• 沒有找到結果。

Calculus Midterm Exam

N/A
N/A
Protected

Academic year: 2022

Share "Calculus Midterm Exam"

Copied!
4
0
0

加載中.... (立即查看全文)

全文

(1)

Calculus Midterm Exam

April 24, 2006 You must show all your work to obtain any credits.

1. (a) (6 points) Expand f (x) = ln(1 + x) in powers of x.

Solution: Since f (0) = 0, and f(k)(x) = (−1)k−1(k − 1)!

(x + 1)k , f(k)(0) = (−1)k−1(k − 1)! for each k ≥ 1. f (x) =

k=1

(−1)k−1 k xk.

(b) (6 points) Find a powers series representation for the improper integral Z x

0

ln(1 + t)

t dt.

Solution: Using part (a), we haveln(1 + t) t =

k=1

(−1)k−1

k tk−1. Therefore, Z x

0

ln(1 + t) t dt = Z x

0

k=1

(−1)k−1

k tk−1dt =

k=1

(−1)k−1 k2 xk.

2. Let f (x) =ex− 1 x .

(a) (5 points) Expand f (x) in powers of x.

Solution: Since dkex

dxk |x=0 = ex|x=0 = 1 for each k ≥ 0, ex =

k=0

xk

k!, and f (x) = ex− 1

x =

k=1xk−1 k! .

(b) (5 points) Show that

n=1

n

(n + 1)! = 1.

Solution: Using part (a), we get f0(x) =

k=1

(k − 1)xk−2 k! =

n=0

nxn−1

(n + 1)!=

n=1

nxn−1

(n + 1)!, where we have set n = k − 1 in the second last equality. Now f0(x) = ex

x −ex− 1

x2 , we get 1 = f0(1) =

n=1

n (n + 1)!.

3. (10 points) Find an equation for the tangent plane and scalar parametric equations for the normal line to the surface z3+ xyz − 2 = 0 at (1, 1, 1).

Solution: Let f (x, y, z) = z3+ xyz − 2 = 0. Thus,f = (yz, xz, xy + 3z2), and, at (1, 1, 1),f = (1, 1, 4). Therefore, (x − 1) + (y − 1) + 4(z − 1) = 0 is an equation for the tangent plane, while (x, y, z) = (t + 1,t + 1, 4t + 1) are parametric equations for the normal line to the surface at (1, 1, 1).

(2)

Calculus Midterm Exam

April 24, 2006

4. (a) (6 points) Find the unit tangent vector and the principal unit normal vector for the curve r(t) = 2ti + lntj − t2k.

Solution: Since r0(t) =¡ 2,1

t, −2t¢

, kr0(t)k = 2t +1

t = 2t2+ 1

t , the unit tangent vector is T(t) = r0(t)

kr0(t)k=

³ 2t

2t2+ 1, 1

2t2+ 1, − 2t2 2t2+ 1

´

=

³ 2t

2t2+ 1, 1

2t2+ 1, 1 2t2+ 1−1

´

. Furthermore,

T0(t) =

³ 2

2t2+ 1 8t2

(2t2+ 1)2, − 4t

(2t2+ 1)2, − 4t (2t2+ 1)2

´

= 2

(2t2+ 1)2

³

1 − 2t2, −2t, −2t

´ ,

and kT0(t)k = 2

(2t2+ 1), the principal unit normal vector is N(t) = T0(t) kT0(t)k

= 1

(2t2+ 1)

³

1 − 2t2, −2t, −2t

´ .

(b) (6 points) Find the length of the plane curve y = x3/2for 0 ≤ x ≤ 4.

Solution: The length is Z 4

0

q

1 + (dy/dx)2dx = Z 4

0

r 1 +9x

4 dx = 8

27(1 +9x 4 )3/2|40

= 8 27

h

103/2− 1 i

.

5. (a) (6 points) Find the directional derivative of f (x, y) = e2x(cos y − sin y) at (12, −π2) in the direction of the greatest increase of f .

