1092 Calculus B 01-03 Final Exam Solution
June 24, 2021 1. Consider the differential equation ydy
dt =
√ 1 − y2.
(a) (6 pts) If the constant function y(t) = a is a solution of the equation, find the value of a.
(b) (10 pts) Given that y(0) = 1
2, solve y(t) for t near 0.
考慮微分方程式 ydy dt =
√
1 − y2。
(a) (6 pts) 如果常數函數 y(t) = a 是方程式的解,求a 的值。
(b) (10 pts) 給定 y(0) = 1
2,求 y(t), 當t在0附近。
Solution:
(a) Let
√
1 − y2=0. Then y = ±1. Hence
a = ±1 (6%).
(b) Now suppose y ≠ ±1 (1%). Rewrite the equation as y dy
√ 1 − y2
=dt (1%) and integrate on both sides with respect to t, we get
−
√
1 − y2=x + c. (3%).
Since y(0) = 1/2, we get that
c = −
√
1 − (1/2)2 = −
√ 3
2 (2%).
Hence
y = ±
√
1 − (x −
√
3/2)2 (1%)
=
√
1 − (x −
√
3/2)2 (since y(0) =1
2, we should choose +) (2%).
2. (16 pts) Solve
⎧⎪
⎪⎪
⎨
⎪⎪
⎪⎩
y′(t) = t(t2−y(t)), y(0) = 3.
(16 pts) 解
⎧⎪
⎪⎪
⎨
⎪⎪
⎪⎩
y′(t) = t(t2−y(t)), y(0) = 3.
Solution:
y′+t y =t3 rearrange the eqn I(t) =e∫ t dt=e12t2 find IF(5%)
e12t2y′+e12t2t y =e12t2t3 multiply the eqn by I(t) = e12t2 (2%) d (e12t2y)
dt =e12t2t3 rewrite the left-hand side
∫
d (e12t2y)
dt dt =∫ e12t2t3dt integrate both sides (2%) e12t2y =e12t2(t2−2) + c evaluate the integrals (3%)
3 = − 2 + c use y(0) = 3
5 =c find c(2%)
y(t) =t2−2 + 5 e−t22 the solution(2%)
Page 2 of 7
3. (8 pts) Suppose that X and Y are independent and E(X) = 1, E(Y ) = 2, Var(X) = 5, Var(Y ) = 7.
Compute E((2X − Y )2).
(8 pts) 已知 X,Y 為兩個獨立的隨機變數,而且 E(X) = 1,E(Y ) = 2, Var(X) = 5, Var(Y ) = 7。
求 E((2X − Y )2)。
Solution:
E((2X`Y )2)
=E(4 X2−4 X Y + Y2) expand (2%)
=4 E(X2) −4 E( X Y ) + E(Y2) E(a X + b Y ) = a E(X) + b E(Y ) (1%)
=4 (Var(X) + (E(X))2) −4 E(X) ⋅ E(Y ) Var(X) = E(X2) − (E(X))2 (2%)
+ (Var(Y ) + (E(Y ))2) X, Y idep. ⇒E(X Y ) = E(X) ⋅ E(Y ) (1%)
=4 (5 + 12) −4 ⋅ 1 ⋅ 2 + (7 + 22) input the values (2%)
=27
4. Suppose that X is a random variable with probability density function fX(t) = ce−t2+8t20 for some constant c > 0.
(a) (8 pts) Find the constant c such that ∫
∞
−∞
fX(t)dt = 1. (Hint: ∫
∞
−∞
e−t2dt =√ π) (b) (8 pts) Find E(X) and Var(X).
(c) (6 pts) Suppose that X1, ⋯, Xn are independent random variables and Xi ∼X for 1 ≤ i ≤ n.
Use Chebyshev’s inequality to estimate the size of n such that we can derive P ( ∣X1+ ⋯ +Xn
n −E(X)∣ < 0.05 ) > 0.9.
已知隨機變數 X 的機率密度函數為 fX(t) = ce−t2+8t20 ,其中 c > 0 是一個常數。
(a) (8 pts) 求常數 c 使得 ∫
∞
−∞
fX(t)dt = 1. (提示: ∫
∞
−∞
e−t2dt =√ π) (b) (8 pts) 求 E(X) 和 Var(X).
