Scribd
Upload
99 views

General Second Derivative Test

The document discusses the Second Derivative Test for functions of n variables. It begins by reviewing the test for 1 and 2 variables from Stewart's Calculus textbook. It then presents the generalized test for n variables, showing the pattern that emerges. The test classifies critical points as local minima/maxima based on the signs of the determinants of successive principal minors of the Hessian matrix evaluated at the critical point. An example demonstrates applying the test to classify the critical point of a function of 3 variables.

Uploaded by

Cure Master Tattoo
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
99 views

General Second Derivative Test

The document discusses the Second Derivative Test for functions of n variables. It begins by reviewing the test for 1 and 2 variables from Stewart's Calculus textbook. It then presents the generalized test for n variables, showing the pattern that emerges. The test classifies critical points as local minima/maxima based on the signs of the determinants of successive principal minors of the Hessian matrix evaluated at the critical point. An example demonstrates applying the test to classify the critical point of a function of 3 variables.

Uploaded by

Cure Master Tattoo
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/ 4

The Second Derivative Test in n variables.

Brian D. Sittinger
February 19, 2010

1 Introduction.
Stewart’s Calculus textbook does a fine job of addressing of addressing the
differential calculus of functions of n variables f : Rn → R in the cases of
n = 2 or 3. (As such, it’s usually easy to guess how these formulas generalise
for arbitrary n.) However, it is curious that Stewart only covers the second
derivative test in the case of n = 2 variables. After reviewing the cases of n = 1
and 2, I’ll address what happens in general.

2 Second Derivative Test in 1 or 2 variables


Before stating the second derivative test as mentioned in Stewart, recall that for
a function y = f (x), the second derivative test uses concavity of the function
at a critical point to determine whether we have a local maximum or minimum
value at the said point.

Second Derivative Test, Single variable case:


Suppose that f 00 is continuous near c, where f 0 (c) = 0 (that is, c is a critical
point of f ).
(a) If f 00 (c) > 0, then f has a local minimum at c.
(b) If f 00 (c) < 0, then f has a local maximum at c. ¥

Note that the theorem is silent in the case that f 00 (c) = 0. In such a case, one
has to try a different method to determine the nature of the critical point.

1
Here is the analogous statement for two variables.

Second Derivative Test, Two variable case:


Suppose that the second partial derivatives of f : R2 → R are continuous on a
disk with centre (a, b), where fx (a, b) = 0 and fy (a, b) = 0 (that is, (a, b) is a
critical point of f ).

Let H denote the “Hessian” matrix of second partial derivatives


· ¸
fxx fxy
H= ,
fyx fyy
¯ ¯
¯ fxx fxy ¯
and let D1 = fxx and D2 = det H = ¯ ¯ ¯.
fyx fyy ¯
(a) If D1 (a, b) > 0 and D2 (a, b) > 0, then f has a local minimum at (a, b).
(b) If D1 (a, b) < 0 and D2 (a, b) > 0, then f has a local maximum at (a, b).
(c) If D2 (a, b) < 0, then f has a saddle point at (a, b). ¥
A couple of remarks are in order. First, I write D2 in terms of a determinant
because not only is it easier to remember (which is always a good thing), but
also it helps to point the way to further generalisations. Second, the test is
inconclusive when D2 (a, b) = 0.

Example: Find and classify the critical points of the function


f (x, y) = 3x2 y + y 3 − 3x2 − 3y 2 + 1.
Solution: First, we set the first partial derivatives equal to zero:

fx = 6xy − 6x = 0 and fy = 3x2 + 3y 2 − 6y = 0.


