differential in several variables
TRANSCRIPT
CHAPTER IIDIFFERENTATION IN SEVERAL VARIABLES
Department of Foundation Year,
Institute of Technology of Cambodia
2015–2016
CALCULUS II () ITC 1 / 43
Contents
1 Function of Two or More Variables
2 Limits and Continuity
3 Partial Derivatives
4 Chain Rules for Functions of Several Variables
5 Directional Derivatives and Gradient Vectors
6 Tangent Planes and Normal Lines
7 Extrema of Function of Several Variables
8 Lagrange Multipliers
CALCULUS II () ITC 1 / 43
Contents
1 Function of Two or More Variables
2 Limits and Continuity
3 Partial Derivatives
4 Chain Rules for Functions of Several Variables
5 Directional Derivatives and Gradient Vectors
6 Tangent Planes and Normal Lines
7 Extrema of Function of Several Variables
8 Lagrange Multipliers
CALCULUS II () ITC 2 / 43
Introduction to Functions of Several Variables
Note
A real-valued function on subset D of Rn is a function whose rangeis R. That is,
f : D ⊂ Rn −→ R(x1, x2, . . . , xn) 7−→ y = f(x1, x2, . . . , xn)
Special cases for n = 2 and n = 3 will be mainly concerned since theyhelp to visualise their geometrical meaning.
CALCULUS II () ITC 3 / 43
Introduction to Functions of Several Variables
Some Operations on Rn: Letx = (x1, x2, . . . , xn), y = (y1, y2, . . . , yn) ∈ Rn and α ∈ R. We define
Addition:x+ y = (x1 + y1, x2 + y2, . . . , xn + yn)
Scalar multiplication:
αx = (αx1, αx2, . . . , αxn)
Inner product:
x · y = 〈x, y〉 = xy = x1y1 + x2y2 + · · ·+ xnyn
CALCULUS II () ITC 4 / 43
Introduction to Functions of Several Variables
In this study we use only Euclidean Norm, that is ifx = (x1, x2, . . . , xn) ∈ Rn, then norm of x is defined by
‖x‖ =
(n∑
i=1
x2i
)1/2
=√x2
1 + x22 + · · ·+ x2
n.
If x = (x1, x2, . . . , xn), y = (y1, y2, . . . , yn) ∈ Rn, then norm of thedifference x and y (or Euclidean distance between x and y)is defined by
‖x− y‖ =
[n∑
i=1
(xi − yi)2
]1/2
.
CALCULUS II () ITC 5 / 43
Level curve, level surface and level hypersurfce
Definition 1 (Level curve, level surface and level hypersurfce)
The set of points (x1, x2, . . . , xn) in Rn where a function of nindependent variables has a constant value f(x1, x2, . . . , xn) = c iscalled a level hypersurface of f .In particular,
the set of points in the plane where a function f(x, y) has aconstant value f(x, y) = c is called a level curve of f .
the set of points in the space where a function f(x, y, z) has aconstant value f(x, y, z) = c is called a level surface of f .
Note that if n ≥ 4, the set of points satisfying the equationf(x1, x2, . . . , xn) = c is called level hypersurface.
CALCULUS II () ITC 6 / 43
Example of level curves
Figure: Level curves show the lines ofequal pressure (isobars) measured inmillibars
Figure: Level curves show the lines ofequal temperature (isotherms)measured in degree Fahrenheit.
CALCULUS II () ITC 7 / 43
Level Curves
CALCULUS II () ITC 8 / 43
Contents
1 Function of Two or More Variables
2 Limits and Continuity
3 Partial Derivatives
4 Chain Rules for Functions of Several Variables
5 Directional Derivatives and Gradient Vectors
6 Tangent Planes and Normal Lines
7 Extrema of Function of Several Variables
8 Lagrange Multipliers
CALCULUS II () ITC 9 / 43
Limit of a Function of Several Variables
Definition 2
Let f : D ⊂ Rn → Rm and a ∈−D. Then we say that the limit of f(x)
equals L as x approaches a, written as
limx→a
f(x) = L,
if given any ε > 0, there exists a δ > 0 such that
‖f(x)− L‖Rm < ε whenever ‖x− a‖Rn < δ.
Theorem 1
If a limit exists, it is unique.
