This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| Khinchin–Kolmogorov theorem | |
|---|---|
| Name | Khinchin–Kolmogorov theorem |
| Field | Probability theory |
| Introduced | 1920s |
| Named after | Aleksandr Khinchin; Andrey Kolmogorov |
Khinchin–Kolmogorov theorem The Khinchin–Kolmogorov theorem is a fundamental result in probability theory connecting sequences of independent random variables, almost sure convergence, and series of variances; it establishes criteria guaranteeing almost sure convergence of normalized sums under moment and independence conditions. The theorem sits at the crossroads of limit theorems developed by Aleksandr Khinchin, Andrey Kolmogorov, Emil Borel, Émile Lévy, and Paul Lévy and plays a central role in modern probability theory alongside results of Andrey Markov, Kolmogorov's zero–one law, Borel–Cantelli lemma, and the Strong law of large numbers.
The theorem asserts that for a sequence of independent real-valued random variables {X_n} with zero means and variances Var(X_n)=σ_n^2, the series Σ σ_n^2 < ∞ implies almost sure convergence of the series Σ X_n, and conversely under suitable triangular array conditions. This statement refines and complements the Three-series theorem, the Kolmogorov three-series theorem, and classical convergence criteria appearing in works by S. N. Bernstein, Paul Erdős, Norbert Wiener, and Andrey Kolmogorov. It links directly to criteria used in proofs by William Feller in his textbooks and to martingale convergence theorems developed by Joseph Doob and Donald Burkholder.
The theorem emerged from parallel advances by Aleksandr Khinchin and Andrey Kolmogorov in the 1920s and 1930s during formative years for probability theory. Influences include earlier contributions from Émile Borel and A. Ya. Khinchin's contemporaries such as Nikolai Luzin and Semyon Stechkin; subsequent expositions appear in monographs by Kolmogorov and later in treatises by William Feller, Kai Lai Chung, and Patrick Billingsley. Debates on priority and formulation inspired commentaries by Paul Lévy and were contextualized within Soviet mathematical institutions including the Steklov Institute of Mathematics and the Moscow State University school that produced figures like Andrei Kolmogorov and Aleksandr Khinchin. The theorem influenced later developments by Kolmogorov's consistency theorem and researches by Alexander G. Kac and Mark Kac in relation to limit distributions.
Variants of the theorem relax independence to forms of weak dependence studied by Sergey Bernstein and Boris Gnedenko, or replace variance summability by moment conditions connected to the Kolmogorov–Feller central limit theorem and to inequalities by Khintchine and Rosenthal. Extensions incorporate martingale differences tied to Joseph Doob's martingale convergence theorem, and versions for triangular arrays relate to Lindeberg–Feller theorem contexts and to criteria used by André Weil and Oscillatory integral analysts. Further variants appear in ergodic settings influenced by John von Neumann, George Birkhoff, and in functional forms treated by Paul Halmos and Norbert Wiener.
Proofs combine truncation techniques from the Kolmogorov three-series theorem, variance bounds leveraging the Chebyshev inequality and Khintchine inequality, and maximal inequalities akin to those introduced by Paul Lévy and Doob; they use orthogonality-like properties of independent increments echoing methods of S. N. Bernstein and Andrey Kolmogorov. A standard approach: truncate X_n to control large jumps (in the spirit of Lévy and Cramér), apply variance summability to obtain L^2-boundedness, then invoke almost sure convergence via Borel–Cantelli lemmas developed by Émile Borel and Émile Picard and maximal inequalities due to Kolmogorov and Doob. Alternative proofs employ martingale convergence techniques associated with Doob, martingale transforms studied by Burkholder, and modern functional-analytic treatments from Alain-Sol Sznitman and K. R. Parthasarathy.
The theorem underpins convergence results used in statistical mechanics frameworks studied by Lars Onsager and Richard Feynman, in stochastic process theory influenced by Norbert Wiener and Andrey Kolmogorov, and in empirical process theory connected to David Blackwell, Thomas Ferguson, and Bradley Efron. It informs theorems about random series representations of processes like the Brownian motion and is applied in the analysis of Fourier series with random coefficients in works by Salem, Zygmund, and Kaczmarz. In modern probability, it is invoked in proofs concerning stability of stochastic recurrence equations studied by H. Kesten and in limit behaviour in ergodic theory researched by D. Ornstein and S. Kakutani.
Standard examples include independent Gaussian X_n with variances σ_n^2 summable, yielding almost sure convergence as in classical constructions by Norbert Wiener and Paul Lévy; another is Rademacher series treated by Salem and Zygmund, where coefficients' square summability ensures convergence. Counterexamples arise when Σ σ_n^2 = ∞: Kolmogorov constructed sequences demonstrating divergence even with diminishing tail behaviour, paralleling constructions by Paul Erdős and André Weil showing necessity of variance conditions without stronger moment control. Pathological triangular arrays exhibiting dependence illustrate failures of the theorem unless replaced by martingale or mixing hypotheses explored by V. V. Petrov and Iosif Pinelis.
Category:Probability theorems