0
$\begingroup$

I'm trying to identify which sets are vector spaces. I know that a set $V$ is a vector space if its elements (vectors) have addition and scalar multiplication so that the result is also within the set $V$. I'm trying to identify the following sets.

  1. Set of real $m\times n$ matrices $\mathbb{R}^{m\times n}$

If we do operations for two matrices it always yields and matrix within $\mathbb{R}^{m\times n}$. So this is a vector space.

  1. Line $\{(x,4x) \in \mathbb{R}^2 \: | \: x\in\mathbb{R}\}$

EDIT: Is a vector subspace of $\mathbb{R}^2$.

  1. $\mathbb{R}^n = \{(x_1, \ldots, x_n) \ \vert \ x_i \in \mathbb{R}\}$

Clearly a vector space because for all $n$-vectors in it the operations hold.

  1. Curve $\{(x,x^2) \in \mathbb{R}^2 \: | \: x\in\mathbb{R}\}$

Addition does not hold on this set $(1,1) + (2,4) = (3,5) \notin V$ so it is not a vector space.

  1. Line $\{(x,3x-2) \in \mathbb{R}^2 \: | \: x\in\mathbb{R}\}$. Same as the other line and not an vector space.

EDIT: Under the vector space axioms addition $\mathbf a + \mathbf b \in V$ when $\mathbf a, \mathbf b \in V$ and $\lambda \mathbf a \in V$ thus if $\mathbf b = -1\cdot \mathbf a$ then $\mathbf a + \mathbf b = \mathbf 0 \notin V$. So additive indentity does not hold and the set is not a vector space.

  1. Set of matrices in $\{ M\in\mathbb{R}^{n\times n}\: | \: \det(M)=1\}$.

\begin{align*} \text{det}\left(\begin{bmatrix} 1 & 0 \\ 0 & 1 \end{bmatrix} + \begin{bmatrix} 1 & 0 \\ 0 & 1 \end{bmatrix}\right) = 4 \neq 1\end{align*}

OR

\begin{align*} \text{det}\left(\lambda \begin{bmatrix} 1 & 0 \\ 0 & 1 \end{bmatrix} \right) = \lambda \neq 1 \text{ when } \lambda \neq 1\end{align*}

EDIT: So addition from said set yields a matrix with non 1 determinant. Not a vector space.

$\endgroup$
  • 1
    $\begingroup$ Your reasoning that $\tt 3.$ isn't a vector space isn't correct, $\mathbf{x}+\mathbf{y}=(3,12)$ has the form $(\alpha,4\alpha)$, so it is certainly in that set. $\endgroup$ – Workaholic Feb 21 '16 at 12:40
  • $\begingroup$ Number 2. seems a vector subspace of $\mathbb{R}^2$ to me. Take two elements in the set, $(\lambda, 4\lambda)$ and $(\mu, 4\mu)$, then $(\lambda, 4\lambda) + (\mu, 4\mu) = (\lambda+\mu , 4(\lambda+\mu))$ which is of the required form. Also if $k\in \mathbb{R}$ then $k(\lambda, 4\lambda)= (k\lambda, 4k\lambda)$ is of the required form. An easy intuition is that this set is a straight line that contains the 0-vector, so it should indeed be a vector space. $\endgroup$ – Martingalo Feb 21 '16 at 12:40
  • $\begingroup$ Reasoning for 5 is wrong although deduction is correct. 5-is not a vector space as the additive identity is not present in the set. $\endgroup$ – Qwerty Feb 21 '16 at 12:42
  • $\begingroup$ 6 is clearly not a vector space. Take 2 identity matrices, add them and take their determinant $\endgroup$ – Qwerty Feb 21 '16 at 12:44
  • $\begingroup$ To qualify as a vector space , don't miss the fact that the additive identity should also be a part of the space $\endgroup$ – Qwerty Feb 21 '16 at 12:47
1
$\begingroup$

Hints:

  1. The deduction is correct, although you should actually check all the vector space axioms to conclude that it is one.

  2. Your reasoning that $\tt 3.$ isn't a vector space isn't correct, $\mathbf{x}+\mathbf{y}=(3,12)$ has the form $(\alpha,4\alpha)$, so it is certainly in that set. $-$ Edit 1: That's correct, assuming that you already know that $\mathbf{R}^2$ is a vector space.

  3. Same comment as in 1.

  4. Your reasoning is correct.
  5. Check instead whether that set contains an additive identity.
  6. I assume that the field you're working over is $\mathbf{R}$. So take an element $c\in\mathbf{R}$, and suppose that $M$ is in that set, does it follow that $\det(cM)=1$? (i.e. does it follow that it is closed under scalar multiplication?) $-$ Edit 1: You showed that it isn't closed under addition, hence not a vector space, and that's correct.
$\endgroup$
1
$\begingroup$

Number (2) seems a vector subspace of $\mathbb{R}^2$ to me. Take two elements in the set, $(\lambda, 4\lambda)$ and $(\mu, 4\mu)$, then $(\lambda, 4\lambda) + (\mu, 4\mu) = (\lambda+\mu , 4(\lambda+\mu))$ which is of the required form. Also if $k\in \mathbb{R}$ then $k(\lambda, 4\lambda)= (k\lambda, 4k\lambda)$ is of the required form. An easy intuition is that this set is a straight line that contains the 0-vector, so it should indeed be a vector space.

For the last one (6) just use the following fact about determinants of matrices. If $A$ i an $n\times n$-matrix and $\lambda\in \mathbb{R}$ a real scalar, then $$\det (\lambda A) =\lambda^n \det(A) = \lambda^n,$$ where the latter follows if $A$ is in the set you defined.

Now, if $\lambda \neq 1$ then $\det(\lambda A) = \lambda^n \neq 1$ and hence $\lambda A$ is not in the set you defined, and hence it cannot be a vector space.

A last comment, since $a,b\in V$ then $a+b\in V$ and $a\in V$, $\lambda \in \mathbb{R}$ then $\lambda a\in V$, this in particular implies that $0\in V$ (the zero vector should always belong to the space, take $a\in V$ then $(-1)*a =-a\in V$ and hence $a-a=0\in V$). Then, since checking this is fairly quick, you may do this for number (5), where $(0,0)\notin \{(x,3x-2)\in \mathbb{R}^2|x\in \mathbb{R}\}$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.