Scribd
Upload
79 views

Topic 4 - Part2

1) The document discusses using the Hessian matrix and its determinant to test for relative extrema of multi-variable objective functions. The determinant must be positive for a minimum and negative for a maximum. 2) An example shows finding the minimum of a 3-variable function by taking partial derivatives, solving for critical points, and checking that the Hessian determinant is positive at those points. 3) A second example applies the method to maximize profit for a firm producing two goods, finding the profit-maximizing output levels and using the Hessian determinant to prove it is a maximum.

Uploaded by

Teo Sheng
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
79 views

Topic 4 - Part2

1) The document discusses using the Hessian matrix and its determinant to test for relative extrema of multi-variable objective functions. The determinant must be positive for a minimum and negative for a maximum. 2) An example shows finding the minimum of a 3-variable function by taking partial derivatives, solving for critical points, and checking that the Hessian determinant is positive at those points. 3) A second example applies the method to maximize profit for a firm producing two goods, finding the profit-maximizing output levels and using the Hessian determinant to prove it is a maximum.

Uploaded by

Teo Sheng
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

MOHD FAISOL MD.

SALLEH BEEQ2013 A201 TOPIC 4 (PART 2)

8 4.1 Objective functions with vi. Use Hessian determinant for second order  Chap 11.4, 11.6
more than two variables condition test.
4.2 Economic application
vii. Solve optimization problem for the case of more 3
than one choice variable in Theory of the Firm
 Maximization of profit or output

Hessian Matrix and its determinant

The Hessian (H) for a function is a square matrix of second-order partials and cross-partials of the function. Accordingly, for
each of the matrix, we can compute its determinant (which is a scalar).

We can use its determinant to test for relative extremum of a given multi-choice variable objective function. It is known as
determinant test.

Example 1: General form of objective function

For a two-choice variable objective function 𝒁 = Three-choice variable objective function 𝒁 = 𝒇(𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 ).
𝒇(𝒙𝟏 , 𝒙𝟐 ). 𝑻𝒉𝒆𝒏, 𝒅𝒁 = 𝒁𝟏 . 𝒅𝒙𝟏 + 𝒁𝟐 . 𝒅𝒙𝟐 + 𝒁𝟑 . 𝒅𝒙𝟑
𝑻𝒉𝒆𝒏, 𝒅𝒁 = 𝒁𝟏 . 𝒅𝒙𝟏 + 𝒁𝟐 . 𝒅𝒙𝟐

Hessian (H) Its determinant (|H|) Hessian (H) of 3x3 Its determinant (|H|)
of 2x2 𝑍11 𝑍12 𝑍11 𝑍12 𝑍13 𝑍11 𝑍12 𝑍13
𝑍 𝑍12 = | |
𝑍21 𝑍22 =[𝑍21 𝑍22 𝑍23 ] = |𝑍21 𝑍22 𝑍23 |
=[ 11 ]
𝑍21 𝑍22 𝑍31 𝑍32 𝑍33 𝑍31 𝑍32 𝑍33

1
MOHD FAISOL MD. SALLEH BEEQ2013 A201 TOPIC 4 (PART 2)

F.O.C: (both S.O.C: F.O.C: (both for S.O.C:


for minimum/ For minimum (i.e positive definite H minimum/ For minimum (i.e positive definite H
maximum) matrix) requires: maximum) matrix) requires:
|H1 | > 0, |H2 | > 0 |H1 |> 0, |H2 | > 0, |H3 | > 0
𝑍1 = 𝑍2 = 0 𝑍1 = 𝑍2 = 𝑍3 = 0
For maximum (i.e negative definite H For maximum (i.e negative definite H
matrix) requires: matrix) requires:
|H1 |< 0, |H2 | > 0 |H1 |< 0, |H2 | > 0, |H1 | < 0
(𝑎𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑒 𝑖𝑛 𝑠𝑖𝑔𝑛𝑠 𝑏𝑒𝑔𝑖𝑛𝑛𝑖𝑛𝑔 𝑤𝑖𝑡ℎ < (𝑎𝑙𝑡𝑒𝑟𝑛𝑎𝑡𝑒 𝑖𝑛 𝑠𝑖𝑔𝑛𝑠 𝑏𝑒𝑔𝑖𝑛𝑛𝑖𝑛𝑔 𝑤𝑖𝑡ℎ <
0) 0)

Example 2: A numerical 3-choice variables objective function

𝑍 = 2𝑥12 + 𝑥1 𝑥2 + 4𝑥22 + 𝑥1 𝑥3 + 𝑥32 + 2. Find the extreme value(s) of the function.

Step 1: Specify the objective function.

𝑍 = 2𝑥12 + 𝑥1 𝑥2 + 4𝑥22 + 𝑥1 𝑥3 + 𝑥32 + 2.

Step 2: Partially differentiate the function with respect to each of its choice variables

𝑍1 = 4𝑥1 + 𝑥2 + 𝑥3
𝑍2 = 𝑥1 + 8𝑥2
𝑍3 = 𝑥1 + 2𝑥3

2
MOHD FAISOL MD. SALLEH BEEQ2013 A201 TOPIC 4 (PART 2)

Step 3: Check for the conditions for relative extremum (i.e F.O.C and S.O.C) using determinant test:

F.O.C:
𝑍1 = 4𝑥1 + 𝑥2 + 𝑥3 = 0 . . . (1)
𝑍2 = 𝑥1 + 8𝑥2 = 0 … (2)
𝑍3 = 𝑥1 + 2𝑥3 = 0 …(3)

Solve the 3-homogenous equation system by matrix approach, for critical value of 𝑥1 , 𝑥2 , and 𝑥3 .