(c) (6 pts) 假設 X1, ⋯, Xn 為互相獨立的隨機變數而且 Xi ∼X, 1 ≤ i ≤ n。 利用 Chebyshev 不等 式,估計 n 需要多大,我們才能保證
P ( ∣X1+ ⋯ +Xn
n −E(X)∣ < 0.05 ) > 0.9.
Solution:
(a) We have
1 = c∫
∞
−∞
e−(t2−8t)/20dt
=c∫
∞
−∞
e−(t−4)2−1620 dt (3%)
=ce45∫
∞
−∞
e−(t−4)220 dt (1%)
=ce45 ⋅
√π ⋅
√
20 (3%).
Hence
c = e−4/5⋅ 1 2√
5π (1%).
(b) Follows form (a), we have that
X ∼ N(4, 10) (2%).
Hence
E(X) = 4 (3%), Var(X) = 10 (3%).
(c) Let
Z ∶= X1+ ⋯ +Xn
n .
Since Xi∼X for all i and X1, ⋯, Xn are independent, we have E(Z) = E(X) = 4 (1%) Var(Z) = Var(X)
n =
10
n (1%).
Page 4 of 7
By the Chebyshev inequality, we want to find n so that P (∣Z − E(X)∣ < 0.05) ≥ 1 − Var(Z)
(0.05)2 (2%)
=1 − 10/n
(0.05)2 (1%)
>0.9, this is equivalent to
0.1 > 10/n
1/400 ⇐⇒n > 40000 (1%).
5. Let f (t) =
⎧⎪
⎪
⎨
⎪⎪
⎩
c ⋅ te−t, t > 0 0, t ≤ 0.
(a) (8 pts) Find constant c such that f (t) is a probability density function.
(b) (4 pts) Suppose that X and Y are independent with probability density functions fX =fY =f . Let Z = X + Y . Find the plane region D in the xy-plane such that the distribution function of Z, FZ(t) = P (X + Y ≤ t), is ∬
D
fX(x)fY(y)dxdy.
(c) (10 pts) Find the probability density function of Z in part (b).
令 f(t) =
⎧⎪
⎪
⎨
⎪⎪
⎩
c ⋅ te−t, t > 0 0, t ≤ 0.
(a) (8 pts) 求常數 c 使得 f (t) 是一個機率密度函數。
(b) (4 pts) 已知隨機變數 X 和 Y 是獨立的而且它們的機率密度函數為 fX =fY =f。 令 Z = X+Y 。求 xy 平面的區域D,使得 Z的分配函數, FZ(t) = P (X +Y ≤ t), 是重積分∬
D
fX(x)fY(y)dxdy。 (c) (10 pts) 求(b)小題中 Z的機率密度函數。
Solution:
(a)
∫
∞
−∞
f (t)dt = c∫
∞ 0
te−tdt = c (−te−t∣
∞ 0 + ∫
∞ 0
e−tdt) = c(0 − e−t∣
∞ 0
) =c.
(3 pts for applying integration by parts. 3 pts for correct result. Students do not need to use the definition of improper integrals to compute this integration.)
Let c = 1. Then ∫
∞
−∞
f (t)dt = 1 and f (t) is a probability density function.
(2 pts for c = 1.)
(b) FZ(t) = P (X + Y ≤ t) =∬
DfX(x)fY(y)dxdy, where D = {(x, y)∣x + y ≤ t, x ≥ 0, y ≥ 0}.
If t ≤ 0, then D is an empty set or a point {(0, 0)}.
If t > 0, then D is the triangle bounded by x = 0, y = 0, and x + y = t.
(4 pts for D = {(x, y)∣x + y ≤ t, x ≥ 0, y ≥ 0} or D = {(x, y)∣x + y ≤ t}. It is o.k. if students don’t discuss cases t > 0 and t ≤ 0. )
(c) For t ≤ 0, FZ(t) = 0 For t > 0,
Solution 1:
FZ(t) =∬
D
xe−xye−ydxdy
⎧⎪
⎪⎪
⎪
⎨
⎪⎪
⎪⎪
⎩
u = x + y v = y
ÔÔÔÔÔÔÔÔ ∬
Ω
(u − v)ev−u⋅v ⋅ e−v∣
∂(x, y)
∂(u, v)∣dudv
= ∫
t 0 ∫
u 0
(u − v)v ⋅ e−udvdu,
where Ω is the region in the uv-plane bounded by u = t, u = v, and v = 0, and ∣∂(x, y)
∂(u, v)∣ =1.
