Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Given that $P$ is a projection linear map, s.t. $P^2=P$. I can show that for some vector space $V$ this implies that $V=\text{Ran}(P) \bigoplus \text{Ker}(P)$. Basically, $v=(I-P)v+Pv$ and $P(I-P)v=0$. Then $\text{Ker}(P)\cap \text{Ran}(P) = \{0\}$

I don't know whether I am failing to take into account the infinite dimensional $V$. Does this have any effect? How might I take into account of this?

Thanks in advance!

share|improve this question
    
Doesn't your argument show that any vector of $V$ can be represented as a sum of something from $\mathrm{Im}(P)$ plus something from $\ker(P)$? And also that $\mathrm{Im}(P)\cap\ker(P)=\{0\}$? Looks like a direct sum decomposition to me! –  Jyrki Lahtonen Sep 11 '11 at 19:17

1 Answer 1

up vote 2 down vote accepted

In order to show that $V = W\oplus Z$, you need to show that $W\cap Z=\{\mathbf{0}\}$ and that $V=W+Z$. The problem of finite dimension does not arise, unless you attempt to show one of the two steps by using dimension arguments.

Implicit in your argument is that $\mathrm{Ran}(I-P) = \mathrm{Ker}(P)$. This does not require $V$ to be finite dimensional: for $P(I-P)(v) = (P-P^2)(v) = Pv - P^v = Pv-Pv = \mathbf{0}$; and if $w\in\mathrm{Ker}(P)$, then $w = w-\mathbf{0} = w-P(w) = (I-P)(w)$. No dimension argument used.

Your first step, showing that $v=(I-P)v+Pv$ shows that $V\subseteq \mathrm{Ker}(P)+\mathrm{Im}(P)$; the other inclusion is trivial. No dimension arguments used.

And your second step is noting that $\mathrm{Ran}(I-P)\cap\mathrm{Ran}(P)=\{\mathbf{0}\}$; again, no dimension arguments are used: if $v$ lies in the intersection, then $v=Pw$ for some $w$, and $v=(I-P)z$ for some $z$. Then $w = Pw = P^2w = P(Pw) = Pv = P(I-P)z = \mathbf{0}$, so $v=P\mathbf{0}=\mathbf{0}$.

Since nowhere are any assumptions made on the dimension of $V$, the argument holds for any vector space, whether finite dimensional or infinite dimensional.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.