Multivariable calculus · Topic 1

Solution manual

Vectors, dot products, and projections

24 problems · Worked solutions for every part · Exact answers and geometric explanations

Companion worksheet: vector_algebra_dot_products_practice.tex

1. Vector algebra and coordinates

Coordinates describe points; displacement vectors describe changes in position. Subtract the starting point’s coordinates from the ending point’s coordinates.

1. Position, displacement, distance, and midpoint

Given: P=(−2,1,3)P=(-2,1,3) and Q=(4,−3,6)Q=(4,-3,6).

(a) Position vectors

A position vector starts at the origin, so its components are the coordinates of its endpoint: 𝒑=⟨−2,1,3⟩,𝒒=⟨4,−3,6⟩. \mathbf p=\langle-2,1,3\rangle,\qquad \mathbf q=\langle4,-3,6\rangle.

(b) Displacement vectors

Subtract in the order “end minus start”: PQ→=𝒒−𝒑=⟨4−(−2),−3−1,6−3⟩=⟨6,−4,3⟩. \overrightarrow{PQ}=\mathbf q-\mathbf p =\langle4-(-2),-3-1,6-3\rangle=\langle6,-4,3\rangle. Reversing the endpoints reverses the vector: QP→=−PQ→=⟨−6,4,−3⟩. \overrightarrow{QP}=-\overrightarrow{PQ}=\langle-6,4,-3\rangle.

(c) Distance

Distance is the magnitude of the displacement: dist⁡(P,Q)=62+(−4)2+32=61. \operatorname{dist}(P,Q)=\sqrt{6^2+(-4)^2+3^2}=\sqrt{61}.

(d) Midpoint

Average corresponding coordinates: M=(−2+42,1−32,3+62)=(1,−1,92). M=\left(\frac{-2+4}{2},\frac{1-3}{2},\frac{3+6}{2}\right) =\left(1,-1,\frac92\right).

2. Vector arithmetic in components

Given: 𝒂=2𝒊−𝒋+3𝒌\mathbf a=2\mathbf i-\mathbf j+3\mathbf k and 𝒃=−𝒊+4𝒋−2𝒌\mathbf b=-\mathbf i+4\mathbf j-2\mathbf k.

(a) Component form

Read the coefficients of 𝒊,𝒋,𝒌\mathbf i,\mathbf j,\mathbf k in order: 𝒂=⟨2,−1,3⟩,𝒃=⟨−1,4,−2⟩. \mathbf a=\langle2,-1,3\rangle,\qquad \mathbf b=\langle-1,4,-2\rangle.

(b) Linear combination

Scale first, then subtract componentwise: 3𝒂−2𝒃=⟨6,−3,9⟩−⟨−2,8,−4⟩=⟨8,−11,13⟩. 3\mathbf a-2\mathbf b =\langle6,-3,9\rangle-\langle-2,8,-4\rangle =\langle8,-11,13\rangle.

(c) Magnitude of the sum

First add the vectors: 𝒂+𝒃=⟨1,3,1⟩\mathbf a+\mathbf b=\langle1,3,1\rangle. Therefore ‖𝒂+𝒃‖=12+32+12=11. \lVert\mathbf a+\mathbf b\rVert=\sqrt{1^2+3^2+1^2}=\sqrt{11}.

3. Solving a vector equation

Given: 2𝒙+𝒂=3𝒃2\mathbf x+\mathbf a=3\mathbf b, with 𝒂=⟨1,−2,3⟩\mathbf a=\langle1,-2,3\rangle and 𝒃=⟨3,0,−1⟩\mathbf b=\langle3,0,-1\rangle.

Isolate the unknown vector just as you would isolate a scalar: 𝒙=3𝒃−𝒂2. \mathbf x=\frac{3\mathbf b-\mathbf a}{2}. 3𝒃−𝒂=⟨9,0,−3⟩−⟨1,−2,3⟩=⟨8,2,−6⟩. 3\mathbf b-\mathbf a =\langle9,0,-3\rangle-\langle1,-2,3\rangle =\langle8,2,-6\rangle. Thus 𝒙=⟨4,1,−3⟩\boxed{\mathbf x=\langle4,1,-3\rangle}.

Substitution check: 2⟨4,1,−3⟩+⟨1,−2,3⟩=⟨9,0,−3⟩=3𝒃. 2\langle4,1,-3\rangle+\langle1,-2,3\rangle =\langle9,0,-3\rangle=3\mathbf b.

4. Unit vectors and prescribed magnitude

Given: 𝒖=⟨−2,3,6⟩\mathbf u=\langle-2,3,6\rangle.

