)and (iii) for the roots a_n+1, a_n+2,…, a_(n+m)of (z)=O, Then the curve defined by the equation gives the boundary of an extremal domain G(0; a_1,…,a_n). The inverse function of the corresponding extremal function is As a corollary of this theorem, we have R(0,G)≤4~1/n)|a| for a_k=ae~((2πk/n)i), k=1, 2,…, n.
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(1): 37-63. https://doi.org/10.12386/A1955sxxb0003
    Abstract ( )   Knowledge map   Save
    The aim of this paper is to give a second proof that the Pontrjagin classes reduced mod 3 of a closed differenfiable manifold are topological invariants of that manifold. The proof is different from a preceeding one in that Steenrod powers do not come explicitely in the considerations. A comparison of the two proofs which exhibits besides some general aspects about Steenrod powers will be given later.
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(1): 65-80. https://doi.org/10.12386/A1955sxxb0004
    Abstract ( )   Knowledge map   Save
    This note gives answer to problem 103 in G. Birkhoff's "Lattice Theory" (revised edition, 1948). The problem is:Let G be the additive group of all real vectors (a, b), ordered by defining (a, b)≧(c, d) to mean "either a>c, or a=c and b≧d." Is it possible to define multiplication in G so as to make it into an ordered ring?The answer is positive. Our results and discussions are summarized below.In §1, it is pointed out that G can be made into an ordered ring by each of the following types of multiplication: (a, b) (c, d) = (0, acβ), β≧0.(Ⅰ) (a, b) (c, d) = (acα, g(ac)), α>0.(Ⅱ) (a, b) (c, d) = (acα,bcα),α>0.(Ⅲ) (a, b) (c, d) = (acα, adα), α>0.(Ⅳ)(a, b) (c, d)= (acα, acβ+adα+bcα),α>0,β arbitrary.(Ⅴ) where α, β are arbitrary constants within restrictions given in each case, and g(x) in (Ⅱ) is any one of a class of functions defined as follows:Assuming the axiom of choice, and hence assuming that the set of real numbers can be arranged into a Well-ordered set W. Define g(x) on W by transfinite induction: Suppose that g(x) has been consistently defined for all numbers preceding aλ in W. (D_1) If aλ can be "rationally expressed" by a finite set of numbers preceding it in W, i.e. if aλ=r_1 aλ_1 +…+ r_k aλ_k(r_1, ..., r_k rational; k finite) for some aλ_1, …, aλ_k preceding aλ in W, then take any one of such expressions, say the above one, and define g(aλ) =r_1 g(aλ_1)+... + r_kg(aλ_k) . (D_2) If no such expression exists, then take a real number β_λ arbitrarily and define g(aλ) = aλβ_λ.The consistency of the above definition is proved. And it is pointed out that what we have defined is the widest class of real functions satisfying the condition g(a+b)=g(a)+g(b) . (including linear functions "ax" as special cases.)In §2, we consider the question: How many possible types of multiplication are there on G?Suppose that G has been made into an ordered ring, in which (1,0)~2= (α,β), (0,1)~2= (Υ, δ), (1,0) (0,1)=(ζ, η), (0,1) (1,0)=(θ,κ). It is shown that Υ=δ=ζ=θ=0, αη=η~2, αk=k~2, α≧0,and β≧0 when α=0 . Then the discussion proceeds along several cases:When α=η=k=0, β≧0, the multiplication of any two vectors in G satisfies the formula: (α,b) (c,d)=(0, αcβ). (Ⅰ)When α>0,η=κ=0, the multiplication of any two vectors in G satisfies the formula: (a, b) (c,d)=(acα, g(ac)); (Ⅱ) where g(x) is any one of the functions defined in §1 and satisfying g(1)=β.In other cases, only partial results are obtained, the complete determination of the multiplication formulas seems difficult.In §3, we examine the isomorphism or non-isomorphism between the ordered rings having G as additive group. It is found that a large part (nearly all) of them fall into six classes of isomorphic ordered rings. (The automorphism groups of five of these classes are also determined.) As a consequence of these considerations, some of the unsatisfactory points (e.g. the use of the axiom of choice in defining (Ⅱ).) in the above sections become algebraically nonessential.
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(1): 81-89. https://doi.org/10.12386/A1955sxxb0005
    Abstract ( )   Knowledge map   Save
    In a previous paper [Ⅰ] we proved the theorem, concerning the behaviour of the sumnability (c, r) for negative indices of the Fourier series of a monotonic function with an infinite limit. It is natural to inquire what happens when the indices are positive.Let be the Fourier series of an L-integrable function f(θ), and σ_n~r(θ) the r-th Cesaro mean of the series at the point θ.Define φ(t) = 1/2 [f(θ + t) + f(θ - t)]. We shall be concerned with behaviour at a single point θ, which we may suppose to be θ=0. And then σ_n~r(O) is the r-th Cesaro mean of the Fourier series of the even function φ(t) at the point t=0.Theorem (i) If r≥1 and limφ(t) = + ∞, then (ii) If0
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(1): 91-108. https://doi.org/10.12386/A1955sxxb0006
    Abstract ( )   Knowledge map   Save
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(1): 109-115. https://doi.org/10.12386/A1955sxxb0008
    Abstract ( )   Knowledge map   Save
    The primitive recursive functions may be defined as the functions definable from the initial function Sx=x + 1 by means of a finite number of substitutions and the following scheme of recursive definition: where u is a parameter. There are several particular cases of this scheme. Rosza Peter and R. M. Robinson proved that if we suitably add some new initial functions we may define all the primitive recursive functions by means of some particular scheme alone. The present paper is to extend the result to other particular schemes. The results may be summarized in the following table in which we list the new initial functions needed for each particular scheme. Those marked with were obtained by R. M. Robinson.In the table "a" denotes a preassigned fixed natural number.
  • Acta Mathematica Sinica, Chinese Series. 1955, 5(1): 117-135. https://doi.org/10.12386/A1955sxxb0009
    Abstract ( )   Knowledge map   Save
    Among various axiom systems for the traditional two-valued logical system, the one given by Hilbert-Bernays in Grundlagen der Mathematik perhaps is the best. It possesses the following advantages: first, it divides the axioms into five groups according to the connectors involved and shows the essential properties of each of them; second, it makes the distinction between the three important logical systems (the traditional system, the intuitionistic system and the minimalkalkul) quite clear. However, it has a shortage that no group of axioms is sufficient. By the sufficiency of groups of axioms we mean that if a proposition can be deduced in the whole system then it can be deduced by means of the groups of axioms involving the same connectors only. The present paper is to remedy this shortage. To meet various requirements we give several systems. The difference of them is either by the number of axioms in each group or the method (adding or strengthening axioms) for distinguishing the three logical systems mentioned above. They may be shown in the following table: Among the main results we have:If we add the proposition "CCCpqpp" to an axiom system of the intuitionistic system we get an axiom system of the traditional system. If every group of its axioms is sufficient the same for the result system.If in a consistent system we can deduce the following proposition and rules: CpCqp; Cαβ→CCβΥCαΥ, CCΥαCΥβ; β, CαCβΥ→CαΥ then any finite number of the rest organic axioms of the form "Cαβ" may be combined into a single organic axiom, provided α may turn into an asserted proposition after a suitable substitution.