The Student Room Group

Vector operations(year 1 uni material)

Let c be a non-zero vector and suppose that a and b satisfy a·c=b·c(∗) Does it follow that a=b? Give a geometrical meaning to equation (∗).

i said yes it does follow by using the scalar product on both sides of the equation but this is wrong.
the answer was to multiply both sides by c/|c|^2 and that proceeds to say that vector a and b have the same projections on c, i.e., Proj c a=Proj c b.

can someone explain how they arrived at the final answer and why they used c/|c|^2?
(edited 3 years ago)
Original post by localmemelord
Let c be a non-zero vector and suppose that a and b satisfy a·c=b·c(∗) Does it follow that a=b? Give a geometrical meaning to equation (∗).

i said yes it does follow by using the scalar product on both sides of the equation but this is wrong.
the answer was to multiply both sides by c/|c|^2 and that proceeds to say that vector a and b have the same projections on c, i.e., Proj c a=Proj c b.

can someone explain how they arrived at the final answer and why they used c/|c|^2?


The statement is false and can be shown by choosing a,b,c to be the three unit vectors in R3\mathbb{R}^3.

Anyway, I find their reasoning not too clear.

I would just note that

ac=bc    (ab)c=0 \mathbf{a}\cdot \mathbf{c} = \mathbf{b} \cdot \mathbf{c} \implies(\mathbf{a}-\mathbf{b}) \cdot\mathbf{c} = 0

which means that ab\mathbf{a} - \mathbf{b} is perpendicular to c\mathbf{c}. It does not necessarily need to be zero, so we do not necessarily have a=b\mathbf{a} = \mathbf{b}
Original post by localmemelord
Let c be a non-zero vector and suppose that a and b satisfy a·c=b·c(∗) Does it follow that a=b? Give a geometrical meaning to equation (∗).

i said yes it does follow by using the scalar product on both sides of the equation but this is wrong.
the answer was to multiply both sides by c/|c|^2 and that proceeds to say that vector a and b have the same projections on c, i.e., Proj c a=Proj c b.

can someone explain how they arrived at the final answer and why they used c/|c|^2?



If c is a unit vector, then (a.c) is the length of the projection of a on c, and then c(a.c) is the actual projected vector.

The "/|c^2|" is what you get when you adjust for c not being a unit vector (i.e. scaling both occurrences of c in c(a.c) to be unit vectors).
Original post by RDKGames
The statement is false and can be shown by choosing a,b,c to be the three unit vectors in R3\mathbb{R}^3.

Anyway, I find their reasoning not too clear.

I would just note that

ac=bc    (ab)c=0 \mathbf{a}\cdot \mathbf{c} = \mathbf{b} \cdot \mathbf{c} \implies(\mathbf{a}-\mathbf{b}) \cdot\mathbf{c} = 0

which means that ab\mathbf{a} - \mathbf{b} is perpendicular to c\mathbf{c}. It does not necessarily need to be zero, so we do not necessarily have a=b\mathbf{a} = \mathbf{b}

can you expand on why it doesnt necessarily need to be zero
Original post by localmemelord
can you expand on why it doesnt necessarily need to be zero

To take a really simple case: suppose c = i and we want a.c = 0. If x, y, z are the components of a (i.e. a = xi + yj + zk), then a.c = x. So requiring that a.c = 0 only tells us with x = 0; we are left with a completely free choice for y and z. Explicitly, we could have c = j or c = k, or any linear combination of these solutions.

On a more generic level, in an n-dimensional space, the condition a.c = 0 is equivalent to a single condition of the form a1c1+a2c2+...ancn=0a_1c_1 + a_2c_2 + ... a_nc_n = 0, and requiring this to be true can only reduce the dimension of possible values for c by 1. I.e. a.c = 0 defines an n-1 dimensional subspace of Rn\mathbb{R}^n.
Original post by localmemelord
can you expand on why it doesnt necessarily need to be zero

Let θ\theta be the angle between (ab)(\mathbf{a} - \mathbf{b}) and c\mathbf{c}.

(ab)cabccosθ(\mathbf{a}-\mathbf{b}) \cdot \mathbf{c} \equiv |\mathbf{a} - \mathbf{b}| |\mathbf{c}| \cos \theta

Recall this??

So if this is zero, then either ab=0|\mathbf{a} - \mathbf{b}| = 0 (in which case a=b\mathbf{a} = \mathbf{b}), OR we simply need cosθ=0\cos \theta = 0.

Hence it is not necessary that we must have the first condition.
Original post by RDKGames
Let θ\theta be the angle between (ab)(\mathbf{a} - \mathbf{b}) and c\mathbf{c}.

(ab)cabccosθ(\mathbf{a}-\mathbf{b}) \cdot \mathbf{c} \equiv |\mathbf{a} - \mathbf{b}| |\mathbf{c}| \cos \theta

Recall this??

So if this is zero, then either ab=0|\mathbf{a} - \mathbf{b}| = 0 (in which case a=b\mathbf{a} = \mathbf{b}), OR we simply need cosθ=0\cos \theta = 0.

Hence it is not necessary that we must have the first condition.

that makes sense because for (a-b).c to be equal to zero, (a-b) has to be perpendicular to c (cos90 = 0)or the length of (a-b) has to be zero. also since c is a non zero vector, c cannot have a zero-length.

one more thing, in the above comment you said their reasoning is unclear, why do you think they used c/|c|^2?
and finally what does it mean for something to have a projection on something else, for e.g. projection of a on c?
Original post by localmemelord
that makes sense because for (a-b).c to be equal to zero, (a-b) has to be perpendicular to c (cos90 = 0)or the length of (a-b) has to be zero. also since c is a non zero vector, c cannot have a zero-length.

one more thing, in the above comment you said their reasoning is unclear, why do you think they used c/|c|^2?
and finally what does it mean for something to have a projection on something else, for e.g. projection of a on c?

I have explained this; if you don't know what projection means (for vectors), this is the relevant wiki page:

https://en.m.wikipedia.org/wiki/Vector_projection

Some of that page is a bit abstract but the diagram showing projection and rejection should be helpful.
Original post by DFranklin
I have explained this; if you don't know what projection means (for vectors), this is the relevant wiki page:

https://en.m.wikipedia.org/wiki/Vector_projection

Some of that page is a bit abstract but the diagram showing projection and rejection should be helpful.

thanks!

Quick Reply

Latest