During my class in real analysis, my teacher mentioned an alternative method to prove Riesz's theorem for $C([0,1])$. He just mentioned the method, but did not prove it. (A usual way of doing this is provided here: Riesz representation theorem for $C([0,1])$)
Step 1: Let $\Lambda$ be a nonnegati...
I discovered that the following integrals are equal:
$$ \int_0^1sx^{s-1}\exp\bigg(\frac{t}{\log(x)}\bigg)~dx=\int_0^1\exp\bigg(\frac{st}{\log(x)}\bigg)~dx $$
Let $f^s(x)=x^s,$ then the LHS can be written as $$ \int_0^1\frac{d}{dx}\bigg(x^s\bigg)\exp\bigg(\frac{t}{\log(x)}\bigg)~dx$$
It reminds me...
Is there a difference between a combinatorial family and the sequence generated by a generating function, or are these the same thing?
Basically trying to understand this video (youtube.com/watch?v=R6szFb4tIow) but the conceptual explanation in the first 90 seconds isn't very clear and I can't find other resources on the subject.
I remember a video, perhaps a Numberphile episode where a mathematician was described who would simply move in to the home of a collaborator with which they were engaged to so that they could work more closely together, essentially sleeping on their sofa and going to work with them and keeping a ...
But when you restrict categories you also restrict the morphisms between objects, which I believe means the universal property that defines the filtered colimit is not the same
I've only seen the contructive approach to colimits done on Sets
And it's not obvious to me that it would work with other categories
I can't really see an obvious application of these topics in Galois theory or Frobenius algebras, but I don't wanna talk about the first application we've ever seen (in matrix exponentials when solving ODEs). I want to study algebraic applications. Does anyone know anything of this in representation theory, group theory, ring theory or something like that? Thanks!
@BalarkaSen literally any application, like "you can apply complex numbers to solve integrals" or "you can use group theory to solve number theory problems"
I'm not sure about how much time the presentation will take, but I think the article is more important than the presentation (possibly we'll just send our individual projects and, if we want to, read another one's)
Learn how the Smith normal form works over PIDs (this is closely related to primary decomposition, just a bit more organized). Use it to prove the invariant factor decomposition of finitely generated modules over PID's (a sweeping generalization of fundamental theorem of finitely generated abelian groups).
Alternatively you can gun for smaller, cuter applications. Jordan normal form proves that finite-order matrices over $\Bbb C$ (algebraically closed fields of characteristic 0 in general) are diagonalizable. Can you prove a generalization: Every representation of a finite group over $\Bbb C$ is unitary?
Smith normal form can be used to efficiently compute abelianization of a given group using its presentation. Do it for $\text{SL}_2(\Bbb Z)$. Can you show this otherwise/interpret this?
You can use the rational canonical form to come up with necessary sufficient conditions for two operators to commute. I forget what this is called but you can find this in standard texts, or try to prove it yourself.
I can give more ambitious things you could do with these decomposition theorems that I don't know the answer to but maybe I shouldn't.
In particular there's a deep theorem in integer programming called Parry's theorem that for a long time I thought could be done using invariant factor decomposition. I couldn't write a proof though.
A nice consequence of the Smith normal form is that a homomrphism $f\colon\mathbb{Z}^n\rightarrow\mathbb{Z}^n$ has finite cokernel iff $\det(f)\neq0$ and, in that case, $|\coker(f)|=|\det(f)|$. This is useful to calculate the order of free abelian groups in terms of a presentation.
@EdwardEvans May sound so, but we need something to help those hard working people operating computers to work out that problem to numbers to the order $10^8$ or ahead....
@EdwardEvans And I am yearning to possess a Fields Medal in my name (lol)
If $f,g$ are linear maps, $\dim\ker(g\circ f)=\dim\ker f+\dim\ker(g\vert_{\im f})\le\dim\ker f+\dim\ker g$. Assume $f$ is an endomorphism s.t. $m_f=\prod_{i=1}^r(X-\lambda_i)$ with $\lambda_i$ distinct eigenvalues of $f$. Then $\dim V=\dim\ker\bigcirc_{i=1}^r(f-\lambda_i\id)\le\sum_{i=1}^r\dim\ker(f-\lambda_i\id)\le\dim V$, so $V$ is the direct sum of the eigenspaces of $f$ and $f$ is diagonalizable.
