Symplectic geometry is a branch of differential geometry and differential topology that studies symplectic manifolds; that is, differentiable manifolds equipped with a closed, nondegenerate 2-form. Symplectic geometry has its origins in the Hamiltonian formulation of classical mechanics where the phase space of certain classical systems takes on the structure of a symplectic manifold.
== Introduction ==
A symplectic geometry is defined on a smooth even-dimensional space that is a differentiable manifold. On this space is defined a geometric object, the symplectic form, that allows for the...
Do any of you guys know about this article: A NEW UPPER BOUND FOR THE SUM OF DIVISORS FUNCTION by Christian Axler. It's from Cambridge. I was hoping someone could help me get it.
If $T \subset G$ generates group $G$ and $S \subset H$ generates group $H$, then what generates $G \times H$ does $T \times S$ do the trick? (doesn't have to be minimal)
Recall if $A \subsetneq B$ then it must be true that $B \supsetneq A$
However, we came up with something where $A \subsetneq B$ but $(B \cap A) \subsetneq A$. We call this a right subset
Likewise, there is also $B \subsetneq A$ but $(B \cap A) \supsetneq A$. We call this a left subset
It is easy to check how the "and" operator in the boolean algebra is broken when one expands the $\subsetneq, \cap$ and noting that will produce a contradiction when evaluating the truth value of $A \subsetneq B \implies (B \cap A) \subsetneq A$
and hence, for a set theory to accomodate non antisymmetric subset operations, you need to either change the formal system, or the axiom of specification
I m taught that "totally bounded and Lebesgue number property iff compact" for metric spaces. And from my previous chat, $\Bbb R$ with discrete metric satisfies Lebesgue number property but is not compact.
@Asaf I googled a little and found this: We call Lebesgue spaces the spaces such that every open covering has a Lebesgue number: they turn out to be the spaces X such that every continuous function on X is uniformly continuous. Such spaces are usually called Atsuji spaces.arxiv.org/abs/math/9602203 Maybe some references from this paper will be useful to find out more. — Martin SleziakApr 9 '12 at 10:02
In case it helps, a usual keyword for searching might be Atsuji space.
@Silent Think of a cover of $\Bbb Q \cap [0, 1]$ coming from a cover of $[0, 1] \setminus \{\sqrt{2}\}$ by open intervals which shrink arbitrarily in length near $\sqrt{2}$.
Without knowing about the posts on meta linking to it, you might have been surprised why it has been undeleted (and now has one delete point). The other post linking there is in the reopen request thread and it was linked a few times in chat.
Of course, if you can think of some way of improving the question, the likelihood that it gets deleted again would be smaller.
Consider the minimization problem described this paper. Let $f_{\lambda}$ be the minimizer. As a part of extending my work, I am able to show the following facts
$$\lim_\limits{\lambda \to 0}\|f_{\lambda}\|_{L^2} = 0$$ and $$\lim_\limits{\lambda \to \infty}\|f_{\lambda}\|_{L^2} = 0$$
My problem...
I'm trying to understand the machinery of root systems for the purpose of classifying complex semisimple Lie algebras. During this process i lost the overview, espacially when it came to highest weight representations and all that, so i'm looking for a short summary:
Do isomorphic root systems ...
@ShineOnYouCrazyDiamond it's surprisingly hard to have axioms not describing infinite structures
For example there is no first order theory (regardless of the language!) such that every finite group is a model of the theory but there are no infinite models
Problem
Let's observe in a closed interval $[a,b] \subset \mathbb{R}$ real-valued and continuous vectorspace $\mathcal{F}([a,b],\mathbb{R})$. Where $\langle ., .\rangle$ is some scalar product. This scalar product only satisfies axioms when $\overline{v},\overline{w},\overline{u} \in V$ and $c \...
unfortunately $0$ is not usually an acceptable prime, while $-p$ definitely is a prime element of $\Bbb Z$
one situation where you should try to understand the indexing set is when you are dealing with directed systems
one of my favourite examples is that a of a quasi-local algebra
(it appears in a physics situation, as an example for any bounded open subset $U$ of minkowski space one has an algebra of "local observables" $\mathscr A(U)$ that live in $U$, further if $U, V$ are space-like seperated one has that $[\mathscr A(U), \mathscr A(V)]=0$ in $\mathscr A(U\cup V)$, which encodes causality)
@s.harp is there a simple example of a unitary operator whose spectrum is the whole unit circle?
