Professional Documents
Culture Documents
Notes: A set of vectors is said to be orthogonal iff each vector in the set is orthogonal to each
of the other vectors in the set. An orthonormal set is an orthogonal set in which each vector has
length 1. So to check whether a set is orthogonal, we check that each pair of distinct vectors from
the set has zero dot product. To check whether a set is orthonormal we check: each pair of distinct
vectors has zero dot product; and the dot product of each vector with itself is 1.
Orthonormals sets of vectors are always independent. More generally, orthogonoal sets of nonzero
vectors are always independent.
Problem 3b, §6.3, p308. Is the given set of vectors orthogonal with respect to the Euclidean
inner product on R3 , in other words with respect to the usual dot product on R3 ?
2 2 1 2 1 2 1 2 2
{( , − , ), ( , , − ), ( , , )}
3 3 3 3 3 3 3 3 3
Solution. Writing
2 2 1 2 1 2 1 2 2
u1 = ( , − , ), u2 = ( , , − ), u3 = ( , , )
3 3 3 3 3 3 3 3 3
we need to show that u1 and u2 are orthogonal, u1 and u3 are orthogonal, and u2 and u3 are
orthogonal, in other words that each of these pairs has zero dot product. So just compute:
2 2 1 2 1 2 1
u1 · u2 = ( , − , ) · ( , , − ) = (4 − 2 − 2) = 0
3 3 3 3 3 3 9
2 2 1 1 2 2 1
u1 · u3 = ( , − , ) · ( , , ) = (2 − 4 + 2) = 0
3 3 3 3 3 3 9
2 1 2 1 2 2 1
u2 · u3 = ( , , − ) · ( , , ) = (2 + 2 − 4) = 0
3 3 3 3 3 3 9
For what it’s worth, this set is also orthonormal, since u1 , u2 , and u3 each have length 1.
form an orthogonal basis for R4 with the Euclidean inner product, in other words with the usual
dot product; then use the formula (this is formula (1) p301)
u · v1 u · v2 u · vn
u= v1 + v2 + . . . + vn
||v1 ||2 ||v2 ||2 ||vn||2
u · v1 u · v2 u · vn
= v1 + v2 + . . . + vn
v1 · v1 v2 · v2 vn · vn
Note: If the given basis were orthonormal (so if all basis vectors had length 1), and not merely
orthogonal (basis vectors are pairwise orthogonal), then the denominators in the formula shown
above would all be 1 and the formula would reduce to the simpler one established in class (and
shown in Theorem 6.3.1 p299). The denominators serve the purpose of normalizing the basis vec-
tors, i.e replacing the given basis vectors by scalar multiples that have length one. For instance,
we could rewrite the first term as follows:
u · v1 1 1
2
v1 = (u · v1 ) v1
||v1 || ||v1 || ||v1 ||
Thus the specified formula (1) from p301 in the text is a pretty easy generalization from the formula
in Thm 6.3.1, which we proved in class.
v1 · v2 = −2 − 2 + 6 − 2 = 0
v1 · v3 = 1 − 2 + 0 + 1 = 0
v1 · v4 = 1 + 0 + 0 − 1 = 0
v2 · v3 = −2 + 4 + 0 − 2 = 0
v2 · v4 = −2 + 0 + 0 + 2 = 0
v3 · v4 = 1 + 0 + 0 − 2 = 0
The given four vectors form an orthogonal set. Since none of these vectors is the zero vector, this
orthogonal set must be independent. It also has the correct number of vectors to be a basis for
R4 , so it forms a basis for R4 , hence an orthogonal basis for R4 .
Therefore
1 5 1
u = (1, 1, 1, 1) = v1 + v2 + v3 + 1v4
7 21 3
1 5 1
= (1, −1, 2, −1) + (−2, 2, 3, 2) + (1, 2, 0, −1) + (1, 0, 0, 1)
7 21 3
1
= (3 − 10 + 7 + 1, −3 + 10 + 14 + 0, 6 + 15 + 0 + 0, −3 + 10 − 7 + 21)
21
1
= (21, 21, 21, 21)
21
= (1, 1, 1, 1)
Problem 16a, §6.3, p308. Let R2 have the Euclidean inner product. Given
use the Gram-Schmidt process to transform the basis {u1 , u2 } into an orthonormal basis. Draw
both sets of basis vectors in the xy-plane.
