let vectors u,v,w be in R^n and suppose that the set (2u+w,u-v,w) is a basis for a subspace W. Show that the set (u,v,w) is a basis for W by showing that it satisfies the two conditions in the definition of a basis
{v1,v2,...,vn} is linearly independent {v1,v2,...,vn} span W.
Got no clue right now. If there were numbers for (u,v,w) i'd be set..but this isn't the case.
anyone who can solve is god.
Why don't we proceed with the trickier part~ showing that u,v, and w are linearly indpendent?
*independent
Well clearly then the vectors 2u+w, u-v and w cannot be represented as a linear combination of each other as it forms a basis. So knowing this, how would we show that u,v,w are linearly independent?
I suggest we start with an assumption that a linear combination of u, v, and w is equal to the zero vector... \[\Large a\color{red}{\vec u}+ b\color{green}{\vec v}+c\color{blue}{\vec w}= \vec0\]
Correct me if i'm wrong but then a(2u+w)+b(u-v)+c(w) =0 implies that a=b=c=0?
Of course, you're right, but that equation has nothing to do with \[\Large a\color{red}{\vec u}+ b\color{green}{\vec v}+c\color{blue}{\vec w}= \vec0\]
Now, let's rewrite this equation, but this time, in terms of the basis of W 2u + w u - v w
\[\Large a\color{red}{\vec u}+ b\color{green}{\vec v}+c\color{blue}{\vec w}=p(2\color{red}{\vec u}{+\color{blue}{\vec w}})+q(\color{red}{\vec u}-\color{green}{\vec v}) + r(\color{blue}{\vec w})= \vec0\] Catch me so far?
Yep. Makes perfect sense.
Now, this will yield a system of equations... we just have to express p,q, and r in terms of a,b, and c. You can solve it as a systems, or you can use your incredible intuition to see a few 'details'...
*system
For instance, you'll see that the only element of the basis of W that involves v at all is the second one, (u - v) And it has a -v in it, so you can bet that q must be -b. \[\Large a\color{red}{\vec u}+ b\color{green}{\vec v}+c\color{blue}{\vec w}=p(2\color{red}{\vec u}{+\color{blue}{\vec w}})-b(\color{red}{\vec u}-\color{green}{\vec v}) + r(\color{blue}{\vec w})= \vec0\]
And now, you tentatively have a \(\large -b\color{red}{\vec u}\) involved, so you have to sort of "neutralise" that with the other element of the basis which involves u, the first one, (2u + w) 2pu - bu = au u(2p - b) = au 2p - b = a 2p = a+b \[\large p = \frac{a+b}{2}\]
Yea i'm following. This is getting interesting now.
Well, stop following, and take the lead. ^.^ We now have \[\Large a\color{red}{\vec u}+ b\color{green}{\vec v}+c\color{blue}{\vec w}=\left(\frac{a+b}{2}\right)(2\color{red}{\vec u}{+\color{blue}{\vec w}})-b(\color{red}{\vec u}-\color{green}{\vec v}) + r(\color{blue}{\vec w})= \vec0\] Can you now find r (in terms of a,b, and c) ?
i'm doing that right now!
I'm waiting :)
is it \[c-(a+b)/2\]?
Very good :D \[\Large a\color{red}{\vec u}+ b\color{green}{\vec v}+c\color{blue}{\vec w}=\left(\frac{a+b}{2}\right)(2\color{red}{\vec u}{+\color{blue}{\vec w}})-b(\color{red}{\vec u}-\color{green}{\vec v}) + \left[c-\frac{a+b}{2}\right](\color{blue}{\vec w})= \vec0\]
Crud.. let me fix that... \[\large a\color{red}{\vec u}+ b\color{green}{\vec v}+c\color{blue}{\vec w}=\left(\frac{a+b}{2}\right)(2\color{red}{\vec u}{+\color{blue}{\vec w}})-b(\color{red}{\vec u}-\color{green}{\vec v}) + \left[c-\frac{a+b}{2}\right](\color{blue}{\vec w})= \vec0\]
Okay, that was the tricky part, the easy part comes now :D
YAY
Remember that our goal was to show that a,b, and c are all 0. Right?
Yea...
Ok. would that mean that (a+b)/2=0,-b=0 and c-(a+b)/2=0?
Remembering as well that (2u + w) , (u-v) , and w are linearly independent. That, combined with this equation... \[\Large \left(\frac{a+b}{2}\right)(2\color{red}{\vec u}{+\color{blue}{\vec w}})-b(\color{red}{\vec u}-\color{green}{\vec v}) + \left[c-\frac{a+b}{2}\right](\color{blue}{\vec w})= \vec0\] Is all we need to show that :) and yes, you're correct... again :D
Oh!
So... solving this system... \[\frac{a+b}2 = 0\]\[-b=0\]\[c-\frac{a+b}2=0\]
And don't skip it... I know you know what's going to happen, but solve anyway :P
a=-b,-b=0,2c=a+b
which means a=-b=0
and 2c=0 which implies c=0!?
as a=b=0
Okay... I'm satisfied. xD So, since assuming au + bv + cw = 0 leads to a = b = c = 0 We can conclude that u, v, and w are linearly independent :)
Nice Job!
We are not finished here, however. We still have to show that u, v, and w span W In other words, any vector in W should be expressable as a linear combination of u, v, and w. But I hope you can see that that's a ridiculously simple task... don't you think? ;)
would we then figure out that: au+bw+cv=(x,y,z)? or isn't it obvious that since these vectors are linearly independent to each other then any vector in W can be expressed as a linear combination of these vectors?
We've basically had enough of linear independence (we've shown all the linear independence needed, namely, that u, v, and w are independent) Now, the task that remains is to show that any element of W can be written as a linear combination of u, v, and w.
Anyway, let's see... let \[\large \nu \in W\] then... Since this set (2u + w) , (u - v), and w spans W, then... \[\Large \nu = k(2\color{red}{\vec u}+\color{blue}{\vec w}) + m(\color{red}{\vec u }-\color{green}{\vec v}) + n\color{blue}{\vec w}\] For some scalars k, m , and n.
Oh oops, that what i meant..for a moment i thought u,v,w were a basis.
So, can you continue from here?
It's just a matter of distributing the scalars, and separating the coefficients of u, v, and w.
let me have a go at that!
can i ask if this is right?
u(2k+m)+v(-m)+w(k+n)=v Therefore v=uX+vY+wZ, where X=2k+m, Y=-m, Z=k+n?
Hence this shows that (u,v,w) spans W?
I shouldn't have used v, maybe I should have used z instead :D but yeah, you got it correctly :D But I would place the scalars to the left of the vector, and not to the right, but that's a technicality ;) Nicely done :D
So, recap, we showed that the set (u,v,w) is linearly independent, and also, that it spans W. Therefore...? ^.^
You are the definition of a maths god! Thank you so much.
it satisfies the definition of a basis
That's the final bit, lol. As for the "maths god" hehe... I don't think so... ;) If you're planning on sticking around, see if you can help others out with their woes :D Welcome to Openstudy I got to sign off now, so have fun ^.^ ------------------------------------------------- Terence out
Thanks mate!
Join our real-time social learning platform and learn together with your friends!