
A stationary Gaussian process defined by a linear stochastic differential equation with mean-reverting drift.The Ornstein–Uhlenbeck process is a mathematical model for a quantity that tends to drift back toward a central value, but is constantly jostled by random fluctuations. Imagine a ball rolling in a shallow bowl: gravity pulls it toward the bottom, but if you keep tapping it randomly, it never settles exactly—it wanders around the center. This process is the continuous-time version of that idea. It is the simplest stationary Gaussian process, meaning its statistical properties do not change over time, and it is often used as a building block for more complex random models.
The process is named after Leonard Ornstein and George Uhlenbeck, who introduced it in 1930 to describe the velocity of a particle undergoing Brownian motion, accounting for friction and random collisions.
Mathematically, it is defined by a stochastic differential equation: the change in the variable over a tiny time step equals a term that pulls it back to the mean (proportional to how far it is from the mean) plus a random noise term. The pull is linear, so the process is Gaussian (its values follow a bell-shaped distribution). The strength of the pull and the intensity of the noise determine how quickly it reverts and how much it fluctuates. Over long times, the process reaches a stationary distribution, meaning its average and variance become constant.
It is widely used in finance to model interest rates and commodity prices, in physics for describing noisy relaxation, and in neuroscience for modeling membrane potentials of neurons.
The Ornstein–Uhlenbeck process is the only stationary Gaussian Markov process in continuous time, making it a unique and fundamental object in probability theory.




