Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
MATHEMATICS-IV 10MAT41
SYLLABUS
ENGINEERING MATHEMATICS – IV
SUBJECT CODE: 10 MAT 41
m
PART-A
o
Unit-I: NUMERICAL METHODS - 1
Numerical solution of ordinary differential equations of first order and first
degree; Picard‟s method, Taylor‟s series method, modified Euler‟s method,
s.c
Runge-kutta method of fourth-order. Milne‟s and Adams - Bashforth
predictor and corrector methods (No derivations of formulae).
PART-B
m
differential equation leading to Bessel function of first kind. Orthogonal
property of Bessel functions. Series solution of Legendre‟s differential
equation leading to Legendre polynomials, Rodrigue‟s formula.
o
Unit-VI: PROBABILITY THEORY – 1
s.c
Probability of an event, empherical and axiomatic definition, probability
associated with set theory, addition law, conditional probability,
multiplication law, Baye‟s theorem.
Index
m
Unit 2: Numerical Methods – II…………………....35-44
o
s.c
Unit 3: Complex Variables – I……………………...45-52
o m
s.c
c
vtu
w.
ww
Unit I
Numerical Methods I
m
The fundamental laws of physics, mechanics, electricity and thermodynamics are
usually based on empirical observations that explain variations in physical properties
and states of the systems. Rather than describing the state of physical system directly,
the laws are usually couched in terms of spatial and temporal changes. The following
o
table gives a few examples of such fundamental laws that are written in terms of the
rate of change of variables (t = time and x= position)
s.c
Physical Law Mathematical Variables and
Expression Parameters
The above laws define mechanism of change. When combined with continuity laws for
energy, mass or momentum, differential equation arises. The mathematical expression
w.
But, many ordinary differential equations arising in real-world applications and having lot
of practical significance cannot be solved exactly using the classical analytical methods.
These ode can be analyized qualitatively. However, qualitative analysis may not be able
to give accurate answers. A numerical method can be used to get an accurate
approximate solution to a differential equation. There are many programs and packages
available for solving these differential equations. With today's computer, an accurate
solution can be obtained rapidly. In this chapter we focus on basic numerical methods
for solving initial value problems.
Analytical methods, when available, generally enable to find the value of y for all
values of x. Numerical methods, on the other hand, lead to the values of y
corresponding only to some finite set of values of x. That is the solution is obtained as a
table of values, rather than as continuous function. Moreover, analytical solution, if it
can be found, is exact, whereas a numerical solution inevitably involves an error which
should be small but may, if it is not controlled, swamp the true solution. Therefore we
must be concerned with two aspects of numerical solutions of ODEs: both the method
itself and its accuracy.
m
The general form of first order differential equation, in implicit form, is
dy
F ( x, y, y / ) 0 and in the explicit form is f ( x, y ) . An Initial Value Problem (IVP)
dx
consists of a differential equation and a condition which the solution much satisfies (or
o
several conditions referring to the same value of x if the differential equation is of higher
order). In this chapter we shall consider IVPs of the form
dy
f ( x, y), y( x0 ) y0 . (1)
s.c
dx
Assuming f to be such that the problem has a unique solution in some interval
containing x0, we shall discuss the methods for computing numerical values of the
solution. These methods are step-by-step methods. That is, we start from y0 y ( x0 )
and proceed stepwise. In the first step, we compute an approximate value y1 of the
solution y of (1) at x = x1 = x0 + h. In the second step we compute an approximate value
c
y2 of the solution y at x = x2 = x0 + 2h, etc. Here h is fixed number for example 0.1 or
0.001 or 0.5 depends on the requirement of the problem. In each step the computations
vtu
are done by the same formula.
4. Milne’s Method
5. Adams – Bashforth Method
dy
Consider an IVP f ( x, y), y ( x0 ) y0 . Let us approximate the exact solution y(x) to
dx
a power series in ( x x0 ) using Taylor’s series. The Taylor’s series expansion of y(x)
about the point x x0 is
( x x0 ) / ( x x0 ) 2 // ( x x0 ) 3 /// ( x x0 ) 4 / V
y ( x) y ( x0 ) y ( x0 ) y ( x0 ) y ( x0 ) y ( x0 )
1! 2! 3! 4!
(1)
dy
From the differential equation, we have y / ( x) f ( x, y ) . Differentiating this
dx
successively, we can get y // ( x), y /// ( x), y / V ( x), etc. Putting x x0 and y y 0 , the values of
y // ( x0 ), y /// ( x0 ), y / V ( x0 ), etc. can be obtained. Hence the Taylor’s series (1) gives the
values of y for every value of x for which (1) converges.
m
On finding the value of y1 for x x1 from (1), y // ( x), y /// ( x), y / V ( x), etc. can be evaluated at
x x1 from the differential equation
Problems:
o
1. Find by Taylor’s series method the value of y at x = 0.1 and 0.2 five places of
dy
s.c
decimals for the IVP x 2 y 1, y(0) 1 .
dx
Soln:
Given x0 0, y0 1 and f ( x, y) x2 y 1
y / ( x) x 2 y 1 y / (0) 1
dx
y // ( x) 2 xy x2 y/ y // (0) 0
Putting the values of y(0), y / (0), y // (0), y /// (0), y iv (0) in (1), we get
x2 x3 x4
y( x) 1 x( 1) (0) (2) ( 6)
2 6 24
x3 x4
y ( x) 1 x
3 4
Hence y(0.1) = 0.90033 and y(0.2) = 0.80227.
2. Employ Taylor’s series method to obtain approximate value of y at x = 0.1 and 0.2 for
dy
the differential equation 2 y 3e x , y(0) 0. Compare the numerical solution
dx
m
obtained with the exact solution.
Soln:
Given x0 0, y 0 0 and f ( x, y) 2 y 3e x
o
Taylor’s series expansion about the point x 0 is
s.c
2 3 4
x // x /// x /V
y ( x) y(0) xy / (0) y (0) y (0) y (0) (2)
2 6 24
Putting the values of y(0), y / (0), y // (0), y /// (0), y iv (0) in (2), we get
9 2 21 3 45 4
y ( x) 0 3 x x x x
2 6 24
9 2 7 3 15 4
ww
3x x x x
2 2 8
Hence,
y(0.1) = 3(0.1)+4.5(0.1)2+3.5(0.1)3+1.875(0.1)4
= 0.3486875
and
y(0.2) = 3(0.2)+4.5(0.2)2+3.5(0.2)3+1.875(0.2)4
= 0.8110.
Exact Soln:
dy
The given differential equation can be written as 2y 3e x which is Leibnitz’s linear
dx
m
differential equation.
2d x 2x
Its I.F. is I.F = e e
o
Therefore the general solution is,
2x
3e x (e 2x x
s.c
ye )dx c 3e c
y 3e x ce 2 x (3)
0. 75x10 -5
ww
dy
3. Using Taylor’s series method solve x2 y, y (0) 1at 0.1 ≤ x ≤ 0.4. Compare
dx
the values with the exact solution.
Soln:
Given x0 0, y0 1 and f ( x, y) x2 y
m
Taylor’s series expansion about the point x 0 is
2 3 4
x // x /// x /V
y ( x) y(0) xy / (0) y (0) y (0) y (0) (4)
2 6 24
o
It is given that y (0) 0
s.c
dy
y / ( x) x 2 y y / (0) (0) 2 1 1
dx
y // ( x) 2x y/ c y // (0) 0 y / (0) 0 ( 1) 1
Putting the values of y(0), y / (0), y // (0), y /// (0), y iv (0) in (4), we get
x2 x3 x4
w.
y ( x) 1 x
2 6 24
Hence,
y (0.1) 1 (0.1)
2 6 24
= 0.9051625
= 0.8212667.
= 0.7491625.
m
= 0.6896.
o
Exact Soln:
dy
s.c
The given differential equation can be written as y x 2 a linear differential equation.
dx
dx
Its I.F. is I.F = e ex
ye x
c
e x ( x 2 )dx c (x2 2 x 2)e x c
vtu
y (x2 2 x 2) ce x
(5)
y(0.1) = 0.9051625,
ww
y(0.2) = 0.8212692,
y(0.3) = 0.7491817 and
y(0.4) = 0.6896799
m
0.1 0.9051625 0.9051625, 0
o
0.3 0.7491625 0.7491817 0. 192x10 -4
s.c
0.4 0.6896 0.6896799 0. 799x10 -4
Euler‟s Method :
h
yn 1 yn f ( xn , y n ) f ( xn 1 , y nE 1 ) , n 0,1,2,3, (2)
2
ww
Remark:
1. The formulae (1) and (2) are also known as Euler’s Predictor – Corrector formula.
2. When Modified Euler’s method is applied to find the solution at a give point, we
first find the solution at that point by using Euler’s method and the same will be
used in the calculation of Modified Euler’s method. Also Modified Euler’s method
has to be applied repeatedly until the solution is stationary.
Problems:
dy y2
1. Solve , y (0) 1 by Euler’s method by choosing h = 0.1 and h = 0.05. Also
dx 1 x
solve the same problem by modified Euler’s method by choosing h = 0.05. Compare the
numerical solution with analytical solution.
Soln:
m
Analytical solution is : -
dy dx 1
log(1 x) c
o
y2 1 x y
s.c
Using the condition y(0) = 1, we get c = 1.
1
Hence the analytical solution is y y(0.2) = 0.84579
1 log(1 x)
y n2
Now by Euler’s method, we have y n 1 yn 0 .1
1 xn
y1 y (0.1) 1 0.1
1 0
c
(1) 2
0.9
vtu
(0.9) 2
y2 y (0.2) 0.9 0.1 0.82636
1 0.1
y n2
Now taking h = 0.05, Euler’s method is y n
w.
1 yn 0.05
1 xn
(1) 2
y1 y (0.05) 1.0 0.05 0.95
1 0
ww
(0.95) 2
y2 y (0.1) 0.95 0.05 0.90702
1 0.05
(0.90702) 2
y3 y (0.15) 0.90702 0.05 0.86963
1 0.1
(0.86963) 2
y4 y (0.2) 0.86963 0.05 0.83675
1 0.15
Error = 0.84579 – 0.83675 = 0.00904
Note that when h = 0.1, Error was 0.01943, which is more.
y n2
Euler’s Formula is y n 1 yn 0.05 , n = 0,1,2 and 3
1 xn
y n2 y nE 1
Modified Euler Formula is y n 1 yn 0.025 , n = 0,1,2 and 3
1 xn 1 xn 1
m
Stage – I: Finding y1 = y(0.05)
o
(1) 2
y1E y (0.05) 1.0 0.05 0.95
1 0
s.c
From Modified Euler’s formula, we have
(1) 2 (0.95) 2
y1(1) y (0.05) 1.0 0.025 0.95351
1 0 1 0.05
(1) 2 (0.95351) 2
y1( 2 ) y (0.05) 1.0 0.025c 0.95335
1 0 1 0.05
(1) 2 (0.95335) 2
vtu
y1(3) y (0.05) 1.0 0.025 0.95336
1 0 1 0.05
(1) 2 (0.95336) 2
y1( 4 ) y (0.05) 1.0 0.025 0.95336
1 0 1 0.05
Hence y1 y(0.05) 0.95336
(0.95336) 2
y 2E y (0.1) 0.95336 0.05 0.91008
ww
1 0.05
From Modified Euler’s formula, we have
(0.95336) 2 (0.91008) 2
y 2(1) y (0.1) 0.95336 0.025 0.91286
1 0.05 1 0.1
(0.95336) 2 (0.91286) 2
y 2( 2 ) y (0.1) 0.95336 0.025 0.91278
1 0.05 1 0.1
(0.95336) 2 (0.91278) 2
y 2( 2 ) y (0.1) 0.95336 0.025 0.91278
1 0.05 1 0.1
m
(0.91278) 2
y 3E y (0.15) 0.91278 0.05 0.87491
1 0.1
From Modified Euler’s formula (for n = 2), we have
o
(0.91278) 2 (0.87491) 2
y 3(1) y (0.15) 0.91278 0.025 0.87720
1 0.1 1 0.15
s.c
( 2) (0.91278) 2 (0.87720) 2
y 3 y (0.15) 0.91278 0.025 0.87712
1 0.1 1 0.15
(0.91278) 2 (0.87712) 2
y 3(3) y (0.15) 0.91278 0.025 0.87712
Hence y3 y (0.15)
c
0.87712
1 0.1 1 0.15
vtu
Stage – IV: Finding y4 = y(0.2)
Error= 0.84579-0.84550=0.00029
Recall that the error from Euler’s method is 0.00904
2. Solve the following IVP by Euler’s modified method at 0.2 x 0.8 with h = 0.2:
dy
log 10 ( x y), y(0) 2 .
dx
Soln:
Given Data is: x 0 0, y 0 2, h 0.2 and f ( x, y ) log 10 ( x y )
To Find: y1 y( x1 ) y(0.2), y 2 y ( x 2 ) y (0.4), y 3 y ( x 3 ) y (0.6)
m
& y 4 y( x 4 ) y(0.8)
Stage – I: Finding y1 = y(0.2)
o
y1E y(0.2) 2.0 0.2 log 10 0 2 2.0602
s.c
Now from Modified Euler’s formula (for n = 0), we have
y1( 2) y(0.2) c
2.0 0.1 log 10 (0 2) log 10 (0.2 2.0655 2.0656
vtu
y1(3) y(0.2) 2.0 0.1 log 10 (0 2) log 10 (0.2 2.0656 2.0656
y2(1) y(0.4) 2.0656 0.1 log 10 (0.2 2.0656) log 10 (0.4 2.1366 2.1415
y2( 2) y(0.4) 2.0656 0.1 log 10 (0.2 2.0656) log 10 (0.4 2.1415 2.1416
y2(3) y(0.4) 2.0656 0.1 log 10 (0.2 2.0656) log 10 (0.4 2.1416 2.1416
m
y3(1) y(0.6) 2.1416 0.1 log 10 (0.4 2.1416) log 10 (0.6 2.2226 2.2272
y3( 2) y(0.6) 2.1416 0.1 log 10 (0.4 2.1416) log 10 (0.6 2.2272 2.2272
o
Hence y3 y (0.6) 2.2272
s.c
Stage – IV: Finding y4 = y(0.8)
y4( 2) y(0.8) 2.2272 0.1 log 10 (0.6 2.2272) log 10 (0.8 2.3217 2.3217
xn yn
m
0.4 y2 y(0.4) 2.1416
0.6
o
y3 y (0.6) 2.2272
s.c
dy
3.Using modified Euler’s method solve the IVP sin x cos y, y(2.5) 0 at
dx
x=3.5 in two steps, modifying the solution thrice at each stages. Here x is in radians.