CALCULUS II () ITC 10 / 43
Limit of a Function of Several Variables
Theorem 2
Suppose that limx→a
f(x) and limx→a
g(x) both exist and that k is a scalar.
Then
limx→a
[f(x)± g(x)] = [ limx→a
f(x)]± [ limx→a
g(x)]
limx→a
[kf(x)] = k[ limx→a
f(x)]
limx→a
[f(x)g(x)] = [ limx→a
f(x)][ limx→a
g(x)]
limx→a
[f(x)/g(x)] = [ limx→a
f(x)]/[ limx→a
g(x)] provided limx→a
g(x) 6= 0 and
both f and g are real-valued functions.
If f(x) ≤ g(x) for all x, then limx→a
f(x) ≤ limx→a
g(x), where f and g
are real-valued functions.
If ‖f(x)− L‖ ≤ g(x) for all x and if limx→a
g(x) = 0, then
limx→a
f(x) = L.
CALCULUS II () ITC 11 / 43
Limit of a Function of Several Variables
Note
To show that the limit does not exist, we need to show that thefunction approaches different values as x approaches a along differentpaths in Rn.
Definition 3
Technique for Showing that lim(x,y)→(a,b)
f(x, y) Doesn’t Exist
If f(x, y) approaches two different numbers as (x,y) approaches (a, b)along two different paths, then lim
(x,y)→(a,b)f(x, y) = L does not exist.
CALCULUS II () ITC 12 / 43
Definition of Limits
Example 4
Show that
a). lim(x,y)→(0,0)
x2 − y2
x2 + y2b). lim
(x,y)→(0,0)
xy
x2 + y2
doesn’t exist.(Hint: Consider different parts C1 : y = 0, C2 : x = 0, C3 : y = x)
Theorem 3
Let f : D ⊂ Rn → Rm be a vector-value functions,f = (f1, f2, . . . , fm)and L = (L1, L2, . . . , Lm). Then lim
x→af(x) = L if and only if
limx→a
fi(x) = Li for i = 1, 2, . . . ,m.
CALCULUS II () ITC 13 / 43
Continuity
Definition 5
Let f : D ⊂ Rn 7−→ Rm. We say that the function f is continuous ata point a in D if
limx→a
f(x) = f(a).
We say that f is a continuous function on D if it is continuous atevery point in its domain D.
Theorem 4 (Algebraic properties)
Let f, g : D ⊂ Rn → Rm be continuous vector-value function and letα ∈ R be a scalar. Then
f + αg and fg are continuous.
If both f and g are real-valued functions and if g(x) 6= 0, then f/gis continuous.
CALCULUS II () ITC 14 / 43
Continuity
Theorem 5
Let f : D ⊂ Rn 7−→ Rm, where f = (f1, . . . , fm). Then f is continuousat a ∈ D (respectively f is continuous on D) if and only if itscomponent functions fi : D → R are all continuous at a (respectively fiare continuous on D).
CALCULUS II () ITC 15 / 43
Contents
1 Function of Two or More Variables
2 Limits and Continuity
3 Partial Derivatives
4 Chain Rules for Functions of Several Variables
5 Directional Derivatives and Gradient Vectors
6 Tangent Planes and Normal Lines
7 Extrema of Function of Several Variables
8 Lagrange Multipliers
CALCULUS II () ITC 16 / 43
Partial derivatives of a function of several variables
Definition 6
Let f : D ⊂ Rn 7−→ R. If y = f(x) = f(x1, x2, . . . , xn), then the firstpartial derivative of f with respect to xi, i ∈ {1, 2, . . . , n}, is definedby fxi(x) or ∂f
∂xi(x) and
fxi(x) =∂
∂xif(x) = lim
∆xi→0
f(x1, . . . , xi + ∆xi, . . . , xn)− f(x1, . . . , xn)
∆xi
provided the limit exists.
Note that if the function∂f
∂xihas a partial derivative with respect to
xj , we denote the partial derivative by
∂
∂xj
(∂f
∂xi
)=
∂2f
∂xj∂xi= fxixj .