(a) Magnitude and unit vector

‖𝒖‖=4+9+36=7. \lVert\mathbf u\rVert=\sqrt{4+9+36}=7. Divide by this positive magnitude to preserve direction: 𝒖̂=𝒖7=⟨−27,37,67⟩. \widehat{\mathbf u}=\frac{\mathbf u}{7} =\left\langle-\frac27,\frac37,\frac67\right\rangle.

(b) All parallel vectors of magnitude 14

Every such vector has the form c𝒖c\mathbf u. Its magnitude is 7|c|7|c|, so 7|c|=14⇒c=2orc=−2. 7|c|=14\quad\Longrightarrow\quad c=2\ \text{or}\ c=-2. The two answers are 2𝒖=⟨−4,6,12⟩(same direction), 2\mathbf u=\langle-4,6,12\rangle\quad\text{(same direction)}, −2𝒖=⟨4,−6,−12⟩(opposite direction). -2\mathbf u=\langle4,-6,-12\rangle\quad\text{(opposite direction)}. These exhaust the possibilities because every parallel vector is a scalar multiple of 𝒖\mathbf u.

5. Testing parallelism

Given: 𝒗=⟨2,−3,1⟩\mathbf v=\langle2,-3,1\rangle. A single scalar must multiply all three components.

(a) First vector

𝒘1=⟨−4,6,−2⟩=−2𝒗. \mathbf w_1=\langle-4,6,-2\rangle=-2\mathbf v. The vectors are parallel and point in opposite directions.

(b) Second vector

𝒘2=⟨6,−9,3⟩=3𝒗. \mathbf w_2=\langle6,-9,3\rangle=3\mathbf v. The vectors are parallel and point in the same direction.

(c) Third vector

For 𝒘3=⟨4,−6,3⟩\mathbf w_3=\langle4,-6,3\rangle, the first two components suggest a multiplier of 22, but 2𝒗=⟨4,−6,2⟩≠𝒘3. 2\mathbf v=\langle4,-6,2\rangle\ne\mathbf w_3. The third component fails, so the vectors are not parallel.

6. A parameter that makes vectors parallel

We need a scalar cc such that ⟨k,4,−2⟩=c⟨3,−2,1⟩. \langle k,4,-2\rangle=c\langle3,-2,1\rangle. The second component gives 4=−2c4=-2c, hence c=−2c=-2. The third component confirms this value: −2=c-2=c. The first component then gives k=3c=−6. k=3c=-6. Indeed, ⟨−6,4,−2⟩=−2⟨3,−2,1⟩\langle-6,4,-2\rangle=-2\langle3,-2,1\rangle. Thus k=−6k=-6 is the only value, and the vectors point in opposite directions.

7. Recovering an endpoint from a midpoint

Given: M=(2,−1,4)M=(2,-1,4) is the midpoint of ABAB, and A=(−3,2,1)A=(-3,2,1).

(a) The missing endpoint

In position-vector notation, 𝒎=(𝒂+𝒃)/2\mathbf m=(\mathbf a+\mathbf b)/2, so 𝒃=2𝒎−𝒂\mathbf b=2\mathbf m-\mathbf a: 𝒃=⟨4,−2,8⟩−⟨−3,2,1⟩=⟨7,−4,7⟩. \mathbf b=\langle4,-2,8\rangle-\langle-3,2,1\rangle =\langle7,-4,7\rangle. Therefore B=(7,−4,7)\boxed{B=(7,-4,7)}.

(b) Displacement from A to B

AB→=𝒃−𝒂=⟨10,−6,6⟩=2⟨5,−3,3⟩. \overrightarrow{AB}=\mathbf b-\mathbf a =\langle10,-6,6\rangle=2\langle5,-3,3\rangle.

(c) Unit vector toward B

The displacement has magnitude 100+36+36=243\sqrt{100+36+36}=2\sqrt{43}. Therefore the desired unit vector is AB→‖AB→‖=⟨5,−3,3⟩43. \frac{\overrightarrow{AB}}{\lVert\overrightarrow{AB}\rVert} =\frac{\langle5,-3,3\rangle}{\sqrt{43}}. Its squared magnitude is (25+9+9)/43=1(25+9+9)/43=1, and the positive scale factor preserves the direction toward BB.

8. Collinearity and the middle point

Given: P=(1,−2,0)P=(1,-2,0), Q=(3,1,−1)Q=(3,1,-1), and R=(k,7,−3)R=(k,7,-3).