Soln:
Given x0 2.5, y 0 0, h
c 0.5 and f ( x, y ) sin x cos y
vtu
To Find: y1 y( x1 ) y(3.0) and y2 y ( x2 ) y(3.5)
y1(1) y (3.0) 0.0 0.25 sin( 2.5) cos(0) sin( 3.0) cos(0.7992) 0.6092
y1( 2) y (3.0) 0.0 0.25 sin( 2.5) cos(0) sin( 3.0) cos(0.6092) 0.6399
y1(3) y (3.0) 0.0 0.25 sin( 2.5) cos(0) sin( 3.0) cos(0.6399) 0.6354
y 2(1) y (3.5) 0.6354 0.25 sin( 3.0) cos(0.6354) sin( 3.5) cos(1.10837) 0.89572
y 2( 2) y (3.5) 0.6354 0.25 sin( 3.0) cos(0.6354) sin( 3.5) cos(0.89572) 0.94043
m
y 2(3) y (3.5) 0.6354 0.25 sin( 3.0) cos(0.6354) sin( 3.5) cos(0.94043) 0.93155
o
Hence y2 y(3.5) 0.93155
s.c
4. Using modified Euler’s method obtain the solution of the differential equation
dy
x y with the initial condition y = 1 at x = 0 for the range 0 < x 0.6 in
dx
steps of 0.2.
Soln:
Given x0 0, y 0 1, h
c
0.2 and f ( x, y ) x y
vtu
To Find: y1 y ( x1 ) y (0.2), y 2 y( x2 ) y(0.4) and y3 y ( x3 ) y(0.6)
x Y1 f ( xn , y n ) yn hf ( x n , y n ) Y2 f ( xn 1 , y n 1 ) h
yn 1 yn Y1 Y2
2
w.
1.6350 1.5253
0.6 1.6350 1.8523 1.9610 1.8849
1.9729 1.8861
1.9734 1.8861
The solution is: y(0.2) 1.2309, y(0.4) 1.5253 and y(0.6) 1.8861
The Taylor’s series method to solve IVPs is restricted by the difficulty in finding the
higher order derivatives. However, Runge – Kutta method do not require the
calculations of higher order derivatives. Euler’s method and modified Euler’s method are
Runge – Kutta methods of first and second order respectively.
m
dy
Consider the IVP f ( x, y), y( x0 ) y0 . Let us find the approximate value of y at
dx
x x n 1 , n = 0,1,2,3,….. of this numerically, using Runge – Kutta method, as follows:
o
First let us calculate the quantities k1 , k 2 , k 3 and k 4 using the following formulae.
s.c
k1 hf ( xn , y n )
h k1
k2 hf x n , yn
2 2
h k2
k3 hf x n , yn
2
c 2
k4 hf x n h, y n k3
vtu
Finally, the required solution y is given by
1
yn 1 yn k1 2k 2 2k 3 k4
6
Problems:
w.
1. Apply Runge – Kutta method, to find an approximate value of y when x = 0.2 given
dy
that x y, y(0) 1 .
dx
Soln:
ww
1
R – K method (for n = 0) is: y1 y(0.2) y0 k1 2k 2 2k 3 k4 -------- (1)
6
Now
k1 hf ( x0 , y 0 ) 0.2 [0 1] = 0.2
h k1 0.2 0.2
k2 hf x0 , y0 0.2 0 1 = 0.2400
2 2 2 2
h k2 0.2 0.24
k3 hf x0 , y0 0.2 0 1 = 0.2440
2 2 2 2
m
1
y1 y(0.2) 1 0.2 0.24 0.244 0.2888 = 1.2468
6
Hence the required approximate value of y is 1.2468.
o
dy y2 x2
2.Using Runge – Kutta method of fourth order, solve , y (0) 1 at
y2 x2
s.c
dx
x = 0.2 & 0.4.
Soln:
y2 x2
Given: x0 0, y 0 1, h 0.2 and f ( x, y )
c y2 x2
h k2
k3 hf x0 , y0 0.2 f 0.1,1.0936 = 0.1967
2 2
1
y1 y (0.2) 1 0.2 2(0.19672) 2(0.1967) 0.1891
6
= 1+0.19599
= 1.19599
Hence the required approximate value of y is 1.19599.
m
h k1
k2 hf x1 , y1 0.2 f 0.3, 1.2906 = 0.1795
2 2
h k2
o
k3 hf x1 , y1 0.2 f 0.3, 1.2858 = 0.1793
2 2
s.c
k4 hf x1 h, y1 k3 0.2 f 0.4, 1.3753 = 0.1688
1
y2 y(0.4) 1.19599 0.1891 2(0.1795) 2(0.1793) 0.1688
6 c
= 1.19599 + 0.1792
vtu
= 1.37519
3. Apply Runge – Kutta method to find an approximate value of y when x = 0.2 with
dy y
h = 0.1 for the IVP 3x , y(0) 1 . Also find the Analytical solution and compare
w.
dx 2
with the Numerical solution.
Soln:
y
Given: x0 0, y0 1, h 0.1and f ( x, y) 3x
2
ww
h k1
k4 hf x0 , y0 0.1 f 0.1, 1.0666563 = 0.0833328
2 2
1
y1 y(0.1) 1.0 0.05 2(0.06625) 2(0.0666563) 0.0833328
m
6
= 1.0 + 0.0665242
= 1.0665242
o
Hence the required approximate value of y is 1.0665242.
s.c
Stage – II: Finding y 2 y(0.2)
k1 hf ( x1 , y1 )
c
0.1 f (0.1, 1.0665242) = 0.0833262
vtu
h k1
k2 hf x1 , y1 0.1 f 0.15, 1.04 = 0.1004094
2 2
h k2
k3 hf x1 , y1 0.1 f 0.15, 1.0485 = 0.1008364
2 2
= 1.0665242 + 0.1006976
= 1.1672218
Hence the required approximate value of y is 1.1672218.
Exact Solution
dy y
The given DE can be written as 3x which is a linear equation whose solution is:
dx 2
x
y 6 x 12 13e 2 .
The Exact solution at x = 0.1 is y(0.1) = 1.0665242 and at x = 0.2 is y(0.2) = 1.1672218
m
Both the solutions and the error between them are tabulated as follows:
o
s.c
x y y xn yn yn Absolute
(Exact) (Numerical) (Exact) (Numerical)
Error
0.1
0.2
0.3
c
1.0665243 1.0665242
1.1672219 1.1672218
1.3038452 1.3038450
0.1 1.0665243 1.0665242 0.0000001
vtu
0.4 1.4782359 1.4782357
0.2 1.1672219 1.1672218 0.0000001
0.5 1.6923304 1.6923302
0.6 1.9481645 1.9481643
0.7 2.2478782 2.2478779
0.8 2.5937211 2.5937207
0.9 2.9880585 2.9880580
1.0 3.4333766 3.4333761
w.
Multi-step Methods:
ww
To solve a differential equation over an interval (xn, xn+1), using previous single-step
methods, only the values of y at the beginning of interval is required. However, in the
following methods, four prior values are needed for finding the value of yn+1 at a given
value of x. Also the solution at yn+1 is obtained in two stages. This method of refining an
initially crude estimate of yn+1 by means of a more accurate formula is known as
Predictor–Corrector method. A Predictor Formula is used to predict the value of yn+1
and then a Corrector Formula is applied to calculate a still better approximation of yn+1.
Now we study two such methods namely (i) Milne’s method and (ii) Adams – Bashforth
method.
dy
Given f ( x, y), y( x0 ) y0 . To find an approximate value of y at x=x0+nh by
dx
Milne’s method, we proceed as follows: Using the given value of y ( x0 ) y 0 , we
compute y( x1 ) y ( x0 h) y1 , y( x2 ) y ( x0 2h) y 2 and y ( x3 ) y ( x0 3h) y3 using
m
Taylor’s series method.
o
I Stage: Predictor Method
s.c
4h
y 4( P ) y0
2 f1 f 2 2 f 3
3
Then we compute f 4 f ( x 4 , y 4( P ) )
y 4(C ) y2
h
3
f2
c
4 f3 f4
vtu
Then, an improved value of f4 is computed and again, corrector formula is applied to
find a better value of y4. We repeat the step until y4 remains unchanged.
dy
Given f ( x, y), y( x0 ) y0 . Using the given value of y ( x0 ) y 0 , we first compute
dx
y ( x 1 ) y ( x0 h) y 1 , y ( x 2 ) y ( x0 2h) y 2 and y ( x 3 ) y ( x0 3h) y 3 using
Taylor’s series method.
ww
h
y1( P ) y0 55 f 0 59 f 1 37 f 2 9f 3
24
h
y1(C ) y0 9 f1 19 f 0 5f 1 f 2
24
m
is applied to find a better value of y1. This step is repeated until y1 remains unchanged
and then proceeds to calculate y2 as above.
Problems:
o
dy
1. Use Milne’s method to find y(0.3) for the IVP x2 y2, y(0) 1
dx
Soln:
s.c
First, let us find the values of y at the points x = -0.1, x = 0.1 and x = 0.2 by using
Taylor’s series method for the given IVP.
cx 2 // x 3 ///
y ( x) y(0) xy / (0) y (0) y (0) -------- (1)
2 6
Given
vtu
y / ( x) x2 y2 y / (0) 0 1 1
// / //
y ( x) 2 x 2 yy y (0) 2 1 1 2
/// // / 2 ///
y ( x) 2 2 yy 2( y ) y (0) 2 4 2 8
Using the values of y(0), y/(0), y//(0) and y///(0) in (1), we get
4x 3
w.
y ( x) 1 x x2
3
Putting x = -0.1, x = 0.1 and x = 0.2 in the above expression, we get
Given:
x0 0.1, y 0 0.9087 and f 0 0.8357
x1 0, y1 1 and f1 1
x2 0.1, y2 1.1113 and f 2 1.2449
x3 0.2, y3 1.2507 and f 3 1.6043
To Find: y4 y( x4 ) y(0.3)
m
II Stage: Corrector Method
h
y 4(C )
o
y(0.3) y2
f2 4 f3 f4
3
0.1
y 4(C ) y (0.3) 1.1113 1.2449 (4 1.6043) 2.1555
s.c
3
= 1.4386
Now, we compute f 4 f (0.3,1.4386) 2.1596
0.1
y 4(C ,1) y(0.3) 1.1113 1.2449 (4 1.6043) 2.1596
3 c
y(0.3) = 1.43869
vtu
Hence, the approximate solution is y(0.3) = 1.43869
dy
2. Given x y 2 , y (0) 0 , y(0.2) = 0.02, y(0.4) = 0.0795 and y(0.6) = 0.1762.
dx
Compute y(1) using Milne’s Method.
Soln:
w.
