The Student Room Group

Linear Algebra

V is a 2 dimensional vector space with basis {a,b}. Find all inner products, (-,-) which satisfy (a,a)=(b,b)=1

I started by considering (u,v) and then writing u and v in terms of the basis and using the rules for inner produts to expand it as much as possible but I end up with (u,v)=c+d(a,b) where c and d are scalars determined u and v. Is this the right way to go and if so how can I get rid of the (a,b) term?
Reply 1
Original post by james22
V is a 2 dimensional vector space with basis {a,b}. Find all inner products, (-,-) which satisfy (a,a)=(b,b)=1

I started by considering (u,v) and then writing u and v in terms of the basis and using the rules for inner produts to expand it as much as possible but I end up with (u,v)=c+d(a,b) where c and d are scalars determined u and v. Is this the right way to go and if so how can I get rid of the (a,b) term?


I'm not entirely sure what they want but it seems that you'll only be able to find a smallish restriction on the inner product. Consider the fact that an arbitrary vector v1=αa+βbv_1 = \alpha a + \beta b has a positive definite inner product with itself.
Original post by james22
(u,v)=c+d(a,b) where c and d are scalars determined u and v.
Is this the right way to go and if so how can I get rid of the (a,b) term?


Why would you want to get rid of the a,b\langle a, b \rangle? A hint is that you get an inner product for every scalar of the field.
Also, you have assumed that a,b=b,a\langle a, b \rangle = \langle b, a \rangle, but you have no been told you have a real vector space.
Reply 3
Original post by jack.hadamard
Why would you want to get rid of the a,b\langle a, b \rangle? A hint is that you get an inner product for every scalar of the field.
Also, you have assumed that a,b=b,a\langle a, b \rangle = \langle b, a \rangle, but you have no been told you have a real vector space.


But I don't want to have the inner product defined as a function of itself. Also we only do inner products over real v spaces at the moment.
Reply 4
Original post by james22
But I don't want to have the inner product defined as a function of itself. Also we only do inner products over real v spaces at the moment.


Follow my suggestion above. You'll be able to find a restriction on the value of a,b\langle a,b\rangle
Reply 5
Original post by atsruser
Follow my suggestion above. You'll be able to find a restriction on the value of a,b\langle a,b\rangle


I get (a,b)>0, is this all I can say?
Reply 6
Original post by james22
I get (a,b)>0, is this all I can say?


You shouldn't get that. Can you put up your working? (in latex, if possible)
Reply 7
Original post by james22
I get (a,b)>0, is this all I can say?


Hmm. The restriction that I *thought* that I'd found was a,b>1\langle a,b \rangle > 1 but I did that without checking; now I've looked at it again, I can't see what lead me to that conclusion and I can't get it to come out again (and can't find the bit of paper I did it on either). So maybe I made a mistake.
Reply 8
Original post by james22
I get (a,b)>0, is this all I can say?


OK, I think I finally figured this out. Hint: Cauchy Schwarz inequality.
Reply 9
Original post by atsruser
OK, I think I finally figured this out. Hint: Cauchy Schwarz inequality.


I never though of that, using this I get 0<(a,b)<=1 however I still want to get rid of teh inner product on the RHS so I think that maybe proving that (u,v)=c+dx where )<x<=1 is an inner product satisfying the conditions given is the way to go.

Quick Reply

Latest