AMATH 460: Mathematical Methods

for Quantitative Finance

7.1 Lagranges Method
Kjell Konis
Acting Assistant Professor, Applied Mathematics
University of Washington
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

1 / 29

Outline

Optimal Investment Portfolios

Relative Extrema of Functions of Several Variables

Lagranges Method

Example

Minimum Variance Portfolio

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

2 / 29

Outline

Optimal Investment Portfolios

Relative Extrema of Functions of Several Variables

Lagranges Method

Example

Minimum Variance Portfolio

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

3 / 29

Investment Portfolios
Portfolio of n assets
Let wi be the proportion of the portfolio invested in asset i
Have constraint
n
X

wi = 1

i=1

Can take long and short positions = no constraints on individual wi

Let i be the expected rate of return on asset i
Let i2 be the risk of asset i
Let ij be the correlation between assets i and j
Expected rate of return and risk of the portfolio:
Expected Return =

n
X

w i i

i=1

Risk =

n
X

wi2 i2 + 2

i=1
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

wi wj i j ij

1i<jn
4 / 29

Investment Portfolios: Matrix Notation

Let w = (w1 , . . . , wn ) and = (1 , . . . , n )
The expected rate of return can be written in matrix notation as
Return =

n
X

wi i = w T

i=1

The risk can be written as

Risk = w T w
is the covariance matrix of the n assets

12

2 1 21

=
..

1 2 12 1 n 1n
22
..
.

..
.

n 1 n1 n 2 n2

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

2 n
..
.
n2

5 / 29

Optimal Investment Portfolios

Given , , and investor selected w , can compute
portfolio return
portfolio risk

Two notions of optimality

For a target expected return, choose w to minimize portfolio risk
For a target level of risk, choose w to maximize expected return

Both notions are constrained optimization problems that can be

solved using Lagrange multipliers

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

6 / 29

Optimal Investment Portfolios

Minimum variance optimization
n asset case
minimize:
subject to:

w T w
eTw = 1
T w = P

2 asset case
minimize:
subject to:

12 w12

+ 21 2 w1 w2 + 22 w22
w1 + w2 = 1
1 w1 + 2 w2 = P

Maximum expected return optimization

n asset case
maximize:
subject to:

2 asset case

maximize:
T w
subject to:
eTw = 1
w T w = P2

Kjell Konis (Copyright 2013)

1 w1 + 2 w2
w1 + w2 = 1
12 w12 + 21 2 w1 w2 + 22 w22 = P2

7.1 Lagranges Method

7 / 29

Outline

Optimal Investment Portfolios

Relative Extrema of Functions of Several Variables

Lagranges Method

Example

Minimum Variance Portfolio

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

8 / 29

Relative Extrema of Single Variable Functions

A local minimum (maximum) of a function f is a point x0 where
f (x0 ) () f (x )

x (x0 , x0 + )

for some  > 0

A local extrema is a point that is a local minimum or maximum
If f is twice differentiable and f 00 is continuous
Any local extremum is a critical point of f : f 0 (x0 ) = 0
Can classify critical points using second derivative test
f 0 (x0 ) < 0
f 0 (x0 ) > 0
f 0 (x0 ) = 0
Kjell Konis (Copyright 2013)

local maximum
local minimum
anything possible
7.1 Lagranges Method

9 / 29

Relative Extrema of Functions of n Variables

A local minimum (maximum) of a function f : Rn R is a point
x0 Rn where
f (x0 ) () f (x )

x : kx x0 k < 

Every local extremum is a critical point: Df (x0 ) = 0

If f is twice differentiable and has continuous second order partial
derivatives
D 2 f (x0 ) is a symmetric matrix with real eigenvalues
Second order conditions
All eigenvalues of D 2 f (x0 ) > 0
All eigenvalues of D 2 f (x0 ) < 0
D 2 f (x0 ) has eigenvalues
D 2 f (x0 ) singular
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

local minimum
local maximum
saddle point
anything can happen
10 / 29

Finding Extrema: Functions of 2 Variables

Find the local extrema of f (x , y ) = x 2 + xy + y 2
Df (x , y )

2x + y x + 2y

Df (0, 0) = 0 0

(0, 0) is a critical point

D f (x , y ) =

"

2 1
1 2

Can use R to compute the eigenvalues

> A <- matrix(c(2, 1, 1, 2), 2, 2)
> eigen(A)$values
[1] 3 1
Since both eigenvalues are greater than 0 = (0, 0) a local minimum
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

11 / 29

Finding Extrema: Functions of 2 Variables (Take 2)

Find the local extrema of f (x , y ) = x 2 xy y 2
Df (x , y )

2x y x 2y

Df (0, 0) = 0 0

(0, 0) is a critical point

D f (x , y ) =

"

2 1
1 2

Can use R to compute the eigenvalues

> A <- matrix(-c(2, 1, 1, 2), 2, 2)
> eigen(A)$values
[1] -1 -3
Since both eigenvalues are less than 0 = (0, 0) a local maximum
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

12 / 29

Finding Extrema: Functions of 2 Variables

Find the local extrema of f (x , y ) = x 2 + 3xy + y 2
Df (x , y )

2x + 3y 3x + 2y

Df (0, 0) = 0 0

(0, 0) is a critical point

D f (x , y ) =

"

2 3
3 2

Can use R to compute the eigenvalues

> A <- matrix(c(2, 3, 3, 2), 2, 2)
> eigen(A)$values
[1] 5 -1
One positive and one negative eigenvalue = (0, 0) a saddle point
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

13 / 29

Finding Extrema: Functions of 2 Variables

3

Find the local extrema of f (x , y ) = 2xy (1 y 2 ) 2

First order condition


Df (x , y ) =
Df (0, 0) =

2y


2x + 3y


0 0

y2

= (0, 0) is a critical point

Second order condition

0
D 2 f (x , y ) =
2

2
36y

1y 2

D 2 f (0, 0) =

0 2
2 3

Compute the eigenvalues of the Hessian at the critical point

> eigen(matrix(c(0, 3, 3, 2), 2, 2))$values
[1] 4.162278 -2.162278
One positive and one negative eigenvalue = (0, 0) a saddle point
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

14 / 29

Outline

Optimal Investment Portfolios

Relative Extrema of Functions of Several Variables

Lagranges Method

Example

Minimum Variance Portfolio

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

15 / 29

Lagranges Method
Problem:
maximize:
subject to:

f (x1 , x2 , . . . , xn )
g1 (x1 , x2 , . . . , xn ) = 0
g2 (x1 , x2 , . . . , xn ) = 0
..
.

(1)

gm (x1 , x2 , . . . , xn ) = 0
18th -century mathematician Joseph Louis Lagrange proposed the
following method for the solution
Form the function
F (x1 , . . . , xn , 1 , . . . , m ) = f (x1 , . . . , xn ) +

m
X

i gi (x1 , x2 , . . . , xn )

i=1

Optimal value for problem (1) occurs at one of the critical points of F
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

16 / 29

Lagranges Method
Terminology:
The function F (x1 , . . . , xn , 1 , . . . , m ) is called the Lagrangian
The column vector = (1 , . . . , m ) is called the Lagrange
multipliers vector
Necessary Condition:
Let x = (x1 , x2 , . . . , xn )

Let g(x ) = g1 (x ), g2 (x ), . . . , gm (x ) be a vector-valued function

of the constraints

The gradient D g(x ) must have full rank at any point where
the constraint g(x ) = 0 is satisfied, that is

rank Dg(x ) = m
Kjell Konis (Copyright 2013)

x where g(x ) = 0

7.1 Lagranges Method

17 / 29

Partial Derivatives of the Lagrangian

D F (x , ) has n + m variables, compute gradient in 2 parts


D F (x , ) = Dx F (x , ) D F (x , )
Recall Lagrangian:
F (x , ) = f (x1 , . . . , xn ) +

m
X

i gi (x1 , x2 , . . . , xn )

i=1

The partial derivatives are

m
X
F
f
gi
=
+
i
xj
xj i=1 xj

f
f
Gradient of f : Df (x ) =
...
x1
xn


Kjell Konis (Copyright 2013)

7.1 Lagranges Method

F
= gi (x )
i


18 / 29

Partial Derivatives of the Lagrangian

Gradient of g(x ):

g1
x1
g2
x1

g1
x2
g2
x2

gm
x1

gm
x2

Dg(x ) = .
.
.

