MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $U : \mathbb R^d \to \mathbb R^d$ be a smooth vector field, and let $F_t : \mathbb R \times \mathbb R^d \to \mathbb R^d$ be the corresponding smooth flow, defined by the differential equation $$\tfrac{d}{dt} F_t(x) = U(F_t(x)).$$ Let $f : \mathbb R^d \to \mathbb R$ be an integrable function, and consider the integral $$\int_{\mathbb R^d} f( F_t(x) ) ~dx.$$ To calculate this, I would make the change of variables $y = F_t(x)$, making the ansatz that there exists some function $\rho(y)$ so that the integral equals $$\int_{\mathbb R^d} f(y) \rho(y) ~dy.$$ From what I can gather from the paper I'm reading, it's common knowledge that this function $\rho$ has the form $$\rho(y) = \tfrac{1}{J_t(F_t^{-1}(y))} = \exp \left(-\int_0^t (\operatorname{div} U)(F_{s-t}(y) ) ~ds \right),$$ where of course $J_t$ stands for the Jacobian of the flow.

Could you please point me to a reference for this formula?

share|cite|improve this question
up vote 4 down vote accepted

This is the Liouville formula. It is explained nicely in Ordinary Differential Equations by Arnold.

share|cite|improve this answer
Bravo, I knew it would be simple. Thanks, Andrey. – Tom LaGatta Jun 8 '10 at 4:19
You're welcome. – Andrey Rekalo Jun 8 '10 at 4:26
For future finders of this question: the precise location Section 27.6 (Liouville's Theorem) on page 195 of Arnold's Ordinary Differential Equations (the green book, translated by Richard Silverman, not the yellow one translated by Cooke). – Tom LaGatta Jun 8 '10 at 4:33

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.