Linear transformations and orthogonal basis

Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
4 replies · 2K views
stunner5000pt
Messages
1,447
Reaction score
5
Let {E1,E2,...En} be an orthogonal basis of Rn. Given k, 1<=k<=n, define Pk: Rn -> Rn by [itex]P_{k} (r_{1} E_{1} + ... + r_{n} E_{n}) = r_{k} E_{k}.[/itex] Show that [itex]P_{k} = proj_{U} ()[/itex] where U = span {Ek}

well [tex]\mbox{proj}_{U} \vec{m}= \sum_{i} \frac{ m \bullet u_{i}}{||u_{i}||^2} \vec{u}[/tex]
right?
here we have Pk transforming linear combination of the orthogonal basis into rk Ek same index as the subscript of P

would it turn into
[tex]\mbox{proj}_{U} \vec{m}= \frac{ m \bullet E_{1}}{||E_{1}||^2}\vec{E_{1}} + ... + \frac{ m \bullet E_{n}}{||E_{n}||^2}\vec{E_{n}}[/tex]
and the whole stuff in front of each Ei can be interpreted as the Ri, a scalar multiple yes?
 
Physics news on Phys.org
You can approach it with fewer symbols by simply noting that the orthogonal projection of a vector is the unique vector that is orthogonal to (the vector minus the projection fo the vector). That is, iff v is a vector and u is its orthogonal projection onto some subspace U, then (v - u) . u = 0, and u is in U.
 
0rthodontist said:
You can approach it with fewer symbols by simply noting that the orthogonal projection of a vector is the unique vector that is orthogonal to (the vector minus the projection fo the vector). That is, iff v is a vector and u is its orthogonal projection onto some subspace U, then (v - u) . u = 0, and u is in U.

why is u = 0?
 
u is not 0
(v-u).u is 0 (i.e., v-u is orthogonal to u)
 
ok so (v-u).u = 0
but how does this equal to the Pk?