CALCULUS II () ITC 17 / 43
Partial Derivatives of Functions of Two Variables
Definition 7 (Partial Derivatives of a Function of Two Variables)
Let z = f(x, y). Then the partial derivative of with respect to xis
fx =∂f
∂x= lim
h→0
f(x+ h, y)− f(x, y)
h
and the partial derivative of with respect to y is
fy =∂f
∂y= lim
h→0
f(x, y + h)− f(x, y)
h
CALCULUS II () ITC 18 / 43
Partial Derivatives
Definition 8
A function f : D ⊂ Rn → R is said to be of class Ck if all itspartial derivatives of order ≤ k are continuous.
A function f : D ⊂ Rn → R is said to be of class C∞ if f hascontinuous partial derivatives of all orders.
A function f : D ⊂ Rn → Rm is said to be of class Ck if each ofcomponent functions is of class Ck.
A function f : D ⊂ Rn → Rm is said to be of class C∞ if each ofcomponent functions is of class C∞.
CALCULUS II () ITC 19 / 43
Partial Derivatives
The function obtained by differentiating f successively with respect toxi1 , xi2 , . . . xir at x is denoted by
∂kf
∂xir∂xir−1 . . . ∂xi1= fxi1
xi2...xir
where i1 + · · ·+ ir = k.
It is called a kth-order partial derivative of f .
Theorem 6
Let f : D ⊂ Rn → R be a Ck function. Then
fxi1xi2
...xir(x) = fxj1
xj2...xjr
(x)
where i1 + i2 + · · ·+ ir = j1 + j2 + · · ·+ jr = k.
CALCULUS II () ITC 20 / 43
Contents
1 Function of Two or More Variables
2 Limits and Continuity
3 Partial Derivatives
4 Chain Rules for Functions of Several Variables
5 Directional Derivatives and Gradient Vectors
6 Tangent Planes and Normal Lines
7 Extrema of Function of Several Variables
8 Lagrange Multipliers
CALCULUS II () ITC 21 / 43
Chain rules for functions of several variables
Theorem 7 (Chain rules for function of several variables)
Let y = f(x1, x2, . . . , xn), where f is differentiable function ofxi, i = 1, 2, . . . , n. If each xi, i = 1, 2, . . . , n is a differentiable function ofm variables t1, t2, . . . , tm, then y is a differentiable function oft1, t2, . . . , tm and
∂y
∂tj=
n∑i=1
∂y
∂xi
∂xi∂tj
.
for j = 1, 2, . . . ,m.In particular, if xi, i = 1, 2, . . . , n is a function of a single variable t,then we have
dy
dt=
n∑i=1
∂y
∂xi
dxidt.
CALCULUS II () ITC 22 / 43
Chain rules for functions of several variables
Theorem 9 (Chain rule: Implicit Differentiation)
If the equation F (x1, x2, . . . , xn, y) = 0 defines y implicitly as adifferentiable function of xi, i = 1, 2, . . . , n, then
∂y
∂xi= −Fxi(x1, x2, . . . , xn, y)
Fy(x1, x2, . . . , xn, y), Fy(x1, x2, . . . , xn, y) 6= 0
for i = 1, 2, . . . , n.
Example 10
Given xy3 + zy − 5z3 + 6 = 0. Find ∂z/∂x and ∂z/∂y.
CALCULUS II () ITC 23 / 43
Contents
1 Function of Two or More Variables
2 Limits and Continuity
3 Partial Derivatives
4 Chain Rules for Functions of Several Variables
5 Directional Derivatives and Gradient Vectors
6 Tangent Planes and Normal Lines
7 Extrema of Function of Several Variables
8 Lagrange Multipliers
CALCULUS II () ITC 24 / 43
Gradient
Definition 11
For a real-valued function f(x1, x2, . . . , xn), the gradient of f at apoint a, denoted by ∇f(a), is the vector
∇f(a) =
(∂f
∂x1(a),
∂f
∂x2(a), . . . ,
∂f
∂xn(a)
).
Theorem 8
Let D ⊂ Rn be open, and suppose f : D → R is differentiable at a ∈ D.Then the directional derivative of f at a exists for all directions (unitvectors) u and, moreover, we have
Duf(a) = ∇f(a) · u.
CALCULUS II () ITC 25 / 43
The Directional Derivative
Example 12
Find the directional derivative of f(x, y) = 4− 2x2 − y2 at the point
(1, 1) in along unit vector(
12 ,√
32
).