This yields four critical points (0, 0), (0, 2), (1, 1), (−1, 1). Next, we compute the
matrix of second partial derivatives:
· ¸
6y − 6 6x
H= .
6x 6y − 6
Now, we apply the second derivative test to each critical point.
¯ ¯
¯ −6 0 ¯
1. At (0, 0), D1 (0, 0) = −6 < 0, D2 (0, 0) = ¯¯ ¯ = 36 > 0. Hence, f
0 −6 ¯
has a local maximum at (0, 0).
¯ ¯
¯ 6 0 ¯
¯
2. At (0, 2), D1 (0, 2) = 6 > 0, D2 (0, 0) = ¯ ¯ = 36 > 0. Hence, f has a
0 6 ¯
local minimum at (0, 0).
¯ ¯
¯ 0 ±6 ¯
3. At (±1, 1), D2 (±1, 1) = ¯ ¯ ¯ = −36 < 0. Hence, f has saddle
±6 0 ¯
points at (±1, 1). ¥

2
3 Second Derivative Test in 3 or more variables
By using the Hessian matrix, stating the second derivative test in more than 2
variables is not too difficult to do. Before stating the general theorem, we will
first state it in 3 variables (so the pattern is clear) and work an example.
Second Derivative Test, Three variable case:
Suppose that the second partial derivatives of f : R3 → R are continuous
on a ball with centre (a, b, c), where fx (a, b, c) = 0 and fy (a, b, c) = 0, and
fz (a, b, c) = 0 (that is, (a, b, c) is a critical point of f ).

Let H denote the Hessian matrix of second partial derivatives


 
fxx fxy fxz
H =  fyx fyy fyz  ,
fzx fzy fzz
¯ ¯
¯ ¯ ¯ fxx fxy fxz ¯
¯ fxx fxy ¯ ¯ ¯
and let D1 = fxx , D2 = ¯¯ ¯, and D3 = det H = ¯ fyx fyy fyz ¯.
¯ ¯ ¯
fyx fyy ¯ fzx fzy fzz ¯
(a) If D1 (a, b, c) > 0, D2 (a, b, c) > 0, and D3 (a, b, c) > 0, then f has a local
minimum at (a, b, c).
(b) If D1 (a, b, c) < 0, D2 (a, b, c) > 0, and D3 (a, b, c) < 0, then f has a local
maximum at (a, b, c).
(c) In any other case where D3 (a, b, c) 6= 0, f has a saddle point at (a, b, c). ¥
Example: Find and classify the critical points of the function
f (x, y, z) = x2 + y 2 + 7z 2 − xy − 3yz.
Solution: First, we set the first partial derivatives equal to zero:

fx = 2x − y = 0, fy = 2y − x − 3z = 0, and fz = 14z − 3y = 0.

This yields exactly one critical point (0, 0, 0). Next, we compute the matrix of
second partial derivatives:
 
2 −1 0
H =  −1 2 −3  .
0 −3 14
Now, we apply the second derivative ¯ test to ¯the critical point (0, 0, 0).
¯ 2 −1 ¯
D1 (0, 0, 0) = 2 > 0, D2 (0, 0, 0) = ¯¯ ¯ = 3 > 0, and
¯ ¯ −1 2 ¯
¯ 2 −1 0 ¯
¯ ¯
D3 (0, 0, 0) = ¯¯ −1 2 −3 ¯¯ = 24 > 0.
¯ 0 −3 14 ¯
Hence, f has a local minimum at (0, 0, 0). ¥

3
Now, we can state the general version of the second derivative test. Observe
how this theorem indeed does summarise the forms of this test given in the
specific cases. For brevity, let ~x = (x1 , x2 , ..., xn ), and use the gradient of f to
concisely list the partial derivatives.
Second Derivative Test, the general n variable version:
Suppose that the second partial derivatives of f : Rn → R are continuous on a
ball with centre ~c, where ∇f (~c) = ~0 (that is, ~c is a critical point of f ).

Let H denote the Hessian matrix of second partial derivatives, and for each
k = 1, 2, ..., n, let Dk denote the determinant of the Hessian in the variables x1 ,
x2 , ..., xk . Assume that |H(~c)| 6= 0.
(a) If Dk (~c) > 0 for all k = 1, 2, ..., n, then f has a local minimum at ~c.
(b) (−1)k · Dk (~c) > 0 for all k = 1, 2, ..., n, then f has a local maximum at ~c.
(c) Otherwise, f has a saddle point at ~c. ¥
This theorem is usually proved using the quadratic approximation of the (mul-
tivariable) Taylor Series for f centred at ~x = ~c and understandably involves a
good amount of Linear Algebra. The curious student may consult an Advanced
Calculus textbook for a proof of this theorem.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy