Brownian motion is a central object of probability theory. Roughly speaking, we could perform a space-time rescaling of a symmetric nearest neighbor random walk on the integer lattice such that the limiting process has normally distributed increments and continuous paths.
We provide different principles of construction: Firstly, via the Kolmogorov–Chentsov theorem on continuous modifications and secondly via a Fourier series with random coefficients in the fashion of Paley–Wiener or of Lévy. Finally, we show the functional central limit theorem (invariance principle) which states that the properly rescaled partial sum process of centered square integrable i.i.d. random variables converges in path space to Brownian motion.
KeywordsFiltration Covariance Convolution
- 35.Donsker MD (1951) An invariance principle for certain probability limit theorems. Memoirs of the American Mathematical Society, vol 6 Google Scholar
- 40.Dvoretzky A, Erdős P, Kakutani S (1961) Nonincrease everywhere of the Brownian motion process. In: Proceedings of the 4th Berkeley symposium on mathematics, statistics and probability, vol II. University of California Press, Berkeley, pp 103–116 Google Scholar
- 125.Paley REAC, Wiener N (1987) Fourier transforms in the complex domain. American Mathematical Society colloquium publications, vol 19. American Mathematical Society, Providence. Reprint of the 1934 original Google Scholar
- 156.Smirnov NV (1939) Sur les écarts de la Courbe de distribution empirique. Mat Sb 2:3–16. Russian with French summary Google Scholar
- 173.Yosida K (1995) Functional analysis. Classics in mathematics. Springer, Berlin. Reprint of the sixth edition from 1980 Google Scholar