Professional Documents
Culture Documents
14.7 Extreme Values and Saddle Points
14.7 Extreme Values and Saddle Points
§Partial Derivatives
Extreme Values and Saddle Points
Department of Mathematics
BITS Pilani, Goa
−1
0 200
100
200
300 0
Theorem
If a function f (x, y ) has a local extrema at an interior
point (a, b) of it’s domain and if the first derivatives exist
then we have
fx (a, b) = fy (a, b) = 0
Definition
An interior point of the domain of a function f is called a
critical point of the function if
1 either both the partial fx , fy vanish at the point.
2 or at least one of the partial derivatives fx and fy does
not exist at the point.
Definition
An interior point of the domain of a function f is called a
critical point of the function if
1 either both the partial fx , fy vanish at the point.
2 or at least one of the partial derivatives fx and fy does
not exist at the point.
Not every critical point of a differentiable function is a
local extremum.
Definition
An interior point of the domain of a function f is called a
critical point of the function if
1 either both the partial fx , fy vanish at the point.
2 or at least one of the partial derivatives fx and fy does
not exist at the point.
Not every critical point of a differentiable function is a
local extremum.
Just as in the one variable situation we could have two
dimensional analogue of “inflection points”. These
analogues are called “saddle points” of the function.
Dr. Tarkeshwar Singh (Dept. of Maths) Mathematics I December 27, 2022 8 / 40
Saddle Points
Definition
A differentiable function f (x, y ) is said to have a saddle
point at a critical (a, b) if for every disk centered at (a, b)
there are points (x, y ) such that f (x, y ) > f (a, b) and
points (x, y ) such that f (x, y ) < f (a, b).
Definition
A differentiable function f (x, y ) is said to have a saddle
point at a critical (a, b) if for every disk centered at (a, b)
there are points (x, y ) such that f (x, y ) > f (a, b) and
points (x, y ) such that f (x, y ) < f (a, b).
The corresponding point (a, b, f (a, b)) in the surface
z = f (x, y ) is called a saddle point of the surface.
−20
−5
0 2 4
0
5 −4 −2
20
5
0
−4 0
−2 0 2 4 −5
5
−20
−4 0
−2 0 2 4 −5
0
200
−200 0
−100 0
100
200
−200
f (x, y ) = y sin x.
f (x, y ) = y sin x.
fx = y cos x, fy = sin x.
f (x, y ) = y sin x.
fx = y cos x, fy = sin x.
f (x, y ) = y sin x.
fx = y cos x, fy = sin x.
0
−1,000
0
800
500 600
400
200
0
Theorem
Let f be a function such that the first and second partial derivatives of
the function exist at a point (a, b) and fx (a, b) = fy (a, b) = 0 then:
f has a local maximum at (a, b) if fxx (a, b) < 0 and
fxx fyy − fxy2 > 0 at the point (a, b).
Theorem
Let f be a function such that the first and second partial derivatives of
the function exist at a point (a, b) and fx (a, b) = fy (a, b) = 0 then:
f has a local maximum at (a, b) if fxx (a, b) < 0 and
fxx fyy − fxy2 > 0 at the point (a, b).
f has a local minimum at (a, b) if fxx (a, b) > 0 and
fxx fyy − fxy2 > 0 at the point (a, b).
Theorem
Let f be a function such that the first and second partial derivatives of
the function exist at a point (a, b) and fx (a, b) = fy (a, b) = 0 then:
f has a local maximum at (a, b) if fxx (a, b) < 0 and
fxx fyy − fxy2 > 0 at the point (a, b).
f has a local minimum at (a, b) if fxx (a, b) > 0 and
fxx fyy − fxy2 > 0 at the point (a, b).
f has a saddle point at (a, b) if fxx fyy − fxy2 < 0 at (a, b).
Theorem
Let f be a function such that the first and second partial derivatives of
the function exist at a point (a, b) and fx (a, b) = fy (a, b) = 0 then:
f has a local maximum at (a, b) if fxx (a, b) < 0 and
fxx fyy − fxy2 > 0 at the point (a, b).
f has a local minimum at (a, b) if fxx (a, b) > 0 and
fxx fyy − fxy2 > 0 at the point (a, b).
f has a saddle point at (a, b) if fxx fyy − fxy2 < 0 at (a, b).
The test is inconclusive at (a, b) if fxx fyy − fxy2 = 0 at (a, b) in
this case we have to look at higher derivatives to determine the
behaviour.
whereas fxx > 0 means the surface is curved “upward” and hence
has a local minimum at this point.
whereas fxx > 0 means the surface is curved “upward” and hence
has a local minimum at this point.
f (x, y ) = 2x 2 + 3xy + 4y 2 − 5x + 2y
f (x, y ) = 2x 2 + 3xy + 4y 2 − 5x + 2y
f (x, y ) = 2x 2 + 3xy + 4y 2 − 5x + 2y
fx = 4x + 3y − 5
fy = 3x + 8y + 2
f (x, y ) = 2x 2 + 3xy + 4y 2 − 5x + 2y
fx = 4x + 3y − 5
fy = 3x + 8y + 2
f (x, y ) = x 2 − y 2 − 2x + 4y + 6
f (x, y ) = x 2 − y 2 − 2x + 4y + 6
f (x, y ) = x 2 − y 2 − 2x + 4y + 6
f (x, y ) = x 2 − y 2 − 2x + 4y + 6
f (x, y ) = x 3 − y 3 − 2xy + 6.
f (x, y ) = x 3 − y 3 − 2xy + 6.
fx = 3x 2 − 2y
fy = −3y 2 − 2x
f (x, y ) = x 3 − y 3 − 2xy + 6.
fx = 3x 2 − 2y
fy = −3y 2 − 2x
So the critical points are: (0, 0), and (−2/3, 2/3).
f (x, y ) = x 3 − y 3 − 2xy + 6.
fx = 3x 2 − 2y
fy = −3y 2 − 2x
So the critical points are: (0, 0), and (−2/3, 2/3).
Definition
The maximum (or minimum value ) of a function in a closed and
bounded region R is called an absolute maximum (minimum) of the
function in that region.
Definition
The maximum (or minimum value ) of a function in a closed and
bounded region R is called an absolute maximum (minimum) of the
function in that region.
Definition
The maximum (or minimum value ) of a function in a closed and
bounded region R is called an absolute maximum (minimum) of the
function in that region.
Definition
The maximum (or minimum value ) of a function in a closed and
bounded region R is called an absolute maximum (minimum) of the
function in that region.
Tx = 2x + y − 6 = 0
Ty = 2y + x = 0
x = 4, y = −2
Value at this point is: T (4, −2) = −10
(4,-2) -10
(3,0) -7
(9/2,-3) -9.25
(0,0) 2
(5,-5/2) -9.25
(0,-3) 11
(5,-3) -9
(5,0) -3
So the absolute maximum is 11 attained at the point (0,-3) and the
absolute minimum is -10 attained at the point (4,-2).
fx = (4 − 2x) cos y = 0
fy = −(4x − x 2 ) sin y = 0
Solving these two gives us (2, 0). The value at this point is 4.
(2,0) √4
(2,−π/4) 2√2
(2,π/4) 2 2
(1,0) 3
(3,0) 3√
(1, −π/4), (1, π/4), (3, −π/4), (3, π/4) 3/ 2
So the absolute maximum √ is at the point (2,0) and the value is 4, and
absolute minimum is 3/ 2 at the points
(1, −π/4), (1, π/4), (3, −π/4), (3, π/4).