Partial Derivative

Download as pdf or txt
Download as pdf or txt
You are on page 1of 71

Geometrical Interpretation of

CMVT:
A geometrical interpretation may be given for the
y t=c
theorem as follows. t=b

Let a plane curve be represented


Parametrically by the equations
x=g(t) and y=f(t), where a<t<b. t=a
0 x
The slope of the curve for a given t is
(1)
(2)The constant

is the slope of the straight line joining the points


on the curve corresponding to t=a and t=b,
respectively.
The theorem says that two slopes (1) and (2)
are equal for at least one value of t between a
and b.
Q1. in CMVT, if , show
that c is independent of x.

Solution. Since both f(x) and g(x) are continuous


on [a, b] and f’(x) and g’(x) exist in (a, b) and
g’(x)≠ 0 in (a, b), therefore we can apply CMVT.
According to CMVT, we get
=>a+b = 2c
=>c= (a+b)/2
Therefore c is independent of x.
Partial Differentiation:
Consider u= f(x,y). Here x and y are independent
variables and u is dependent variable. In
practical many situation arises where a variable
depends on more than one independent
variable.
The differentiation u w.r.t. x treating y as a
constant is called the partial derivative of u w.r.t. x
and denoted as

Analytically, if exists.

Similarly, if exists.
Homogenous Function:
Definition: A function f(x,y) is said to be
homogenous of degree n if it can be express in
the form
or

For example, Let z = x2+xy+y2


= x2(1 + y/x + y2/x2)
=x2.Ф(y/x)
Or, z = y2(x2/y2 + x/y + 1)
= y2.Ф(x/y)
In either cases z is homogenous of degree 2.
For more than 2 variables, it can be expressed as,
f(x,y,z)= xn.Ф1(y/x, z/x)
Or yn.Ф2(x/y, z/y)
Or zn. Ф3(x/z, y/z)

EULER’S THEOREM: If f(x,y) be homogenous


function of x and y of degree n, then
Proof: Since f(x,y) is homogenous function of degree n.

Let f(x,y) =

= xn.Ф(t), where t = y/x.


So,

And

Therefore,
Implicit Function:
Definition: An implicit function x and y is an
equation of the form f(x,y)=0 which is not
necessarily be solved for one of the variables in
terms of others explicitly that is say x in terms of y.
For example, X3+y3-3axy=0 is an implicit function of
x and y.
In contract, when y is directly expressed in terms of
x, say y=f(x), then y is called as explicit function.
Partial Differentiation of an implicit
function:
Let f(x,y,z)=0 represents an equation of an implicit
function of three variables x, y and z.
(a) Determination of :
Assuming z=f(x,y) then differentiating partially
f(x,y,z)=0 with respect to x and y respectively, we
get.
and
Provided
(b) Determination of :

Assuming x=f(y,z) then by differentiating


f(x,y,z) partially with respect to y and z
respectively , we get .

and

Provided
(c) Determination of :

Assuming y=f(x,z) then by differentiating


partially f(x,y,z)=0 w.r.t. x and z respectively, we
get.

and

Provided
Example 1. Find

Solution: Let, f(x, y,z)= ………………………(1)


Differentiating (1) partially w.r.t. x, we get

Differentiating (1) partially w.r.t. z, we get

Differentiating (1) partially w.r.t. y, we get


Total Derivative:
Let u = f(x,y) be a function of two independent variables x and y, which has
continuous partial derivatives w.r.t. x and y. Let be differentiable
function of t, then the composite function is differentiable function
of t and

Here , is called total derivative of u. Equation (1) is called chain rule for two
independent variables.
In general form of this rule for functions of several independent variables like
can be written as,
Corollary 1. If u=f(x,y) is a function of two
independent variables x and y, y is a function of
x, y=g(x), then chain rule (1) can be written as
follows,

Corollary 2. If u=f(x,y) is a function of two


independent variables x and y, and if
x=Ф1(t1, t2), y=Ф2(t1, t2) then the composite
function u=f[Ф1(t1, t2), Ф2(t1, t2)] is a function of
t1 and t2.
If x, y and f are differentiable function, then we get,

Generalizing, chain rule (4) for

We get,

………………………………
Example.1 Find as total derivative for
u= 3x+xy-y2 where, x=cos3t, y=sin3t.
Solution: Here u= 3x+xy-y2

