Topics in Geometric Mechanics: Week 14

$ %% not provided in align* environment \def\newsavebox#1{} \def\intertext#1{\text{#1}} \def\let#1#2{} \def\makeatletter{} \def\makeatother{} \newenvironment{theorem}{\textbf{Theorem.}\rm}{} %%% Local Variables: %%% mode: latex %%% TeX-master: t %%% End: %% sets \def\R{\mathbf{R}} \def\Z{\mathbf{Z}} \def\C{\mathbf{C}} \def\CP{\C{}P} \def\Q{\mathbf{Q}} \def\sphere#1{\mathbf{S}^{#1}} \def\rp#1{\R P^{#1}} \def\set#1{\left\{ #1 \right\}} \def\st{\,\mathrm{s.t.}\,} \def\extalg#1#2{\Lambda^{#1}(#2)} \def\symalg#1#2{\mathsf{S}^{#1}(#2)} \def\tenalg#1#2{\mathsf{T}^{#1}(#2)} \def\hom#1#2{\mathrm{Hom}(#1;#2)} \def\extproduct{\wedge} \def\symmetricproduct{\cdot} \def\tensorproduct{\otimes} \def\image#1{\mathrm{Im}\,#1} \def\kernel#1{\mathrm{Ker}\,#1} \def\Trace[#1]#2{\mathrm{Tr}\ifx#1\empty\else{}_{#1}\fi\ifx#2\empty\else\,(#2)\fi} \def\Trce#1{\mathrm{Tr}(#1)} \def\orbit#1#2{{#1}\cdot{#2}} \def\stab#1#2{\mathrm{stab}_{#1}(#2)} \def\Span{\mathrm{span}} \def\cinfty#1{C^{\infty}(#1)} %% Lie groups \def\GL#1{\mathrm{GL}(#1)} \def\SL#1{\mathrm{SL}(#1)} \def\Symp#1{\mathrm{Sp}(#1)} \def\Orth#1{\mathrm{O}(#1)} \def\SOrth#1{\mathrm{SO}(#1)} \def\Unitary#1{\mathrm{U}(#1)} \def\SUnitary#1{\mathrm{SU}(#1)} \def\Torus#1{\mathbf{T}^{#1}} \def\H#1{\mathbf{H}^{#1}} \def\Diagonal#1{\mathbf{Diag}(#1)} \def\diag#1{\mathbf{diag}\left(#1\right)} \def\Liegp#1{\mathrm{#1}} %% Lie algebras \def\Mat#1#2{\mathrm{Mat}_{{#1} \times {#1}}(#2)} \def\Matrect#1#2#3{\mathrm{Mat}_{{#1} \times {#2}}(#3)} \def\gl#1{\mathfrak{gl}(#1)} \def\spl#1{\mathfrak{sl}(#1)} \def\symp#1{\mathfrak{sp}(#1)} \def\orth#1{\mathfrak{o}(#1)} \def\sorth#1{\mathfrak{so}(#1)} \def\unitary#1{\mathfrak{u}(#1)} \def\sunitary#1{\mathfrak{su}(#1)} \def\torus#1{\mathfrak{t}^{#1}} \def\liealg#1{\mathfrak{#1}} \def\liebracket#1#2{[#1,#2]} \def\Ad#1#2{\mathrm{Ad}_{#1}{#2}} \def\ad#1#2{\mathrm{ad}_{#1}{#2}} \def\coAd#1#2{\mathrm{Ad}^*_{#1}{#2}} \def\coad#1#2{\mathrm{ad}^*_{#1}{#2}]} \def\pb#1#2{\left\{{#1},{#2}\right\}} \def\coorbit#1{\mathcal{O}_{#1}} %% operators \def\D#1{\,\mathrm{d}{#1}\,} \def\Dat#1#2{\,\mathrm{d}_{#1}{#2}\,} \def\lieder#1#2{{\sf{L}}_{#1}{#2}} \def\covder#1#2{\,\frac{D #2}{d{#1}}} \def\liebracket#1#2{\left[#1,#2\right]} \def\crossproduct{{\boldsymbol{\times}}} \def\ip#1#2{\langle #1, #2 \rangle} \def\tlangle{\langle\hspace{-1.1mm}\langle} \def\trangle{\rangle\hspace{-1.1mm}\rangle} \def\dotp#1#2{\tlangle #1, #2 \trangle} \def\interiorproduc#1#2{\iota_{#1}{#2}} \def\norm#1{|#1|} \def\lt{<} \def\gt{>} \def\ddt#1#2{\frac{d{#2}}{d{#1}}} \def\ddtat#1#2#3{\left.\ddt{#1}{#2}\right|_{#1=#3}} \def\dndtn#1#2#3{\frac{d^{#3}{#2}}{d{#1}^{#3}}} \def\wint#1#2{\int\limits_{#1}^{#2}} \def\didi#1#2{\frac{\partial{#1}}{\partial{#2}}} \def\dindin#1#2#3{\frac{\partial^{#3}{#1}}{\partial{#2}^{#3}}} \def\eulerlagrange#1#2#3#4{\ddt{#1}{}\didi{#4}{#3} - \didi{#4}{#2}} \def\imagpart#1{\mathrm{Im}(#1)} \def\realpart#1{\mathrm{Re}(#1)} \def\equivclass#1{\left[#1\right]} \def\bivector#1{{\mathcal #1}} \def\bivectorwargs#1#2#3{\bivector{#1}(#2, #3)} \def\compose{\circ} %% labels, etc. \def\labelenumi{{\bf\Alph{enumi}.}} \def\solutioncolor{blue} \def\solutiontopsep{3mm} \def\solutionbotsep{0mm} \newenvironment{solution}{\vspace{\solutiontopsep}\noindent\textbf{Solution}. \textcolor{\solutioncolor}\bgroup}{\egroup\vspace{\solutionbotsep}} %% redo quote environment \def\quote{} \def\endquote{} %% redo enumerate environment \makeatletter \let\oldenumerate\enumerate \let\oldendenumerate\endenumerate \def\enumerate{\bgroup\oldenumerate\setlength{\itemsep}{3mm}\setlength{\topsep}{3mm}} \def\endenumerate{\oldendenumerate\egroup} \makeatother %% utility used in l11.muse \def\cmag#1#2#3{\begin{bmatrix} (d \iota_{#1} + \iota_{#1} d) \Omega \end{bmatrix} ({#2}, {#3}) &= \begin{bmatrix}\lieder{#1}{\Omega}\end{bmatrix}({#2}, {#3})} \def\museincludegraphicsoptions{% width=0.25\textwidth% } \def\museincludegraphics{% \begingroup \catcode`\|=0 \catcode`\\=12 \catcode`\#=12 \expandafter\includegraphics\expandafter[\museincludegraphicsoptions] } \def\musefigurewidth{!} \def\musefigureheight{!} \newsavebox{\mypicboc}% \def\includefigure#1#2#3#4{% \edef\fileending{#4}% \def\svgend{svg}% \def\pdftend{pdf_t}% \def\pdftexend{pdf_tex}% \def\bang{!}% \begin{lrbox}{\mypicboc}% \ifx\fileending\pdftend% \input{#3.#4}% \else% \ifx\fileending\pdftexend \input{#3.#4}% \else% \ifx\fileending\svgend% \input{#3.#4}% \fi\fi% \includegraphics{#3.#4}% \fi% \end{lrbox}% \edef\thiswidth{#1}\edef\thisheight{#2}% \ifx\thiswidth\bang% \ifx\thisheight\bang% \resizebox{!}{!}{\usebox{\mypicboc}}% \else% \resizebox{!}{\thisheight}{\usebox{\mypicboc}}% \fi% \else% \ifx\thisheight\bang% \resizebox{\thiswidth}{!}{\usebox{\mypicboc}}% \else% \resizebox{\thiswidth}{\thisheight}{\usebox{\mypicboc}}% \fi\fi} $
Navigation
MTH-696a
Course outline
Assignments & Solutions
Tutorials & Solutions
Schedule
Lecture Notes
Sources
Home Page of Leo Butler

