Proof regarding orthogonal projections onto spans

In summary, the problem is to show that Px*y is equal to Pu*y plus Pv*y if and only if the space U is orthogonal to the space V for all y in R^n. To solve this, we can consider a simple case where U and V are each 1-dimensional lines in R^3 and Px, Pu, and Pv represent orthogonal projections onto these lines. By expressing y in terms of the basis vectors of U and V, we can see that the theorem holds true for all possibilities of y being in U, V, or neither. This can be generalized by considering the composition of y in the span of U and V and expressing it in terms of the basis vectors.
  • #1
samuelr0750
3
0

Homework Statement



Let U be the span of k vectors, {u1, ... ,uk} and Pu be the orthogonal projection onto U. Let V be the span of l vectors, {v1, ... vl} and Pv be the orthogonal projection onto V. Let X be the span of {u1, ..., uk, v1, ... vl} and Px be the orthogonal projection onto X.

Show Px*y = Pu*y + Pv*y if and only if the space U is orthogonal to the space V (for all y in R^n).

I'm having trouble on both sides of this if and only if proof. Any help? thanks.

Homework Equations



See above

The Attempt at a Solution



I'm a bit lost - this seems intuitive but I'm having trouble processing it...
 
Physics news on Phys.org
  • #2
Just to get some insight, start with a simple case. Take R3 as your underlying space and let U and V each be 1 dimensional. So each is a line. Pu is on the U line and Pv is on the V line. What is the subspace X spanned by the basis vectors of U and V? ? If y is in X can you see why the theorem would be true in this case? Consider 3 possibilities: y is in U, y is in V or y is in neither.

To generalize, consider the composition of y in X. The u's and v's are the basis of X, so you can express y in terms of those basis vectors. If you do that, and think about the situation in the above paragraph, perhaps you can get started.
 

FAQ: Proof regarding orthogonal projections onto spans

What is an orthogonal projection?

An orthogonal projection is a mathematical operation that projects a vector onto a subspace in such a way that the projected vector is perpendicular (orthogonal) to the subspace. This means that the projected vector is the shortest distance from the original vector to the subspace.

What is a span in relation to orthogonal projections?

In the context of orthogonal projections, a span refers to the set of all possible linear combinations of a given set of vectors. This set forms a subspace, and the orthogonal projection of a vector onto this subspace is the vector that lies within the span and is closest to the original vector.

How do you calculate the orthogonal projection onto a span?

The calculation of an orthogonal projection onto a span involves finding the orthogonal projection matrix, which is the matrix that projects any vector onto the given span. This can be done by using the Gram-Schmidt process to find an orthonormal basis for the span, and then using that basis to construct the projection matrix.

What is the purpose of orthogonal projections onto spans?

Orthogonal projections onto spans have various applications in mathematics, computer science, and engineering. They are used to solve systems of linear equations, perform data compression, and find the best fit for a set of data points, among other things.

Can orthogonal projections onto spans be extended to higher dimensions?

Yes, orthogonal projections onto spans can be extended to higher dimensions. The concept of a span and the Gram-Schmidt process can be generalized to any number of dimensions, allowing for the calculation of orthogonal projections onto higher dimensional subspaces.

Back
Top