Also, x= cos3t
y= sin3t
= (3+y).(-3sin3t) + (x-2y).(3cos3t)
= (3+sin3t)(-3sin3t) + (cos3t-2sin3t)(3cos3t)
= -9sin3t-3sin23t + 3cos23t - 6sin3t.cos3t
= -9sin3t-3sin23t+3(1-sin23t) – 3(2sin3t.cos3t)
= -9sin3t – 6sin23t + 3 - 3sin6t
= -9sin3t + 3cos6t – 3sin6t.
Example 2. If z=f(x,y), prove that if x=eu+e -v,
y=e-u- ev then
Solution: Here we can write the chain rule as z

x y

u
v
Fig.1.2

Now (1)-(2), we get,


Hence proved.
Laplace Operator or Laplacian:
The Laplace operator is a second order differential
operator in the n-dimensional Euclidean space,
defined as the divergence of the gradient .Thus
if f is a twice-differentiable real-valued function,
then the Laplacian of f is defined by

Equivalently, the Laplacian of f is the sum of all the


unmixed second partial derivatives in the Cartesian
coordinates xi
… … … … … (2)
Notes:
(1) Let x1, x2,… … xn be a system of Cartesian
coordinates on a n-dimensional Euclidean space,
and let i1, i2,… … in be the corresponding basis of
unit vectors.
The divergence of a continuously differentiable
vector field f = g1 i1 + g2 i2 + … …+ gn in is defined
to be the scalar-valued function:
(2) The gradient (or gradient vector field) of a
scalar function is denoted where
(the nabla symbol) denotes the vector
differential operator, del. The notation grad(f) is
also used for the gradient. The gradient of f is
defined to be the vector field whose components
are the partial derivatives of f. That is:
…. …. … … … (4)

Here the gradient is written as a row vector,


but it is often taken to be a column vector.
(3) Vector differential operator called nabla
symbol.

(4) Laplace operator or Laplacian called tebla


symbol.
Applications of Partial Differentiation:
1. Jacobians and their Applications.
2. Taylor’s Theorem for functions of two
variables.
3. Error and Approximation.
4. Maxima and Minima.
5. Saddle points.
6. Lagrange’s method of undetermined
multipliers.
1. Jacobians and their Applications:
Defination1: (2-D case): If u = f(x, y) and v = g(x, y)
be differentiable functions of two independent
variables x and y, then the jaconian of u and v w.r.t.
x and y is denoted by and is defined
as follows,
Defination2:(3-D Case): If u = f(x, y, z), v = (x, y, z)
and w = (x, y, z) be differentiable functions of three
independent variables x, y and z, then the Jacobian
of u, v, w w.r.t. x, y, z is denoted by
and is defined as follows,
Thus in general (n-D Case) if
u1 = f1(x1, x2, … …, xn)
u2 = f2(x1, x2, … …, xn),
… … … … … … … … … …,
un = fn(x1, x2, … …, xn),
be differentiable functions of n-independent
variables x1, x2, … …, xn then
Example1: Find when x = r.cosθ, y = r.sinθ.
Solution: Here x = r.cosθ and y = r.sinθ.
Now we have,

= r.cos2θ+r.sin2θ
=r(cos2θ+sin2θ)
=r.
Example2: If x= r.sinθ.cosφ, y = r.sinθ.sinφ,
z = r.cosθ, find .
Solution: Here given that,
x= r.sinθ.cosφ, y = r.sinθ.sinφ, z = r.cosθ.
Now we have,

= r2sinθ.
Properties of Jacobians:
1). If J is the Jacobian of u, v w.r.t. x, y and J’ is the
Jacobian of x, y w.r.t. u, v, then JJ’ = 1 that is,

Proof: Let, u = f(x, y) and v = g(x, y) ………………..(1)


Solving (1) for x and y, we get
x = φ(u, v) and y = ψ(u, v) ……………….(2)
Differentiating (1) partially w.r.t. u and v gives,
Now,

Using (3) to (6), we get

Hence proved.
In general, the property can be written as

If u1 = f1(x1, x2, … … xn), u2 = f2(x1, x2, … … xn), … … …


… un = fn(x1, x2, … … xn).

