Stochastic Differential Equations

A stochastic differential equation describes the evolution of a system subject to a deterministic drift and a random fluctuation driven by Brownian motion, defining a diffusion process.

Definition

A stochastic differential equation specifies the differential of a process as a drift coefficient times a time increment plus a diffusion coefficient times a Brownian increment, and its solution is a diffusion process whose law is governed by the associated second-order differential operator.

Scope

This topic covers the interpretation of stochastic differential equations as Ito integral equations, existence and uniqueness of strong solutions under Lipschitz and growth conditions, the distinction between strong and weak solutions, the diffusion's generator and its link to the Fokker-Planck and backward Kolmogorov equations, the Feynman-Kac and Girsanov theorems, and numerical schemes such as the Euler-Maruyama and Milstein methods.

Core questions

• How is a stochastic differential equation interpreted as an Ito integral equation?
• What conditions guarantee existence and uniqueness of a solution?
• How is the diffusion's generator linked to partial differential equations?
• How are solutions approximated numerically and with what accuracy?

Key theories

Existence and uniqueness of strong solutions

Under Lipschitz continuity and linear growth of the drift and diffusion coefficients, the stochastic differential equation has a unique strong solution that is a continuous Markov diffusion, established by a Picard-type iteration using the Ito isometry.

Feynman-Kac and the generator

The diffusion's infinitesimal generator is a second-order elliptic operator, its transition density solves the Fokker-Planck equation, and the Feynman-Kac formula represents solutions of parabolic partial differential equations as expectations of functionals of the diffusion.

Clinical relevance

Stochastic differential equations model asset prices, interest rates, and volatility in finance, the noisy dynamics of physical, chemical, and biological systems, and population and epidemic models with environmental randomness, while their numerical solution by Euler-Maruyama and related schemes enables Monte Carlo pricing and simulation.

History

Ito introduced stochastic differential equations in the 1940s to construct diffusion processes whose generators are prescribed elliptic operators, Stroock and Varadhan reframed the subject through the martingale problem in the 1960s and 1970s, and the numerical analysis of these equations was systematised by Kloeden and Platen in the 1990s.

Key figures

• Kiyosi Ito
• Bernt Oksendal
• Daniel Stroock
• Srinivasa Varadhan

Seminal works

• oksendal2003

Frequently asked questions

What does a stochastic differential equation describe?

It describes a process that moves under a predictable drift plus random kicks from Brownian motion, producing a diffusion whose probability distribution evolves according to an associated partial differential equation.

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

A strong solution is built on a given Brownian motion and filtration, whereas a weak solution only requires the existence of some Brownian motion and process with the prescribed law; weak solutions can exist when strong ones do not.