Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Background

I'm working to simplify the Lagrangian formalism of classical field theory for the situation of a vector bundle with a bundle metric and a metric connection. Particularly, I want to specify the Euler-Lagrange equations and the Noether theorem for this case.


Given a vector bundle $(E,\pi,M,\mathbb R^n)$ with a bundle metrig $g$ and a metric connection $\nabla$. Let $J^1E$ be the 1-Jet bundle associated to $E$. Is there a canonical way to identify an element $j \in J^1E$ with an element $(\phi,\nabla \phi) \in E\times(E\otimes TM^*)$? I would also be grateful for some bibliography on that subject.

What i know is that there exists a 1:1 correpondence between sections of $J^1E \to E$ and connections on $E$. Furthermore the connection leads uniquley to a splitting $TE=VE \oplus HE$ of the Tangent Bundle of $E$.

share|improve this question

1 Answer 1

up vote 4 down vote accepted

Is there a canonical way to identify an Element... ?

Yes: an element $j\in J^1E$ is the same as subspace $R\subset T_{\phi}E$ of dimension $\dim(M)$ transversal to $VE$. Since your metric connection gives a splitting $T_\phi E=V_\phi E\oplus H_\phi E$ and since $V_{\phi}E\cong E_{\pi(\phi)}$ and $H_\phi E\cong T_{\pi(\phi)}M$ canonically, you may interpret $R$ as the graph of a linear map $T_{\pi(\phi)}M\to E_{\pi(\phi)}$, hence as an element in $E\otimes T^*M$.

A reference which might be useful: Symmetries and Conservation Laws for Differential Equations of Mathematical Physics

Edit: (in response to the comment) I assume your definition of jet is as follows: two sections $\phi,\tilde\phi$ of the bundle have the same 1st jet at $p\in M$ iff their values and their first derivatives coincide at $p$ (one then checks that this is independent of the coordinates). Geometrically this means that the two sections are tangent (picture them as submanifolds in the total space), so the plane $R$ is their tangent space at $p$. In local coordinates $(x_1,\ldots,x_m,\phi_1,\ldots,\phi_n)$ the plane $R$ is spanned by the vectors $\partial_{x_1}+\sum\partial_{x_1}(\phi_j)e_j,\ldots,\partial_{x_m}+\sum\partial_{x_m}(\phi_j)e_j$ and your tensor is $\sum \partial_{x_k}(\phi_j)e_j\otimes dx_j$.

share|improve this answer
    
Thank you! What i still don't understand is that an element $j\in J^1E$ builds an subspace $R \subset T_\phi E$. With a given local trivialization $\psi$, $j$ could be represented by a point in the corresponding horizontal space $HE^\psi_\phi \subset T_\phi E$ of $\psi$. Why does the set of those points in $T_\phi E$ according to the set of all trivializations build a subspace? I tried to prove that, but what I need is that the trivializations build a vector space themselves, which is obviously wrong (take [$M=\mathbb R^n, E=TM, \alpha = id =-\beta$ local trivialisations] as a counterexample). –  Tobias Ohrmann Sep 12 '12 at 22:59
    
Thank you very much, this helped me a lot, unfortunately i'm not autherized to click you answer up :) –  Tobias Ohrmann Sep 14 '12 at 14:29
    
You're welcome. –  Michael Bächtold Sep 14 '12 at 19:38
    
I still have one question: Does such an element $(\phi,\nabla \phi)$ exist for every $j_x=(x_i,y_j,y^j_i) \in (J^1E)_x$? As the condition, which have to be fulfilled for such an $\phi$, i get: $y_i^j-\sum_{k=1}^n \Gamma_{ik}^j y_k = \partial_{x_i}\phi^j - \sum_{k=1}^n\Gamma_{ik}^j \phi^k$ So, if $j$ is holonomic, one can choose the prolonged section of $j$ and the condition is fulfilled. Is holonomy also a necessary condition for the existence of such a section $\phi$? –  Tobias Ohrmann Sep 15 '12 at 14:07
    
I'm not sure I understand the question. The identification we were talking about is bijective: to every section of $J^1 E$ corresponds a section of $E\oplus(E\otimes T^*M)$. By holonomic section in $J^1 E$ you mean a section which is the prolongation of a section in $E$? –  Michael Bächtold Sep 15 '12 at 17:37

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.