Given:
x0 0, y 0 0 and f 0 f ( x0 , y 0 ) 0
ww
To Find: y4 y( x4 ) y(0.8)
4h
y 4( P ) y(0.8) y0 2 f1 f2 2 f3
3
4(0.2)
0
(2 0.1996) 0.3937 2 2 0.56895
3
= 0.30491
Now we compute f 4 f (0.8,0.30491) 0.7070
m
h
y 4(C ) y(0.8) y2
f2 4 f3 f4
3
0.2
y 4(C ) y(0.8) 0.0795 0.3937 4 0.56895 0.7070
3
o
= 0.3046
s.c
Now f 4 f (0.8,0.3046) 0.7072
0.2
y 4(C ,1) y(0.8) 0.0795 0.3937 4 0.56895 0.7072
3
y(0.8) = 0.3046
c
vtu
Stage - II: Finding y(1.0)
Given:
x1 0.2, y1 0.02 and f1 f ( x1 , y1 ) 0.1996
x2 0.4, y2 0.0795 and f 2 f ( x2 , y2 ) 0.3937
x3 0.6, y3 0.1762 and f 3 f ( x3 , y3 ) 0.56895
w.
To Find: y5 y ( x5 ) y (1.0)
ww
4h
y5( P ) y(1.0) y1 2 f2 f3 2 f4
3
4(0.2)
0.02
(2 0.3937) 0.56895 2 0.7072
3
= 0.45544
Now we compute f 5 f (1.0,0.45544) 0.7926
h
y5(C ) y(1.0) y3 f3 4 f4 f5
3
0.2
y5(C ) y(1.0) 0.56895 0.56895 4 0.7072 0.7926
3
= 0.4556
m
Now f 5 f (1.0,0.4556) 0.7024
o
0.2
y5(C ,1) y(1.0) 0.56895 0.56895 4 0.7072 0.7924
3
s.c
y(1.0) = 0.4556
dy
3. Given x 2 1 y , y(1) 1, y(1.1) 1.233, y(1.2) 1.548, y(1.3) 1.979 .Evaluate
dx
y(1.4) by Adam’s – Bashforth method. c
Soln:
vtu
Given:
x 3 1, y 3 1 and f 3 2
x 2 1.1, y 2 1.233 and f 2 2.70193
x 1 1.2, y 1 1.548 and f 1 3.66912
x0 1.3, y 0 1.979 and f 0 5.03451
w.
To Find: y1 y( x1 ) y(1.4)
h
y1( P ) y(1.4) y0 55 f 0 59 f 1 37 f 2 9f 3
24
(0.1)
1.979
(55 5.03451) (59 3.66912) (37 2.70193) (9 2)
24
= 2.57229
Now we compute f1 f (1.4,2.57229) 7.0017
h
y1(C ) y(1.4) y0 9 f1 19 f 0 5f 1 f 2
24
0.1
1.979
(9 7.0017) (19 5.03451) (5 3.66912) 2.70193
24
y(1.4) = 2.57495
0.1
m
y1(C ,1) 1.979
(9 7.0069) (19 5.03451) (5 3.66912) 2.70193
24
= 2.57514
Again f1 f (1.4, 2.57514) 7.0073
0.1
o
y1(C , 2) 1.979 (9 7.0073) (19 5.03451) (5 3.66912) 2.70193
24
s.c
y(1.4) = 2.57514
dy
4. Given x2 y, y (0) 1. Find y(0.4) by Adam’s method.
dx
Soln:
c
First, let us find the values of y at the points x = 0.1, x = 0.2 and x = 0.3 by using
Taylor’s series method for the given IVP.
vtu
Taylor’s expansion of y(x) about the point x = 0( = x0) is
x 2 // x 3 ///
y ( x) y(0) xy / (0) y (0) y (0) -------- (2)
2 6
Given
w.
y / ( x) x2 y y / (0) 0 1 1
// / //
y ( x) 2x y y (0) 0 ( 1) 1
y /// ( x) 2 2 y // y /// (0) 2 1 1
Using the values of y(0), y/(0), y//(0) and y///(0) in (2), we get
ww
x2 x3
y ( x) 1 x
2 3
Let x 3 0, y 3 1 and f 3 1
x 2 0.1, y 2 0.9051 and f 2 0.8951
m
(0.1)
0.7492 55 ( 0.6592) 59 ( 0.7812) 37 ( 0.8951) 9 ( 1)
24
= 0.6896507
Now we compute f1 f (0.4,0.6896507) 0.5296507
o
II Stage: Corrector Method
s.c
h
y1(C ) y(0.4) y0 9 f1 19 f 0 5f 1 f 2
24
0.1
0.7492 9 ( 0.5297) 19 ( 0.6592) 5 ( 0.7812) 0.895125
24
y(0.4) = 0.6896522
c
vtu
w.
ww
o m
s.c
c
vtu
w.
ww
Unit II
Numerical Methods II
Introduction:
In this unit we discuss numerical solution of simultaneous first order ODEs
m
And also second order ODEs as an extension of some of the earlier discussed method for solving
ODEs of first order.
o
Picard’s method:
s.c
Consider the following system of equations:
dy
f ( x, y, z )................1
dx
dz
g ( x, y, z ).................2
dx
With the initial condition y(x0)=y0
that is, y y0 and z z0 at x x0
c
vtu
we have to find successive approximations for y and z interms of x.
from 1 and 2
dy f ( x, y, z )dx
where y=y0 at x=x 0
dz=g(x,y,z)dx
where, z=z 0 at x=x 0
w.
z x
dz g ( x, y, z )dx
z0 x0
x
y y0 f ( x, y, z )dx
x0
x
y y0 f ( x, y, z )dx
x0
x
z z0 g ( x, y, z )dx
x0
x
z z0 g ( x, y, z )dx
x0
Problems
m
1. Use picard‟s method to find y(0.1) and z(0.1) given that
dy dz
=x+z, =x-y2and y(0)=2, z(0)=1
dx dx
Soln. we have a system of two equation and we need to find two approximations for y
o
and z as functions of x.
dy
x z; y (0) 2
s.c
dx
dz
x y 2 ; z (0) 1
dx
y x
dy (x z )dx;
2 0
z x
1
dz
0
(x
x
y 2 )dx c
vtu
y 2 (x z )dx (1)
0
x
z 1 (x y 2 )dx (2)
0
The first approximation for y and z are obtained by replacing the initial values of y and z in the
RHS of (1) and (2).
x x
( x z 2 )dx
w.
y1 2 ( x 1)dx z1 1
0 0
x 2
x2
y1 2 x z1 1 4 x
2 2
x x
y2 2 ( x z1 )dx z2 1 ( x y12 )dx
ww
0 0
x 2 x
x x2
2 ( x (1 4 x ))dx 1 ( x (2 x ))dx
0
2 0
2
x3 3x 2 3x 2 x4 x5
2 x 1 4x x3
6 2 2 4 20
3(0.1)2 0.13
y(0.1) 2 (0.1) z (0.1) 0.584
2 6
2.0852
2. Find the second approximation for the solution of the following system equations by applying
dx dy
picard‟s method =(x+y)t, =(x-t)y; x=0, y=1 at t=0.
dt dt
Soln:we have by data
dx=(x+y)tdt; x=0,t=0; dy=(x-t)y; y=1,t=0
x t y t
m
dx (x y )tdt and dy ( x t ) ydt
0 0 1 0
t
x (x y )tdt
o
0
t
y 1 ( x t ) ydt
s.c
0
t t
x1 tdt y1 1 tdt
0 0
2 2
t c t
x1 y1 1
2 2
t t
vtu
x2 ( x1 y1 )tdt y2 1 ( x1 t ) y1dt
0 0
t2 t2 t3 t4 t5
x2 y2 1
2 2 6 8 20
m
y ( x0 ) y0 ; z ( x0 ) z0
We compute y(x 0 +h) and z(x 0 +h)
First let us calculate the quantities k1 , k 2 , k 3 and k 4 using the following formulae.
o
k1 hf ( x0 , y0 , z0 ) l1 hg ( x0 , y0 , z0 )
s.c
h k1 l h k1 l1
k2 hf x0 , y0 , z0 1 l2 hg x0 , y0 , z0
2 2 2 2 2 2
h k2 l
k3 hf x0 , y0 , z0 2 h k2 l2
2 2 2 l3 hg x0 , y0 , z0
2 2 2
k4 c
hf x0 h, y0 k3 , z0 l3 l4 hg x0 h, y0 k3 , z0 l3
vtu
Finally, the required solution y is given by
1
y( x0 h) y0 k1 2k2 2k3 k4
6
And
1
w.
z ( x0 h) z0
l1 2l2 2l3 l4
6
1. Apply Runge – Kutta method to find an approximate value of y when x = 0.2 with
dy y
h = 0.1 for the IVP 3x , y(0) 1 . Also find the Analytical solution and compare
dx 2
with the Numerical solution.
ww
Soln:
y
Given: x0 0, y 0 1, h 0.1and f ( x, y) 3x
2
Stage – I: Finding y1 y(0.1)
1
R – K method (for n = 0) is: y1 y(0.1) y0 k1 2k 2 2k 3 k4 -------- (4)
6
h k1
k2 hf x0 , y0 0.1 f 0.05, 1.025 = 0.06625
2 2
h k1
k3 hf x0 , y0 0.1 f 0.05, 1.033125 = 0.0666563
2 2
h k1
k4 hf x0 , y0 0.1 f 0.1, 1.0666563 = 0.0833328
2 2
m
Using the values of k1 , k 2 , k 3 and k 4 , we get
o
1
y1 y(0.1) 1.0 0.05 2(0.06625) 2(0.0666563) 0.0833328
6
= 1.0 + 0.0665242
s.c
= 1.0665242
h k2
k3 hf x1 , y1 0.1 f 0.15, 1.0485 = 0.1008364
2 2
= 1.1672218
Hence the required approximate value of y is 1.1672218.
2 Apply Runge – Kutta method, to find an approximate value of y when x = 0.2 given
dy
that. x y, y (0) 1
dx
Soln: Given: x0 0, y 0 1, h 0.2 and f ( x, y ) x y
1
R – K method (for n = 0) is: y1 y(0.2) y 0 k1 2k 2 2k 3 k 4 -------- (1)
6
Now
m
k1 hf ( x0 , y 0 ) 0.2 [0 1] = 0.2
h k1 0.2 0.2
k2 hf x0 , y0 0.2 0 1 = 0.2400
2 2 2 2
o
h k2 0.2 0.24
k3 hf x0 , y0 0.2 0 1 = 0.2440
2 2 2 2
s.c
k4 hf x0 h, y 0 k3 0.2 0 0.2 1 0.2440 = 0.2888
y1
1
y(0.2) 1
6
c
0.2 0.24 0.244 0.2888 = 1.2468
Hence the required approximate value of y is 1.2468.
vtu
Numerical solution of secand order ODEs by Picard’s method and Runge kutta method
Let y '' g ( x, y, y ' ) with intial condition y ( x0 ) y0 , y' ( x0 ) y0' be the second order
w.
differential equations.
dz
Now let y ' z this gives y ''
dx
dz
The second order differential equations assumes the form g ( x, y, z ) with the
dx
ww
Milne‟s Method:
dy
Given f ( x, y), y( x0 ) y0 . To find an approximate value of y at x=x0+nh by
dx
Milne’s method, we proceed as follows: Using the given value of y ( x0 ) y 0 , we
compute y( x1 ) y ( x0 h) y1 , y( x2 ) y ( x0 2h) y 2 and y ( x3 ) y ( x0 3h) y3 using
m
Taylor’s series method.
o
I Stage: Predictor Method
s.c
4h
y 4( P ) y0
2 f1 f 2 2 f 3
3
Then we compute f 4 f ( x 4 , y 4( P ) )
y 4(C ) y2
h
3
f2
c
4 f3 f4
vtu
Then, an improved value of f4 is computed and again, corrector formula is applied to
find a better value of y4. We repeat the step until y4 remains unchanged.
dy
1 Use Milne’s method to find y(0.3) for the IVP x2 y2, y(0) 1
dx
w.
Soln:
First, let us find the values of y at the points x = -0.1, x = 0.1 and x = 0.2 by using
Taylor’s series method for the given IVP.
x 2 // x 3 ///
y ( x) y(0) xy / (0) y (0) y (0) -------- (1)
2 6
Given
y / ( x) x2 y2 y / (0) 0 1 1
// / //
y ( x) 2 x 2 yy y (0) 2 1 1 2
/// // / 2 ///
y ( x) 2 2 yy 2( y ) y (0) 2 4 2 8
Using the values of y(0), y/(0), y//(0) and y///(0) in (1), we get
4x 3
y ( x) 1 x x2
3
Given:
x0 0.1, y 0 0.9087 and f 0 0.8357
x1 0, y1 1 and f1 1
m
x2 0.1, y2 1.1113 and f 2 1.2449
x3 0.2, y3 1.2507 and f 3 1.6043
To Find: y4 y( x4 ) y(0.3)
o
I Stage: Predictor Method
4h
y 4( P ) y(0.3) y 0 2 f1 f 2 2 f 3
3
s.c
4(0.1)
0.9087 (2 1) 1.2449 2 1.6043
3
= 1.4372
Now we compute f 4 f (0.3,1.4372) 2.1555
II Stage: Corrector Method
y 4(C )
h
y(0.3)
3
y2
f2 4 f3 f4
c
vtu
0.1
y 4(C ) y (0.3) 1.1113 1.2449 (4 1.6043) 2.1555
3
= 1.4386
Now, we compute f 4 f (0.3,1.4386) 2.1596
0.1
y 4(C ,1) y(0.3) 1.1113 1.2449 (4 1.6043) 2.1596
w.
3
y(0.3) = 1.43869
dx
Compute y(1) using Milne’s Method.
Soln:
To Find: y4 y( x4 ) y(0.8)
I Stage: Predictor Method
4h
y 4( P ) y(0.8) y 0 2 f1 f 2 2 f 3
3
4(0.2)
0 (2 0.1996) 0.3937 2 2 0.56895
3
= 0.30491
m
Now we compute f 4 f (0.8,0.30491) 0.7070
II Stage: Corrector Method
h
y 4(C ) y(0.8) y 2 f2 4 f3 f4
o
3
0.2
y 4(C ) y(0.8) 0.0795 0.3937 4 0.56895 0.7070
3
s.c
= 0.3046
Now f 4 f (0.8,0.3046) 0.7072
Again applying corrector formula with new f4, we get
0.2
y 4(C ,1) y(0.8) 0.0795 0.3937 4 0.56895 0.7072
3
y(0.8) = 0.3046
3
= 0.45544
Now we compute f 5 f (1.0,0.45544) 0.7926
II Stage: Corrector Method
h
y5(C ) y(1.0) y3 f3 4 f4 f5
3
0.2
y5(C ) y(1.0) 0.56895 0.56895 4 0.7072 0.7926
3
= 0.4556
Now f 5 f (1.0,0.4556) 0.7024
y(1.0) = 0.4556
o m
c s.c
vtu
w.
ww
UNIT-III
COMPLEX ANALYSIS-1
INTRODUCTION:
An extension of the concept of real numbers to accommodate complex numbers was evolved
while considering solutions of equations like x 2 1 0 . This equation cannot be satisfied for
m
any real value of x. In fact, the solution of the equation x 2 1 0 is of the form x 1 . The
square root of -1 cannot be a real no. because the square o any real no. is nonnegative. Similarly,
there are any number of algebraic equations whose solutions involve square roots of negative
numbers.
o
FUNCTION OF A COMPLEX VARIABLE:
s.c
If z = x+iy is a complex variable, then w = f(z) is called function of a complex variable.
W = f(z) = u+iv where u = u(x,y), v = v(x,y). Hence for every point of (x,y) in z-plane, there
corresponds (u,v) in w-plane
z
c
Complex value function f(z) defined in the neighbourhood of a point z0 is said to have limit L as
z 0 if for all ε>0 however small, there exists a positive real number δ such that
vtu
f ( z) L whenever z z0 ,
i.e., lim f ( z ) L
z z0
CONTINUITY:
A function f(z) is said to be continuous at a point z0 if lim f ( z) f ( z0 )
z z0
w.
DIFFERENTIABILITY:
f ( z) f ( z0 )
A function f(z) is said to be differentiable at a point z0 if lim exists and this is
z z0 z z0
unique
ww
ANALYTIC FUNCTION:
A function f(z) is said to be analytic at a point z0 if f(z) is differentiable at z0 as well as at all
points in a neighbourhood of z0.
f ( z z) f ( z)
i.e., f ( z ) lim exists and unique for all points in complex region.
z 0 z
NOTE:
Analytic function is also called as regular function or holomorphic function.
Statement: If w = f(z) = u+iv is analytic function at the point z = x+iy, then there exists partial
u v v u
derivatives and satisfy the equations & called C-R equations in cartesian
x y x y
form.
m
Proof: By data f(z) is analytic at a point z = x+iy, there by definition of analytic function,
f ( z z) f ( z)
f ( z ) lim ……………(1)exists and unique.
z 0 z
We have f(z) = u(x,y) +iv(x,y)
o
f ( z z ) u ( x x, y y ) iv( x x, y y )
z x iy
s.c
z x i y
Substituting the above in (1) we get
u ( x x, y y) iv( x x, y y) u ( x, y) iv( x, y)
f ( z) lim …………(2)
z 0 x i y
Since z 0 , we have 2 possibilities.
cCase(i): If δz is only real, then δy = 0
i.e., if z 0 then x 0
vtu
equn(2) becomes
u( x x, y ) u ( x, y ) v( x x, y ) v ( x, y )
f ( z) lim i
x 0 x x
u v
f ( z) i ................(3)
x x
w.
1 u ( x, y y ) u ( x , y ) v ( x, y y ) v ( x , y )
f ( z) lim lim
ww
i y 0 y y 0 y
u v
f ( z) i ................(4)
y y
Comparing real and imaginary parts of equations (3) and (4) we get
u v v u
& ………….proved
x y x y
Statement: If w = f(z) = u+iv is analytic function at the point z re i , then there exists partial
u 1 v v 1 u
derivatives and satisfy the equations & called C-R equations in polar
r r r r
m
form.
o
f ( z ) lim
z 0 z
We have f(z) = u(r,θ) +iv(r,θ)
f ( z z ) u (r r , ) iv(r r , )
s.c
z re i irei
Substituting the above in (1) we get
u (r r, ) iv(r r , ) u (r , ) iv(r , )
f ( z) lim i i
…………(2)
z 0 re ire
c Since z 0 , we have 2 possibilities.
i u (r r , ) u (r , ) i v( r r , ) v(r , )
f ( z) e lim ie lim
r 0 r r 0 r
i u v
f ( z) e i ................(3)
r r
w.
e i u (r , ) u (r , ) e i v(r , ) v(r , )
ww
f ( z) lim i lim
ir 0 ir 0
i
e u v
f ( z) i ................(4)
ir
Comparing real and imaginary parts of equations (3) and (4) we get
u 1 v v 1 u
& ………….proved
r r r r
HARMONIC FUNCTION:
A function u is said to be harmonic function if it satisfies the Laplace equation.
2 2
u u
i.e., 2
o in Cartesian form
x y2
2 2
u 1 u 1 u
o in polar form.
m
2
r r r r2 2
Theorem:
Statement: The real and imaginary parts of an analytic function are harmonic.
o
Proof: Let f(z) = u(x,y)+iv(x,y)
Since f(z) is analytic, satisfies C-R equations
s.c
u v v u
i.e., ……..(1) & ………….(2)
x y x y
differentiating (1) partially w.r.t x and (2) w.r.t y , we get
2 2 2 2
u v v u
&
x 2
x y y x y2
Therefore
u
x2
2 2
y2
u c
vtu
2 2
u u
2
0 u satisfies Laplace equation.
x y2
Hence real part u is harmonic.
differentiating (1) partially w.r.t y and (2) w.r.t x , we get
2 2 2 2
u v v u
2
& 2
x y x x x y
w.
2 2
v v
Equating the above equations 2
y x2
2 2
v v
2
0 vsatisfies Laplace equation.
x y2
ww
PROBLEMS:
1. Prove that u e 2 x x cos 2 y y sin 2 y is harmonic. Find the analytic function f(z)
Soln:
u
e 2 x cos 2 y 2e 2 x x cos 2 y y sin 2 y
x
2
u
4e 2 x cos 2 y 4e 2 x x cos 2 y y sin 2 y
x2
u
e 2 x 2 x sin 2 y sin 2 y 2 y cos 2 y
y
m
2
u
2
e2x 4 x cos 2 y 4 cos 2 y 4 y sin 2 y
y
o
2 2
u u
Hence e2 x 4 x cos 2 y 4 cos 2 y 4 y sin 2 y +
x2 y 2
4e 2 x cos 2 y 4e 2 x x cos 2 y
s.c
y sin 2 y = 0
Therefore u is harmonic.
u v u u
f ( z) i i ..........byC Re quations
x x x y
f ( z) e 2 x cos 2 y 2e 2 x x cos 2 y y sin 2 y ie2 x
c 2 x sin 2 y sin 2 y 2 y cos 2 y
Put x = z and y=0
vtu
f ( z) e 2 z 1 2 z ie2 z 0 e 2 z 1 2 z
Integrating we get
e2z e2z
f ( z) (1 2 z ) 2 c
2 4
2z
f ( z) ze ic
w.
ww
2)Find the analytic function f(z) where imaginary part is e x ( x sin y y cos y )
x
So ln : v e x sin y y cos y
v
e x x sin y y cos y e x sin y
x
v
e x x cos y cos y y sin y
y
m
u v
we have f 1 ( z ) i
x x
u v
By C-R equation
o
x y
v v
:f 1 ( z ) i
y x
s.c
f 1 ( z) ex x cos y cos y y sin y i x sin y y cos y sin y
Put x z, y 0
f 1 ( z) ez z 1 i 0
( z 1)e z
Integrating w.r.t z
f ( z) (z 1)e z
c
(1)e z c
vtu
f ( z) ze z c
w.
ww
ex x2 y 2 cos y 2 xy sin y
3)Find the analytic function where real part is
Solution:
u ex x2 y 2 cos y 2 xy sin y
u
e x x 2 y 2 cos y 2 xy sin y e x 2 x cos y 2 y sin y
m
x
u
( z.0) e z z 2 2 z
y
u
e x 2 y cos y ( x 2 y 2 ) sin y 2 x sin y 2 xy cos y
o
y
u
ez 0 0
s.c
z ,o
y
u v
f 1 ( z) i z ,o
x x
u u
i z ,o By C-R equations
x y
z2 2z ez
c 0
vtu
f 1 ( z) z2 2z ez
f ( z) z 2 2 z e z dz
z2 2z ez 2z 2 ez (2)e z c
f ( z) z 2e z c
w.
ww
2 2
2 2
2 2
f (z ) 4 f 1 ( z)
4) If f(z)=u(x,y)+iv(x,y) is analytic, Show that x y
Solution:
Let f(z) =u+iv is analytic:
f 1 ( z ) exist
m
u v
f 1 ( z) i
x x
2 2
u v
f 1 ( z)
o
x x
2 2
1 2 u v
f ( z)
s.c
x x
2
Also f 1 ( z ) u 2 v2
Diff partially (2) w.r.t x
u v
f ( z ) 2u 2v
x x x c
Differentialing again w.r.t x we get
2 2 2 2 2
vtu
2 u v v
f ( z) 2 u v
x2 x2 x x2 x
Similarly we can obtain
2 2 2 2 2
2 u u v v
f ( z) 2 u 2 v 2
y2 y y y y
w.
(3) (4) )
2 2 2 2 2 2 2 2 2 2
2 u u v v u v u v
f ( z) 2 u v
x2 y2 x2 y2 x2 y2 x x y y
Since u & v are harmonic & using C-R equation
ww
u v v u
& , we get
x y x y
2 2 2 2
2 u v
f ( z) 2 2 2
x2 y2 x x
2 2
u v
4
x x
2
4 f 1 ( z)
UNIT-IV
COMPLEX ANALYSIS-2
CONFORMAL TRANSFORMATION:
Let c1 and c2 be any two curves in the z-plane intersecting at z0, suppose the transformation w =
m
f(z) transforms the curves c1 and c2 to the curves c1 and c2 respectively which intersect at a
point w0 = f(z0) in the w-plane. Then the transformation is said to be conformal if the angle
between c1 and c2 is equal to the angle between c1 and c2 in both magnitude and direction.
o
SOME STANDARD TRANSFORMATION:
1. Discuss the transformation w z 2
s.c
w z 2 .............(1)
u iv (x iy ) 2
= (x 2 -y 2 )+i2xy
c
u=x 2 -y 2 ; v=2xy............(2)
case(i): consider a line parallel to y-axis i.e., x=a, a is a constant
vtu
eqn(2) becomes
u=a 2 -y 2 , v = 2ay
or v 2 = 4a 2 y 2 =4a 2 (a 2 -u ) = -4a 2 (u a 2 )
This represents parabola in the w-plane along negative u-axis vertex at the point
(a 2 , 0)
w.
w ez
u iv e(x iy )
e x (cos y i sin y )
u e x cos y,
v e x sin y..............(1)
m
Case(i): consider a straight line parallel to x-axis i.e., x = a, a is a constant
Therefore eqn(2 becomes)
u 2 v 2 e 2a r 2
o
This represents a circle with centre origin and radius r in the w-plane
Case(ii): consider a straight line parallel to x-axis i.e., y = b, b is a constant
Therefore eqn(2 becomes)
s.c
v
tan b. m( say)
u
v um
This represents a straing line passing thro‟ the origin in the w-plane
CONCLUSION: Straight line parallel to coordinate axes in z – plane transforms circle with
c
centre origin and straight line passing thro‟ origin respectively in w-plane under the
transformation w e z
vtu
k2
3. Discuss the transformation w z ,z 0 . What are the points at which the
z
transformation is not conformal.
k2
f ( z) z
w.
z
k2
f ( z) 1 0
z2
Only if z k
k2
Therefore the transformation w z ,z 0 is not conformal at z k
ww
z
k2
w z
z
k2
w re i
re i
\
k2
w r (cos i sin ) (cos i sin )
r
k2 k2
w r cos i r sin
r r
k2
u r cos ........(1)
r
k2
v r sin ........(2)
r
Eliminating from (1)and (2)
m
u v
2
cos , sin ,
k k2
r r
o
r r
s.c
u2 v2
2 2
1................(3)
k2 k2
r r
r r
1..................(4)
(2k cos ) 2 . (2k sin ) 2 ..