(1 pt for substitution u = x + y, v = y, or u = x + y, v = x. 1 pt for ∣∂(x, y)
∂(u, v)∣ =1. 1 pt for Ω.
2 pts for the iterated integral.)
Page 6 of 7
Hence fZ(t) = FZ′(t) =∫
t
0 e−t⋅ (t − v)vdv = 1 6t3e−t. (2 pts for fZ(t) = FZ′(t) =∫
t 0
e−t⋅ (t − v)vdv. 3 pts for∫
t 0
e−t⋅ (t − v)vdv = 1 6t3e−t.) Solution 2:
fZ(t) =∫
∞
−∞
fX(t − s)fY(s) ds or ∫
∞
−∞
fX(s)fY(t − s) ds. (3 pts) Hence fZ(t) =∫
∞
−∞
fX(t − s)fY(s) ds =∫
t 0
(t − s)es−tse−sds (2 pts for the upper and lower limits of the integration 0, t.)
Thus fZ(t) =∫
t
0 s(t − s)e−tds = 1 6t3e−t. (2 pt for fZ(t) =∫
t 0
s(t − s)e−tds. 3 pts for ∫
t 0
s(t − s)e−tds = 1 6t3e−t.)
6. The number of flaws in a carpet appears to be Poisson distributed at a rate of one every 6m2. (a) (8 pts) Find the probability that a 3m by 4m carpet contains more than 2 flaws.
(b) (8 pts) There are two carpets with sizes 2m by 4m and 2m by 5m. Find the probability that these two carpets together contain less than 2 flaws.
假設地毯上的瑕疵個數呈 Poisson 分配,而且平均每 6平方公尺有一個瑕疵。
(a) (8 pts) 求一條 3公尺寬 4公尺長的地毯有超過 2個瑕疵的機率。
(b) (8 pts) 有兩條地毯,各是 2公尺寬 4公尺長和 2公尺寬 5公尺長。求這兩條地毯上總共的瑕疵數少
於 2的機率。
Solution:
(a) The average number of flaws in a 3m×4m carpet is 1 × (3 × 4
6 ) =2. (2 pts)
Let F be the number of flaws in the carpet. Then F is Poisson distributed with E(F ) = 2.
Hence P (F = 0) = e−2, P (F = 1) = 21
1!e−2 =2e−2, P (F = 2) = 22
2!e−2=2e−2.
(1 pt for P (F = 0) = e−2. 1 pt for P (F = 1) = 2e−2. 1 pt for P (F = 2) = 2e−2.) Thus P (F > 2) = 1 − P (F ≤ 2) = 1 − P (F = 0) − P (F = 1) − P (F = 2) = 1 − 5 ⋅ e−2.
(2 pts for P (F > 2) = 1 − P (F ≤ 2) = 1 − P (F = 0) − P (F = 1) − P (F = 2). 1 pt for the final answer.)
(b) Solution 1:
Together the average number of flaws in the two carpets is 1 ×(2 × 4 + 2 × 5)
6 =3. (4 pts) Let F be the number of flaws in the two carpets. Then P (F < 2) = P (F = 0) + P (F = 1) = e−3+3e−3 =4e−3.
(1 pt for P (F < 2) = P (F = 0)+P (F = 1). 1 pt for P (F = 0) = e−3. 1 pt for P (F = 1) = 3e−3. 1 pt for the final answer.)
Solution 2: Let F1 be the number of flaws in the 2 × 4 carpet. P (F1 = 0) = e−86 and P (F1 =1) = 4
3e−43. (2 pts)
Let F2 be the number of flaws in the 2 × 5 carpet. P (F2 =0) = e−106 , P (F2=1) = 5
3e−53. (2 pts)
P (F1+F2 ≤1) = P (F1=0, F2=0) + P (F1=1, F2=0) + P (F1 =0, F2 =1) (2 pts)
=e−43e−53 + 4
3e−43e−53 +e−435
3e−53 =4e−3. (2 pts)
Page 8 of 7