Scribd
Upload
5 views4 pages

K-K-T Conditions in Optimization Analysis

The document discusses several approaches to solving constrained optimization problems: 1. Using the KKT conditions to solve a utility maximization problem subject to two constraints, finding the optimal solution when x* ≠ 0 and y* ≠ 0. 2. The substitution method, where one constraint is used to express one choice variable in terms of the other before maximizing the objective function. 3. A general approach equates the tangency points of the budget constraint and indifference curves to find the optimal solution. All three approaches yield the same optimal solution when applied to a sample constrained maximization problem.

Uploaded by

Nicholas Thigpen
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
5 views4 pages

K-K-T Conditions in Optimization Analysis

The document discusses several approaches to solving constrained optimization problems: 1. Using the KKT conditions to solve a utility maximization problem subject to two constraints, finding the optimal solution when x* ≠ 0 and y* ≠ 0. 2. The substitution method, where one constraint is used to express one choice variable in terms of the other before maximizing the objective function. 3. A general approach equates the tangency points of the budget constraint and indifference curves to find the optimal solution. All three approaches yield the same optimal solution when applied to a sample constrained maximization problem.

Uploaded by

Nicholas Thigpen
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

Notes on K-K-T and Optimization

Julio Acuna

October 10, 2018

1 K-K-T

Maximize utility subject to two constrains:

max xy s.t., 100 x + y, 40 x, x, y 0

Solution:

• Set up the Lagrangean:

L =xy + 1 (100 x y) + 2 (40 x)

• Write down FOC and K-K-T condtions:

@L
=y 1 2  0 & x⇤ (y 1 2) =0 (1)
@x

@L
=x 1  0 & y ⇤ (x 1) =0 (2)
@y
@L
= 100 x y 0 & ⇤
1 (100 x y) = 0 (3)
@ 1
@L
= 40 x 0 & ⇤
2 (40 x) = 0 (4)
@ 2

• Analyze cases:

1
Case 1:

x⇤ = 0 and y ⇤ 6= 0

1. From equation 3, 100 x y 0 =) 100 y ⇤ . But non-satiation and increasing


marginal utility imply 100 = y ⇤ (why!? ). Moreover, ⇤
1 > 0 (why?).

2. From equation 4, 40 x 0 =) (40 x) 6= 0 and ⇤


2 = 0.

3. From equation 2, x⇤ 1 < 0, since x⇤ = 0. Thus, y ⇤ ( 1) 6= 0. Contraction! y ⇤ = 100


and 1 > 0. The product cannot be 0!.

4. Since we reached a contraction, this is not an optimal solution.

Case 2:

x⇤ 6= 0 and y ⇤ = 0

1. From equation 4, 40 x⇤ . Same conditions give us x⇤ = 40 and ⇤


2 > 0 (why?)

2. From equation 3, 100 x⇤ 0 =) 100 40 and ⇤


1 = 0, since ⇤
1 (100 x⇤ ) = 0

3. From equation 2, x ⇤
1  0 and y ⇤ (x 1) = 0 =) 0 (40 0), no contradictions.

4. From equation 1, ⇤
2  0, but x⇤ (y 1 2) = 0 =) 40 (0 0 2) 6= 0.
Contradiction!

5. This is not an optimal solution.

1.0.1 Case 3

x⇤ 6= 0 and y ⇤ 6= 0. This is your typical lagrangean. No contradictions here and obviously


the optimal solution (WHY?).

2
1.0.2 What if you never reach a contradiction?

Plug in your solution set in the objective function, and pick the one with the highest value
for the objective function!

2 Optimization with restrictions

3 Substitution method

max x0.25
1 x2
0.75
s.t. 100 2x1 4x2 = 0

3.1 Solve for x

From the budget constrain, express x1 as a function of x2 :

50 2x2 = x1

3.2 Replace in objective function

max (50 2x2 )0.25 x0.75


2

3.3 Find FOC and equate to zero (why?)


@f 0.75
= 0.25 ( 2) (50 2x2 ) x0.75
2 + 0.75 (50 2x2 )0.25 x2 0.25 = 0
@x2
0.75
0.5 (50 2x2 ) x0.75
2 = 0.75 (50 2x2 )0.25 x2 0.25

150
0.5x2 = 0.75 (50 2x2 ) =) x2 =
8

3.4 Find x1

Use budget constrain

3
100
50 2x2 = x1 =) = x1
8

4 General approach

max x0.25
1 x2
0.75
s.t. 100 2x1 4x2 = 0

4.1 Equate tangency points


0.25x1 0.75 x0.75
2 1
0.25 =
0.75x0.25
1 x2 2
3
x2 = x1
2

4.2 Use budget constrain


100
100 2x1 6x1 = 0 =) x1 =
8

5 Clearly, the three approaches yield the same solution

(solve the lagrangean)

You might also like