Let z c1
case(i):
i.e., r c1
ww
m
u2 v2
1
(2k cos c 2 ) 2 . (2k sin c 2 ) 2
u2 v2
i.e., 1
A2 B 2
o
Where A= 2k cos c2 and B = 2k sin c2
This represents a hyperbola in w-plane with foci at ( A2 B 2 ,0) i.e., ( 2k ,0)
s.c
CONCLUSION: The circle centred at origin with radius constant in z-plane transforms to an
ellipse with foci ( 2k ,0) in w-plane and straight line passing thro‟ origin in z-plane transforms
to a hyperbola with foci ( 2k ,0) in w-plane. Since both of these conics have the same foci
( 2k ,0) , they are called confocal conics.
2 2 2
y 4a y
4a 2 ( a 2 u )
v2 4a 2 (u a 2 )
This represent parabola in the w-plane along-ve u-axis vertex at the pt( a 2 ,0)
ww
Case(ii): Consider a straight line 11el to y-a xis i.e, y=b, b is a constant
u x2 b2 & v 2 xb
(1)=) 2 2 2
v 4x b 4b (u b 2 )
2
This represents parabola in the w-plan along the u-axis, vertex at the pt b2 , 0
Hence, we conclude that straight line 11el to Co-ordinate axes in z-plane transforms parabolas in
w-plane under w z 2
az b
The transformation w , where a,b,c,d are real or complex constants such that
cz d
ad bc 0 is called a bilinear transformation
CROSS RATIO:
m
( z1 z 2 )( z 3 z 4 )
Cross ratio of four points z1 , z 2 , z 3 , z 4 defined by
( z 2 z 3 )( z 4 z1 )
INVARIANT POINTS OR FIXED POINTS
o
If a point z maps ont itself i.e., w = z then the point is called invariant point or a fixed point of
the bilinear transformation
s.c
PROPERTIES OF BLT
( w1 , w2 , w3 , w4 )= (w, i, 1, 0 )
Using cross ratio of 4 points
( w i )(1 0) ( z 0)( i 1)
(i 1)(0 w) (0 i )( 1 z )
ww
w i (1 i ) 2 z
.
w i z 1
w i z
2.
w z 1
wz iz w i 2 zw
iz 1
w
z 1
This is the required BLT
2. Find the BLT that maps the points (∞, i, 0) of z-plane onto the points (-1, -i, 1) of w-
plane respectively.Also find the invariant points.
Soln.
Given ( z1 , z 2 , z 3 , z 4 )= (z , ∞, i, 0)
( w1 , w2 , w3 , w4 )= (w,-1, -i, 1)
Using cross ratio of 4 points and z 2 &1/ z2 0
m
( w 1)( 1 i ) ( 1)(i )
( 1 i )(1` w) (1)( z )
( w 1)(1 i ) i
(1 i )(1 w) z
o
zwi zi wz z i 1 wi w
( w 1)(i 1) z (i 1)(1 w)
s.c
z 1
w
z 1
This is the required BLT
put w z c
z 1
z
z 1
vtu
z2 2z 1 0
z 1 2
Are invariant points
COMPLEX INTEGRATION:
w.
Consider xy-plane is taken as complex plane then the point p(x,y) on this curve corresponding to
the complex number z=x-iy
The equation z=z(t) where t is parameter is reoffered to as the equation of curve in the complex
form.
Ex: as t varies over the interval and x=a cast y=asin t then the complex form of the equation of
ww
circle is
z x iy
z a cos t i sin t
a(cast i sin t
z ae it 0 t 2
Represent a circle leaving center at origen. And radies is equal to a.
Consider a continuous function f z of complex variable z=x=iy defined at all points of curve
from p to Q
Divide the curve in to n equal parts by taking points p= p( z 0 ), p1 ( z1 ), p 2 ( z 2 )...
p k 1 ( Z k 1 )........p n ( z n ) Q on the came C
Let “D” is the region of complex plane and f(z)=u(x,y)+iv(x,y) be complex valued function
defind on „D‟ Let C be the curve in „D‟ then (z=x+iy dz=dx+idy)
m
f ( z )dz. a( x, y ) iv( x, y )(dx idy)
c c
o
f ( z )dz (udx vdy ) i (vdx udy )
c c c
Is called the line integral of f(z) along the carve C
PROBLEMS:
s.c
1. Evaluate z 2 dz Along the straight line from z=0 to z=3+i
c
3 i
Solution ; a) z 2 dz z 2 dz
c z 0
Here z is varies from o to 3+i
Z=x+iy
c
vtu
Z=0=) (0.0) (3,1)
y y0 y1 y 0
Equation of the line gaining in
x x0 x1 x0
x
y
3
dx 3.dy
w.
f ( z) z 2 )( x iy) 2 x2 y2 i 2 xy
(3.1) (3.1)
z 2 dz (x 2 y 2 )dx 2 xydy i 2xydx 1x 2 y 2 dy
c ( 0.0 ) ( 0.0 )
Convert these integral in to ltu variable y or x respects to integral wi y from 0 to1 and x from
0to3
ww
1 1
2 2 2
z dz (9 y y )3dy 2(3 y ) y dy 2(3y) y3dy i9 y 2 y 2 )dy
c 0 0
1 1
(24 y 2 6 y 2 )dy i 18 y 2 8 y 2 dy
y 0 0
1 1
i 18 y 2 dy i 26 y 2 dy
0 0
m
1 1
2
y y3 26
18 26i 6 i
3 3 3
0 0
26
o
z 2 dz i 6
c 3
Along the given path.
s.c
( 2, 40
2. Evaluate (2 y x 2) dx 93x y ) dy
along the parabola x=2t, y= t 2 3
( 0.3)
3 t4 1
8t 3t 2 12t
2 0
1
)8 3 12 (0)
2
w.
1 33
) 16 6 24 2 )
2
Cauchy’s Theorem:
Statement: If f(Z) is analytic at all points inside and on a simple Closed Curve C then
ww
f ( Z )dz 0
c
Proof:
Let f(Z) dz = u+iv
f ( z )dz (u iv)(dx idy)
If M(x,y) and N(x,y) are two real Valued functioin then we have Green‟s Theorem
n m
Mdx Ndy dxdy
x y
Apply this theorem to 1
v u u v
f ( z )dz dxdy i ( )dxdy
R
x y x y
By C R Equation
m
f ( z )dz 0
c
o
Corollary –II:
If c1 c2 are two simple closed Curves Such that c2 lies entirely with c1 and if f(z) is analytic on
s.c
c1 c2 and in the region bounded by c1 c2 then f ( z )dz f ( z )dz
c1 c2
Proof:
The Point p c1 and Q on c2 then the Curve PQRSTUQP as a simple closed curve on boundary C
Hence by Cachy‟s theorem
f ( z )dz 0
c
Since C is the union of arc PRSP,PQ,QTUQ,AND QP THEN THEOREM BECOMES
vtu
f ( z )dz f ( z )dz f ( z )dz f ( z )dz 0
c1 c2 c3 c4
Therefore
f ( z )dz f ( z )dz
c2
Statement:
If f(z) is analytic inside and on a simple closed curve C and if a is any point with in C then
1 f ( z)
f(a) = dz
2 i z a
Proof:
ww
Since „a‟ is a Point with in c we shall enclose it by a circle c1 with z-a at a centre and r as a
radius such that c1 lies entirely with in c
The function f(z)/z-a is analytic inside and angular region between C and c1
By Cauchy‟s theorem
f ( z )dz f ( z )dz ………………1
c c2
2
f ( z) f (a re i ) i
dz .ire d
z a 0 re i
Hence is r is very small but r should greater than 0
2
f ( z)
dz i f (a)d
z a 0
= if(a)[2a]
m
= 2πif(a)
There fore
1 f ( z)
f(a) = dz .
2 i z a
o
Problems :
s.c
1. Verify Cauchy’s theorem for the function f(z) = z 2 where ‘C’is squair leaving vertices
(0,0) (1,0)91,1)(0,1)
z 2 dz z 2 dz z 2 dz z 2 dz 0
oA AB BC co
y 0
AB
2!
z 2 dz 1
AB
3
0
z 2 dz (x2 2ix 1)dz
ww
1
BC
2
z 2 dz i
BC
3
0
z 2 dz ( y 2 )idy
1
CO
i
z 2 dz
CO
3
Adding (1),(2) ,(3), (4)
1 2i 2 i
z 2 dz 1 i
C
3 3 3 3
0
cz
2. Evaluate dz over each contour C z 1 1
C z iy
f ( z)
m
2 i dz.
c (z a)
f ( z) ez , z 1 1
Soln: we have f(a) a 1, r 1
o
f ( z)
dz 0
c (z a)
s.c
when z 1 1
z2 z 1
3. Evalute dz over c: z =3
c ( z 2) 3
Soln: The point z=2 lies inside circle
Causley’s integral formula is c
f ( z) 2 i (n)
n 1
dz f (a)
vtu
( z a) n!
f ( z) z2 z 1, f ( z) 2z 1
f ( z) 2
z2 z 1 2 i 2
dz f (2)
( z 2)3 n!
2 i 2
w.
2 i
2!
TAYLOR’S THEOREM:
If f(z) is analytic inside and on the boundary of the annular region ‘R’ bounded by two
ww
concentric circle C1 ,C2 with centered ‘a’ and radius r1 , r2 then for every ‘Z’
an( z a) n a n ( z a) n
n 0 n 1
F(z) = dw
1 f ( w) 1 f ( w)
an dw ; a n
2 i c1 ( w a) n 1 2 i c 2 ( w a) n 1
1
1. Show that 1 (k 1)( z 1) k where ( z 1) 1
z2 k 1
z 1 1
a 1
( z 1) 2 ( z 1) 3
f ( z) f ( 1) ( z 1) f ( 1) f ( 1) f ( 1)
2
1
f ( z) f ( 1) 1
z2
m
2
f ( z) f ( 1) 2
z3
6
f ( z) f ( 1) 6
z4
o
24
f ( z) f ( 1) 24
z5
s.c
1 ( z 1) 2
1 ( z 1)2 6 .........
z2 2
Soln: 1 (k 1)( z 1) k
k 1
2. Expand f ( z )
c 1
( z 1)( z 2)
in terms of Laurent’s series valid in the regions i)
vtu
z 1 1 ii) z 1 1 .
Solution:
1
Given f ( z ) is proper fraction
( z 1)( z 2)
1 2
f ( z) .............(1)
( z 1) z 2
w.
u
Case(i): If z 1 1 , put z-1 = u then 1
1
1 2
(1) Becomes f ( z)
( z 1) z 1 1
ww
1 2
u (1 u )
1 1
(1 u ) 1 2 1 u u 2 ..........
u u
1
2 1 ( z 1) ( z 1) 2 ...........
z 1
Case(ii): If z 1 1 , put z-1 = u
1
i.e., u 1 1
u
1 2
(1) Becomes f ( z)
( z 1) z 1 1
1 2 1 2
u u 1 u 1
u (1 )
u
2
1 2 1 1 1 2 1 1
(1 ) 1 .........
m
u u u u u u u
2
1 2 1 1
1 ...........
z 1 u z 1 z 1
o
z 1
3. Expand f ( z ) in terms of Laurent’s series valid in the regions i) z 3
( z 2)( z 3)
s.c
ii) 2 z 3 .
Solution:
We shall first resolve f(z) into partial fraction
1 2
i.e., f ( z ) .............(1)
( z 2) z 3
z3 2
c
case(i): z 3 this implies that z 2 also.
z
vtu
i.e., 1, 1, 1
1, or
3 2 z z
Hence we have to write (1) in the form
1 2
f ( z)
2 3
z 1 z 1
z z
1 1
w.
