On the problem of Goldbach's type (Q2277012)

From MaRDI portal





scientific article; zbMATH DE number 4193810
Language Label Description Also known as
default for all languages
No label defined
    English
    On the problem of Goldbach's type
    scientific article; zbMATH DE number 4193810

      Statements

      On the problem of Goldbach's type (English)
      0 references
      1992
      0 references
      Let \(P\) and \(P'\) be pseudo-primes, i.e. positive integers with no more than \(m\) prime factors, where \(m\) is some preassigned bound. Let \(N\) be an even integer. From Chen's Theorem [\textit{J. Chen}, Sci. Sin. 16, 157--176 (1973; Zbl 0319.10056)], it is known that the equation (1) \(N=P+P',\) where the total number of prime factors of \(P\) and \(P'\) does not exceed 3, is solvable for every sufficiently large \(N\). In ``The dispersion method in binary additive problems'', (Providence 1963, cf. a review of the Russian original in Zbl 0099.03104), p. 4, \textit{Yu. V. Linnik} remarked, ``the discovery of the appropriate formula for (1) would substantially facilitate the solution of the binary Goldbach problem''. Under the extended Riemann hypothesis, in 1967 \textit{R. J. Miech} [J. Reine Angew. Math. 233, 1--27 (1968; Zbl 0191.05001), pp. 1--2, Theorem] proved that the number of solutions of the equation (2) \(N=n(r)+n(s)\) where, for each \(r\), \(n(r)\) is an integer that is the product of \(r\) distinct primes, all of which are greater than a certain parameter, satisfies an asymptotic relation for nearly every integer \(N\). Miech's result, according to Miech himself, ``is a heuristic one as far as the dispersion method is concerned; its utility is limited by its probabilistic form''. Let \(p,p_ i,q_ j\) denote positive odd primes. Set \[ C_ N=\prod_{p>2}(1-(p-1)^{-2})\prod_{2<p| N}(p-1)(p-2)^{-1}. \] In this paper, without invoking any unproved hypothesis, the author obtains the following results. Theorem 1. Let \(\zeta\),\(\delta\) be fixed numbers with \(0<\zeta <1\), \(0<\delta <1\). For every sufficiently large even integer \(N\) and any \(s\geq 1\), \(r\geq 2\), \[ \#(a :\;a=N-q_ 1...q_ s,\;(a,N)=1,\;a=p_ 1\dots p_{r-1}\text{ or } p_ 1\dots p_ r,\;\exp (\ln^{\zeta}N)<q_ 1<\dots <q_ s,\;\exp (\ln^{\delta}N)<p_ 1<\dots <p_{r-1}<p_ r\} \] \[ >\frac{0.77(1-\zeta)^{s-1}(1-\delta)^{r-2}}{(s-1)!(r-2)!}C_ NN \ln^{-2}N(\ln \ln N)^{s+r-3}. \] Theorem 2. For every sufficiently large even integer \(N\) and any \(s\geq 1\), \(r\geq 2\), \[ \#\{a :\;a=N-q_ 1\dots q_ s,\;(a,N)=1,\;a=p_ 1\dots p_{r-1}\text{ or } p_ 1\dots p_ r,\;q_ 1<\dots <q_ s,\;p_ 1<\dots <p_{r-1}<p_ r\} \] \[ >\frac{0.77}{(s- 1)!(r-2)!}C_ NN \ln^{-2}N(\ln \ln N)^{s+r-3}. \] The results here imply Chen's Theorem \((s=1,r=2)\) and the author's [J. Reine Angew. Math. 414, 117--130 (1991; Zbl 0702.11065), Theorems 1 and 2] \((s=1,r\geq 3)\) as special cases.
      0 references
      Goldbach type theorems
      0 references
      pseudo-primes
      0 references
      Chen's Theorem
      0 references
      Miech's result
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references

      Identifiers