Recall

We left off with:

Recall - Momentum map

Let $\liealg{g}$ be a Lie algebra, and $\liealg{h}$ a Lie algebra of smooth hamiltonians on $(M, \pb{}{})$. An homomorphism \begin{align*} \Psi &: \liealg{g} \to \liealg{h} && \xi \mapsto h_{\xi} \\ \text{induces}\\ \psi &: M \to \liealg{g}^* && \ip{\psi(x)}{\xi} = h_{\xi}(x). \end{align*} We call $\psi$ a momentum map.

If $\psi$ is a momentum map, then it is a Poisson map.

Convexity

Assumption: $(M^{2n}, \omega)$ is a compact symplectic manifold.

(Atiyah-Guillemin-Sternberg) Let $T=\Torus{n}$ have an effective Hamiltonian action on $M$ with momentum map $\psi : M \to \liealg{t}^*$. Then $\psi(M)$ is a compact, convex subset of $\liealg{t}^*$.

(Kirwan) If $G$ is a compact Lie group with an effective Hamiltonian action on $M$ with momentum map $\psi$, and $T \lhd G$ is a maximal torus, then $\psi(M) \cap \liealg{t}^*_+$ is a compact, convex polytope.

(Delzant) There is a distinguished family $D$ of convex polytopes which are the image of a torus momentum map. Given such a polytope, one can reconstruct $(M, \omega)$, the $T$-action and $\psi$.

Let $x \in M$ have the stabilizer subgroup $T_x \lhd T$ of dimension $k \leq n$. There are symplectic coordinates $(x_1, y_1, \ldots, x_k, y_k, \ldots, x_n, y_n)$ on a neighbourhood of $x$ and a basis $t_1, \ldots, t_n$ of $\liealg{t}^*$ such that \begin{align*} \psi(p)-\psi(x) &= \sum_{a=1}^k \frac{1}{2} (x_a^2 + y_a^2) t_a + \sum_{b=k+1}^n x_b t_b. \end{align*}

Example

Let $(M, \omega)$ be $\sphere{2}$ with its unit area form. Let $T = \Torus{1}$ act by rotation about the vertical $z$-axis. The momentum map is $\psi = z$.

The momentum map.
The momentum map.

Let $M = \sphere{2} \times \sphere{2}$ with $\omega$ the sum of the unit area forms, and $T = \Torus{2}$ acting by rotating about each vertical axis.

The image of the momentum map.
The image of the momentum map.

Liouville-Arnold Theorem

Let $(M^{2n}, \omega)$ be a symplectic manifold, $f_1, \ldots, f_n$ be Poisson commuting, independent functions.

Let $T$ be a connected, compact component of a regular level set $F_c = \set{f_1=c_1, \ldots, f_n=c_n}$. Then $T \cong \Torus{n}$ and there is a neighbourhood $U \supset T$ and coordinates $(\theta, I) : U \to \Torus{n} \times \R^n$ such that \begin{align*} 1. &&& \omega|U = \sum_{i=1}^n \D{\theta_i} \wedge \D{I_i}, \\ 2. &&& f_i|U = F_i(I), \\ 3. &&& X_{f_i} = \sum_{j=1}^n \didi{F_i}{I_j} \didi{}{\theta_j}. \end{align*}

Simple Pendulum

Take \[ H = \frac{1}{2} p^2 + \cos(x). \]

Pendulum.
Pendulum.