I'm thinking about the diagonal operator on $\ell^2$ corresponding to a countable dense subset of the unit circle, it has the full circle as spectrum, but it's not unitary
Problem
Let's observe in a closed interval $[a,b] \subset \mathbb{R}$ real-valued and continuous vectorspace $\mathcal{F}([a,b],\mathbb{R})$. Where $\langle ., .\rangle$ is some scalar product. This scalar product only satisfies axioms when $\overline{v},\overline{w},\overline{u} \in V$ and $c \...
apart from that there is one more problem in your question, namely you have a vector space $\mathcal F$ but then write that whenever $u,v,w$ are in $V$ you have [....]
in this case: what is $V$ supposed to be? do you mean $\mathcal F$?
@s.harp The adjoint here is multiplication with $\overline{e^{ix}}$ so being unitary is clear. I'm missing something easy because I don't see the spectrum
@AlessandroCodenotti ie you get a sequence of operators given by multiplications taht are locally constant (except on measure 0 sets) and this sequence approximates $e^{ix}$ in the strong operator topology
I mean let $T$ be this operator. We want complex numbers $\lambda$ such that $\lambda-T$ is not invertible. What goes wrong when $\lambda$ is on the unit circle?
@AlessandroCodenotti Then $\lambda -T$ is a multiplication operator by a continuous function that has a $0$, if you can show that this is not invertible you are done
@Tuki Your question is something very central to mathematics. I cannot express it elegantly, but if you have L(u) = f, then any expression you write down with f is equal to that same expression but with f replaced by L(u)
this is because f and L(u) are the same thing
As an example, you now that 4 times 3 is 12
and that $4=2^2$, so you also get that $2^2 \cdot 3$ is 12
but Isn't my situation more like this, if I have equality $5=5$ then $\iff 5 \cdot 5 = 5 \cdot 5$ (both sides multiplied by 5 which leaves statement still true)
I do operation to both sides, but how I can be sure this operation holds the equality?
If we assume that $\langle Lf,v \rangle = \langle u, v \rangle$ is true for some $v \in \mathcal{F}$. Does It imply that $Lf=u$? On what condition they are equivelent ($\iff$) then?
$ \langle Lf,v \rangle = \langle u,v \rangle \implies Lf=u $ on what condition?
Well if I try to show that $\langle Lf,v \rangle = \langle u , v \rangle \implies Lf=u$, axiom 4 says that $\langle Lf,v \rangle \ge 0$ and $\langle u,v \rangle \ge 0$ or 0 if $v=0$
remember that you are in a different situation now, you want to show that $Lf = u$. What you have is $\langle Lf ,v \rangle = \langle u ,v\rangle$ for all $v\in \mathcal F$
@Tuki right, so try to use what you have (namely $\langle Lf , v\rangle = \langle u, v\rangle$ for all $v\in\mathcal F$) to get what you want (namely $\langle Lf -u , Lf -u\rangle = 0$)
@ShineOnYouCrazyDiamond @ShineOnYouCrazyDiamond Lol. Yep, let me just break RSA (and probably ECC too, while I'm at it). Just give me a few hours. Nevermind, someone who had access was kind enough to give it to me
Here's a weird question. Say $M$ is a manifold and $F : PM \to \Bbb R$ is a functional on the pathspace of $M$. What restrictions on $F$ do I need to impose to say there exists a 1-form $\omega$ on $M$ such that $F[\gamma] = \int_\gamma \omega$?
