In a (physics) paper I have seen a stochastic differential equation $$ \frac{\mathrm{d} x}{\mathrm{d} t} = f(x, t) + W(x) $$ where $W$ is described as Brownian noise written with $\left\langle W(x) \right\rangle = 0$ and $\left\langle W(x) W(x^\prime) \right\rangle = |x-x^\prime|$. By taking a derivative with respect to $x$ the author arrives to $$ \frac{\mathrm{d} v}{\mathrm{d} x} = \frac{\mathrm{d} f}{\mathrm{d} x} + w(x) $$ where $v = \mathrm{d} x / \mathrm{d} t$ and the noise term $$ w(x) = \frac{\mathrm{d} W}{\mathrm{d} x} $$ has $\left\langle w(x) \right\rangle = 0$ and $\left\langle w(x) w(x^\prime) \right\rangle = \delta(x-x^\prime)$.

This makes sense to me as white noise ($w$ here) can be understood as the formal derivative of Brownian motion and the delta-correlation seems right. Although I can't figure out how one gets this $|x-x^\prime|$ correlation out of Brownian noise or transforms between these two.

Also in an another paper a similar equation is written with $\left\langle \mathrm{d} W \right\rangle = 0$ and $\left\langle | \mathrm{d} W |^2 \right\rangle = \mathrm{d} x$ which perplexes me even further. How should I properly deal with these?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.