Scribd
Upload
52 views

Lagrange PDF

The document summarizes the method of Lagrange multipliers for finding extrema of a function subject to equality constraints. It provides: 1) An example of using Lagrange multipliers to find extrema of f(x,y)=2xy+y^2 on the unit circle constraint g(x,y)=x^2+y^2=1. This yields a system of equations with solutions giving the extrema. 2) An explanation of why the method works - that the gradients of the objective function f and constraint g must be parallel at an extremum. 3) An example of using Lagrange multipliers to find extrema of a function f of 4 variables subject to two

Uploaded by

carolina
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
52 views

Lagrange PDF

The document summarizes the method of Lagrange multipliers for finding extrema of a function subject to equality constraints. It provides: 1) An example of using Lagrange multipliers to find extrema of f(x,y)=2xy+y^2 on the unit circle constraint g(x,y)=x^2+y^2=1. This yields a system of equations with solutions giving the extrema. 2) An explanation of why the method works - that the gradients of the objective function f and constraint g must be parallel at an extremum. 3) An example of using Lagrange multipliers to find extrema of a function f of 4 variables subject to two

Uploaded by

carolina
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/ 2

system mentioned above has these three equations

fx (x, y) = 2y = gx (x, y) = 2x
fy (x, y) = 2x + 2y = gy (x, y) = 2y
g(x) = x2 + y 2 = c = 1

Lagrange multipliers
Math 131 Multivariate Calculus

in the three unknowns x, y, and . The first two


equations give

D Joyce, Spring 2014

=
Constraints and Lagrange multipliers.
Weve seen how to find extrema for a function
when were looking in an open subset of Rn ,
namely, find the critical points, then determine
which give extrema, perhaps by using the secondderivative test. But what if the extrema occur on
the boundaries of that open subset.
For example, suppose f is defined on the closed
unit disk, that is, when kxk 1, and its defined by
f (x, y) = 2xy +y 2 . How do you determine if theres
an extremum on the boundary? The boundary is
the unit circle, which is defined by kxk = 1. In
this context, the equation kxk = 1 is called a constraint when we look for extreme values of f (x) for
x which satisfy the equation. Lagrange developed
a technique, now called the method of Lagrange
multipliers, to solve this problem.

x+y
x
y
=
= + 1,
x
y
y

and the equation = 1 + 1 has the two solutions

1 = 12 (1 + 5) and 2 = 12 (1 5). These lead to


the four solutions of the system. Two turn out to
be maxima and two minima.
Why it works. Lets look at the planar case.
Suppose x is a point satisfying g(x, y) = c. The
gradient f (x) points in the direction in R2 in
which f grows fastest, while g(x) is normal to
the curve g(x, y) = c. So, if these directions, f (x)
and g(x), arent the same, then f will increase in
one direction along the curve and decrease in the
opposite direction. A sketch helps here. So the
only way an extremum can occur at x is if f (x)
is some multiple, denoted here, of g(x).
More than one constraint. A variant of this
method works when there is more than one constraint. Suppose that g1 (x) = c1 and g2 (x) = c2
are two constraints that have to be satisfied. Then
solve the system of equations

The method. First, well see how it works, then


well see why it works. Suppose we want to find the
extreme values of a function f : Rn R subject
to the constraint g(x) = c, where g is a function
g : Rn R and c is a constant. Introduce a new
variable , and solve the system of equations

f (x) = 1 g1 (x) + 2 g2 (x)


g1 (x) = c1
g2 (x) = c2

f (x) = g(x)
g(x) = c

to find critical points.

Example 2. Consider the function f : R4 R


to get what we call critical points subject to the defined by f (x) = f (w, x, y, z) = w2 x2 +y 2 z 2 . Find
constraint. Then determine which of these give the the extrema on the plane of points in R4 which
extreme values of f .
satisfy
g1 (x) = w + x + y + z = 2
g2 (x) = w + x y z = 1

Example 1. Consider the function f (x) = 2xy+y 2


subject to the constraint g(x, y) = x2 +y 2 = 1. The
1

Since f (x) = (2wx2 , 2w2 x, 2yz 2 , 2y 2 z), g1 (x) =


(1, 1, 1, 1), and g2 (x) = (1, 1, 1, 1), we have
the system of six equations in six unknowns
2wx2
2wx
2yz 2
2y z
w+x+y+z
w+xyz

=
=
=
=
=
=

1 + 2
1 + 2
1 2
1 2
2
1

We can eliminate 1 and 2 from the first two


equations to get the equations 2wx2 = 2w2 x and
2yz 2 = 2y z . From the first of these, either w = 0
or x = 0 or w = x. From the second, either y = 0
or z = 0 or z = y. Thats nine combination cases
to consider in all.
Lets look at just one of them, say w = 0 and
z = y. The last two equations become x + 2y = 2
and x = 1. In this case we get the unique solution
(w, x, y, z) = (0, 1, 32 , 32 ).
The other eight cases will also yield some solutions. Among all these will be a global minimum,
but as f takes arbitrarily large values on this plane,
f wont have a maximum.
Math 131 Home Page at
http://math.clarku.edu/~djoyce/ma131/

You might also like