Note: J’ is known as inverse transformation


Jacobian of J and vice versa.
2). Chain rule of Jacobian (Jacobians of
composite functions): If u, v are functions of r, s
and r, s are functions of x, y then

Proof: Here, u and v are composite functions of x


and y.
Differentiating u and v partially w.r.t. x and y, we get
Similarly,

Now,

using (1) to (4)


Hence proved.
In general, the property can be written as

If u1, u2, … …, un are functions of r1, r2, … …, rn are


functions of x1, x2, … …, xn.
3). If functions u, v, w of three independent
variables x, y, z are functionally related or
dependent iff
4). If the functions u1, u2, …, un of the variables x1,
x2, …, xn be defined by the relations u1 = f1(x1), u2 =
f2(x1,x2), u3 = f3(x1, x2, x3) and so on un = fn(x1, x2, …
…, xn), then

Proof: Given, u1 = f1(x1).

u2 = f2(x1,x2).
and so on.
Now,

=
Example: If x = u(1-v), y = uv, evaluate
and . Also verify jj’=1.

Solution: Here, x = u(1-v) and y = uv.

Now, x=u(1-v), y=uv solving them we get,


u = x+y and v=y/(x+y)
Hence verified.
2. Taylor’s Theorem for functions of two
variables.
Taylor’s series is useful to approximate a function in
the neighborhood of a point. It is widely used in
many disciplines like engineering, science, etc. to
establish continuous models. It is also the mainstay
of many numerical techniques.
Statement: If f(x, y) and all its partial derivatives
upto the nth order are finite and continuous for all
points (x, y), where a ≤ x ≤ a+h, b ≤ y ≤ b+k, then
Alternate forms of Taylor’s series:
1.) In suffix notation, we get

2.) Putting x = a and y = b in (1), we have

3.) Putting a+h = x and b+k = y in (2), we get


Which expands f(x,y) in powers of (x-a) and (y-b)
and is known as Taylor’s series or Taylor’s
expansion or Taylor’s series expansion of f(x,y)
about the point (a,b).
Example: Use Taylor’s series to expand sinx.cosy in
powers of (x-π/3) and (y-π/4).
Solution: Here, x-a = x - π/3. and y-b =y - π/4
=> a = π/3 => b = π/4
Let, f(x, y) = sinx.cosy => f(π/3, π/4) = √3/2√2.
fx(x, y) = cosx.cosy => fx(π/3, π/4) = 1/2√2.
fy(x, y) = -sinx.siny => fy(π/3, π/4) = -√3/2√2.
fxx (x, y) = -sinx.cosy => fxx(π/3, π/4) = -√3/2√2.
fxy (x, y) = -cosx.siny => fxy (π/3, π/4) = - 1/2√2.
fyy (x, y) = -sinx.cosy => fyy(π/3, π/4) = -√3/2√2.
... … … ……… …………… ………
Now, the Taylor’s series in powers of (x - π/3) and (y
- π/4) is given by
3.Errors and Approximations:
Let u be some physical quantity of interest depends
on two measurable quantities x and y by the
relation u = f(x, y). It is obvious that any change ∆x
in x and ∆y in y will propagate the overall error ∆u in
u and is given by ∆u = ∆f = f(x+∆x, y+∆y) – f(x,y).
Now , expanding f(x+∆x, y+∆y) using Taylor’s series,
we have
For small values of ∆x and ∆y, we have
approximately

This formula is useful for finding overall error in


f(x,y) if the small errors in x and y occurs. It can
easily be generalized for functions more than two
variables approximately as
Where f is the function of x, y, z, t, …. …
As we know that the error occurs in ∆x and ∆y may
be positive or negative, therefore, the maximum
error in f if given by

And in general,

If f is a function of x, y, z, t, … … …
Note: (i) df in (1) and (2) above known as
approximate error in f while dx, dy, … … are known
as actual (Absolute) errors in x, y, …., respectively.

(ii) dx/x, dy/y, … …, are known as relative errors


(proportinal error) in x, y,…. …. ,respectively.

(iii) (dx/x).100 and (dy/y).100 are known as


percentage errors in x and y respectively.
Example1: Using differentials find the
approximate value of √8(65)1/3.
Solution: Let, f(x, y) = x1/2y1/3.
Choosing x = 9, y = 64 with decrease 1 in x and
increase 1 in y.
Now we have,

Here ∆x = -1 and ∆y = 1, putting these values we


have,
Thus, change in x and y, decrease the value of the
function by .

Therefore,
Example2: The two resistors x and y are connected
in parallel so that the total resistance R is given by
xy/(x+y). If x and y are measured to be 200Ω and
300Ω respectively, with the increase of 1.5Ω in x
and decrease of 4Ω in y, find the change in R.
Solution: According to the given question, we get
x=200Ω

y=300Ω
Here,

Similarly,

Now ,

Here, x = 200Ω, y = 300Ω, ∆x = 1.5Ω, ∆y = -4Ω


putting these values we get,

Therefore the decrease in R is 0.1Ω.


4.Maxima and minima of a function of
two variables:
Definition: Consider a function f(x, y) of two
variables x and y and a point (a, b) and the
neighborhood points (a+h, b+k). The value f(a, b) is
said to be
(i) maximum if f(a, b) > f(a+h, b+k) and
(ii) minimum if f(a, b) < f(a+h, b+k).
Extremum point: A point (a, b) which either a
maximum or a minimum is called an extremum
point.
Theorem1: The necessary conditions for f(a, b) to
be an extremum value of “f” are that
fx(a, b) = 0 and fy(a, b) = 0
The point (a, b) is called a stationary point and the
value f(a, b) is called the stationary value.
Theorem2: The sufficient conditions for the
stationary value f(a, b) to be an extremum value as
follows:
Let fxx(a, b) = A, fxy(a, b) = B, fyy(a, b) = C , then
1. If AC-B2 > 0 and A (or C) < 0, then f(a, b) is a
maximum value.
2. If AC-B2 > 0 and A (or C) > 0, then f(a, b) is a
minimum value.
3. If AC-B2 < 0, then f(a, b) is neither a maximum
nor a minimum value. But it a saddle point.
4. If AC-B2= 0 then no conclusion can be drown
about maximum or minimum and further
investigation is needed.
Note: Every extremum value is a stationary value
but converse is not true.
Stationary point or Critical point: A point (a, b) is
said to be stationary point of “f” if
(i) at (a, b) both fx and fy are zero, or
(ii) it is a boundary point of the domain.
Working rules for finding extremum values of a
function f(x, y):
1. Find .
2. Find points (x, y) for which .
This gives the stationary points.
3. At each stationary point, evaluate values of
Saddle point or minimax point: In mathematics, a
saddle point is a point in the domain of a function
of two variables which is a stationary point but not
a local extremum. At such a point, in general, the
surface resembles a saddle that curves up in one
direction, and curves down in a different direction
(like a horse saddle or a mountain pass). In terms
of contour lines, a saddle point can be recognized,
in general, by a contour that appears to intersect
itself. For example, two hills separated by a high
pass will show up a saddle point, at the top of the
pass, like a figure-eight contour line.
Example1: Examine the function f(x, y) =
2x4+y2-x2-2y for maxima and minima.
Solution: Here, f(x, y) = 2x4+y2-x2-2y.
fx = 8x3-2x, and fy = 2y-2.
For stationary points, we have fx =0, fy = 0.
Therefore, 8x3-2x = 0 and 2y-2 = 0.
Which implies points (0, 1), (1/2, 1), (-1/2, 1)
Now A = fxx = 24x2-2, B = fxy = 0, C = fyy = 2.
Case1: At the point (0, 1), AC-B2 = (-2)(2)= -4<0.
Therefore (0. 1) is a saddle point.
Case2: At the point (1/2, 1), AC – B2 = (4)(2)-0=8>0
and A=4>0.
Therefore (1/2, 1) is a minimum point and the
minimum value of the function is -9/8.
Case3: At the point (-1/2, 1), AC-B2 = (4)(2)-0= 8>0
and A=4>0.
Therefore (-1/2, 1) is a minimum point and the
minimum value of the function is -9/8.
6.Lagrange’s Method of Undetermined
Multipliers:
This method is more useful for finding the
stationary points of the function subject to the
given constraint with the introduction of unknown
constant(s) λ, known as Lagrange Multiplier.
The main drawback of this method is that it
gives no information regarding the nature of the
stationary points. This fact is to be determined
from the physical or geometrical consideration of
the problem and which in some cases may be
difficult to comprehend.
Let, u=f(x, y, z), be a function of three
independent variables x, y, z which are connected
by the relation g(x, y, z) = 0. [ g(x, y, z)=0 is called
constraint.].
The Lagrange’s function is given by
F(x, y, z) = f(x, y, z)+λg(x, y, z)……………………(1)
Where λ is the arbitrary constant.
Then dF = 0 implies that,
fx +λgx=0, fy +λgy=0, fz +λgz=0 ……………………(2)