Step 2. Compute
u2 · v1 −4 2 −2 6
v1 = v1 = − (1, −3) = ( , )
v1 · v1 10 5 5 5
−2 6 −2 6 12 4 4
u2 − ( , ) = (2, 2) − ( , ) = ( , ) = (3, 1)
5 5 5 5 5 5 5
Call this result (or some scalar multiple of it, if you’d prefer) v2 :
4
v2 = (3, 1)
5
Or one could use v2 = (3, 1). By now, we have an orthogonal basis {v1 , v2 } for R2 .
Step 3. Normalize v1 and v2 (by multiplying each by the reciprocal of its length) and call the new
vectors w1 and w2 .
1 1
w1 = v1 = √ (1, −3)
||v1 || 10
1 1
w2 = v2 = √ (3, 1)
||v2 || 10
.
1 −3 3 1
The desired orthonormal basis is {w1 , w2 } = {( √ , √ ), ( √ , √ )}. It is easy to check
10 10 10 10
that w1 and w2 are orthogonal and have length 1. I’ll leave it to you to do a sketch.
Problem 18, §6.3, p308. Let R4 have the Euclidean inner product, in other words the usual
dot product. Given
use the Gram-Schmidt process to create from the basis {u1 , u2 , u3 , u4 } an orthonormal basis.
u2 · v1 −2 −2 −4 −2
v1 = v1 = (0, 2, 1, 0) = (0, , , 0)
v1 · v1 5 5 5 5
Subtract this result from u2 and call this result (or some scalar multiple of it) v2 :
−4 −2 4 2 −1 2
v2 = u2 − (0, , , 0) = (1, −1, 0, 0) + (0, , , 0) = (1, , , 0)
5 5 5 5 5 5
If you prefer, use v2 = (5, −1, 2, 0). (But stick to whatever choice you make.) Note that v2 and v1
are orthogonal. By now we have an orthogonal basis {v1 , v2 } for span{u1 , u2 }.
u3 · v1 u3 · v2
Step 3. Compute each of the projections v1 and v2 :
v1 · v1 v2 · v2
u3 · v1 4 4 8 4
v1 = v1 = (0, 2, 1, 0) = (0, , , 0)
v1 · v1 5 5 5 5
u3 · v2 3 1 −1 1
v2 = (5, −1, 2, 0) = ( , , , 0)
v2 · v2 30 2 10 5
Subtract both of these results from u3 and call the result (or some scalar multiple of it) v3 :
8 4 1 −1 1 1 1
v3 = (1, 2, 0, −1) − (0, , , 0) − ( , , , 0) = ( , , −1, −1)
5 5 2 10 5 2 2
If you prefer, use v3 = (1, 1, −2, −2), a choice that’s easier to work with. (But stick to whatever
choice you make.) Note that v3 is orthogonal to each of the vectors v1 and v2 .
By now, we have an orthogonal basis {v1 , v2 , v3 } for span{u1 , u2 , u3 }.
u4 · v1 u4 · v2 u4 · v3
Step 4. Compute each of the projections v1 , v2 , and v3 :
v1 · v1 v2 · v2 v3 · v3
u4 · v1
v1 = 0v1 = (0, 0, 0, 0)
v1 · v1
u4 · v2 5 5 −1 1
v2 = (5, −1, 2, 0) = ( , , , 0)
v2 · v2 30 6 6 3
u4 · v3 −1 −1 −1 1 1
v3 = (1, 1, −2, −2) = ( , , , )
v3 · v3 10 10 10 5 5
5 −1 1 −1 −1 1 1
v4 = u4 − (0, 0, 0, 0) − ( , , , 0) − ( , , , )
6 6 3 10 10 5 5
−5 1 −1 1 1 −1 −1
= (1, 0, 0, 1) + ( , , , 0) + ( , , , )
6 6 3 10 10 5 5
4 4 −8 4
=( , , , )
15 15 15 5
Or, for a choice that’s easier to work with, use the scalar multiple (1, 1, −2, 3) as v4 . Note that
the chosen v4 is, in fact, orthogonal to each of the vectors v1 , v2 , v3 . (This is a good check against
mistakes!) By now, we have an orthogonal basis {v1 , v2 , v3 , v4 } for span{u1 , u2 , u3 , u4 }, in other
words for R4 .
Step 5. Normalize each of the vectors v1 , . . . , v4 (multiply each by the reciprocal of its length) and
call the new vectors w1 , . . . , w4 :
1 1
w1 = v1 = √ (0, 2, 1, 0)
||v1 || 5
1 1
w2 = v2 = √ (5, −1, 2, 0)
||v2 || 30
1 1
w3 = v3 = √ (1, 1, −2, −2)
||v3 || 10
1 1
w4 = v4 = √ (1, 1, −2, 3)
||v4 || 15