Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I have some smooth function $g(x) \colon \mathbb{R}^{n}_{+} \to \mathbb{R}_+$ such that $G_{t} = \{ x \in \mathbb{R}^n_+ \mid g(x) \leqslant t \}$ is compact. I consider a function $$ f(t) = \int\limits_{G_t}a(x)dx_1 \wedge ... \wedge dx_n $$ I want to find its derivative.

In this article http://amath.colorado.edu/pub/wavelets/papers/BEYLKI-1984.pdf author uses the represenation of the form $dx_1 \wedge ... \wedge dx_n = dg(x) \wedge \Omega$ to reduce an integral of the form $dx$ to an iterated integral. Is it possible to do something similar here? I think the answer is $$ f'(t) = \int\limits_{ \{ x\mid g(x)=t \}} a(x) \Omega $$

share|improve this question
add comment

1 Answer

up vote 2 down vote accepted
  1. Let's make a substitution: $$ (x_1,x_2,...,x_n) \mapsto (g(x),x_2,x_3,...,x_n) $$ with Jacobian $J = \frac{1}{\partial_{1} g(x)}$. Then we can rewrite $f(t)$ as $$ f(t) = \int\limits_{0}^{t} \frac{a(x_1(\tau;y);y) d\tau}{\partial_{1}g(x_1(\tau;y);y)} \int\limits_{\mathbb{R}^{n-1}_+}dy $$ where $y = (y_1,...,y_{n-1})$ and $x_1(\tau;y)$ is defined by $g(x_1(\tau;y);y) = \tau$. It is easy to differentiate $f(t)$ now: $$ f'(t) = \int\limits_{\mathbb{R}^{n-1}_+} \frac{a(x_1(t;y);y)dy}{\partial_1 g(x_1(t;y);y)} $$
  2. We can consider $M_t = \{ x \in \mathbb{R}^n_+ \mid g(x) = t \}$ as a smooth manifold with the atlas $\{ (M_t,\varphi) \}$ such that $\varphi \colon M_t \to \mathbb{R}^{n-1}_+$ and $\varphi(x_1,...,x_n) = (x_2,...,x_n)$. Then a form $\omega$ defined by $$ \omega = \frac{a(x_1(t;y);y)}{\partial_1 g(x_1(t;y);y)} dy_1 \wedge ... \wedge dy_{n-1} $$ is a pullback via $(\varphi^{-1})^{*}$ of a form $\alpha$ defined by $$ \alpha = \frac{a(x)}{\partial_1 g(x)} dx_2 \wedge ... \wedge dx_n $$ Hence $$ f'(t) = \int\limits_{M_t} \alpha $$
  3. Since $dg \wedge \omega_1 = dg \wedge \omega_2$ implies $$ \int\limits_{M_t} \omega_1 = \int\limits_{M_t} \omega_2 $$ and $dg \wedge \alpha = a(x) dx_1 \wedge ... \wedge dx_n$ than for any form $\beta$ such that $dg \wedge \beta = a(x)dx$ we have $$ f'(t) = \int\limits_{M_t} \beta $$ For example, we can choose $\beta$ defined by $$ \sum\limits_{k=1}^{n} (-1)^{k-1} \frac{a(x)\partial_k g(x)}{| \nabla g(x) |^2} dx_1 \wedge ... \wedge \overline {dx_k} \wedge ... \wedge dx_n $$
share|improve this answer
add comment

Your Answer

 
discard

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.