(a) The parameter value

Use two displacements with the same starting point: PQ→=⟨2,3,−1⟩,PR→=⟨k−1,9,−3⟩. \overrightarrow{PQ}=\langle2,3,-1\rangle,\qquad \overrightarrow{PR}=\langle k-1,9,-3\rangle. For collinearity, require PR→=cPQ→\overrightarrow{PR}=c\overrightarrow{PQ}. The second and third components both force c=3c=3. Hence k−1=3(2)=6⇒k=7. k-1=3(2)=6\quad\Longrightarrow\quad \boxed{k=7}. All components then agree: PR→=⟨6,9,−3⟩=3PQ→\overrightarrow{PR}=\langle6,9,-3\rangle=3\overrightarrow{PQ}.

(b) Which point is between the others?

PQ→=13PR→. \overrightarrow{PQ}=\frac13\overrightarrow{PR}. Starting at PP, reaching QQ takes one third of the displacement to RR. Thus QQ lies between PP and RR.

2. Dot products and angles

The dot product measures alignment. For nonzero vectors, divide it by the product of the magnitudes to obtain the cosine of their angle. Keep the sign: a negative dot product can give an obtuse or straight angle.

9. Classifying angles by the dot product

Given: 𝒂=⟨1,−2,2⟩\mathbf a=\langle1,-2,2\rangle. Each pair below consists of nonzero, nonparallel vectors.

(a) First pair

For 𝒃=⟨2,1,3⟩\mathbf b=\langle2,1,3\rangle, 𝒂⋅𝒃=1(2)+(−2)(1)+2(3)=6>0. \mathbf a\cdot\mathbf b=1(2)+(-2)(1)+2(3)=6>0. The angle is acute.

(b) Second pair

For 𝒃=⟨2,3,2⟩\mathbf b=\langle2,3,2\rangle, 𝒂⋅𝒃=1(2)+(−2)(3)+2(2)=0. \mathbf a\cdot\mathbf b=1(2)+(-2)(3)+2(2)=0. The angle is right.

(c) Third pair

For 𝒃=⟨−1,1,−1⟩\mathbf b=\langle-1,1,-1\rangle, 𝒂⋅𝒃=1(−1)+(−2)(1)+2(−1)=−5<0. \mathbf a\cdot\mathbf b=1(-1)+(-2)(1)+2(-1)=-5<0. The angle is obtuse. No inverse-cosine calculation is needed.

10. Computing vector angles

(a) A familiar exact angle

For 𝒂=⟨1,1,0⟩\mathbf a=\langle1,1,0\rangle and 𝒃=⟨0,1,1⟩\mathbf b=\langle0,1,1\rangle, the dot product is 11, and both magnitudes are 2\sqrt2. Therefore cos⁡θ=122=12,θ=π3=60.0∘. \cos\theta=\frac{1}{\sqrt2\sqrt2}=\frac12, \qquad \boxed{\theta=\frac\pi3=60.0^\circ}.

(b) An obtuse angle

For 𝒖=⟨2,−1,2⟩\mathbf u=\langle2,-1,2\rangle and 𝒗=⟨1,3,−1⟩\mathbf v=\langle1,3,-1\rangle, 𝒖⋅𝒗=2−3−2=−3,‖𝒖‖=3,‖𝒗‖=11. \mathbf u\cdot\mathbf v=2-3-2=-3, \qquad \lVert\mathbf u\rVert=3,\quad \lVert\mathbf v\rVert=\sqrt{11}. Thus cos⁡θ=−3311=−111, \cos\theta=\frac{-3}{3\sqrt{11}}=-\frac1{\sqrt{11}}, θ=cos⁡−1(−111)≈107.5∘. \boxed{\theta=\cos^{-1}\!\left(-\frac1{\sqrt{11}}\right)\approx107.5^\circ}. The negative dot product agrees with an angle greater than 90∘90^\circ. Taking its absolute value would give the wrong vector angle.

11. Classifying an angle as a parameter changes

Given: 𝒖(t)=⟨t,2,−1⟩\mathbf u(t)=\langle t,2,-1\rangle and 𝒗=⟨3,−1,4⟩\mathbf v=\langle3,-1,4\rangle. 𝒖(t)⋅𝒗=3t−2−4=3t−6. \mathbf u(t)\cdot\mathbf v=3t-2-4=3t-6. Both vectors are always nonzero. They can never be parallel: matching the second components would require multiplier −2-2, which would make the third component −8-8, rather than −1-1. Thus angles 00 and π\pi never occur.

(a) Right angle

3t−6=0⇒t=2. 3t-6=0\quad\Longrightarrow\quad \boxed{t=2}.

(b) Acute angle

3t−6>0⇒t>2. 3t-6>0\quad\Longrightarrow\quad \boxed{t>2}.

