Stochastic Differential Equations

A stochastic differential equation describes the evolution of a system driven by both a deterministic trend and Brownian noise, and its solutions, the diffusion processes, model continuous random dynamics across science and finance.

Definition

A stochastic differential equation is an equation for a process whose infinitesimal change is a drift term times the time increment plus a diffusion term times a Brownian increment, interpreted through the Ito integral, whose solutions are diffusion processes.

Scope

The topic covers the formulation of stochastic differential equations with drift and diffusion coefficients driven by Brownian motion, the distinction between strong and weak solutions and between pathwise and distributional uniqueness, existence and uniqueness under Lipschitz and linear-growth conditions, the Markov and diffusion property of solutions with their generators, standard examples such as geometric Brownian motion and the Ornstein-Uhlenbeck process, and numerical schemes such as the Euler-Maruyama method.

Core questions

• How is a differential equation driven by Brownian noise given rigorous meaning?
• What is the difference between strong and weak solutions and the corresponding notions of uniqueness?
• Under what conditions does a unique solution exist?
• How are the resulting diffusions described by their generators and simulated numerically?

Key concepts

• drift and diffusion coefficients
• strong and weak solutions
• pathwise uniqueness
• diffusion generator
• Euler-Maruyama scheme

Key theories

Existence and uniqueness of solutions

When the drift and diffusion coefficients are Lipschitz continuous and grow at most linearly, the stochastic differential equation has a unique strong solution, obtained by a Picard iteration that parallels the deterministic theory but uses the Ito integral and isometry.

Diffusions and their generators

Solutions of stochastic differential equations are Markov diffusion processes whose infinitesimal generator is a second-order differential operator built from the drift and diffusion coefficients, linking the probabilistic dynamics to parabolic and elliptic partial differential equations.

Clinical relevance

Stochastic differential equations model asset prices and interest rates in quantitative finance, the velocity of particles under friction and noise in physics, population sizes and chemical concentrations under random fluctuation in biology and chemistry, and noisy control systems in engineering, with their numerical solution central to Monte Carlo simulation of these models.

History

Ito introduced stochastic differential equations in the 1940s as the rigorous form of equations driven by white noise, and the existence, uniqueness, and diffusion theory was developed by Ito, Watanabe, Stroock, and Varadhan; their applications expanded dramatically with the rise of mathematical finance from the 1970s.

Key figures

• Kiyosi Ito
• Bernt Oksendal
• Shinzo Watanabe
• Leonard Ornstein

Seminal works

• oksendal2003

Frequently asked questions

What is the difference between a strong and a weak solution?

A strong solution is built on a given Brownian motion and filtration, so the solution is a function of that specific noise, whereas a weak solution provides only a process with the correct distribution on some probability space; the two come with correspondingly different notions of uniqueness.

How are stochastic differential equations solved numerically?

Schemes such as the Euler-Maruyama method discretize time and replace Brownian increments by simulated Gaussian steps; they converge to the true solution as the step size shrinks, though at rates that reflect the irregularity of the noise.