How Do You Calculate Orthogonal Projections in Polynomial Subspaces?

  • Thread starter Thread starter Benny
  • Start date Start date
  • Tags Tags
    Subspaces
Benny
Messages
577
Reaction score
0
Hi, can someone shed some light on the following question. It's been bothering me for a while and I'd like to know where I went wrong. Here is what I can remember of the question.

The following is an inner product for polynomials in P_3(degree <= 3): \left\langle {f,g} \right\rangle = \int\limits_{ - 1}^1 {f\left( x \right)g\left( x \right)} dx

Let W be the subspace of the vector space P_3, spanned by {x^2, x^3}.

Find the orthogonal projection of a polynomial p\left( x \right) = a_0 + a_1 x + a_2 x^2 + a_3 x^3 \in P_3 onto W. Find also the polynomial q\left( x \right) \in W which minimises the integral \int\limits_{ - 1}^1 {\left( {3 + 5x - q\left( x \right)} \right)^2 } dx.

I think that q(x) is some kind of projection onto W. I kind of drew an analogy with 'distance' when I did this question. But obviously something's wrong with that approach. Does anyone have suggestions as to how to find q?
 
Physics news on Phys.org
The first question is a really basic one, just compute the orthogonal projection of p onto W using the definition of orthogonal projection. The second question looks like it has nothing to do with linear algebra, it's just high school calculus, at least, that's how I solved it.
 
I included the first question to provide a context for the second question. I know that it's just using the definition.

Looking at the problem from another perspective - Suppose that U is a subspace of some real vector space V spanned by two unit vectors b and c and d is just some element of V (not necessarily in the span of b and c). Then the projection of d onto U is e = <d,b>b + <d,c>c. The vector orthogonal to that projection is simply d - e. I was thinking that it might have something to do with projections.
 
It would be interesting to see if there were an algebraic approach to solving the second question. However, the analytic approach is simple and obvious, albeit inelegant. If you need to just do the problem, you can do it using calculus. If you want to know if there is an algebraic solution, I can't think of one off the top of my head
 
AKG said:
It would be interesting to see if there were an algebraic approach to solving the second question. However, the analytic approach is simple and obvious, albeit inelegant. If you need to just do the problem, you can do it using calculus. If you want to know if there is an algebraic solution, I can't think of one off the top of my head

The orthogonal projection of p\left( x \right) = a_0 + a_1 x + a_2 x^2 + a_3 x^3 \in P_3 on W is, of course, a_2x^2+ a_3x^3.

3+ 5x- q(x)= 0 when q(x)= 3- 5x. If q(x)= 3+ 5x, the integral is 0. That's the minimum isn't it?
 
Halls, q must lie in W = span{x^2, x^3}.
 
There are two things I don't understand about this problem. First, when finding the nth root of a number, there should in theory be n solutions. However, the formula produces n+1 roots. Here is how. The first root is simply ##\left(r\right)^{\left(\frac{1}{n}\right)}##. Then you multiply this first root by n additional expressions given by the formula, as you go through k=0,1,...n-1. So you end up with n+1 roots, which cannot be correct. Let me illustrate what I mean. For this...
Back
Top