0% found this document useful (0 votes)
113 views14 pages

Lagrange Multipliers for Optimization

This document discusses Lagrange multipliers, which can be used to find the extreme values (maximum and minimum) of a function subject to one or more constraints. It provides the derivation and method for using Lagrange multipliers with one or two constraints. Examples are included to demonstrate finding the maximum/minimum of a function subject to one or two constraints. Exercises are provided at the end to apply the Lagrange multiplier method.

Uploaded by

NguyễnXuân
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
0% found this document useful (0 votes)
113 views14 pages

Lagrange Multipliers for Optimization

This document discusses Lagrange multipliers, which can be used to find the extreme values (maximum and minimum) of a function subject to one or more constraints. It provides the derivation and method for using Lagrange multipliers with one or two constraints. Examples are included to demonstrate finding the maximum/minimum of a function subject to one or two constraints. Exercises are provided at the end to apply the Lagrange multiplier method.

Uploaded by

NguyễnXuân
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

Chapter 3: Partial derivatives

Lecture 4
Lagrange Multipliers
Lagrange Multipliers: Problem
Q: Find extreme values of a function f(x,y,z) subject to the constraint
g(x,y,z)=k

Recall
Gradient vector is perpendicular to the tangent plane to the
surface 𝑆: 𝐹 𝑥, 𝑦, 𝑧 = 𝑘

2
Lagrange Multipliers: Derivation
Maximize (or minimize) f(x,y) subject to the constraint g(x,y)=k

Find largest value c such that level


curve f(x,y)=c intersects g(x,y)=k.

These curves touch each other and so


they have a common tangent line.

Thus, they have the same normal line at this point: ∇f(a,b)=  ∇g(a,b)
For functions of three variables, the two level surfaces have the
same normal vector: ∇f(a,b,c)=  ∇g(a,b,c)
3
𝑆: 𝑔 𝑥, 𝑦, 𝑧 = 𝑘
𝑓 ℎ𝑎𝑠 𝑎𝑛 𝑒𝑥𝑡𝑟𝑒𝑚𝑒 𝑣𝑎𝑙𝑢𝑒 𝑎𝑡 𝑃(𝑎, 𝑏, 𝑐)

𝐶: 𝑟 𝑡 =< 𝑥 𝑡 , 𝑦 𝑡 , 𝑧 𝑡 > 𝑙𝑖𝑒𝑠 𝑜𝑛 𝑆 𝑎𝑛𝑑 𝑝𝑎𝑠𝑠𝑒𝑠 𝑡ℎ𝑟𝑜𝑢𝑔ℎ 𝑃


𝑟 𝑡0 =< 𝑎, 𝑏, 𝑐 >

ℎ 𝑡 = 𝑓(𝑥 𝑡 , 𝑦 𝑡 , 𝑧 𝑡 )
𝑓 ℎ𝑎𝑠 𝑎𝑛 𝑒𝑥𝑡𝑟𝑒𝑚𝑒 𝑣𝑎𝑙𝑢𝑒 𝑎𝑡 (𝑎, 𝑏, 𝑐)

0 = ℎ′ 𝑡0 = ∇𝑓 𝑎, 𝑏, 𝑐 . 𝑟′ 𝑡0
∇𝑓 𝑎, 𝑏, 𝑐 ⊥ 𝑟′ 𝑡0
∇𝑔 𝑎, 𝑏, 𝑐 ⊥ 𝑟′ 𝑡0 , for any C

∇f(a,b,c)=  ∇g(a,b,c)
4
Method of Lagrange Multipliers for
Constraint Optimization Problems