4 1 1 𝑥1 0
𝑥
[1 8 0] [ 2 ] = [0]
1 0 2 𝑥3 0
𝑥̅1 0
[𝑥̅2 ] = [0]
𝑥̅3 0

S.O.C:

Hessian (H) of 3x3

𝑍11 𝑍12 𝑍13 4 1 1


=[𝑍21 𝑍22 𝑍23 ]=[1 8 0]
𝑍31 𝑍32 𝑍33 1 0 2

3
MOHD FAISOL MD. SALLEH BEEQ2013 A201 TOPIC 4 (PART 2)

𝑍11 𝑍12 𝑍13 4 1 1


𝑍 𝑍23 8 0 1 1 4 1
|H1 |= 4 > 0, |H2 | = | 22 |=| | = 16 > 0, |H3 | = |𝑍21 𝑍22 𝑍23 | = |1 8 0| = 1. | | + 2. | |=
𝑍32 𝑍33 0 2 8 0 1 8
𝑍31 𝑍32 𝑍33 1 0 2
−8 + 62 = 54 > 0

Since |H1 |> 0, |H2 | > 0, |H3 | > 0 are all positive, the H is said to be positive definite.

Step 4: Conclusion

Positive definite H indicates the function achieves its relative minimum when evaluate at its critical value (0,0,0). The value
of the extreme i.e minimum is 𝑍 = 2 unirs.

Example 3: Economic application

A firm is a monopolistic producer of two goods. The prices are related to quantities 𝑄1 and 𝑄2 according to the demand
functions 𝑃1 = 50 − 𝑄1 and 𝑃2 = 95 − 3𝑄2 . If the total cost function is 𝑇𝐶 = 𝑄12 + 3𝑄1 𝑄2 + 𝑄22 , find the profit of the firm
and prove that the profit is at a maximum.

Step 1: Derive the firm’s total revenue function.

TR = PQ
TR1 = (50 − Q1 )Q = 50Q1 − Q21 and TR 2 = (95 − 3Q 2 )Q = 95Q 2 − 3Q22
∴ TR = TR1 + TR 2
TR = 50Q1 − Q21 + 95Q 2 − 3Q22

4
MOHD FAISOL MD. SALLEH BEEQ2013 A201 TOPIC 4 (PART 2)

Step 2: Find the firm’s total profit function.

π = TR − TC
π = 50Q1 − Q21 + 95Q 2 − 3Q22 − (Q21 + 3Q1 Q 2 + Q22 )
π = 50Q1 − 2Q21 + 95Q 2 − 4Q22 − 3Q1 Q 2

Step 3: Calculate the values of 𝑄1 and 𝑄2 which maximize profit.

FOC:
𝜋1 = 50 − 4𝑄1 − 3𝑄2 = 0 … … … (1)
𝜋2 = 95 − 3𝑄1 − 8𝑄2 = 0 … … … (2)

To eliminate 𝑄1 , 𝑠𝑜 (1) × 3 𝑎𝑛𝑑 (2) × 4 to be (3) and (4), respectively, then (3) - (4)
𝜋1 = 150 − 12𝑄1 − 9𝑄2 = 0 … . . (3)
𝜋2 = 380 − 12𝑄1 − 32𝑄2 = 0 … . (4)

−230 + 23𝑄2 = 0
230
𝑄̅2 = = 10 units
23

Substitute 𝑄2 = 10 into (1) to get 𝑄1


50 − 4𝑄1 − 3(10) = 0
20 − 4𝑄1 = 0
20
𝑄̅1 = = 5 units
4

5
MOHD FAISOL MD. SALLEH BEEQ2013 A201 TOPIC 4 (PART 2)

The value of the profit is π(5, 10) = 50Q1 − 2Q21 + 95Q 2 − 4Q22 − 3Q1 Q 2 = 50(5) − 2(25) + 95(10) − 4(100) −
3(5)(10) = 600 units.

Step 4: Using the second order condition (SOC), prove that the profit level is at a maximum.

SOC
𝜋11 = −4; 𝜋22 = −8; 𝜋12 = 𝜋21 = −3

Since 𝜋11 = −4 < 0 ; 𝜋22 = −8 < 0

and (𝜋11 )(𝜋22 ) > (𝜋12 )2


(−4)(−8) > (−3)2
32 > 9

Since 𝜋11 < 0, 𝜋22 < 0 and (𝜋11 )(𝜋22 ) > (𝜋12 )2 , so it is negative definite and the profit level of 600 units is at a
20 230
maximum when 𝑄̅1 = = 5 units and 𝑄̅2 = = 10 units.
4 23

OR using determinant test of Hessian.

|𝐻1 | = −4 < 0

−4 −3
|𝐻2 | = | |
−3 −8
= (−4)(−8) − [(−3)(−3)]
= 32 − 9
= 23 > 0

6
MOHD FAISOL MD. SALLEH BEEQ2013 A201 TOPIC 4 (PART 2)

̅̅̅̅2 | >0, so H is negative definite and the profit level of 600 units when 𝑄̅1 = 20 = 5
̅̅̅1̅| <0 and |𝐻
In conclusion, since |𝐻
4
230
units and 𝑄̅2 = = 10 units is indeed a maximum.
23

You might also like