This procedure is only for one constraint.


If there are 3 constraints, then the Lagrange’s function is
given by,
F(x, y, z) = f(x, y, z)+λ1g1(x, y, z)+λ2 g2(x, y, z)+λ3g3(x, y, z)
………………………………………………...(3)
Where, g1(x, y, z)=0, g2(x, y, z)=0,and gn(x, y, z) =0 are
constraints. Then dF=0 implies that
fx+λ1g1x +λ2g2x + λ3g3x = 0 ………………………………..(4)
fy+λ1g1y +λ2g2y + λ3g3y = 0 ……………………………….(5)
fz+λ1g1z +λ2g2z + λ3g3z = 0 …………………………………(6)
Then solve the equations (4), (5) and (6) to find out the
values of λ1, λ2 and λ3. Put the values of λ1, λ2 and λ3 in
the equations (4), (5) and (6) to get the values of x, y
and z. Finally put the values of x, y and z in the given
function.
Working rule for the Lagrange’s Method
of undetermined multipliers:
Step1: Construct the Lagrange’s function[F(x, y, z)].
Step2: Find the differential of Lagrange’s function
(dF).
Step3: Consider, dF = 0 for stationary values. Here
we will get equations (4), (5) and (6).
Step4: Solve the equations (4), (5) and (6) to get
the values of λ.
Step5: Put the values of λ in equations (4), (5) and
(6) to get the values of x, y and z.
Step6: Finally put the values of x, y and z in the
given function.
Example1: Find the minimum value of x2+y2+z2,
given that ax+by+cz = p. Using Lagrange’s method.

Solution: Let u = x2+y2+z2 and


g(x,y,z) = ax+by+cz-p=0. ………………………..(i)
The Lagrange’s function is given by
F(x, y, z) = (x2+y2+z2)+λ(ax+by+cz-p)
For stationary values, dF=0 gives,
2x+λa=0 ……………………………………………….(ii)
2y+λb=0 ………………………………………………(iii)
2z+λc=0 ……………………………………………….(iv)
Now , (ii).x+(iii).y+(iv).z implies that
2(x2+y2+z2)+λ(ax+by+cz)=0
=> 2u+λp=0, using (i)
=>λ=- 2u/p
Putting this value in (ii), (iii) and (iv), we get
x = au/p, y = bu/p, z = cu/p.
Substituting these values of x, y, and z in the given
function, we have
u= u2(a2+b2+c2)/p2 => u = p2/(a2+b2+c2), neglecting
the trivial case u = 0.
As we know that the perpendicular distance is
always minimum. Therefore, the length of the
perpendicular from origin o on the plane
ax+by+cz=p is given by . Hence
u(min) = p2/(a2+b2+c2).
O(0,0,0)

(x, y, z) ax+by+cz=p
Example2: Find the maximum value of x2y3z4 given
that 2x+3y+4z=a. using Lagrange’s method.
Solution: The Lagrange’s function is given by
F(x, y, z) = (x2y3z4)+λ(2x+3y+4z-a).
For sationary values, dF = 0 gives
2xy3z4+2λ = 0. ……………………………………………..(i)
3x2y2z4+3λ = 0. …………………………………………….(ii)
4x2y3z3+4λ = 0. ……………………………………………..(iii)
Now (i).x+(ii).y+(iii).z, implies that
λ(2x+3y+4z) = -9x2y3z4 => λ = -9x2y3z4 /a, using
2x+3y+4z=a.
Substituting this value of λ in (i), (ii) and (iii), we get,
x = y = z = a/9.
Thus maximum value of x2y3z4 is .

You might also like