..
.

g1
xn
g2
xn

..
.

..

gm
xn

Can express sum in second term in matrix notation

m
X

i=1



gi
= T Dg(x ) j
xj

It follows that
D F (x , ) = Df (x ) + T Dg(x )


Kjell Konis (Copyright 2013)

7.1 Lagranges Method

T 

g(x )

19 / 29

Outline

Optimal Investment Portfolios

Relative Extrema of Functions of Several Variables

Lagranges Method

Example

Minimum Variance Portfolio

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

20 / 29

Example
Want to
4x2 2x3
2x1 x2 x3 = 0
x12 + x22 13 = 0

max/min:
subject to:

Start by writing down the Lagrangian

F (x , ) = f (x ) + 1 g1 (x ) + 2 g2 (x )
= 4x2 2x3 + 1 (2x1 x2 x3 ) + 2 (x12 + x22 13)
Check necessary condition:
"

2 1 1
Dg(x ) =
2x1 2x2
0
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

21 / 29

Derivatives of the Lagrangian

The Lagrangian
F (x , ) = 4x2 2x3 + 1 (2x1 x2 x3 ) + 2 (x12 + x22 13)
Gradient of the Lagrangian
T

21 + 22 x1
4 + 2 x
1
2 2

2 1
D F (x , ) =

2x1 x2 x3
x12 + x22 13

get 1 = 2 for free

Set D F (x , ) = 0 and solve for x and

set

21 + 22 x1 = 0
set
4 1 + 22 x2 = 0
set
2x1 x2 x3 = 0
set
x12 + x22 13 = 0
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

22 / 29

Example (continued)

A little algebra gives

x1 =

2
2

x2 =

3
2

x3 =

7
2

Also know that

x12 +x22

= 13

2
2

2

3
+
2


2

13
= 13
22

2 = 1

The critical points are

= (2, 1), x = (2, 3, 7), f (x ) = 26
= (2, 1), x = (2, 3, 7), f (x ) = 26

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

23 / 29

Outline

Optimal Investment Portfolios

Relative Extrema of Functions of Several Variables

Lagranges Method

Example

Minimum Variance Portfolio

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

24 / 29

Minimum Variance Portfolio

Recall: minimum variance portfolio optimization
minimize:
subject to:

w T w
eTw = 1
T w = P

Lagranges method setup

f (w ) = w T w
"

g(w ) =

"

g1 (w )
T w P = 0
=
g2 (w )
eTw 1 = 0

First, check necessary condition

"

T
Dg(x ) = T
e
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

25 / 29

Derivative of a Quadratic Form

"

a b
Let A =
b c

Let f (x ) = x T Ax = ax12 + 2bx1 x2 + cx22



Then Df (x ) = 2ax1 + 2bx2

2bx1 + 2cx2 = 2x T A


In general, let A be an n n symmetric matrix

The derivative (gradient) of the quadratic form f (x ) = x T Ax is
Df (x ) = 2x T A

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

26 / 29

Minimum Variance Portfolio

The Lagrangian
F (y , ) = w T w + 1 e T w 1 + 2 T w P


Gradient of the Lagrangian

T 

Df (w ) + T Dg(w )

g(w )

2w T + 1 e T + 2 T

eTw 1

D F (w , ) =

T w P

Find the critical point by solving the linear system

2 e
w
0
T

0 0 1 = 1
e
T 0 0
2
P
Kjell Konis (Copyright 2013)

7.1 Lagranges Method

27 / 29

Minimum Variance Portfolio

Further reading:
Second order conditions, e.g., Theorem 9.2 and Corollary 9.1 in
PFME

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

28 / 29

http://computational-finance.uw.edu

Kjell Konis (Copyright 2013)

7.1 Lagranges Method

29 / 29