Solution: Since the directional derivative Duf (p) of f in the direction u at a point p satisfies that Duf (p) = k∇f (p)k cosθ, whereθ ∈ [0,π] is the angle between u andf (p). Therefore, Duf (p) reaches maximum value k∇f (p)k whenθ = 0, or u =kf (p)f (p)k.

Since∇f =¡

2e2x(cos y − sin y), −e2x(sin y − cos y)¢

,f (12, −π2) =¡ 2e, e¢

, and the directional derivative of f at (12, −π2) in the direction 15(2, e) increases fastest with the rate equals to kf (12, −π2)k =√

5e.

(b) (6 points) Find the directional derivative of g(x, y, z) = sin(xyz) at (12,13,π) in the direction of the greatest decrease of g.

Solution: From part (a), we know that the directional derivative Dug(p) = kg(p)k cosθ, reaches minimum value −kg(p)k whenθ =π, or u = k∇g(p)g(p)k.

Since ∇g =¡

yz cos(xyz), xz cos(xyz), xy cos(xyz)¢

, ∇g(12,13,π) =

3 2

¡π

3,π2,16¢

, and the direc- tional derivative of g at (12,13,π) decrease fastest in the direction −1

13π2+1

¡2π, 3π, 1¢with the

rate −kg(12,13,π)k = −

39π2+ 3

12 .

6. (10 points) Find the curvature of the plane curve x(t) = 2e−t, y(t) = e−2t.

Page 2

(3)

Calculus Midterm Exam

April 24, 2006

Solution: Let r(t) =¡

x(t), y(t)¢

2e−t, e−2t¢

. Then r0(t) = −2e−2t¡ et, 1¢

, and ds

dt = kr0(t)k = 2e−2t

p

e2t+ 1. Thus,

the unit tangent vector T (t) = r0(t)

kr0(t)k= − 1

√e2t+ 1

¡et, 1¢

et

√e2t+ 1, − 1

√e2t+ 1

¢.

T0(t) =¡

et

e2t+ 1+ e3t

(e2t+ 1)3/2, e2t (e2t+ 1)3/2

¢= et (e2t+ 1)3/2

¡− 1, et¢ ,

the principal unit normal vector N(t) = T0(t) kT0(t)k

1

√e2t+ 1, et

√e2t+ 1

¢,

and dN

dt = et (e2t+ 1)3/2

¡et, 1¢

. The curvatureκ= kdN

dsk = kdN dt k

.ds dt

= et

(e2t+ 1)3/2 p

e2t+ 1 1 2e−2t

e2t+ 1 = e3t 2(e2t+ 1)3/2.

7. (a) (5 points) A set of points is said to be convex provided that every pair of points in the set can be joined by a line segment that lies entirely within the set. Show that, if k∇f (x)k ≤ M for all x in some convex setΩ, then | f (x1) − f (x2)| ≤ Mkx1− x2k for all x1and x2inΩ.

Solution: For any x1, x2Ω, the line segment r(t) = (1−t)x1+tx2lies entirely inΩand joins x1= r(0) to x2= r(1). Thus, the function g(t) = f (r(t)) is differentiable for t ∈ [0, 1]. Mean Value Theorem implies that there exists a t0∈ (0, 1) such that | f (x2)− f (x1)| = |g(1)−g(0)| =

|g0(t0)(1 − 0)| = |∇f (r(t0)) · r0(t0)| = |∇f (c) · (x2− x1)| ≤ k∇f (c)k · kx2− x1k ≤ Mkx2− x1k, where c = r(t0).

(b) (5 points) Show that if f is differentiable at each point of the line segment ab and f (a) = f (b), then there exists a point c between a and b for whichf (c) ⊥ (b − a).

