On the elementary proof of the prime number theorem with a remainder term (Q1974427)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | On the elementary proof of the prime number theorem with a remainder term |
scientific article |
Statements
On the elementary proof of the prime number theorem with a remainder term (English)
0 references
11 May 2002
0 references
The prime number theorem states that the prime counting function \(\pi(x)\) has the asymptotic value \(x/\log x\) as \(x\to\infty\). This was proved in 1896 by Hadamard and de la Vallée Poussin independently using complex function theory. About a hundred years earlier still Gauss had already predicted that the logarithmic integral \(\text{ li}(x)\) would serve as a better estimate for \(\pi(x)\), and the difference \(\pi(x)-\text{ li}(x)\) is now called the remainder term of the prime number theorem. In 1899 de la Vallée Poussin deduced from a zero-free region that he obtained for the Riemann zeta-function \(\zeta(s)\) the bound \(O(x\exp(-A\sqrt{\log x}))\) for the remainder term; here the positive constant \(A\) may be made explicit. In 1958, with further widening of the said zero-free region, Vinogradov and Korobov improved the bound to \(O_\theta(x\exp(-\log^{\theta}x))\), for any \(\theta<{3\over 5}\). In 1949 Selberg found an elementary proof of the prime number theorem, and his method was modified by several mathematicians to give a bound of the form \(O_A(x/\log^{A}x)\) for the remainder term. This was further improved to \(O_\theta(x\exp(-\log^{\theta}x))\) for any \(\theta<{1\over 7}\) by \textit{H. Diamond} and \textit{J. Steinig} [Invent. Math. 11, 199-258 (1970; Zbl 0201.05402)], and then for any \(\theta<{1\over 6}\) by \textit{A. F. Lavrik} and \textit{A. S. Sobirov} [Sov. Math., Dokl. 14, 1063-1066; translation from Dokl. Akad. Nauk SSSR 211, 534-536 (1973; Zbl 0288.10014)]. Continuing with this modified form of the elementary method, the author has succeeded in establishing the bound for any \(\theta<{1\over 2}\), a result which is nearly as good as that obtained by de la Vallée Poussin using function theory. The proof is based on an identity of \textit{A. Balog} [Acta Math. Acad. Sci. Hung. 37, 285-299 (1981; Zbl 0418.10042)] involving Dirichlet convolutions of the Möbius function and powers of the logarithm function. A recurrence relation is obtained in order to deduce the required bound, and the inductive argument makes use of a result of \textit{E. Wirsing} [J. Reine Angew. Math. 214/215, 1-18 (1964; Zbl 0166.31004)] on the mean-square of the convolution of two measurable functions. Although all the details are given in a well organised and carefully written paper, there is no indication or discussion on what and where the obstacle is that limits the method to \(\theta<{1\over 2}\).
0 references
prime number theorem
0 references
elementary proof
0 references