Large sieve

Large sieve
Name	Large sieve
Field	Analytic number theory
Introduced	1940s–1960s

Large sieve

The large sieve is a collection of analytic techniques in Analytic number theory and Harmonic analysis that provide upper bounds for the distribution of integer sequences with respect to congruence or frequency conditions. Originating in mid‑20th century work by Y. Linnik, K. Roth, and later formalized by P. D. T. A. Elliott, Enrico Bombieri, and H. Davenport, the large sieve connects exponential sum estimates, bilinear forms, and Hilbert space methods to yield powerful inequalities used across Prime number theorem refinements, sieves of Brun, and equidistribution problems. The theory interacts with results of Paul Erdős, Atle Selberg, John von Neumann techniques, and later developments by Henryk Iwaniec, Emmanuel Kowalski, and Henryk Montgomery.

History and development

The historical development traces contributions from Yakov G. Sinai‑era probabilistic ideas, the work of Harold Davenport on exponential sums, and the pioneering sieving heuristics of Viggo Brun that influenced J. E. Littlewood and G. H. Hardy. Early formulation by J. Roth and Y. Linnik led to quantitative statements exploited by A. Selberg, while systematic operator and Hilbert space views were emphasized by A. O. Gel'fond contemporaries and expanded in the 1960s by Enrico Bombieri and Patrick X. Gallagher. Subsequent refinements employed techniques from Paul Erdős combinatorial number theory, methods of Atle Selberg's sieve, and spectral ideas reminiscent of Selberg trace formula treatments by Hillel Furstenberg and Yuri Manin.

Statement and main inequalities

Canonical formulations present a family of inequalities bounding sums over sets of residues or frequencies. Typical statements compare sums over moduli drawn from sets related to Dirichlet characters and Fourier analysis bases familiar to John von Neumann spectral theory. Key inequalities include the original large sieve inequality in the form of a quadratic form bound due to P. D. T. A. Elliott and Enrico Bombieri, the dual large sieve attributed to H. L. Montgomery, and the large sieve with multiplicative characters linked to Dirichlet L-series and results of Bernhard Riemann‑inspired frameworks. Quantitative versions often cite constants and ranges studied by Henryk Iwaniec and Emmanuel Kowalski in their monographs.

Proofs and methods

Proofs rely on orthogonality relations from Fourier analysis and completeness arguments akin to those in Hilbert space theory developed by John von Neumann. The duality principle, attributed to H. L. Montgomery and influenced by Stefan Banach functional analysis, transforms direct estimates into dual quadratic form bounds. Other methods invoke the dispersion method used by Linnik and the large sieve inequality with weights via bilinear form decomposition popularized by Enrico Bombieri and A. I. Vinogradov. Further techniques draw upon trace formula analogues in the style of Atle Selberg and spectral interpretations inspired by Marshall Stone and Gelfand representation theory.

Applications in number theory

The large sieve underpins many results about the distribution of primes and almost primes in arithmetic progressions related to Dirichlet's theorem on arithmetic progressions and refinements of the Prime number theorem in short intervals used by Ivan Vinogradov. It plays a central role in estimates for exponential sums appearing in research of Nikolai Korobov and Willem van der Corput, and in bounding exceptional character sums studied by Dudley H. Lehmer collaborators. The method is instrumental in sieve problems associated with Brun's sieve, work on gaps between primes pursued by Goldston, Pintz, and Yıldırım teams, and in additive problems linked to Vinogradov's theorem and results by Terence Tao and Ben Green. It also contributes to zero density estimates for Dirichlet L-series and results concerning multiplicative functions considered by Elliott and Halász.

Variants and generalizations

Numerous variants include the multiplicative large sieve, the combinatorial large sieve developed in the spirit of Paul Erdős arguments, and harmonic large sieve formulations related to Fourier transform analogues in automorphic forms studies by Henryk Iwaniec and Erez Lapid. Generalizations incorporate weighted inequalities, bilinear form versions inspired by Enrico Bombieri, and geometric large sieve adaptations used in algebraic geometry contexts by Pierre Deligne and Nicholas Katz for equidistribution of Frobenius traces. Probabilistic and random matrix analogues evoke perspectives from Eugene Wigner and intersect with conjectures of Montgomery on pair correlation influenced by Bernhard Riemann.

Examples and numerical illustrations

Concrete examples show bounding the discrepancy of sequences modulo primes as in classical treatments by Harold Davenport and refined numerical bounds by Enrico Bombieri. Numerical illustrations often reference sample computations from works of Henryk Iwaniec and Emmanuel Kowalski showing savings in sieve weights compared to naive estimates used by Brun and Legendre‑style counts. Case studies include distribution of quadratic residues linked to Carl Friedrich Gauss and computations of character sums echoing experiments by D. A. Burgess and later numerical verifications aligned with conjectures by Atle Selberg and N. G. de Bruijn.

Category:Analytic number theory