(c) Obtuse angle

3t−6<0⇒t<2. 3t-6<0\quad\Longrightarrow\quad \boxed{t<2}.

12. A right triangle in space

Given: A=(1,0,1)A=(1,0,1), B=(3,1,1)B=(3,1,1), and C=(0,2,3)C=(0,2,3).

(a) Locate a right angle

To test the angle at AA, use two vectors that both start there: AB→=⟨2,1,0⟩,AC→=⟨−1,2,2⟩. \overrightarrow{AB}=\langle2,1,0\rangle,\qquad \overrightarrow{AC}=\langle-1,2,2\rangle. AB→⋅AC→=2(−1)+1(2)+0(2)=0. \overrightarrow{AB}\cdot\overrightarrow{AC}=2(-1)+1(2)+0(2)=0. Both vectors are nonzero, so the right angle is at AA.

(b) Check with side lengths

AB=22+12=5,AC=(−1)2+22+22=3. AB=\sqrt{2^2+1^2}=\sqrt5,\qquad AC=\sqrt{(-1)^2+2^2+2^2}=3. The remaining displacement is BC→=⟨−3,1,2⟩\overrightarrow{BC}=\langle-3,1,2\rangle, giving BC=14BC=\sqrt{14}. Consequently AB2+AC2=5+9=14=BC2. AB^2+AC^2=5+9=14=BC^2. This confirms that BCBC is the hypotenuse, opposite the right angle at AA.

13. Using magnitudes and an angle

Given: ‖𝒂‖=3\lVert\mathbf a\rVert=3, ‖𝒃‖=4\lVert\mathbf b\rVert=4, and θ=2π/3\theta=2\pi/3. Coordinates are unnecessary.

(a) Dot product

𝒂⋅𝒃=3(4)cos⁡2π3=12(−12)=−6. \mathbf a\cdot\mathbf b=3(4)\cos\frac{2\pi}{3} =12\left(-\frac12\right)=\boxed{-6}.

(b) Magnitude of the sum

‖𝒂+𝒃‖2=32+2(−6)+42=13. \lVert\mathbf a+\mathbf b\rVert^2 =3^2+2(-6)+4^2=13. Hence ‖𝒂+𝒃‖=13\boxed{\lVert\mathbf a+\mathbf b\rVert=\sqrt{13}}.

(c) Magnitude of the difference

‖𝒂−𝒃‖2=32−2(−6)+42=37. \lVert\mathbf a-\mathbf b\rVert^2 =3^2-2(-6)+4^2=37. Hence ‖𝒂−𝒃‖=37\boxed{\lVert\mathbf a-\mathbf b\rVert=\sqrt{37}}. Magnitudes use the nonnegative square root.

14. Recovering a dot product from a sum

Given: ‖𝒖‖=5\lVert\mathbf u\rVert=5, ‖𝒗‖=2\lVert\mathbf v\rVert=2, and ‖𝒖+𝒗‖=33\lVert\mathbf u+\mathbf v\rVert=\sqrt{33}.

(a) Dot product

Expand the squared magnitude: 33=25+2(𝒖⋅𝒗)+4. 33=25+2(\mathbf u\cdot\mathbf v)+4. Therefore 2(𝒖⋅𝒗)=42(\mathbf u\cdot\mathbf v)=4, so 𝒖⋅𝒗=2\boxed{\mathbf u\cdot\mathbf v=2}.

(b) Angle

cos⁡θ=25(2)=15,θ=cos⁡−1(15)≈78.5∘. \cos\theta=\frac{2}{5(2)}=\frac15, \qquad \boxed{\theta=\cos^{-1}\!\left(\frac15\right)\approx78.5^\circ}.

(c) Magnitude of the difference

‖𝒖−𝒗‖2=25−2(2)+4=25,‖𝒖−𝒗‖=5. \lVert\mathbf u-\mathbf v\rVert^2=25-2(2)+4=25, \qquad \boxed{\lVert\mathbf u-\mathbf v\rVert=5}.

15. Which dot products are possible?

Given: ‖𝒂‖=2\lVert\mathbf a\rVert=2 and ‖𝒃‖=3\lVert\mathbf b\rVert=3; proposed dot products are 7,−6,0,37,-6,0,3.

(a) Apply the bound

|𝒂⋅𝒃|≤‖𝒂‖‖𝒃‖=6. |\mathbf a\cdot\mathbf b|\le\lVert\mathbf a\rVert\lVert\mathbf b\rVert=6. Thus 77 is impossible. The values −6,0,3-6,0,3 are possible.

(b) Corresponding exact angles

Divide each allowed dot product by 66:

Dot product cos⁡θ\cos\theta Exact angle
−6-6 −1-1 π\pi
00 00 π/2\pi/2
33 1/21/2 π/3\pi/3

Each is realizable: take 𝒂=⟨2,0,0⟩\mathbf a=\langle2,0,0\rangle, and choose 𝒃\mathbf b, respectively, as ⟨−3,0,0⟩,⟨0,3,0⟩,⟨32,332,0⟩. \langle-3,0,0\rangle,\qquad \langle0,3,0\rangle,\qquad \left\langle\frac32,\frac{3\sqrt3}{2},0\right\rangle. Each choice has magnitude 33 and the required dot product. The angle π\pi is a straight angle, with opposite directions.

16. Equal magnitudes and the zero-vector exception

(a) Prove the equivalence

Subtract the two squared-norm identities: ‖𝒂+𝒃‖2−‖𝒂−𝒃‖2=4𝒂⋅𝒃. \lVert\mathbf a+\mathbf b\rVert^2-\lVert\mathbf a-\mathbf b\rVert^2 =4\mathbf a\cdot\mathbf b. If the two magnitudes are equal, the left side is zero, so 𝒂⋅𝒃=0\mathbf a\cdot\mathbf b=0. Conversely, if the dot product is zero, the squared magnitudes are equal. Because magnitudes are nonnegative, the magnitudes themselves are equal. This proves both directions: ‖𝒂+𝒃‖=‖𝒂−𝒃‖⇔𝒂⋅𝒃=0. \lVert\mathbf a+\mathbf b\rVert=\lVert\mathbf a-\mathbf b\rVert \quad\Longleftrightarrow\quad\mathbf a\cdot\mathbf b=0. Since the given vectors are nonzero, this occurs exactly when their angle is π/2\pi/2.

(b) Why the zero vector has no vector angle

The proposed cosine calculation would be 𝟎⋅𝒃‖𝟎‖‖𝒃‖=00, \frac{\mathbf0\cdot\mathbf b}{\lVert\mathbf0\rVert\lVert\mathbf b\rVert} =\frac00, which is undefined. The zero vector has no direction, so it cannot determine an angle with 𝒃\mathbf b, even though its dot product with 𝒃\mathbf b is zero.

3. Projection and orthogonal decomposition

For a nonzero target vector 𝒃\mathbf b, the worksheet uses proj⁡𝒃𝒂=𝒂⋅𝒃‖𝒃‖2𝒃, \operatorname{proj}_{\mathbf b}\mathbf a =\frac{\mathbf a\cdot\mathbf b}{\lVert\mathbf b\rVert^2}\mathbf b, orth⁡𝒃𝒂=𝒂−proj⁡𝒃𝒂. \operatorname{orth}_{\mathbf b}\mathbf a =\mathbf a-\operatorname{proj}_{\mathbf b}\mathbf a. Both outputs are vectors. The signed scalar projection along 𝒃\mathbf b is the number (𝒂⋅𝒃)/‖𝒃‖(\mathbf a\cdot\mathbf b)/\lVert\mathbf b\rVert. Its sign matters; it is not the perpendicular vector component.

Parallel and perpendicular components A diagonal vector a is the sum of a horizontal vector proj sub b a and a vertical vector orth sub b a. The horizontal component lies along b, and a right-angle marker shows the vertical component is perpendicular to b. Oaprojb aorthb ab
Schematic for a positive scalar projection. Add the parallel component and the perpendicular remainder to recover the original vector.

17. Computing both vector components

Given: 𝒂=⟨4,1,3⟩\mathbf a=\langle4,1,3\rangle and 𝒃=⟨1,2,2⟩\mathbf b=\langle1,2,2\rangle.

(a) Projection and perpendicular remainder

Compute the numerator and denominator: 𝒂⋅𝒃=4+2+6=12,‖𝒃‖2=1+4+4=9. \mathbf a\cdot\mathbf b=4+2+6=12,\qquad \lVert\mathbf b\rVert^2=1+4+4=9. Therefore proj⁡𝒃𝒂=43⟨1,2,2⟩=⟨43,83,83⟩, \operatorname{proj}_{\mathbf b}\mathbf a =\frac43\langle1,2,2\rangle =\left\langle\frac43,\frac83,\frac83\right\rangle, orth⁡𝒃𝒂=𝒂−proj⁡𝒃𝒂=⟨83,−53,13⟩. \operatorname{orth}_{\mathbf b}\mathbf a =\mathbf a-\operatorname{proj}_{\mathbf b}\mathbf a =\left\langle\frac83,-\frac53,\frac13\right\rangle.

(b) Check the decomposition

The components add back to the original vector: ⟨43,83,83⟩+⟨83,−53,13⟩=⟨4,1,3⟩. \left\langle\frac43,\frac83,\frac83\right\rangle +\left\langle\frac83,-\frac53,\frac13\right\rangle =\langle4,1,3\rangle. The remainder is perpendicular to 𝒃\mathbf b: (orth⁡𝒃𝒂)⋅𝒃=83−103+23=0. \left(\operatorname{orth}_{\mathbf b}\mathbf a\right)\cdot\mathbf b =\frac83-\frac{10}3+\frac23=0.

18. A negative scalar projection

Given: 𝒂=⟨1,−4,2⟩\mathbf a=\langle1,-4,2\rangle and 𝒃=⟨2,1,−2⟩\mathbf b=\langle2,1,-2\rangle. Here 𝒂⋅𝒃=2−4−4=−6\mathbf a\cdot\mathbf b=2-4-4=-6 and ‖𝒃‖=3\lVert\mathbf b\rVert=3.

(a) Signed scalar projection

𝒂⋅𝒃‖𝒃‖=−63=−2. \frac{\mathbf a\cdot\mathbf b}{\lVert\mathbf b\rVert} =\frac{-6}{3}=\boxed{-2}. This is a signed number, so a negative answer is valid.

(b) Vector projection and orthogonal component

proj⁡𝒃𝒂=−23⟨2,1,−2⟩=⟨−43,−23,43⟩. \operatorname{proj}_{\mathbf b}\mathbf a =-\frac23\langle2,1,-2\rangle =\left\langle-\frac43,-\frac23,\frac43\right\rangle. orth⁡𝒃𝒂=⟨1,−4,2⟩−proj⁡𝒃𝒂=⟨73,−103,23⟩. \operatorname{orth}_{\mathbf b}\mathbf a =\langle1,-4,2\rangle-\operatorname{proj}_{\mathbf b}\mathbf a =\left\langle\frac73,-\frac{10}3,\frac23\right\rangle. The remainder’s dot product with 𝒃\mathbf b is 14/3−10/3−4/3=014/3-10/3-4/3=0.

(c) Direction

The vector projection is a negative multiple of 𝒃\mathbf b, so it points in the opposite direction. This matches the negative scalar projection. Its magnitude is 22, while its signed scalar projection is −2-2.

19. Projecting onto a unit direction

Given: 𝒂=⟨1,2,3⟩\mathbf a=\langle1,2,3\rangle and 𝒖=⟨3/5,0,4/5⟩\mathbf u=\langle3/5,0,4/5\rangle.

(a) Check the unit vector

‖𝒖‖2=925+0+1625=1, \lVert\mathbf u\rVert^2=\frac9{25}+0+\frac{16}{25}=1, so ‖𝒖‖=1\lVert\mathbf u\rVert=1.

(b) Find both components

𝒂⋅𝒖=35+0+125=3. \mathbf a\cdot\mathbf u=\frac35+0+\frac{12}5=3. Since the denominator ‖𝒖‖2\lVert\mathbf u\rVert^2 is 11, proj⁡𝒖𝒂=3𝒖=⟨95,0,125⟩, \operatorname{proj}_{\mathbf u}\mathbf a=3\mathbf u =\left\langle\frac95,0,\frac{12}5\right\rangle, orth⁡𝒖𝒂=𝒂−3𝒖=⟨−45,2,35⟩. \operatorname{orth}_{\mathbf u}\mathbf a =\mathbf a-3\mathbf u =\left\langle-\frac45,2,\frac35\right\rangle.

(c) Check the squared-norm identity

‖𝒂‖2=1+4+9=14, \lVert\mathbf a\rVert^2=1+4+9=14, ‖proj⁡𝒖𝒂‖2=8125+14425=9, \lVert\operatorname{proj}_{\mathbf u}\mathbf a\rVert^2 =\frac{81}{25}+\frac{144}{25}=9, ‖orth⁡𝒖𝒂‖2=1625+4+925=5. \lVert\operatorname{orth}_{\mathbf u}\mathbf a\rVert^2 =\frac{16}{25}+4+\frac9{25}=5. Thus 14=9+514=9+5, as required. This is the Pythagorean theorem for the two perpendicular vector components.

20. Parallel, perpendicular, and zero inputs

Target: 𝒃=⟨1,−1,2⟩≠𝟎\mathbf b=\langle1,-1,2\rangle\ne\mathbf0.