CALCULUS II () ITC 26 / 43
Contents
1 Function of Two or More Variables
2 Limits and Continuity
3 Partial Derivatives
4 Chain Rules for Functions of Several Variables
5 Directional Derivatives and Gradient Vectors
6 Tangent Planes and Normal Lines
7 Extrema of Function of Several Variables
8 Lagrange Multipliers
CALCULUS II () ITC 27 / 43
Tangent Planes and Normal Lines
So far, you have represented surfaces in space primarily by equations ofthe form
z = f(x, y)
In the development to follow, however, it is convenient to use the moregeneral representation F (x, y, z) = 0. For a surface S given byz = f(x, y), you can convert to the general form by defining F as
F (x, y, z) = f(x, y)− z.
Because f(x, y)− z = 0, you can consider S to be the level surface of Fgiven by
F (x, y, z) = 0.
CALCULUS II () ITC 28 / 43
Tangent Planes and Normal Lines
Let S be a surface given byF (x, y, z) = 0 and let P (x0, y0, z0)be a point on S. Let C be a curve Son through P that is defined by thevector-valued function
r(t) = x(t)i + y(t)j + z(t)k
Then, for all t
F (x(t), y(t), z(t)) = 0.
If F is differentiable and x′(t), y′(t)and z′(t) all exist, then we have
F ′(t) = Fx(x, y, z)x′(t)+Fy(x, y, z)y′(t)+Fz(x, y, z)z′(t) = 0
CALCULUS II () ITC 29 / 43
Tangent Planes and Normal Lines
At point P (x0, y0, z0), the equivalent vector form is
∇F (x0, y0, z0).r′(t0) = 0.
Meaning that the gradient at P is orthogonal to the tangent vector ofevery curve on S through P . So, all tangent lines on S lie in a planethat is normal to ∇F (x0, y0, z0) and contains P , as shown in theFigure.
Definition 13
Let F be differentiable at the point P (x0, y0, z0) on the surface givenby F (x, y, z) = 0 such that ∇F (x0, y0, z0) 6= 0.
The plane through P that is normal to ∇F (x0, y0, z0) is called thetangent plane to S at P .
The line through P having the direction of ∇F (x0, y0, z0) is calledthe normal line to S at P .
CALCULUS II () ITC 30 / 43
Tangent Planes and Normal Lines
Theorem 9 (Equation of Tangent plane)
If F is differentiable at (x0, y0, z0) then an equation of the tangentplane to the surface given by F (x, y, z) = 0 at (x0, y0, z0) is
Fx(x0, y0, z0)(x− x0) + Fy(x0, y0, z0)(y − y0) + Fz(x0, y0, z0)(z − z0) = 0
Theorem 10
If the surface given by equation z = f(x, y), then an equation oftangent line to the the surface at the point (x0, y0, z0) is
fx(x0, y0)(x− x0) + fy(x0, y0)(y − y0)− (z − z0) = 0.
CALCULUS II () ITC 31 / 43
Tangent planes and normal lines
Example 14
Show that the tangent plane to the Ellipsoidx2
a2+y2
b2+z2
c2= 1 at the
point (x0, y0, z0) is given byx0x
a2+y0y
b2+z0z
c2= 1.
CALCULUS II () ITC 32 / 43
Contents
1 Function of Two or More Variables
2 Limits and Continuity
3 Partial Derivatives
4 Chain Rules for Functions of Several Variables
5 Directional Derivatives and Gradient Vectors
6 Tangent Planes and Normal Lines
7 Extrema of Function of Several Variables
8 Lagrange Multipliers
CALCULUS II () ITC 33 / 43
Local Extrema
Definition 15
Let D ⊂ Rn, f : D → R be a function and a ∈ D.f(a) is called a local minimum of f if there is an r > 0 such thatf(a) ≤ f(x) for all x ∈ Br(a) ∩D.f(a) is called a local maximum of f if there is an r > 0 suchthat f(a) ≥ f(x) for all x ∈ Br(a) ∩D.f(a) is called a local extremum of f if there f(a) is a localminimum or a local maximum of f .
CALCULUS II () ITC 34 / 43
Differential fo Vector-Valued Functions
Definition 16
Let f : D ⊂ Rn → Rm be a vector-valued function of n variables. Letx = (x1, . . . , xn) denote a point of Rn and f = (f1, . . . , fm). We definethe matrix of partial derivatives of f , denoted Df , to be the
m× n matrix whose (i, j) entry is∂fi∂xj
. That is, Df =
(∂fi∂xj
).
