First we take identity element from set which is Identity matrix so S=I for which b(σ(x),σ(y))=b(x,y) which is identity transformation in O(V,b) so kernal becomes identity and so it is surjective since it is finite dimensional. Is it correct or not please guide regarding this.
Orthogonal group isomorphisam
80 Views Asked by Bumbble Comm https://math.techqa.club/user/bumbble-comm/detail AtThere are 2 best solutions below
On
We want prove the surjectivity of the map $\Psi$.
We had fix a base $\beta:=\{v_1,...,v_n\}$ on $V$ and the function $\Psi$ maps every $\sigma\in O(V,b)$ to matrix $A_\sigma$ associate with itself in the base $\beta$.
If $A=(a_{ij})\in O(V,B)$ we can define a function
$\sigma’: \beta\to V$
that maps every $v_i\in \beta$ to $ \sigma’(v_i):=\sum_{k=1}^{n}a_{ki}v_k$
In this way there exists a unique linear map $\sigma: V\to V$ that extends $\sigma’$.
Now we want prove that $\sigma\in O(V,b)$.
We know that the matrix associate to $\sigma$ in the base $\beta$ is the matrix that have on the $j$-th column the coefficients of $\sigma(v_j)$ written in the base $\beta$ and so, by construction, the matrix associate to $\sigma$ is $A$. But now for all $x,y\in V $
$b(\sigma(x),\sigma(y))=(Ax)^tB(Ay)=$
$=x^t(A^tBA)y=x^tBy=b(x,y)$
So $\sigma \in O(V,b)$ and oviously $\Psi(\sigma)=A$.
In other words the map $\Psi$ is surjective.

$O(V,b)$ is the set of linear maps $\sigma$ that verify $b(\sigma(x),\sigma(y))=b(x,y)$ for every $x,y\in V$ and $B$ is the matrix associated to $b$ in the canonical base (for example) Now you can prove that when you fix a base on $V$
$b(x,y)=x^tBy$
And so if you consider $A_\sigma$ the matrix associated to $\sigma \in O(V,b)$ in that base, than
$b(\sigma(x),\sigma(y))=x^t(A_\sigma^tBA_\sigma) y=b(x,y)=x^tBy$
So $A_\sigma$ verify the property $ x^t(A_\sigma^tBA_\sigma)y=x^tBy$ for every $x,y\in V$
and in this case it is simple prove that you get that $A_\sigma$ verify the property
$A_\sigma^t BA_\sigma=B$
Now if $b$ is a definite bilinear form than you have the property $y^tBy\neq 0$ for every $y\in V/ \{0\}$ and so if $det(A_\sigma)=0$ than there exist an $x\in V/ \{0\}$ such that $A_\sigma x=0$ but $x^tBx =x^t(A_\sigma^tBA_\sigma)x= x^t(A_\sigma^tB)(A_\sigma x)=0$ and it is impossibile.
So you have also that $det(A_\sigma)\neq 0$ .
Now if you define $O(V,B)=\{A\in M_n(\mathbb{R}) : det(A)\neq 0, A^tBA=B\}$
there exists a map $\Psi: O(V,b) \to O(V,B)$ that maps every $\sigma\in O(V,b)$ to $A_\sigma$
This map is oviously bijective and your sets are also a Groups and so it is important to ask oneself if it is also a morphism of group.