Orthogonal Complements of Linear Transformations with Matrix A

  • Thread starter LosTacos
  • Start date
  • Tags Tags
    Orthogonal
8 replies · 3K views · last reply
LosTacos
Messages
79
Reaction score
0

Homework Statement



Let L: ℝn→ℝm be a linear transformation with matrix A ( with respect to the standard basis). Show that ker(L) is the orthogonal complement of the row space of A.

Homework Equations





The Attempt at a Solution


The ker(L) is the subset of all vectors of V that map to 0. The orthogonal complement of W is the set of all vectors x with property that xw= 0. Would we use the dimension theorem:
dim(ker(L)) = n - dim(range(L)) = n - dim(W) =dim(Wτ). Since Wτ is contained in ker(L).

*Wτ is orthogonal complement.
 
This seems way to obvious. The row space is the subset of all vectors that are linear combinations of rows of A. So the kernel of L is the subset of all vectors in V that map to 0. Then, the orthogonal complement is the set of vectors with the property that x ⋅ w = 0. So, I need to prove that the vector x in the orthogonal complement is the same vector in the row space?
 
LosTacos said:

Homework Statement



Let L: ℝn→ℝm be a linear transformation with matrix A ( with respect to the standard basis). Show that ker(L) is the orthogonal complement of the row space of A.

Homework Equations





The Attempt at a Solution


The ker(L) is the subset of all vectors of V that map to 0. The orthogonal complement of W is the set of all vectors x with property that xw= 0. Would we use the dimension theorem:
dim(ker(L)) = n - dim(range(L)) = n - dim(W) =dim(Wτ). Since Wτ is contained in ker(L).

*Wτ is orthogonal complement.

Let's use W in a meaningful way since you haven't defined what W represents. Let W = ker(L).

The orthogonal complement of W is the set of all vectors x with property that x ##\cdot## w= 0, where w ##\in## W.

Now, assume that x is any vector in the row space of L. How can you write x, knowing what you know about L and its matrix representation?
 
If a is scalars, then ax1 + ax2 + ... + axk = 0
 
That was the linear combination of the row space. That was how I thought to write x in terms of L. And it's equal to 0 because it is the kernel.
 
LosTacos said:
That was the linear combination of the row space. That was how I thought to write x in terms of L. And it's equal to 0 because it is the kernel.
There are several things wrong with this.

1. There are not k rows in the matrix, so it makes no sense to list them as x1, x2, ... , xk.
2. The row space is not in the kernel. Your goal in this problem is to show that the row space is the orthogonal complement of the kernel.
3. A vector in the kernel doesn't have to be 0.
 
Okay, given matrix A and vector x:

A⋅x = 0 means that wk⋅x = 0 for ever row vector wk in R. Therefore, the orthogonal complement of row space is kernel.
 
LosTacos said:
Okay, given matrix A and vector x:
Any old vector x? Is x in the row space of A or is it in the kernel of A?
LosTacos said:
A⋅x = 0 means that wk⋅x = 0 for ever row vector wk in R.
I don't follow this. Ax = 0 means only that x is in the kernel of A. Why does it follow that wk⋅x = 0? Neither wk or x has to be the zero vector.
LosTacos said:
Therefore, the orthogonal complement of row space is kernel.