OK, I guess what I would do is take parametrized neighborhood $U$ of $M$ and choose a section of the pathspace fibration $PM \to M$ over $U$ which exists because $U$ is contractible. Let's say $\gamma_x$ is a path from the origin $0$ (the basepoint) to $x \in U$ for every $x$ under this section. Then $\varphi(x) = F[\gamma_x]$ defines a function $\varphi : U \to \Bbb R$
What constraints are there on $F$ from Stokes' theorem?
And what about from the definition of integration?
First off, $PM$ is really a category (or the morphisms-set of; I'm thinking of a category as a set of morphisms with a partially defined product). and $F$ is a functor. That is your first constraint: additivity.
Secondly fix morphism sets P(x,y). What about $F$ on this? We know that if $F(\gamma) = \int_\gamma \omega$, and if $\gamma'$ is homologous to $\gamma$ by a surface $S$, we have $F(\gamma) - F(\gamma') = \int_{\partial S} \omega = \int_S d\omega$. OK, this is pretty nasty admittedly
Take a curve $\gamma : \Bbb R \to U$ such that $\gamma(0) = x_0$ and $\gamma(t) = x$ with $\gamma'(t) = v$. Then look at uh $F[\gamma_{[0, t]}] - F[\gamma_{[0, t+h]}]$
It's certainly true that $\frac{d}{dt} \int_{\gamma(0)}^{\gamma(t)} \omega = \omega(0)$
So I guess you should assume that $F$ is $C^1$, in the sense that a map $X \to PM$ is smooth if the associated map $X \times [0,1] \to M$ is smooth, and we assume that for all smooth maps to $PM$ that $F\big|_{X \times [0,1]}$ is smooth
Note that $\omega$ need not be unique, and locally I can always make such a choice of $\gamma$'s by the contractibility argument. Maybe what I want is $F = \int_\gamma \omega$ locally? This is getting ugly
@MikeMiller Also, we don't really want $F$ to just be a functional on $PM$, right? We want it to be $Diff(I)$-invariant
Suppose $F = \int_\gamma \omega = \int_\gamma \omega'$. Then $\int_\gamma \omega - \omega' = 0$ for all curves $\gamma$. In particular your argument taking the derivative along a curve shows that $\omega - \omega'$ is zero at all points.
Let $X$ be a complex Frechet space, let $\Omega \subseteq \Bbb{C}$ be open, and let $f : \Omega \to X$ be weakly holomorphic (i.e., $\Lambda f$ is holomorphic in the ordinary sense for every $\Lambda \in X^*$).
Given $\Lambda \in X^*$, why is it true that $\frac{(\Lambda f)(z)-(\Lambda f)(0)}{z} = \frac{1}{2 \pi i} \int_{\Gamma} \frac{(\Lambda f)(\xi)}{(\xi - z ) \xi} d \xi$?
Shouldn't there be a limit as $z \to 0$ on the LHS?
@RyanUnger Yo this adiabatic accessibility is weird
Say $\alpha$ is a 1-form on $\Bbb R^3$. Then there are points arbitrarily close to some point $p$ which cannot be reached from $p$ by curves tangent to $\ker \alpha$ iff $\alpha$ is integrable.
One direction is trivial, and if $\alpha$ is totally non-integrable then I know everything is reachable from everything.
If $\alpha \wedge d\alpha \neq 0$ at a point then $\alpha \wedge d\alpha \neq 0$ on a neighborhood as well. On that nbhd $\alpha$ is totally non-integrable :P
OK I have just proved existence of entropy by using an h-principle
Let $f(x)$ be any continuous, differentiable function that has extrema (use $f(x)=\sin^2(\frac{33}{x}\pi)+sin^2(x \pi)$ as the example function in your answers) and that $f(x) \geq 0$ for all $x$ and has a finite amount of roots (it is not always a polynomial). Let $g(x)$ be a function (written i...
@J.Doe I was thinking about it because I wanted to find roots numerically for functions that are all positive and have other extrema other than the roots (so something like f(x)/f'(x) wouldn't work). Newton's, fixed point, and so on are only valuable if you start near the root, and even then, with a function like the example, it's not guaranteed to work. Bisection is the only one guaranteed to work, but I need to get the equations in a form so that can work.