You are on page 1of 2

Practice Problems 7: Solutions

1. Consider using Lagranges Method to solve the following optimization problem:


minimize: 3x1 4x2 + x3 2x4
subject to: x22 + x23 + x24 = 1
3x21 + x23 + 2x24 = 6


F (x, ) = 3x1 4x2 + x3 2x4 + 1 x22 + x23 + x24 1 + 2 3x21 + x23 + 2x24 6
T
3 + 62 x1

4 21 x2

1 + 21 x3 + 22 x3

DF (x, ) =
21 x4 + 42 x4
2 +

x22 + x23 + x24 1


3x21 + x23 + 2x24 6

Finding a critical point, solving DF (x, ) = 0, is difficult because DF (x, ) is not a linear system.
2. Compute the gradient of the Lagrangian for the maximum expected return portfolio subject to
risk = (20%)2
maximize: T w
subject to: eT w = 1
wT w = (0.20)2


F (w, ) = T w + 1 eT w 1 + 2 wT w 0.04
T
T
+ 1 eT + 22 wT

eT w 1
DF (w, ) =
T
w w 0.04
Note: Unlike the lecture example of finding the minimum variance portfolio, DF (w, ) = 0 is not a
linear system in this case, so the solution must be computed numerically.
3. Second order Taylor polynomial around S0 for C(S) in terms of and :
2

C (S0 ) + (S S0 )

(S S0 ) 2 C
(S S0 )
C
(S0 ) = C (S0 ) + (S S0 ) (S0 ) +
(S0 ) +
(S0 )
S
2
S 2
2

1
1+x
1
f 0 (x) =
(1 + x)2

4. f (x) =

f 00 (x) =

2
(1 + x)3

f 000 (x) =

6
(1 + x)4

...

f (n) (x) =

(1)n n!
(1 + x)n+1

Taylor series expansion:


T (x) =

(x 0)2
(x 0)3 6
(x 0)n (1)n n!
1
1
2
+
+
+

+
+ (x 0)
1+0
(1 + 0)2
2
(1 + 0)3
6
(1 + 0)4
n!
(1 + x)n+1

= 1 x + x2 x3 + + (1)n xn
Radius of convergence:
1
1
R=
= =1
1/k
1
lim |ak |

where lim |ak | /k = lim |(1)k | /k = lim 1 /k = 1


k

Where does T (x) =

1
?
1+x



(1)n n!
rn
rn
rn
(n)

= lim
lim
max |f (z)| = lim
max
= 0 for r <

n (1 r)n+1
n n! z[r,r]
n n! z[r,r] (1 + z)n+1
1
T (x) =
for |x| < 12
1+x
Pt Pt1
Rt Pt1 = Pt Pt1 Rt Pt1 + Pt1 = Pt
Pt1




Pt
(Rt + 1) Pt1
rt = log
= log
= log (Rt + 1)
Pt1
Pt1

5. Rt =

1
2

Pt = (Rt + 1) Pt1

Note: Rt is actually a first order Taylor polynomial of rt around the point a = 0.


Let Rt = 0.01, then rt = log(1.01) 0.00995
6. Find the order of the Taylor polynomial needed to compute e0.25 to six digits of accuracy.
Z
f (x) Pn (x) =
a

(x t)n (n+1)
f
(t)dt
n!

Let a = 0 and x = 0.25


Z
Find n such that f (0.25) Pn (0.25) =
0

n
1
2
3
4
5

0.25

(0.25 t)n t
e dt < 106
n!

Error
3.40 102
2.78 103
1.73 104
8.49 106
3.52 107

We need a 5th order Taylor polynomial to compute e0.25 to six digits of accuracy.

You might also like