The matrix
Df(a) =
(∂fi∂xj
(a)
)is also called Jacobian matrix of f at a.
CALCULUS II () ITC 35 / 43
Hessian Matrix
Definition 17 (Principal Minor)
Let A = (aij)n×n be a square matrix. The determinant
Ak = det((aij)k×k
), 1 ≤ k ≤ n, is called the kth principal minor of
matrix A.
Definition 18 (The Hessian of a Function)
Let D ⊂ Rn and f : D → R be a function of class C2. The Hessian off , denoted by Hf , is the matrix whose (i, j) entry is fxixj .. That is,
Hf (a) =
(∂2f
∂xj∂xi(a)
)n×n
.
We denote Hk = det(
∂2f∂xj∂xi
)k×k
the kth principal minor of Hf for
k = 1, 2, . . . , n.
CALCULUS II () ITC 36 / 43
Local Extrema
Theorem 19
Let f have continuous second partial derivatives on an open region
containing a = (x0, y0) for which
{fx(a) = 0
fy(a) = 0To test for relative extrema of f , consider the quality
H1 = fxx(a) , H2 = fxx(a)fyy(a)− (fxy(a))2.
1 If H2 > 0 and H1 > 0, then f has a local minimum at a = (x0, y0).
2 If H2 > 0 and H1 < 0, then f has a local maximum at a = (x0, y0).
3 If H2 < 0, then a = (x0, y0) is a saddle point.
4 If H2 = 0, the test is inconclusive.
CALCULUS II () ITC 37 / 43
Local Extrema
Example 20
Determine the local extrema/saddle points of the function
f(x, y) = 3x2 + 2y2 − 6x− 4y + 16.
CALCULUS II () ITC 38 / 43
Local Extrema
Theorem 21
Let f have continuous second partial derivatives on an open region
containing a = (x0, y0, z0) for which
fx(a) = 0
fy(a) = 0
fz(a) = 0To test for relative extrema of f , consider the quality
H1 = fxx(a, b), H2 =
∣∣∣∣ fxx(a) fxy(a)fyx(a) fyy(a)
∣∣∣∣ , H3 =
∣∣∣∣∣∣fxx(a) fxy(a) fxz(a)fyx(a) fyy(a) fyz(a)fyz(a) fzy(a) fzz(a)
∣∣∣∣∣∣1 If H1 > 0, H2 > 0 and H3 > 0, then f has a local minimum at a.
2 If H1 < 0, H2 > 0 and H3 < 0, then f has a local maximum at a.
3 If neither case 1 nor 2 holds, then a is a saddle point.
4 If H3 = 0, the test is inconclusive.CALCULUS II () ITC 39 / 43
Local Extrema
Example 22
Determine the local extrema/saddle points of the function
f(x, y, z) = x2 − xy + z2 − 2xz + 6z.
CALCULUS II () ITC 40 / 43
Contents
1 Function of Two or More Variables
2 Limits and Continuity
3 Partial Derivatives
4 Chain Rules for Functions of Several Variables
5 Directional Derivatives and Gradient Vectors
6 Tangent Planes and Normal Lines
7 Extrema of Function of Several Variables
8 Lagrange Multipliers
CALCULUS II () ITC 41 / 43
Lagrange Multipliers
Theorem 23
Let f, g1, . . . , gk : D → R be C1 functions where D ⊂ Rn is open andk < n. Suppose there is an a ∈ D such that
det
(∂gi∂xj
(a)
)6= 0.
If f(a) is local extremum of f subject to the constraints gi(x) = ci fori = 1, 2, . . . , k, then there exist scalars λ1, . . . , λk (called LagrangeMultipliers) such that{
∇f(a) +∑k
i=1 λi∇gi(a) = 0
g(x) = ci i = 1, 2, ..., k.
CALCULUS II () ITC 42 / 43
Local Extrema
Example 24
Use Lagrange Multipliers, determine the local extrema of functionf(x, y) = 5x+ 2y subject to constraint 5x2 + 2y2 = 14.
CALCULUS II () ITC 43 / 43