(a) A parallel input

Here 𝒂=⟨3,−3,6⟩=3𝒃\mathbf a=\langle3,-3,6\rangle=3\mathbf b, so all of 𝒂\mathbf a lies along the target direction: proj⁡𝒃𝒂=⟨3,−3,6⟩,orth⁡𝒃𝒂=𝟎. \operatorname{proj}_{\mathbf b}\mathbf a=\langle3,-3,6\rangle, \qquad \operatorname{orth}_{\mathbf b}\mathbf a=\mathbf0.

(b) A perpendicular input

For 𝒂=⟨2,0,−1⟩\mathbf a=\langle2,0,-1\rangle, 𝒂⋅𝒃=2+0−2=0. \mathbf a\cdot\mathbf b=2+0-2=0. There is no component along 𝒃\mathbf b: proj⁡𝒃𝒂=𝟎,orth⁡𝒃𝒂=⟨2,0,−1⟩. \operatorname{proj}_{\mathbf b}\mathbf a=\mathbf0, \qquad \operatorname{orth}_{\mathbf b}\mathbf a=\langle2,0,-1\rangle.

(c) The zero input

With 𝒂=𝟎\mathbf a=\mathbf0, both components are zero: proj⁡𝒃𝟎=𝟎,orth⁡𝒃𝟎=𝟎. \operatorname{proj}_{\mathbf b}\mathbf0=\mathbf0, \qquad \operatorname{orth}_{\mathbf b}\mathbf0=\mathbf0. These formulas are defined because the target 𝒃\mathbf b is nonzero. A zero input does not cause division by zero here.

21. Rescaling the target vector

Given: 𝒂=⟨2,1,−1⟩\mathbf a=\langle2,1,-1\rangle and 𝒃=⟨1,0,1⟩\mathbf b=\langle1,0,1\rangle.

(a) The original target

𝒂⋅𝒃=1\mathbf a\cdot\mathbf b=1 and ‖𝒃‖2=2\lVert\mathbf b\rVert^2=2, so proj⁡𝒃𝒂=12𝒃=⟨12,0,12⟩, \operatorname{proj}_{\mathbf b}\mathbf a =\frac12\mathbf b=\left\langle\frac12,0,\frac12\right\rangle, orth⁡𝒃𝒂=⟨32,1,−32⟩. \operatorname{orth}_{\mathbf b}\mathbf a =\left\langle\frac32,1,-\frac32\right\rangle.

(b) Replace the target by −3𝒃-3\mathbf b

The numerator becomes −3-3 and the squared magnitude becomes 1818. Therefore proj⁡−3𝒃𝒂=−318(−3𝒃)=12𝒃=proj⁡𝒃𝒂. \operatorname{proj}_{-3\mathbf b}\mathbf a =\frac{-3}{18}(-3\mathbf b)=\frac12\mathbf b =\operatorname{proj}_{\mathbf b}\mathbf a. Subtracting the same projection from 𝒂\mathbf a also gives orth⁡−3𝒃𝒂=orth⁡𝒃𝒂. \operatorname{orth}_{-3\mathbf b}\mathbf a =\operatorname{orth}_{\mathbf b}\mathbf a. Both vector answers are unchanged. Rescaling a nonzero target changes its length or orientation, but it describes the same line through the origin.

22. Prescribing the perpendicular magnitude

Given: 𝒂(t)=⟨t,2,1⟩\mathbf a(t)=\langle t,2,1\rangle and 𝒃=⟨1,0,1⟩\mathbf b=\langle1,0,1\rangle.

(a) Express the orthogonal component

𝒂(t)⋅𝒃=t+1\mathbf a(t)\cdot\mathbf b=t+1 and ‖𝒃‖2=2\lVert\mathbf b\rVert^2=2, so proj⁡𝒃𝒂(t)=⟨t+12,0,t+12⟩. \operatorname{proj}_{\mathbf b}\mathbf a(t) =\left\langle\frac{t+1}{2},0,\frac{t+1}{2}\right\rangle. Subtract from 𝒂(t)\mathbf a(t): orth⁡𝒃𝒂(t)=⟨t−12,2,1−t2⟩. \boxed{\operatorname{orth}_{\mathbf b}\mathbf a(t) =\left\langle\frac{t-1}{2},2,\frac{1-t}{2}\right\rangle}.