1 3 2
f ( z) 21 1
z z z
1 3 9 27 2 4 8
f ( z) 21 ....... 1 .......
z z z2 z3 z z2 z3
ww
1 4 14 46
f ( z) 1 .......
z z z2 z3
2 z
Case(ii): 2 z 1,
3 this implies that 1
z 3
Hence we have to write (1) in the form
1 2
f ( z)
2 z
z 1 31
z 3
1 1
1 2 2 z
f ( z) 1 1
z z 3 3
1 2 4 8 2 z z2 z3
m
f ( z) 1 ....... 1 ............
z z z2 z3 3 3 9 27
1 2 4 8 2 2z 2z 2 2z 3
Thusf ( z ) ......... .............
z z2 z3 z4 3 3 27 81
o
1
4. Find the Taylor’s series expansion of f ( z ) about the point z = -i.
( z 1) 2
s.c
Solution:
1
For f ( z ) , we have
( z 1) 2
1 1 i
f ( i) 2
(1 i) 2i 2
& f (n) ( z)
( 1) n (n 1)!
( z 1) n 2
c
vtu
( 1) n (n 1)! 1 ( 1) n (n 1)! i ( 1) n (n 1)!
f ( n) ( i) .
( i 1) n 2 (1 i ) 2 ( i 1) n 2 ( i 1) n
Therefore the Taylor’s expansion of the given f(z) about the point z = -i is
f ( n) ( i) i ( 1) n (n 1)!
f ( z) f ( i) ( z i) n 1 n
( z i) n
n 1 n! 2 n 1 ( i 1)
w.
2z 3 1
5. Find the Taylor’s series expansion of f ( z ) about the point z = i.
z2 z
Solution:
2z 3 1 2z 1 ( z 1) z 1 1
For f ( z ) (2 z 1) 2( z 1) 2( z 1)
ww
z2 z z2 z z ( z 1) z z 1
This gives
1 1 (n) ( 1) n n! ( 1) n n!
f ( z) 2 and f ( z) ,n 2
z2 ( z 1) 2 zn 1 ( z 1) n 1
Accordingly,
1 2i ( i)(1 2i)(1 i) i i 3
f (i) ( 1 3i)
i(1 i) (1 i)(1 i) 2 2 2
1 1 1 i
f (i) 2 2 1 3
i2 (i 1) 2
2i 2
And
( 1) n n! ( 1) n n! 1 1
f ( n ) (i ) n 1 n 1
( 1) n n! n 1 ,n 2
i (i 1) i (i 1) n 1
Hence, the Taylor’s expansion of the given f(z) about the point z = i is
f ( n ) (i ) i 3 i 1 1
f ( z ) f (i ) ( z i) n 3 ( z i) ( 1) n (n)! n 1 z i) n
n 1 n! 2 2 2 n 2 i (i 1) n 1
o m
c s.c
vtu
w.
ww
o m
s.c
c
vtu
w.
ww
Unit-V
SPECIAL FUNCTIONS
Many Differential equations arising from physical problems are linear but have variable
coefficients and do not permit a general analytical solution in terms of known functions.
Such equations can be solved by numerical methods (Unit – I), but in many cases it is
easier to find a solution in the form of an infinite convergent series. The series solution
m
of certain differential equations gives rise to special functions such as Bessel’s function,
Legendre’s polynomial. These special functions have many applications in engineering.
o
Consider the Bessel Differential equation of order n in the form
s.c
d2y dy
x2 2
x ( x 2 n2 ) y 0 (i)
dx dx
where n is a non negative real constant or parameter.
dx r 0
2
d y
a r (k r )(k r 1) x k r 2
dx 2 r 0
w.
x2 a r (k r )(k r 1) x k r 2
x a r (k r)x k r 1
x2 n2 ar x k r
0
r 0 r 0 r 0
ww
i.e., a r (k r )(k r 1) x k r
a r (k r)x k r
ar x k r 2
n2 ar x k r
0
r 0 r 0 r 0 r 0
Grouping the like powers, we get
a r (k r )(k r 1) (k r) n 2 x k r
ar x k r 2
0
r 0 r 0
a r (k r) 2 n2 xk r
ar x k r 2
0 (iii)
r 0 r 0
Equating the coefficient of xk from the first term and equating it to zero, we get
a0 k 2 n 2 0. Since a0 0, we get k 2 n 2 0, k n
k+1
Coefficient of x is got by putting r = 1 in the first term and equating it to zero, we get
m
which is a contradiction to k = n.
Let us consider the coefficient of xk+r from (iii) and equate it to zero.
i.e, a r (k r ) 2 n 2 a r 2 0.
o
ar 2
ar (iv)
s.c
(k n2 r)2
If k = +n, (iv) becomes
ar 2 ar 2
ar 2 2 2
(n r ) n r 2nr
c
Now putting r = 1,3,5, ….., (odd vales of n) we obtain,
a1
vtu
a3 0 , a1 0
6n 9
i.e., a1 = a5 = a7 = …… = 0
a0 a0 a2 a0
a2 ; a4 ;
4n 4 4(n 1) 8n 16 32(n 1)(n 2)
We shall substitute the values of a1 , a 2 , a3 , a 4 , in the assumed series solution, we get
y ar x k r
x k ( a0 a1 x a 2 x 2 a3 x 3 a 4 x 4 )
r 0
a0 a0
y1 x n a0 x2 x 4
4(n 1) 32(n 1)(n 2)
x2 x4
i.e., y1 a0 x n 1 (v)
2 2 (n 1) 2 5 (n 1)(n 2)
This is a solution of the Bessel’s equation.
x2 x4
m
y2 a0 x n
1 (vi)
2 2 ( n 1) 2 5 ( n 1)( n 2)
The complete or general solution of the Bessel’s differential equation is y = c 1y1 + c2y2,
where c1, c2 are arbitrary constants.
o
Now we will proceed to find the solution in terms of Bessel’s function by choosing
1
s.c
a0 n
and let us denote it as Y1.
2 (n 1)
2 4
xn x 1 x 1
i.e., Y1 n
1
2 (n 1) 2 (n 1) 2 (n 1)(n 2) 2
x
2
n
1
(n 1)
c x
2
2
1
(n 1) (n 1)
x
2
4
1
(n 1)(n 2) (n 1) 2
vtu
We have the result (n) = (n – 1) (n – 1) from Gamma function
Hence, (n + 2) = (n + 1) (n + 1) and
(n + 3) = (n + 2) (n + 2) = (n + 2) (n + 1) (n + 1)
n 2 4
x 1 x 1 x 1
Y1
2 (n 1) 2 (n 2) 2 (n 3) 2
which can be further put in the following form
ww
n 0 2 4
x ( 1) 0
x ( 1)1
x ( 1) 2 x
Y1
2 (n 1) 0! 2 (n 2) 1! 2 (n 3) 2! 2
n 2r
x ( 1) r
x
2 r 0 (n r 1) r! 2
n 2r
r x 1
( 1)
r 0 2 (n r 1) r!
This function is called the Bessel function of the first kind of order n and is denoted by
Jn(x).
n 2r
r x 1
Thus J n ( x) ( 1)
r 0 2 (n r 1) r!
m
and B are arbitrary constants.
o
1. J n ( x ) ( 1 ) n J n ( x ) , where n is a positive integer.
s.c
Proof: By definition of Bessel’s function, we have
n 2r
x 1
J n ( x) ( 1) r
……….(1)
r 0 2 (n r 1) r!
n 2r
x 1
Hence, J n ( x ) ( 1) r
……….(2)
r 0 2 ( n r 1 ) r!
c
But gamma function is defined only for a positive real number. Thus we write (2) in the
following from
vtu
n 2r
x 1
J n( x ) ( 1 )r ………..(3)
r n 2 ( n r 1 ) r!
Let r – n = s or r = s + n. Then (3) becomes
n 2s 2n
x 1
J n( x ) ( 1 )s n
s 0 2 ( s 1 ) ( s n )!
w.
Comparing the above summation with (1), we note that the RHS is Jn(x).
n 2r
r x 1
Proof: By definition, J n ( x) ( 1)
r 0 2 (n r 1) r!
n 2r
x 1
Jn( x ) ( 1 )r
r 0 2 ( n r 1 ) r!
n 2r
n 2r x 1
i.e., ( 1 )r 1
r 0 2 ( n r 1 ) r!
n 2r
n x 1
1 ( 1 )r
2 ( n r 1 ) r!
m
r 0
o
Recurrence Relations:
s.c
Recurrence Relations are relations between Bessel’s functions of different order.
d n
Recurrence Relations 1: x Jn( x ) x n J n 1( x )
dx
From definition,
n 2r 2( n r )
c x 1 x 1
xn Jn( x ) xn ( 1 )r ( 1 )r
r 0 2 ( n r 1 ) r! r 0 2 ( n r 1 ) r!
2( n r )x 2( n r) 1
vtu
d n
x Jn( x ) ( 1 )r
dx r 0 2n 2r
( n r 1 ) r!
n r ( n r )x n 2r 1
x ( 1)
r 0 2n 2r 1
( n r ) ( n r ) r!
(x / 2) n 1 2r
xn ( 1 )r x n J n 1( x )
w.
r 0 ( n 1 r 1 ) r!
d n
Thus, x Jn( x ) x n J n 1( x ) --------(1)
dx
d
Recurrence Relations 2: x n J n ( x) x n J n 1 ( x)
dx
ww
From definition,
n 2r
n n x 1
x Jn( x ) x ( 1 )r
r 0 2 ( n r 1 ) r!
2r
x 1
( 1 )r
r 0 2 ( n r 1 ) r!
d n 2r x 2 r 1
x Jn( x ) ( 1 )r
dx r 0 2n 2r
( n r 1 ) r!
n xn 1 2( r 1 )
x ( 1 )r 1
r 1 2n 1 2( r 1 )
( n r 1 ) ( r 1 )!
Let k = r – 1
n xn 1 2k
x ( 1 )k x n
J n 1( x )
k 0 2n 1 2k
( n 1 k 1 ) k!
d n n
Thus, x Jn( x ) x J n 1( x ) --------(2)
dx
x
Recurrence Relations 3: J n ( x ) J n 1( x ) J n 1( x )
2n
m
d n
We know that x Jn( x ) x n J n 1( x )
dx
Applying product rule on LHS, we get x n J n/ ( x ) nx n 1 J n ( x ) x n J n 1 ( x )
Dividing by xn we get J n/ ( x ) ( n / x )J n ( x ) J n 1 ( x ) --------(3)
o
d n n
Also differentiating LHS of x Jn( x ) x J n 1( x ) , we get
dx
s.c
n
x J n/ ( x ) nx n 1
Jn( x ) x n
J n 1( x )
–n
Dividing by –x we get J n/ ( x )
( n / x )J n ( x ) J n 1 ( x ) --------(4)
Adding (3) and (4), we obtain 2nJ n ( x ) x J n 1 ( x ) J n 1 ( x )
x
i.e., Jn( x ) J n 1( x ) J n 1( x )
2n
1 c
Recurrence Relations 4: J n/ ( x ) J n 1( x ) J n 1( x )
2
Subtracting (4) from (3), we obtain 2 J n/ ( x ) J n 1 ( x ) J n 1 ( x )
vtu
1
i.e., J n/ ( x ) J n 1( x ) J n 1( x )
2
n
Recurrence Relations 5: J n/ ( x ) J n ( x ) J n 1( x )
x
This recurrence relation is another way of writing the Recurrence relation 2.
n
Recurrence Relations 6: J n/ ( x ) J n 1 ( x ) Jn( x )
x
w.
Problems:
2 2
Prove that (a) J 1 / 2 ( x) sin x (b) J 1/ 2 ( x) cos x
x x
By definition,
n 2r
x 1
Jn( x ) ( 1 )r
r 0 2 ( n r 1 ) r!
Putting n = ½, we get
1 / 2 2r
x 1
J1/ 2( x ) ( 1 )r
r 0 2 ( r 3 / 2 ) r!
2 4
x 1 x 1 x 1
J1/ 2( x ) --------(1)
2 (3 / 2) 2 ( 5 / 2 )1! 2 ( 7 / 2 )2!
m
x 2 x2 4 x 4 8
J1/ 2( x )
2 4 3 16 15 .2
x 2 x3 x 5 2 x3 x5
x x
2 x 6 120 x 3! 5!
o
2
J1/ 2( x ) sin x
x
s.c
Putting n = - 1/2, we get
1 / 2 2r
x 1
J 1/ 2( x ) ( 1 )r
r 0 2 ( r 1 / 2 ) r!
2 4
x 1 x 1 x 1
J 1/ 2( x ) --------(2)
2 (1 / 2 ) 2 ( 3 / 2 )1! 2 ( 5 / 2 )2!
2 x2 x4
1
x 2! 4!
2
J 1/ 2( x ) cos x
x
2. Prove the following results :
w.
2 3 x2 3
( a ) J5/ 2( x ) 2
sin x cos x and
x x x
2 3 x2 3
(b) J 5/ 2( x ) cos x sin x
x x2 x
ww
Solution :
x
We prove this result using the recurrence relation J n ( x ) J n 1( x ) J n 1( x ) ------ (1).
