Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $U_1,U_2$ be vector subspaces from $\in \mathbb R^5$.

$$\begin{align*}U_1 &= [(1,0,1,-2,0),(1,-2,0,0,-2),(0,2,1,2,2)]\\ U_2&=[(0,1,1,1,0),(1,2,1,2,1),(1,0,1,-1,0)] \end{align*}$$ (where [] = [linear span])

Calculate a basis from $U_1+U_2$ and a vector subspace $W \in \mathbb R^5$ so that $U_1+U_2=(U_1 \cap U_2) \oplus W$. ($\oplus$ is the direct sum).

I have the following so far. I calculated a basis from $U_1 \cap U_2$ in the previous exercise and got the following result: $(1,0,1,-1,0)$. I've also calculated a basis from $U_1+U_2$ and got that the standard basis from $\mathbb R^5$ is a basis.

So I suppose now I should solve the following:

standard basis from $\mathbb R^5$ = $(1,0,1,-1,0)\oplus W$

I thought I should get 4 additional vectors and they should also respect the direct sum criterion, that their intersection $= \{0\}$, however my colleagues have this:

$W = \{(w_1,w_2,0,w_3,w_4) | w_1,w_2,w_3,w_4 \in \mathbb R\}$. Where did I go wrong?

Many many many thanks in advance!

share|improve this question
1  
Note that your colleagues' $W$ has 4 arbitrary parameters and that $(1,0,1,-1,0)$ is not in $W$. (In fact, $W$ is the span of $e_1, e_2, e_4, e_5$ where $e_i$ is the usual $i$th unit vector). –  David Mitra Dec 18 '11 at 10:24

2 Answers 2

up vote 3 down vote accepted

It seems that you're fine. The $W$ given by your colleagues' has as a basis $\{e_1,e_2, e_4,e_5 \}$ where $e_i$ is the standard $i^{\rm th}$ unit vector in $\Bbb R^5$. Moreover, their $W$ does not contain $ (1,0,1,-1,0)$ (any vector in $W$ has 0 in its third coordinate); thus, this vector together with $e_1$, $e_2$, $e_4$, $e_5$ will give a basis for $\Bbb R^5$. So, then, $\mathbb R^5$ = $(1,0,1,-1,0)\oplus W$.

Your approach would be more or less the same. I imagine your colleagues' interpreted the question as "exhibit the subspace $W$", rather than "find a basis of the subspace $W$".

share|improve this answer
    
Hi David, thanks for the fast reply! Question: would the basis(es?) $\{e_2, e_3, e_4, e _5\},\{e_1, e_2, e_3, e _5\}$ also be valid? Using your reasoning, there's no way to get $(1,0,1,-1,0)$ (always a 0 in one of the coordinates that should be 1). Can I just simply write that (any vector in $W$ has 0 in its third coordinate) or do I have the proof something more? many thanks again! –  Clash Dec 18 '11 at 11:05
    
My reasoning to reply the question would be the following. We have the vector $(1,0,1,-1,0)$ and we want to expand it to basis $\mathbb R^5$. Let's try using the vectors $\{e_2, e_3, e_4, e _5\}$. I then perform gauß and see that I get the standard basis. Then I finalize by writing that W can not be written as a linear combination from $\{e_2, e_3, e_4, e _5\}$. Would this be ok? –  Clash Dec 18 '11 at 11:08
    
@clash The two bases you give will work. You could also say "$W$ is the set of all vectors with 0 third coordinate". –  David Mitra Dec 18 '11 at 11:10
    
@clash for your second comment, you want to say "(1,0,1,-1,0) can not be written as a linear combination from $\{e_2,e_3,e_4,e_5\}$" (not $W$, as you wrote). –  David Mitra Dec 18 '11 at 11:13
    
oops, thanks david! –  Clash Dec 18 '11 at 11:22

You could also find your $W$ like this: all you have to do is complete the matrix

$$ \begin{pmatrix} 1 & * & * & * & * \\ 0 & * & * & * & * \\ 1 & * & * & * & * \\ -1 & * & * & * & * \\ 0 & * & * & * & * \\ \end{pmatrix} $$

in such a way that it has rank $5$. You can do this almost as you want, but maybe an easy strategy is like this: sure enough these firs two columns are linearly independent, aren't they?

$$ \begin{pmatrix} 1 & 1 & * & * & * \\ 0 & 0 & * & * & * \\ 1 & 0 & * & * & * \\ -1 & 0 & * & * & * \\ 0 & 0 & * & * & * \\ \end{pmatrix} $$

So, you keep going the same way. These first three columns are linearly independent too:

$$ \begin{pmatrix} 1 & 1 & 0 & * & * \\ 0 & 0 & 1 & * & * \\ 1 & 0 & 0 & * & * \\ -1 & 0 & 0 & * & * \\ 0 & 0 & 0 & * & * \\ \end{pmatrix} $$

So, let's try again with

$$ \begin{pmatrix} 1 & 1 & 0 & 0 & * \\ 0 & 0 & 1 & 0 & * \\ 1 & 0 & 0 & 1 & * \\ -1 & 0 & 0 & 0 & * \\ 0 & 0 & 0 & 0 & * \\ \end{pmatrix} $$

Still four linearly independent columns. Now, the only risk is with the last one:

$$ \begin{pmatrix} 1 & 1 & 0 & 0 & 0 \\ 0 & 0 & 1 & 0 & 0 \\ 1 & 0 & 0 & 1 & 0 \\ -1 & 0 & 0 & 0 & 1 \\ 0 & 0 & 0 & 0 & 0 \\ \end{pmatrix} $$

If you keep on just putting the next vector of the standard basis of $\mathbb{R}^5$, this matrix has rank four. Never mind: just replace your last column with this one:

$$ \begin{pmatrix} 1 & 1 & 0 & 0 & 0 \\ 0 & 0 & 1 & 0 & 0 \\ 1 & 0 & 0 & 1 & 0 \\ -1 & 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 & 1 \\ \end{pmatrix} $$

This matrix has rank $5$. Hence, your vector $(1,0,1,-1,0)$ and $e_1, e_2, e_3, e_5$ are linearly independent vectors. Hence the sum $[(1,0,1,-1,0)] + [e_1, e_2, e_3, e_5]$ is a direct sum and is equal to $\mathbb{R}^5$. So, you can take $W = [e_1, e_2, e_3, e_5]$. (The solution of this problem is far away from being unique.)

share|improve this answer
    
thanks agusti, this is indeed an easy to way to see it! –  Clash Dec 18 '11 at 12:10

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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