(b) Set its magnitude equal to 3

Its squared magnitude is (t−1)24+4+(1−t)24=(t−1)22+4. \frac{(t-1)^2}{4}+4+\frac{(1-t)^2}{4} =\frac{(t-1)^2}{2}+4. Because magnitudes are nonnegative, the required equation is equivalent to (t−1)22+4=9⇔(t−1)2=10. \frac{(t-1)^2}{2}+4=9 \quad\Longleftrightarrow\quad (t-1)^2=10. Hence t=1+10ort=1−10. \boxed{t=1+\sqrt{10}\quad\text{or}\quad t=1-\sqrt{10}}. For either value the squared magnitude is 10/2+4=910/2+4=9, so the magnitude really is 33. Both solutions are valid.

23. The nearest point on a line through the origin

Given: Q=(4,1,0)Q=(4,1,0), 𝒒=⟨4,1,0⟩\mathbf q=\langle4,1,0\rangle, and a line LL through the origin parallel to 𝒃=⟨1,2,−1⟩\mathbf b=\langle1,2,-1\rangle.

(a) Find the nearest point

𝒒⋅𝒃=4+2+0=6,‖𝒃‖2=1+4+1=6. \mathbf q\cdot\mathbf b=4+2+0=6, \qquad \lVert\mathbf b\rVert^2=1+4+1=6. Thus the foot has position vector 𝒉=proj⁡𝒃𝒒=66𝒃=⟨1,2,−1⟩. \mathbf h=\operatorname{proj}_{\mathbf b}\mathbf q =\frac66\mathbf b=\langle1,2,-1\rangle. The nearest point is H=(1,2,−1)\boxed{H=(1,2,-1)}, which lies on LL at parameter t=1t=1.

(b) Perpendicular component and distance

orth⁡𝒃𝒒=𝒒−𝒉=⟨3,−1,1⟩. \operatorname{orth}_{\mathbf b}\mathbf q =\mathbf q-\mathbf h=\langle3,-1,1\rangle. Therefore dist⁡(Q,L)=32+(−1)2+12=11. \boxed{\operatorname{dist}(Q,L)=\sqrt{3^2+(-1)^2+1^2}=\sqrt{11}}.

(c) Verify perpendicularity

(𝒒−𝒉)⋅𝒃=3−2−1=0. (\mathbf q-\mathbf h)\cdot\mathbf b=3-2-1=0. This also explains why the foot is nearest. For any point on LL with position vector t𝒃t\mathbf b, 𝒒−t𝒃=(𝒒−𝒉)+(1−t)𝒃. \mathbf q-t\mathbf b=(\mathbf q-\mathbf h)+(1-t)\mathbf b. The terms are perpendicular, so the squared distance is ‖𝒒−t𝒃‖2=11+6(1−t)2≥11. \lVert\mathbf q-t\mathbf b\rVert^2=11+6(1-t)^2\ge11. Equality occurs only at t=1t=1, confirming the unique nearest point.

24. Why the orthogonal decomposition is unique

Given: 𝒃≠𝟎\mathbf b\ne\mathbf0, 𝒂=λ𝒃+𝒘\mathbf a=\lambda\mathbf b+\mathbf w, and 𝒘⋅𝒃=0\mathbf w\cdot\mathbf b=0.

(a) Solve for the scalar

Take the dot product with 𝒃\mathbf b: 𝒂⋅𝒃=λ(𝒃⋅𝒃)+𝒘⋅𝒃=λ‖𝒃‖2. \mathbf a\cdot\mathbf b =\lambda(\mathbf b\cdot\mathbf b)+\mathbf w\cdot\mathbf b =\lambda\lVert\mathbf b\rVert^2. Since ‖𝒃‖2>0\lVert\mathbf b\rVert^2>0, division gives λ=𝒂⋅𝒃‖𝒃‖2. \boxed{\lambda=\frac{\mathbf a\cdot\mathbf b}{\lVert\mathbf b\rVert^2}}.

(b) Identify the components and establish uniqueness

Multiplying by 𝒃\mathbf b gives λ𝒃=𝒂⋅𝒃‖𝒃‖2𝒃=proj⁡𝒃𝒂. \lambda\mathbf b =\frac{\mathbf a\cdot\mathbf b}{\lVert\mathbf b\rVert^2}\mathbf b =\operatorname{proj}_{\mathbf b}\mathbf a. The remaining vector must therefore be 𝒘=𝒂−λ𝒃=𝒂−proj⁡𝒃𝒂=orth⁡𝒃𝒂. \mathbf w=\mathbf a-\lambda\mathbf b =\mathbf a-\operatorname{proj}_{\mathbf b}\mathbf a =\operatorname{orth}_{\mathbf b}\mathbf a. The dot-product equation fixes exactly one value of λ\lambda. Once that scalar is fixed, subtraction fixes exactly one remainder 𝒘\mathbf w. Therefore no different pair of parallel and perpendicular components can give the same vector under these conditions.