Given subspaces ##U \& W##, show they are equal | Linear Algebra

JD_PM
Messages
1,125
Reaction score
156
Homework Statement
Show that ##U = span \{ (1, 2, 3), (-1, 2, 9)\}## and ##W = \{ (x, y, z) \in \Bbb R^3 | z-3y +3x = 0\}## are equal.
Relevant Equations
N/A
Show that ##U = span \{ (1, 2, 3), (-1, 2, 9)\}## and ##W = \{ (x, y, z) \in \Bbb R^3 | z-3y +3x = 0\}## are equal.

I have the following strategy in mind: determine the dimension of subspaces ##U## and ##W## separately and then make use of the fact ##dim U = dim W \iff U=W##. For ##U## I would construct a ##2 \times 3## matrix and use Gaussian elimination on it. My issues are

1) Is the dimension of the row space the one of interest? If that is the case, one gets that ##dim U = 2## because the Gaussian elimination yields no zero rows.

2) For ##W## I do not see how to get the dimension, because I am not given the basis elements of ##W##.Your guidance is appreciated.

Thanks! :biggrin:
 
Physics news on Phys.org
JD_PM said:
Homework Statement:: Show that ##U = span \{ (1, 2, 3), (-1, 2, 9)\}## and ##W = \{ (x, y, z) \in \Bbb R^3 | z-3y +3x = 0\}## are equal.
Relevant Equations:: N/A

Show that ##U = span \{ (1, 2, 3), (-1, 2, 9)\}## and ##W = \{ (x, y, z) \in \Bbb R^3 | z-3y +3x = 0\}## are equal.

I have the following strategy in mind: determine the dimension of subspaces ##U## and ##W## separately and then make use of the fact ##dim U = dim W \iff U=W##. For ##U## I would construct a ##2 \times 3## matrix and use Gaussian elimination on it. My issues are

1) Is the dimension of the row space the one of interest? If that is the case, one gets that ##dim U = 2## because the Gaussian elimination yields no zero rows.
It should be obvious that the dimension of U is 2, since the two given vectors span U, and are clearly linearly independent (neither is a multiple of the other). I don't see that you need to be thinking about row space here.
JD_PM said:
2) For ##W## I do not see how to get the dimension, because I am not given the basis elements of ##W##.
For W, you have a single equation in three variables, so there are two degrees of freedom; i.e., the solution space is a plane in ##\mathbb R^3##.

My advice is to find a basis for the W subspace, and then show that it is also a basis for U.
The equation that defines W is z - 3y + 3x = 0, or equivalently,
z = -3x + 3y, where x and y are arbitrary. It doesn't matter which two variables you consider as arbitrary -- once you set anyone variable, the other two variables can be chosen arbitrarily.

From the equation above, you can write this system of equations.
x = x
y = . . . . . y
z = -3x + 3y

A basis fairly leaps out at you from the system above.
 
  • Like
Likes JD_PM
There's a simple way to show that ##U \subseteq W##.

Proving that ##W## is not 3D should be simple as well.
 
  • Like
Likes FactChecker
Can you show that the two vectors in U are also in W? That would probe that dim(W) ##\ge## dim(U) = 2. Then can you find a vector that is not in W? That would prove that dim(W) ##\lt## dim(##\mathbb{R}^3##) = 3.
 
  • Like
Likes JD_PM and PeroK
Thank you all for the helpful replies! :)

Mark44 said:
A basis fairly leaps out at you from the system above.

Indeed! I now see that a possible basis for ##W## is ##\beta = \{ (1, 1, 0), (0, 1, 3)\}##.

FactChecker said:
Can you show that the two vectors in U are also in W? That would probe that dim(W) ##\ge## dim(U) = 2.

That can shown by writing these as a linear combination of the basis vectors of ##W##.

$$(1,2,3) = \underbrace{k_1}_{=1}(1,1,0) + \underbrace{k_2}_{=1}(0,1,3) $$

$$(-1,2,9) = \underbrace{k_1}_{=-1}(1,1,0) + \underbrace{k_2}_{=3}(0,1,3) $$

FactChecker said:
Then can you find a vector that is not in W? That would prove that dim(W) ##\lt## dim(##\mathbb{R}^3##) = 3.

Yes, say ##(1,1,1)## because we cannot write it as a linear combination of the basis vectors of ##W##.

However, why does that prove that dim(W) ##\lt## dim(##\mathbb{R}^3##) = 3?
 
There are no proper 3D subspaces. If a subspace is 3D then it's the entire space.
 
  • Like
Likes JD_PM
PeroK said:
There are no proper 3D subspaces. If a subspace is 3D then it's the entire space.

Ahhh so the fact that we can find a vector ## v \in (x, y, z)## that is not in ##W## means that ##dim(W) < dim( \Bbb R^3) = 3## We would find any ## v \in (x, y, z)## in ##W## otherwise.
 
Mark44 said:
From the equation above, you can write this system of equations.
x = x
y = . . . . . y
z = -3x + 3y

A basis fairly leaps out at you from the system above.

JD_PM said:
Indeed! I now see that a possible basis for ##W## is ##\beta = \{ (1, 1, 0), (0, 1, 3)\}##.
Maybe the first vector works, but I didn't check. The basis that "leaps out" from my work is {<1, 0, -3>, <0, 1, 3>}.

More explicitly, the system of equations I wrote is
x = 1x + 0y
y = 0x + 1y
z = -3x + 3y
The coefficients of x are the first basis vector, and the coefficients of y are the second one.
 
  • Like
Likes JD_PM
JD_PM said:
That can shown by writing these as a linear combination of the basis vectors of ##W##.

$$(1,2,3) = \underbrace{k_1}_{=1}(1,1,0) + \underbrace{k_2}_{=1}(0,1,3) $$

$$(-1,2,9) = \underbrace{k_1}_{=-1}(1,1,0) + \underbrace{k_2}_{=3}(0,1,3) $$
I would prefer that you just plug the numbers into the original equation defining W and show that the equation is satisfied.
JD_PM said:
Yes, say ##(1,1,1)## because we cannot write it as a linear combination of the basis vectors of ##W##.

However, why does that prove that dim(W) ##\lt## dim(##\mathbb{R}^3##) = 3?
Again, I would prefer that you plug the numbers into the original equation defining W and show that it is not satisfied. Since the subspace generated by (1,1,1) is not contained in W, the dimension of W must be less than 3. (I am not sure what theorems are available for you to use to formally prove that.)
 
  • Like
Likes JD_PM and PeroK
Back
Top