2n
3
Putting n = 3/2 in (1), we get J 1 / 2 ( x ) J 5 / 2 ( x ) J3/ 2( x )
x
3
J5/ 2( x ) J3 / 2( x ) J1/ 2( x )
x
3 2 sin x x cos x 2
i .e., J 5 / 2 ( x ) sin x
x x x x
3
Also putting n = - 3/2 in (1), we get J 5/ 2( x) J 1/ 2( x) J 3/ 2( x)
x
3 3 2 x sin x cos x 2
J 5/ 2( x ) J 3/ 2( x ) J 1/ 2( x ) cos x
m
x x x x x
o
dx x
Solution:
s.c
d 2
L.H.S = J n ( x ) J n2 1 ( x ) 2 J n ( x )J n/ ( x ) 2 J n 1 ( x )J n/ 1 ( x ) ------- (1)
dx
We know the recurrence relations
xJ n/ ( x ) nJ n ( x ) xJ n 1 ( x )
------- (2)
xJ n/ 1 ( x )
xJ n ( x ) ( n 1 )J n 1 ( x ) ------- (3)
Relation (3) is obtained by replacing n by n+1 in xJ n/ ( x ) xJ n 1 ( x ) nJ n ( x )
2
Solution :
1
We have the recurrence relation J n/ ( x ) J n 1 ( x ) J n 1 ( x ) -------(1)
2
1 1
Putting n = 0 in (1), we get J 0/ ( x ) J 1( x ) J1( x ) J1( x ) J1( x ) J1( x )
ww
2 2
Thus, J 0/ ( x ) J 1 ( x ) . Differentiating this w.r.t. x we get, J 0// ( x ) J 1/ ( x ) ----- (2)
1
Now, from (1), for n = 1, we get J 1/ ( x ) J0 ( x ) J 2( x ) .
2
Using (2), the above equation becomes
1 1
J 0// ( x ) J 0 ( x ) J 2 ( x ) orJ 0// ( x )
J 2( x ) J0 ( x ) .
2 2
1
Thus we have proved that, J 0// ( x ) J 2( x ) J0 ( x )
2
2
5. Show that (a) J 3 ( x )dx c J 2 ( x ) J 1( x )
x
1 2 2
(b) xJ 02 ( x )dx x J 0 ( x ) J 12 ( x )
2
Solution :
d n n
(a) We know that x Jn( x ) x J n 1( x ) or x n J n 1 ( x )dx x n
J n ( x ) ------ (1)
dx
m
Now, J 3 ( x )dx x 2 x 2 J 3 ( x )dx c x2 x 2 J 3 ( x )dx 2 x x 2 J 3 ( x )dx dx c
x2 x 2J2( x ) 2 x x 2 J 2 ( x ) dx c ( from (1) when n = 2)
2 2
c J2( x )
J 2 ( x )dx c J 2 ( x ) J 1 ( x ) ( from (1) when n = 1)
x x
o
2
Hence, J 3 ( x )dx c J 2 ( x ) J 1 ( x )
x
1 1
(b) xJ 02 ( x )dx J 02 ( x ) x 2 2 J 0 ( x ) J 0/ ( x ). x 2 dx (Integrate by parts)
s.c
2 2
1 2 2
x J 0 ( x ) x 2 J 0 ( x ) J 1 ( x )dx (From (1) for n = 0)
2
1 2 2 d d
x J 0 ( x ) xJ 1 ( x ) xJ 1 ( x ) dx xJ 1 ( x ) xJ 0 ( x ) from recurrence relation (1)
2 dx dx
1 2 2 1 1 2 2
2
x J0 ( x )
2
xJ 1 ( x ) 2 c 2
x J 0 ( x ) J 12 ( x )
vtu
Generating Function for Jn(x)
x
(t 1/ t )
To prove that e 2 tnJn( x )
n
or
x
n (t 1/ t )
If n is an integer then Jn(x) is the coefficient of t in the expansion of e 2 .
w.
Proof:
x
(t 1/ t )
We have e 2 e xt / 2 e x / 2t
( xt / 2 ) ( xt / 2 ) 2 ( xt / 2 ) 3 ( xt / 2 ) ( xt / 2 ) 2 ( xt / 2 ) 3
1 1
1! 2! 3! 1! 2! 3!
ww
x
(t 1/ t )
Thus, e 2 tnJn( x )
n
x
(t 1/ t )
Result: e2 J0( x ) tn ( 1 )n t n
Jn( x )
n 1
Proof :
x
(t 1/ t ) 1
e2 tnJn( x ) tnJn( x ) tnJn( x )
m
n n n 0
t n J n( x ) J0 ( x ) t n J n( x ) J0 ( x ) t n ( 1 )n J n ( x ) t n J n ( x ) { J n ( x ) ( 1 ) n J n ( x )}
n 1 n 1 n 1 n 1
x
(t 1/ t )
o
Thus, e2 J0( x ) tn ( 1 )n t n
Jn( x )
n 1
Problem 6: Show that
s.c
1
(a) J n ( x ) cos( n x sin )d , n being an integer
0
1
(b) J 0 ( x ) cos( x cos )d
0
(c) J 02 2 J 12 2 J 22 J 32 1
Solution :
We know that e
x
2
(t 1/ t )
c
J0( x ) tn ( 1 )n t n
Jn( x )
vtu
n 1
J 0 ( x ) tJ 1 ( x ) t 2 J 2 ( x ) t 3 J 3 ( x ) t 1 J 1 ( x ) t 2
J 2( x ) t 3
J 3 ( x )
n
Since J n ( x ) ( 1 ) J n ( x ) , we have
x
(t 1/ t )
e2 J0 ( x ) J1( x ) t 1 / t J2( x ) t2
----- (1) 1/ t2 J3( x ) t 3 1/ t3
p p
Let t = cos + i sin so that t = cosp + i sinp and 1/t = cosp - i sinp .
w.
Since eixsin = cos(xsin ) + i sin(xsin ), equating real and imaginary parts in (2) we get,
cos( x sin ) J 0 ( x ) 2 J 2 ( x ) cos 2 J 4 ( x ) cos 4 ----- (3)
sin( x sin ) 2 J 1 ( x ) sin J 3 ( x ) sin 3 ----- (4)
These series are known as Jacobi Series.
Now multiplying both sides of (3) by cos n and both sides of (4) by sin n and
integrating each of the resulting expression between 0 and , we obtain
1 J n ( x ), n is even or zero
cos( x sin ) cos n d
0 0 , n is odd
1 0, n is even
and sin( x sin ) sin n d
0 J n ( x ), n is odd
, if p q
Here we used the standard result cos p cos q d sin p sin q d 2
0 0 0 , if p q
From the above two expression, in general, if n is a positive integer, we get
1 1
m
Jn( x ) cos( x sin ) cos n sin( x sin ) sin n d cos( n x sin )d
0 0
o
Integrating the above equation w.r.t from 0 to , we get
cos( x cos )d J 0 ( x ) 2 J 2 ( x ) cos 2 2 J 4 ( x ) cos 4
s.c
0 0
sin 2 sin 4
cos( x cos )d J0 ( x ) 2J 2 ( x ) 2J 4 ( x ) J0 ( x )
0 2 4 0
1
Thus, J 0 ( x ) cos( x cos )d
0
(c) Squaring (3) and (4) and integrating w.r.t. from 0 to and noting that m and n
being integers
c
vtu
cos 2 ( x sin )d J0 ( x ) 2 4 J2( x ) 2 4 J4( x ) 2
0 2 2
Adding, d J 02 ( x ) 2 J 12 ( x ) 2 J 22 ( x ) J 32 ( x )
0
Hence, J 02 2 J 12 2 J 22 J 32 1
w.
xJ n ( x )J n ( x )dx 1 / 2 1
0 Jn ( ) J n 1 ( ) 2 , if
2 2
Proof:
m
d d d( x )
Since u = Jn( x), u / Jn( x ) Jn( x ) J n/ ( x )
dx d( x ) dx
d
Similarly v = Jn( x) gives v / Jn( x ) J n/ ( x ) . Substituting these values in (3), we
dx
o
get
1 J n/ ( )J n ( ) J n ( )J n/ ( )
xJ n ( x )J n ( x )dx 2 2
------- (4)
s.c
0
If and are the two distinct roots of Jn(x) = 0, then Jn( ) = 0 and Jn( ) = 0, and hence
(4) reduces to xJ n ( x )J n ( x )dx 0 .
0
This is known as Orthogonality relation of Bessel functions.
1
c
When = , the RHS of (4) takes 0/0 form. Its value can be found by considering
as a root of Jn(x) = 0 and as a variable approaching to . Then (4) gives
J n/ ( )J n ( )
vtu
Lt xJ n ( x )J n ( x )dx Lt 2 2
0
n
J n/ ( ) J n ( ) J n 1 ( ). Since J n ( ) 0 , we have J n/ ( ) J n 1( )
1 1 / 2 1 2
Thus, (5) becomes Lt xJ n ( x )J n ( x )dx Jn ( ) J n 1( )
0 2 2
LEGENDRE’S POLYNOMIAL
ww
If n is a positive even integer, a0u(x) reduces to a polynomial of degree n and if n is positive odd
integer a1v(x) reduces to a polynomial of degree n. Otherwise these will give infinite series
called Legendre functions of second kind.
Polyninomials u(x),v(x) contain alternate powers of x and a general form of the polynomial that
represents either of them in descending powers of x and can be presented in the form
y f ( x) a n x n a n 2 x n 2 a n 4 x n 4 ......... F ( x)...................(1)
where F ( x) a 0 if n is even
a1 x if n is odd
We note that ar is the coefficient of xr in the series solution of the differential equation and we
have obtained
n(n 1) r (r 1)
ar 2 a r ....................................(2)
(r 2)(r 1)
We plan to express an-2,an-4.....present in (1) in terms of an. Replacing r by (n-2) in (2) we obtain
n(n 1) (n 2)(n 1)
an a n 2 ..
n(n 1)
m
( 4 n 2)
an an 2
n(n 1)
n(n 1)
an 2 an
o
2(2n 1)
Again from (2) on replacing r by (n-4) we obtain
n(n 1) (n 4)(n 3)
s.c
an 2 an 4
(n 2)(n 3)
8n 12
an 2 an 4
n 2)(n 3)
(n 2)(n 3)
an 4 an 2
an 2
4(2n 3)
8n 12
an 4
c
vtu
n 2)(n 3)
n(n 1)(n 2)(n 3)
an 4 a n by u sin g values of a n 2
2.4(2n 1)(2n 3)
U sin g these values in (1) we have
n(n 1) n n(n 1)(n 2)(n 3) n 4
y f ( x) an x n x 2
x .......... G ( x)
2(2n 1) 2.4(2n 1)(2n 3)
w.
whereG ( x) a 0 / a n if n is even
a1 x / a n if n is odd
If the constant an is so choosen such that y-f(x) becomes 1 when x=1, the polynomials are called
Legendre polynomials denoted by Pn(x).
ww
Let us choose
1.3.5.....2n 1
an , that is
n!
1.3.5.....2n 1 n n(n 1) n 2 n(n 1)(n 2)(n 3) n 4
Pn ( x) x x x .......... ...........(3)
n! 2(2n 1) 2.4(2n 1)(2n 3)
We obtain few Legendre polynomials by putting n=0,1,2,3,4,
1 1 1
P0 1, P1 ( x) x, P2 ( x) 3x 2 1 ., .P3 ( x) 5 x3 3x ., P4 ( x) 35 x 4 30 x 2 3 etc....
2 2 8
Rodrigues’ formula
m
nth derivative is a solution of the Legendre' sdifferential equation
Proof. Let (1 x 2 ) y '' 2 xy' n(n 1) y 0..................(1)
differentiation u w.r.t
o
du
u1 n( x 2 1) n 1 2 x i.e ( x 2 1)u1 2nxu
dx
s.c
Differentiating w.r.t x again we have
( x 2 1)u 2 2 xu1 2n( xu1 u )
Differentiate the result n times by applying Lebnitz theorem for the nth derivate
( x 2 1)u 2 n 2 xu1 n 2n xu1 n 2nu1
(x 2 1)u n 2 2 xu n 1 n 2un nu n 0
c
(1 x 2 )u n 2 2 xu n 1 n(n 1)u n 0...........(2)
Comparing (2) with (1) we conclude that un is a solution of the legendre‟s differential
vtu
equation.Also Pn(x) which satisfies the legendre differential equation is also a polynomial of
degree n . Hence un must be the same as Pn(x) for some constant factor k
n
Pn ( x) ku n k x 2 1 n
n n
Pn ( x) n k x 1 x 1 n apply inf leibnitztheoremforRHS
n n n n(n 1) n n
Pn ( x) k x 1 x 1 n n.n( x 1) n 1
x 1 n 1 n(n 1)( x 1) n 2
x 1 n 2 .. x 1 n ...(3)
w.
2!