Solution: Let r(t) = (1 − t)a + tb for t ∈ [0, 1], and set g(t) = f (r(t)). Mean Value Theorem implies that there exists t0∈ (0, 1) such that 0 = f (b) − f (a) = g(1) − g(0) = g0(t0)(1 − 0) =

f (r(t0)) · r0(t0) =∇f (c) · (b − a), where c = r(t0) ∈ ab. Thus, we havef (c) ⊥ (b − a).

8. Let t, n, b = t × n denote the unit tangent, principal normal, and binormal at each point of a regular curve.

(a) (4 points) Show that db/ds ⊥ b.

Solution: Since hb, bi(s) = 1 for each s, we have 0 = d

ds1 = d

dshb, bi = 2hdb

ds, bi. Thus, db

ds ⊥ b.

(b) (4 points) Use the fact thatdb ds = d

ds(t × n) to show that db/ds ⊥ t.

Page 3

(4)

Calculus Midterm Exam

April 24, 2006

Solution: Since db ds = d

ds(t × n) = dt

ds× n + t ×dn

dsn × n + t ×dn

ds = t ×dn

ds, we have hdb

ds, ti = ht ×dn

ds, ti = 0 which implies that db ds ⊥ t.

(c) (4 points) Show that db

dsn, for some scalarτ.

Solution: Since t, n, b form an orthonormal basis of R3 at each r(s) in the curve, and db ds is perpendicular to both b and t from parts (a) and (b), there exists a scalarτsuch that db

dsn.

9. Let f (x, y) = x2+ y2− 2x + 2y + 2 and D = {(x, y) | x2+ y2≤ 8}.

(a) (6 points) Find local extreme values of f in the interior of D, i.e. inside {(x, y) | x2+ y2< 8};

determine the local maxima and minima and the saddle points.

Solution:f = (2x − 2, 2y + 2) = (0, 0) implies that (x, y) = (1, −1). At (1, −1), the Hes- sian matrix Hess( f ) of f ,

µfxx fxy fyx fyy

=

µ2 0 0 2

is positive definite, f (1, −1) = 0 is a local minimum.

(b) (6 points) Use the method of Lagrange to determine the extreme values of f on the boundary of D, i.e. on the circle {(x, y) | x2+ y2= 8}.

Solution: Let g(x, y) = x2+ y2− 8 = 0 be the constraint condition. Solve λ, x, y such that

f (x, y) =λ∇g(x, y), which is equivalent to the system (λ − 1)x = −1 and (λ− 1)y = 1.

Sinceλ−1 6= 0, we get (x, y) =¡

1

λ− 1, 1 λ− 1

¢. Substituting these into the constraint g = 0, we get (x, y) = (−2, 2), or (x, y) = (2, −2), and f (−2, 2) = 18 is the maximum value, while

f (2, −2) = 2 is the minimum value of f on the circle {(x, y) | x2+ y2= 8}.

Page 4

參考文獻

相關文件

Since everyone needs to write the final solutions alone, there is absolutely no need to lend your homework solutions and/or source codes to your classmates at any time.. In order

First, write a program to implement the (linear) ridge regression algorithm for classification (i.e. use 0/1 error for evaluation)?. Use the first 400 examples for training to get g

(a) Consider a binary classification algorithm A majority that returns a constant classifier that always predicts the majority class (i.e., the class with more instances in the data

(A 10% bonus can be given if your proof for either case is rigorous and works for general polynomial regression.).. If gradient boosting is coupled with linear regression

Since everyone needs to write the final solutions alone, there is absolutely no need to lend your homework solutions and/or source codes to your classmates at any time.. In order

Since everyone needs to write the final solutions alone, there is absolutely no need to lend your homework solutions and/or source codes to your classmates at any time.. In order

(*) Implement the fixed learning rate gradient descent algorithm below for logistic regression, ini- tialized with 0?. Run the algorithm with η = 0.001 and T = 2000 on the following

The set of all adherent points of A is denoted by A called the closure of A.. When A is empty, the statement