Problem: Find the maximum and minimum values of f(x,y,z)


subject to the constraint g(x,y,z) = k, where k is a constant

(a) Find all values of x, y, z, and  such that


f(x,y,z) = g(x,y,z)
and
g(x,y,z) = k

(b) Evaluate f at all points found in step (a):


the largest value is maximum, the smallest is minimum

5
Remarks
 If we write the vector equation
∇f(x,y,z)=  ∇g(x,y,z)
in terms of its components, then the equations in step (a)
become
f x ( x, y, z ) =  g x ( x, y, z ), f y ( x, y , z ) =  g y ( x, y , z )
f z ( x, y, z ) =  g z ( x, y, z ), g ( x, y , z ) = k
 This is a system of four equations in the four
unknowns x, y, z, and , but it is not necessary
to find explicit values for .
6
Functions of two variables
 The method of Lagrange multipliers is
similar to the method just described.
 To find the maximum and minimum values
of f(x,y) subject to the constraint g(x,y)=k we
look for values of x, y and  such that
 ∇f(x,y)=  ∇g(x,y) and g(x,y)=k
 This gives three equations in three unknowns:

f x ( x, y) =  g x ( x, y), f y ( x, y) =  g y ( x, y), g ( x, y) = k

7
Example
 The profit from the sale of x units of radiators
for automobiles and y units of radiators for
generators is given by
f(x,y)= -x2 – y2 + 4x + 8y
 Find values of x and y that lead to a maximum
profit if the firm must produce a total of 6 units
of radiators.

8
Solution
 f(x,y)= -x2 – y2 + 4x + 8y
 Set g(x,y)=x+y=6

 Method of Lagrange multipliers yields:

f x ( x, y ) = −2 x + 4 =  g x ( x, y ) = 
f y ( x, y ) = −2 y + 8 =  g y ( x, y ) =  So,this must be
the maximum
 2( x − y ) + 4 = 0  x − y = −2 value

g ( x, y ) = x + y = 6
Verify: 𝑓 2,4 = 20 > 𝑓 3,3 = 12
 x = 2, y = 4
9
Method of Lagrange Multipliers:
Problems with Two Constraints

Problem: Find the maximum and minimum values of f(x,y,z)


subject to two constraints: g(x,y,z) = k and h(x,y,z) = c

(a) Find all values of x, y, z,  and  such that


f(x,y,z) = g(x,y,z) +  h(x,y,z)
and
g(x,y,z) = k, h(x,y,z) = c

(b) Evaluate f at all points found in step (a):


the largest value is maximum, the smallest is minimum

10
Example:
Find extreme values of
f(x,y,z) = x+2y,
subject to constraints
g(x,y,z)=x+y+z = 1
h(x,y,z)= y2 + z2 = 4

(Figure shows the two


constraints)

11
f ( x, y, z ) = x + 2 y  f ( x, y, z ) = 1, 2, 0 
g ( x, y, z ) = x + y + z  g ( x, y, z ) = 1,1,1 
h( x, y, z ) = y 2 + z 2  h( x, y, z ) = 0, 2 y, 2 z 
f = g + h
 1, 2, 0 =   1,1,1  +   0, 2 y, 2 z =  ,  + 2  y,  + 2  z 

1 =  , 2 =  + 2 y, 0 =  + 2 z
1 = 2 y , − 1 = 2 z   −1 = y / z  z = − y
g ( x, y, z ) = x + y + z = 1, h ( x, y , z ) = y 2 + z 2 = 4
 x = 1, 2 y 2 = 4, y =  2 = − z

f (1, 2, − 2) = 1 + 2 2, f (1, − 2, 2) = 1 − 2 2
 max g =1,h = 4 f = f (1, 2, − 2) = 1 + 2 2

12 min g =1,h = 4 f = f (1, − 2, 2) = 1 − 2 2


Remark
 Sometimes, the unconstraint and constraint optimization
problems come together
 Example: Find the extreme values of f ( x, y) = e− xy
on the region D described by the inequality D : x 2 + 4 y 2 1

Solution:
1. Evaluate values of f at critical points inside D : x + 4y 1
2 2

2. Find max, min f subject to constraint


x2 + 4 y 2 = 1
3. Compare values of f from 1) & 2) D

13
Exercises
1. Let f ( x, y ) = (2 x − y )(1 − 2 xy )
a) Find the local maximum and local minimum
values and saddle point(s) of f
b) Find absolute maximum and minimum values of f
on the triangular region with the vertices (0,4),
(-8,0) and (4,0).
2. Using Lagrange Multipliers method, find the
maximum volume of a rectangular box whose
surface area is 1800cm2 and whose total edge
length is 240cm.
14

You might also like