Now convince me JNF doesn't take significantly more effort
@LeakyNun i have a little question. i'm asking to you since you speak french. in linear algebra course i take, we talked about "normes qui provient d'un produit scalaire". for example we defined $N_\infty((a,b)) = \operatorname{max}\{|a|,|b|\}$ and said that "$N_\infty$ ne provient pas d'un produit scalaire". and we gave the requirement as follows: "norme provient d'un produit scalaire <=> il satisfait l'égalité de polarisation"
and i've looked at wikipedia articles etc. but couldn't find any plausible terminology in english. do you know how those are studied in english? @LeakyNun
I'm looking at the chapter where it is first mentioned and he interprets the decomposition of the vector space as module over the polynomial ring in terms of the linear endomorphism to get the rational canonical form
in my edition he does a very simple proof in the linear algebra chapter
just by induction and choosing explicit jordan basis
ill tell you the idea later if you cant find it; trying to get some work done rn
i took one of the classes at the end of my linear algebra course because the instructor was away on a seminar and did this proof and a bunch of applications
Suppose $f(z)$ is analytic at $z = 0$ and satisfies $f(z) = z + f(z)^2$. What is the radius of convergence of the power series expansion of $f(z)$ about $z = 0$?
From the conditions, it's clear that $f(z) = 0$ or $f(z) = 1$. I tried to use $f(z) = z + f(z)^2$ to deduce a (recursive) formula for the $n$-th derivative of $f(z)$, so I could calculate the $n$-th coefficient in the power series expansion, but this didn't seem helpful.
I was reading a solution and it says "Near $0$ the function coincides with one of the branches of $(1 \pm \sqrt{1 - 4z})/2$. The radius of convergence of the power series of either branch is $1/4$, which is the distance to the singularity at $1/4$" But this explanation doesn't make any sense to me at all.
$S = \mathbb{Z}, a * b = a+b^2$
Commutative: $a*b = b*a$
$a*b = a + b^2$ and $b*a = b+a^2$ and they aren't the same at all.
Associative: $(a*b)*c = a*(b*c)$
$(a*b)*c = (a+b^2)* c = a+b^2+c^2$ and $a*(b*c) = a + (b+c^2)^2$ and they aren't the same at all.
therefore it's not a binary operation on t...
In the definition of a metric space we require semipositive-definiteness of the codomain of the metric, does this mean that spaces like Minkowski space used in physics $(\Bbb R^4,\eta)$, $\eta=\text{diag}(-,+,+,+)$ is not a metric space? Or do metric spaces without semipositive-definiteness have a different name?
Found a task in my Professors exercises which I am not sure I understand: Find $lim_{x \rightarrow a+0} \frac{1}{(x-a)^{2k}} \, \, \, k \in \mathbb{N}$ How would one proceed with such a task?
@Thorgott @BalarkaSen, I asked a close professor about it and he said I could study the connectedness of $\mathrm{GL}(n,\,\mathbb{C})$. I haven't had a lot of topology, so it looks fun.
@LeakyNun o que eu quis dizer é que esse artigo tem bastante informação sobre como demonstrar e essencialmente usa matrizes "quase diagonais" (teorema de Jordan)
I saw that the triangular matrices are path-connected and every matrix is triangularizable, but I was trying to think of a good path from a matrix to its conjugate rather than just trying to conjugate the path in the triangular matrices lol
You don't even need linear algebra to see GL(n, C) is path-connected.
It follows from pure topology. GL(n, C) is C^(n^2) minus det = 0, which is a real codimension 2 subvariety. Join any two things in GL(n, C) by a path in C^(n^2) and use transversality.
I got played a few days ago when I told a probability person that obviously two random walks on $\Bbb Z^3$ will intersect each other with 0 probability
Because of transversality ("how can random paths intersect in 3 space with positive probability?")
How do I show this apparently elementary result? Let $A$ be a compact set with positive Lebesgue measure and let $h$ be constant. Define $A + h = \{a + h: a \in A\}$, and suppose $G$ is an open set with $A \subset G$. How do I know that $A + h \subset G$ for $h$ "sufficiently small"? (If I'm missing anything additional, let me know and I'll add in more facts about these sets.)
If the result were false, there'd be a sequence $(a_n)_n$ in $A$ such that $a_n+1/n\not\in G$. Since $A$ is compact, let it WLOG converge to $a\in A$, then $(a_n+1/n)_n$ converges to $a$ as well. $G$ is open, so $G^c$ is closed, hence $a\not\in A$, contradiction.
Is the following claim true?
Let $(X, \mathcal{B}, \mu)$ be a measure space with $\mu$ nonatomic. If $0 < k < \mu(X)$, there exists an increasing sequence $B_1, B_2, \dots \in \mathcal{B}$ (i.e., $B_1 \subset B_2 \subset \cdots $) with $\lim_{n \to \infty}\mu(B_n) = k$.
I'm honestly not sure if...
@BalarkaSen Not quite. I'm looking for an increasing sequence of sets that have such a measure. Also, I edited the post to clarify that I'm not allowed to use that result.
BTW for those of you who have made it through PhD coursework... any tips for studying for a measure theory exam? The scope just seems ridiculously vast, and my prof is definitely not relying on a textbook
I'm more or less used to this after going through my MS degree. I think my undergrad classmates were expecting the homework assignments to be more... straightforward. Had to break it to them that this has been my grad-level class experience.