Take the 2-minute tour ×
Physics Stack Exchange is a question and answer site for active researchers, academics and students of physics. It's 100% free, no registration required.

Consider one parameter transformation: $y = y ( \tilde{y}, \alpha)$ such that lagrangian satisfies: $\tilde{L}(\tilde{y}, \alpha) = L(y ( \tilde{y}, \alpha))$. We say that equation is invariant under this transformation if: $$\tilde{L}(\tilde{y}, \alpha) = L(\tilde{y}) + \frac{dF(\tilde{y},\alpha)}{dt}$$ Show that: $$\int \frac{\delta L}{\delta y_t} \frac{\partial y}{\partial \alpha} dx - \frac{\partial F}{\partial \alpha}$$ is a constant of motion.

Here $\tilde{y}$ is an old field before transformation, so is lagrangian $\tilde{L}$. Space is 2-dimensional $(x, t)$ and the exercise is listed in section concerning KdV equation.

Any hints?

share|improve this question
Which textbook is the exercise from? –  Qmechanic Nov 25 '12 at 0:24
@Qmechanic I don't know. Lecturer gave me a list of homework tasks without specifying the source. –  qoqosz Nov 25 '12 at 8:42
Is it possible to link to a file with the original problem formulation? –  Qmechanic Nov 25 '12 at 10:05
@Qmechanic here is an original in polish: sdrv.ms/YfdHvx –  qoqosz Nov 25 '12 at 10:18
add comment

1 Answer

up vote 1 down vote accepted

Since OP's problem looks like a homework assignment, we will only provide OP with a series of hints rather than a full solution.

I) The Lagrangian $L~=~\int \! dx ~{\cal L}$ is the spatial integral of the Lagrangian density ${\cal L}$. Let us call the position field $q(x,t)$ and the corresponding velocity field $v(x,t)$. The Lagrangian $L[q(\cdot,t),v(\cdot,t)]$ is a functional of the position and the velocity fields, cf. e.g. this Phys.SE answer. The momentum field is the functional derivative

$$ \tag{1} p(x,t)~:=~\frac{\delta L[q(\cdot,t),v(\cdot,t)]}{\delta v(x,t)}.$$

An infinitesimal variation $\delta L$ of the Lagrangian functional $L$ is given by $$\tag{2} \delta L~=~\int \! dx~\left( \frac{\delta L}{\delta q}\delta q+\frac{\delta L}{\delta v}\delta v \right). $$ The exercise basically asks OP to derive a functional version of Noether's Theorem, which he probably knows for ordinary point mechanics.

II) Let $$\tag{3} \delta q~=~\varepsilon Y$$ be an infinitesimal variation of the position field, where $\varepsilon$ is an infinitesimal constant, and where $Y$ is the generator. Correspondingly, the velocity field transforms as $$\tag{4}\delta v~=~\varepsilon \dot{Y}.$$ (The transformation (3) is a so-called vertical transformation. In general, one could also allow horizontal contributions from variation of $x$ and $t$.)

III) Assume that the transformation (2) is a so-called quasi-symmetry $$\tag{5} \delta\left(\left. L \right|_{v=\dot{q}}\right)~=~ \varepsilon\frac{d}{dt} \left(\left. F\right|_{v=\dot{q}}\right),$$ where $F[q(\cdot,t),v(\cdot,t)]$ is some functional. (If $F=0$ is zero, then the quasi-symmetry becomes a symmetry of the Lagrangian $L$.)

IV) The (temporal component of the) bare Noether current $j^0$ for a vertical transformation (3) is simply the momentum field $p$ times the vertical generator $Y$,

$$\tag{6} j^0 ~:=~p Y. $$

The bare Noether charge $Q^0$ is the spatial integral

$$\tag{7} Q^0 ~:=~ \int \! dx~j^0. $$

The full Noether charge $Q$ is improved with (minus) the $F$-functional,

$$\tag{8} Q ~:=~Q^0-F. $$

Formula (8) corresponds to OP's last formula (v1). Noether's Theorem states that $Q$ is a conserved quantity on-shell, i.e. if the equation of motion

$$\tag{9} \left.\frac{d}{dt} \left(\frac{\delta L}{\delta v}\right|_{v=\dot{q}}\right)~=~\left. \frac{\delta L}{\delta q}\right|_{v=\dot{q}} $$

is satisfied. The proof in the functional setting is very similar to the usual proof given in ordinary point mechanics.

share|improve this answer
Thank you for your answer. My main problem is how does the partial derivative $\partial F / \partial \alpha$ appears? I can't see this one. –  qoqosz Nov 25 '12 at 18:36
@qoqosz: The first order contribution is $\alpha~\partial F / \partial \alpha$ in the notation of the exercise. This should be identified with $\varepsilon F$ in my notation. –  Qmechanic Nov 25 '12 at 19:35
Thanks a lot! I must think on this a little more but it's definitely clearer now. –  qoqosz Nov 25 '12 at 23:07
add comment

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.