1
$\begingroup$

Is the following Proof Correct? For the facility of the reader the formal statement of the Gram-Schimdt Procedure is provided below.

Proposition. If $\{w_1,w_2,\dots, w_n\}$ is an orthogonal set of vectors, then the vectors $v_1,v_2,\dots, v_n$ derived from the Gram-Schmidt process satisfy $v_i=w_i,\forall i = 1,2,\dots,n$.

Proof. Define $I_n = \{1,2,\dots,n\}$ and $S_k' = \{v_1,v_2,\dots,v_k\},\forall k\in I_n$. We proceed by recourse to Induction. For $i=1$ it is evident by the Gram-Schmidt procedure that $v_1 = w_1$. Now assume for an arbitrary $k\in I_n\backslash\{n\}$ that the set $S_k'$ satisfies the proposition that $v_i=w_i,\forall i=1,2,\dots,k$.

Then by using the the inductive hypothesis in tandem with the procedure to determine $v_{k+1}$ yields $$v_{k+1} = w_{k+1}-\sum_{i=1}^{k+1-1}\frac{\langle w_{k+1},v_i\rangle}{||v_i||^2}v_i = w_{k+1}-\sum_{i=1}^{k}\frac{\langle w_{k+1},w_i\rangle}{||w_i||^2}w_i = w_{k+1}$$ where the last equality follow from our original hypothesis that $\{w_1,w_2,\dots, w_n\}$ is orthogonal completing the induction.

$\blacksquare$


Theorem. Let $V$ be an inner product space and $S = \{w_1,w_2,\dots.w_n\}$ be a linearly independent subset of $V$. Define $S' = \{v_1,v_2,\dots,v_n\}$, where $v_1 = w_1$ and $$v_k = w_k-\sum_{j=1}^{k-1}\frac{\langle w_k,v_j\rangle}{||v_j||^2}v_j\text{ for }2\leq k\leq n.$$ The $S'$ is an orthogonal set of nonzero vectors such that $\operatorname{span}(S') = \operatorname{span}(S)$.

$\endgroup$
2
$\begingroup$

Yes, other than the fact that there's a typo. What you meant to write was:\begin{align}v_{k+1}&= w_{k+1}-\sum_{i=1}^{k+1-1}\frac{\langle w_{k+1},v_i\rangle}{||v_i||^2}v_i\\&=w_{k+1}-\sum_{i=1}^{k}\frac{\langle w_{k+1},w_i\rangle}{||w_i||^2}w_i\\&=w_{k+1}.\end{align}

$\endgroup$
  • $\begingroup$ Thankyou Sir. Your counsel is appreciated as always. $\endgroup$ – Atif Farooq Jun 11 '18 at 17:15
  • $\begingroup$ @AtifFarooq I'm glad I could help. $\endgroup$ – José Carlos Santos Jun 11 '18 at 17:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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