Math 21C Brian Osserman Practice Exam 3

1 (10 pts.) Find the curvature κ for times t > 0 for the space curve

r(t) = (cos t + t sin t)i + (sin t − t cos t)j + 3k.

Solution:

v(t) = (− sin t + sin t + t cos t)i + (cos t − cos t + t sin t)j = t cos ti + t sin tj,

so
p √
|v(t)| = t2 cos2 t + t2 sin2 t = t2 = t,

and the unit tangent vector is given by

v
T= = cos ti + sin tj.
|v|
So
d
T = − sin tt + cos tj,
dt
and
1 d 1 1
κ= | T| = | − sin tt + cos tj| = .
|v| dt t t

2 (10 pts.) Find the following limit, if it exists:

x+y−4
lim √
(x,y)→(2,2) x+y−2
where the domain is everything with x + y ≥ 0 and x + y 6= 4.
√
Solution: We multiply the top and bottom by x + y + 2 to get
√
x+y−4 (x + y − 4)( x + y + 2)
lim √ = lim √ √
(x,y)→(2,2) x+y−2 (x,y)→(2,2) ( x + y − 2)( x + y + 2)
√
(x + y − 4)( x + y + 2)
= lim
(x,y)→(2,2) x+y−4
√
= lim x+y+2
(x,y)→(2,2)
√
= 2 + 2 + 2 = 4.
3 (10 pts.) Use the formula for implicit dierentiation which we obtained from the
two-variable Chain Rule to nd dy
dx
if y is dened implicitly in terms of x by
xey + sin xy + y = 2.

Solution: The formula was dy

dx
= −Fx
Fy
, if F (x, y) = xey + sin xy + y − 2. We
get
dy −(ey + y cos xy)
= y .
dx xe + x cos xy + 1

4 (15 pts.) Use linear approximation to estimate how accurately the volume V = πr2 h
of a cylinder can be computed if the measurements for radius and height
are o by less than 1%.
Solution: The linear approximation of V at a point (r0 , h0 ) is

L(r, h) = πr02 h0 + (Vr )r0 ,h0 (r − r0 ) + (Vh )r0 ,h0 (h − h0 )

= πr02 h0 + 2πr0 h0 (r − r0 ) + πr02 (h − h0 ),

which gives us the error estimate (if our measurements are r0 , h0 , but the
actual values were r, h)

L(r, h) − V (r0 , h0 ) = 2πr0 h0 (r − r0 ) + πr02 (h − h0 ).

If |r − r0 | < 0.01r0 , |h − h0 | < 0.01h0 , we get

|L(r, h) − V (r0 , h0 )| < 0.02πr02 h0 + 0.01πr02 h0 = 0.03πr02 h0 ,

so the error is estimated to be less than 0.03πr02 h0 = 0.03V (r0 , h0 ), meaning

the error in volume is less than 3% if the error in radius and height are both
less than 1%.

5 (35 pts.) Consider the function

f (x, y) = x2 + xy + y 2 − 6x + 2,

on the rectangular region R with 0 ≤ x ≤ 5, −3 ≤ y ≤ 3.

2
(a) Find all local extreme values and saddle points of f in the interior of
R.

Solution: The function is everywhere dierentiable (also to second

order), so we can rst nd the critical points (which will be where
both partial derivatives vanish), and then use the second derivative
test. We have fx = 2x + y − 6, and fy = x + 2y , so the critical points
have 0 = x+2y, 0 = 2x+y −6. From the rst equality we get x = −2y ,
and from the second we get 0 = −3y − 6, so y = −2. Then x = 4. The
critical point (4, −2) is in the interior of R, so we have to determine if
it is a local maximum, a local minimum, or a saddle point.
fxx = 2, fyy = 2, fxy = 1, so the discriminant is 2 · 2 − 12 = 3 > 0, so
(4, −2) is a local extremum. Since fxx > 0, we see it is a minimum,
and the local minimum value is f (4, −2) = 16 − 8 + 4 − 24 + 2 = −10.
(b) Find the absolute extreme values of f on all of R.
Solution: We have already found the only possible extreme value
on the interior of R. Since this function is continuous on a closed
and bounded domain, it has an absolute maximum and an absolute
minimum, and we can nd them by computing values of f at all the
possible points in the interior and boundary and comparing the results.
We break the boundary up into 4 pieces.
The rst is x = 0, −3 ≤ y ≤ 3. On this segment, f = y 2 + 2, which
is a parabola centered at y = 0. The only critical point of this is at
y = 0, where f = 2, and at the endpoints y = −3 and y = 3, the value
of f is 11.
The next piece is x = 5, −3 ≤ y ≤ 3. On this segment, f = y 2 + 5y − 3,
which has a critical point 2y+5 = 0, or y = − 25 . At this point, f = −37
4
,
and at the endpoints y = −3 and y = 3, the value of f is −9 and 21
respectively.

3
 The third piece is y = −3, 0 ≤ x ≤ 5. On this segment, f = x2 − 9x +
11, which has a critical point at 2x − 9 = 0, or x = 29 . At this point,
f = −37
4
, and at the endpoints x = 0 and x = 5 the value of f is 11
and −9 respectively.
The nal piece is y = 3, 0 ≤ x ≤ 5. On this segment, f = x2 − 3x + 11,
which has a critical point at 2x − 3 = 0, or x = 23 . At this point,
f= 35
4
, and at the endpoints x = 0 and x = 5 the value of f is 11 and
21 respectively.

Comparing the values of f at all the points we have found, the maxi-
mum value is 21 at (5, 3), and the minimum value is −10 at (4, −2).

6 (20 pts.) Suppose that the temperature at the point (x, y, z) on the sphere x2 + y 2 +
z 2 = 1 is T = xyz 2 . Locate the highest and lowest temperatures on the
sphere.
Solution: First notice that if x, y, z are all positive, T is positive, and if
x is negative but y, z are positive, then T is negative. So the maximum
temperature must be positive and the minimum temperature is negative. If
any of x, y, z are equal to 0, the temperature is 0, so can't be at a maximum
or minimum point. Thus, we can restrict our attention to points where none
of x, y, z is equal to 0.
Let g(x, y, z) = x2 + y 2 + z 2 − 1. According to the Langrange multiplier
method, we want to nd x, y, z, λ with g(x, y, z) = 0 and ∇T = λ∇g .

∇T = yz 2 i + xz 2 j + 2xyzj,

and
∇g = 2xi + 2yj + 2zk,

4
so our equations are

x2 + y 2 + z 2 = 1

yz 2 = 2λx

xz 2 = 2λy

2xyz = 2λz.

From the last equation we conclude λ = xy (since z 6= 0). The second

equation then gives yz 2 = 2x2 y , so (since y 6= 0) we get z 2 = 2x2 . The third
equation gives xz 2 = 2xy 2 , so z 2 = 2y 2 . We conclude that x2 = y 2 = z2
2
,
and substituting into the rst equation gives z2
2
+ z2
2
+ z 2 = 1, so 2z 2 = 1,
and z √±12 . Then x = ±y = ±1
2
. The maximum temperature occurs at the
four points ( 12 , 12 , √±12 ) and ( −1
2
, −1
2
±1
,√ 2
), and is equal to 1 1 √1
2 2 ( 2)2
= 18 degrees,
while the minimum temperature occurs at the four points ( −1 ±1
, 1, √
2 2 2
) and
( 12 , −1
2
±1
,√ 2
), and is equal to −1 1 √1
2 2 ( 2)2
= − 18 degrees.