Lévy processes

Lévy processes
Name	Lévy processes
Field	Probability theory
Introduced	1930s
Major contributors	Paul Lévy, Aleksandr Khinchin, Andrey Kolmogorov, André de Finetti, Kai Lai Chung

Lévy processes Lévy processes are stochastic processes with stationary independent increments that arise in Paul Lévy's work and the development of probability theory in the 20th century. They generalize Brownian motion and the Poisson process and underpin modern theories in finance, statistical physics, and queueing theory. Lévy processes connect to foundational results by Aleksandr Khinchin, Andrey Kolmogorov, and applications studied by researchers at institutions such as École Normale Supérieure and Princeton University.

Definition and basic properties

A Lévy process is a right-continuous with left limits stochastic process (càdlàg) starting at zero with stationary independent increments, studied by Paul Lévy and formalized using the frameworks of Andrey Kolmogorov and Aleksandr Khinchin. Key properties include infinite divisibility (linked to Lévy–Khintchine formula and characteristic exponent theory pioneered by Aleksandr Khinchin), stochastic continuity (developed in texts from Kiyoshi Itô and Henry P. McKean), and the Markov property (related historically to work by Andrey Kolmogorov and E. B. Dynkin). Semigroup and generator descriptions tie to the Hille–Yosida theorem and operator methods used at institutions like University of Cambridge and Massachusetts Institute of Technology.

Examples and classes (Brownian motion, Poisson process, stable processes)

Classical examples include Brownian motion (Wiener process) associated with Norbert Wiener and diffusion theory at Princeton University, the Poisson process central to counting processes studied by Siméon Denis Poisson, and stable processes related to work by Paul Lévy and Gennady Samorodnitsky. Other notable classes: compound Poisson processes used in actuarial models at University of Chicago, gamma processes introduced by M. F. M. de Finetti, inverse Gaussian processes linked to Oskar Morgenstern, and infinitely divisible distributions studied by Aleksandr Khinchin and William Feller. Self-similar processes and strictly stable laws connect to results investigated at University of Paris and Columbia University.

Lévy–Khintchine formula and characteristic exponent

The Lévy–Khintchine formula, proved in variants by Aleksandr Khinchin and Paul Lévy, characterizes the characteristic function of an infinitely divisible distribution via a triplet: a drift vector, a covariance matrix (Brownian component), and a Lévy measure. This characteristic exponent appears in analysis by Kiyoshi Itô and in spectral descriptions used in functional analysis at University of Göttingen. The formula underpins estimation techniques developed by statisticians at London School of Economics and links to transform methods taught at Princeton University.

Paths, jump structure, and Lévy measure

Pathwise structure is decomposed into a continuous Gaussian part and a pure-jump part, formalized in decomposition theorems influenced by Kiyoshi Itô's work at Kyoto University. The Lévy measure encodes jump intensity and distribution; its integrability conditions control finite versus infinite activity, with types studied in seminars at École Polytechnique and Stanford University. Sample path regularity, variation, and Blumenthal–Getoor indices are topics advanced by researchers at University of Oxford and Yale University. Decompositions such as the Itô–Lévy decomposition are standard in graduate courses at Massachusetts Institute of Technology.

Fluctuation theory and potential theory

Fluctuation theory treats hitting times, supremum processes, and Wiener–Hopf factorization, building on classical results associated with William Feller and factorization techniques used in queueing theory at Bell Labs. Potential theory for Lévy processes connects to harmonic analysis developed at Institute for Advanced Study and resolvent methods seen in work by E. B. Dynkin. Exit problems, ladder processes, and renewal theoretic aspects have been advanced by groups at University of Cambridge and University of Bath and are central in studies of extremes in statistical physics by researchers at CERN and Los Alamos National Laboratory.

Applications (finance, queuing, physics, insurance)

In finance, Lévy models (jump-diffusion and pure-jump) extend classical models used at Goldman Sachs and in academic work at Columbia Business School and London School of Economics for option pricing and risk management. In queueing theory, compound Poisson and heavy-tailed Lévy inputs model traffic studied in research at Bell Labs and AT&T Labs. In statistical physics, Lévy flights describe anomalous diffusion investigated at Max Planck Institute for Physics and Los Alamos National Laboratory. In actuarial science and insurance mathematics, ruin theory and aggregated claim models using compound Lévy processes are central to methodology at Princeton University and University of Chicago.

Category:Probability theory