We proceed to find k by choosing a suitable value for x. putting x=1 in (3) in RHS becomes zero
expect the last term
1
Pn (1) k .n!.2 n or k
n!2 n
ww
1 n
Pn ( x) ku n , where Pn ( x) n
x2 1 n
n!2
1 dn n
Thus proved that Pn ( x) n n
x 2 1 .....(Rodrigue' sFormula.)
n!2 dx
PROBLEMS:
1. If x3 2 x 2 x 1 aP0 ( x) bP1 ( x) cP2 ( x) dP3 ( x)
Find the values of a, b, c, d
so ln :
Let f ( x) x3 2 x 2 x 1
substitutuing for x 3 , x 2 , x,1int erms of legendre polynomial we have
m
2 4 2 5
f ( x) P3 ( x) P2 ( x) P1 ( x) P0 ( x)
5 3 5 3
hence we have
o
aP0 ( x) bP1 ( x) cP2 ( x) dP3 ( x)
5 2 4 2
P0 ( x) P1 ( x) P2 ( x) P3 ( x)
s.c
3 5 3 5
thus by comparing both sides we obtain
5 2 4 2
a ,b ,c ,d
3 5 3 5
8 10 4
2 Showthat x 4 3 x 2 x c P4 ( x) P2 ( x) P1 ( x) P0 ( x)
35 7 5
so ln :
vtu
Let f ( x) x 4 3x 2 x and we have obtained
8 4 1
x4 P4 ( x) P2 ( x) P0 ( x)
35 7 5
2 1
x2 P2 ( x) P0 ( x)
3 3
substituting thesein f ( x) with x P1 ( x) we have
w.
8 4 1 2 1
f ( x) P4 ( x) P2 ( x) P0 ( x) 3 P2 ( x) P0 ( x) P1 ( x)
35 7 5 3 3
thus
8 10 4
f ( x) P4 ( x) P2 ( x) P1 ( x) P0 ( x)
ww
35 7 5
4. Obtain P3(x) from Rodrigue‟s formula and verify that the same satisfies the Legendre‟s
equation in the standard form.
so ln :
1
P3 ( x) (5 x3 3x)
2
we have Legendre ' s equation
m
(1 x 2 ) y '' 2 xy ' n(n 1) y 0
We haveto verify that
(1 x 2 ) P3'' ( x) 2 xP '3 ( x) 3(3 1) P3 ( x) 0,sin ce n 3
o
Fromthe exp ression of P3 ( x) we get
1
P '3 ( x ) (15 x 2 3) and P3'' ( x) 15 x
s.c
2
(1 x 2 ) P3'' ( x) 2 xP '3 ( x) 12 P3 ( x)
1 1
(1 x 2 )15 x 2 x (15 x 2 3) 12 (5 x 3 3 x)
2 2
3 3 3
15 x 15 x 15 x 3x 30 x 18 x 0
3. showthat P2 (cos )
c
Thus we have verified P3 ( x) satifies Legendre ' s equation
1
(1 3cos 2 )
vtu
4
so ln :
1
We have P2 ( x) (3x 2 1)
2
1
now P2 (cos ) (3cos 2 1)
2
w.
1
P2 (cos ) 3 3cos 2 2
4
thus
1
P2 (cos ) (1 3cos 2 )
ww
Unit VI
Probability Theory I
PROBABILITY:
Random experiment:
m
performed repeatedly giving different results on outcomes are called random experiment.
Eg: tossing a coin, throwing a die.
Sample space:
o
Sample space of random experiment set of all possible outcomes & it is denoted by „S‟ . The
number of elements in the set is denoted by 0(S).
s.c
Event : event is a subset of sample space & is denoted by E
Exhaustive event: The set of events is said to be exhaustive if it includes all possible events.
Mutually exclusive events: Two events A & B are mutually exclusive, if A & B cannot happen
simultaneously A B
A & B are disjoint i.e. p A
c
B 0
vtu
Mutually independent events: Two events A & B are mutually independent if the occurrence of
the event A does not depend on the occurrence of the event B.
Probability: If an event A can happen in M ways out of the possible n-ways (mutually exclusive
& equally likely) then probability of A is denoted by P(A)
m favourable number of cases
P( A)
w.
P( A) P(A)=1
Axioms of probability:
i) for an event A of S, probability lies
ii) between 0 P( A) 1 The numerical value of probability lies between 0 & 1, P(S)=1
iii) P( A B) P( A) P( B) ; A & B are disjoint
A B A B A
P A B P A B A A and B A are disjoint
P A P B A
Add & subtract P A B
m
P A B P A P B A P A B P A B
P A B P A P B P A B
similarly,
o
P A B C P A P B P C P A B P B C P C A P A B C
If A, B, C are mutually exclusive, then
s.c
P A B C P A P B P C
Conditional probability: let A & B are two events, probability of the happening of event B when
the event A has already occurred is called Conditional probability & is denoted by P B A
P B
A
c
Probability of occurance of both A & B
Probability of occurance of given event A
vtu
P A B
P B Multiplication rule of probability
A P A
P A B P B .P A
A
If A & B are Mutually independent event then, P B P B
A
w.
P A B P B .P A
Problems:
1. A boy & girl appeared in interview for 2 vacancy‟s in the same post. The probability of boy
ww
& girl selection is 1 & that of girl selection is 1 ,what is the probability that
7 5
(i) both will be selected
(ii) None of them will be selected
(iii) one of them will be selected
(iv) atleast one of them will be selected.
Sol: let A & B be the events of selection of boy girl respectively
1 1
P A , P B
7 5
1 1 12
Total space
7 5 35
i) Probability of both of them getting selected
(i) P A B P A .P B
m
1 1 1
= . =
7 5 35
(ii) probability of none of them getting selected
o
(ii) P A B P A .P B
1 P A 1 P B
s.c
1 1
= 1 1
7 5
= 0.685
(iii) Probability of one of them will be selected is
P A B P A .P B + P B .P A
c
1 4 1 6
=
vtu
7 5 5 7
4 6
= 0.285
35 35
(iv)probability of atleast one of them will be selected
=1 P A B
=1 0.685
w.
=0.315
Probability that a cricket team wins a match is 3 . If the team wins 3 matches in the
5
ww
3
P C
5
i )Probability that team wins all matches
P W1 P W2 P W3 3
5
3 3 3 27
= . . 0.216
5 5 5 125
m
(ii )probability of team loosing all matches
3 3 3
= 1 1 1
5 5 5
o
2 2 2 8
= 0.064
5 5 5 125
s.c
iii ) probability that it wins atleast one match
=1 0.064
=0.936
iv) =P W1 P W2 P W3 P W1 P W2 P W3 P W1 P W2 P W3
=
3 2 2
5 5 5
c 2
5
3
5
2
5
2
5
2
5
3
5
vtu
12 36
=3 0.2888
125 125
BAYES THEORM:
Statement:
If A1 , A2 ,........... An are partition of a set 's' so that there union in s
& E be any other event, then
w.
E
P P Ai
Ai Ai
P n
E E
P P Ai
Ai
ww
i=1
Proof:
S A1 A2 A3 .............. An
E E S
E E A1 A2 A3 .............. An
= E A1 E A2 ....................... E An
P E P E A1 E A2 ....................... E An
m
since A1 , A2 , A3 ,.............. An are mutually disjoint set
P E P E A1 P E A2 ............... P E An
o
n
P E P E Ai (1)
i 1
s.c
By definition of conditional probability
Ai P Ai E
P
E P E
P E Ai
=
P E c
P E Ai
= n
from (1)
vtu
P E Ai
i 1
E
P P A
Ai Ai
P n
E
P E Ai
i 1
w.
Problem;
1. In a college boys & girls are equal in proportion. It was found that 10 out of 100
boys & 25 out of 100 girls were using same company of 2 wheeles If the student
ww
using that was selected at random. What is the probability of being a boy.
Sol:
E 10
P 0.1; P A 100% 0.5
A 100
E 25
P 0.25; P A 100% 0.5
B 100
Let E be an event of choosing a student from same company of two wheelers
E E
P E P A .P P B .P
A B
= 0.5 0.1 0.5 0.25
P E =0.35
probability of choosing a student from boy
m
E
P A .P
A A
P
E P E
o
1 0.1
=
0.35
=0.2857
s.c
2. In a school 25% of the students failed in first language, 15% of the students failed
in second language and 10% of the students failed in both. If a student is selected at
random find the probability that
(i) He failed in first language if he had failed in the second language.
(ii)
(iii)
c
He failed in second language if he had failed in the first language.
He failed in either of the two languages.
vtu
Solution: Let A be set of students failing in the first language and B be the set of students failing
in the second language./ We have by data
P(A) = , P(B) = , P(A∩B) =
(i) P(A/B) = = =
(ii) P(B/A) = =
w.
3. The probability that a team wins a match is 3/5. If this team play 3 matches in a
tournament, what is the probability that the team
(i) Win all the matches
(ii) Win atleast one match
(iii) Win atmost one match
(iv) Lose all the matches
m
= 1-(8/25) =17/25
iii) Probability of winning atmost one match.
= P(L1) P(L2) P(L3)+ P(W1) P(L2) P(L3)+ P(L1) P(W2) P(L3)+ P(L1) P(L2) P(W3)
= =
o
IV) Probability of losing all the matches
s.c
= P(L1) P(L2) P(L3) =
4. The odds that a book will reviewed favourably by 3 independent critics are 5 to 2, 4
to 3 and 3 to 4. Find the probability that majority of the reviews will be favourable.
Solution: Let E1 , E2, E3 be the events of favourable review by the three critics respectively.
c
Therefore P( ) = 5/7 , P(E2)=4/7 , P(E3) = 3/7
Hence P( ) = 2/7 , P( )=3/7 , P( ) =4/7
vtu
Majority of the reviews are favourable means that atleast two of three reviews should
be favourable and if E denotes this event then we have
P(E) = P( ) P(E2) P( ) + P( ) P(E2) P(E3)+ P( ) P( P(E3)+ P( ) P(E2) P(E3)
=
5. Three machines A, B, C produces 50%, 30% and 20% of the items in factory.
w.
The
percentage of defective outputs are 3, 4, 5. If an item is selected at random. What is
the probability that it is defective? What is the probability that it is from A?
= =0.4054
6. In a college where boys and girls are equal proportion, it was found that 10 out of 100
boy and 25 out of 100 girls were using the same brand of a two wheeler. If a student
using that was selected at random what is the probability of being a boy?
m
Now, P(E) = P(B) P(E/B)+ P(G) P(E/G) = 0.175
We have to find P(B/E) AND BY Baye‟s theorem
P(B/E) = = = 0.2857
o
c s.c
vtu
w.
ww
Unit VII
Probability Theory II
In a practical situation, one may be interested in finding the probabilities of all the events
and may wishes to have the results in a tabular form for any future reference. Since for
m
an experiment having n outcomes, totally, there are 2n totally events; finding
probabilities of each of these and keeping them in a tabular form may be an interesting
problem.
o
Thus, if we develop a procedure, using which if it is possible to compute the probability
of all the events, is certainly an improvement. The aim of this chapter is to initiate a
s.c
discussion on the above.
Also, in many random experiments, outcomes may not involve a numerical value. In
such a situation, to employ mathematical treatment, there is a need to bring in numbers
into the problem. Further, probability theory must be supported and supplemented by
c
other concepts to make application oriented. In many problems, we usually do not show
vtu
interest on finding the chance of occurrence of an event, but, rather we work on an
experiment with lot of expectations
Let S denote the sample space of a random experiment. A random variable means it is
a rule which assigns a numerical value to each and every outcome of the experiment.
Thus, random variable may be viewed as a function from the sample space S to the set
ww
Note that all the outcomes of the experiment are associated with a unique number.
Therefore, f is an example of a random variable. Usually a random variable is denoted
m
by using upper case letters such as X, Y, Z etc. The image set of the random variable
may be written as f ( S ) {0, 1, 2, 3} .
A random variable is divided into
o
Discrete Random Variable (DRV)
s.c
Continuous Random Variable (CRV).
If the image set, X(S), is either finite or countable, then X is called as a discrete random
variable, otherwise, it is referred to as a continuous random variable i.e. if X is a CRV,
then X(S) is infinite and un – countable.
c
vtu
Example of Discrete Random Variables:
1. Let X denote the random variable equals the speed of a moving car, say, from a
destination A to another location B, then it is known that speedometer indicates
the speed of the car continuously over a range from 0 up to 160 KM per hour.
m
Therefore, X is a CRV, Continuously Varying Random Variable.
2. Let X denotes the monitoring index of a patient admitted in ICU in a good
hospital. Then it is a known fact that patient’s condition will be watched by the
o
doctors continuously over a range of time. Thus, X is a CRV.
3. Let X denote the number of minutes a person has to wait at a bus stop in
s.c
Bangalore to catch a bus, then it is true that the person has to wait anywhere
from 0 up to 20 minutes (say). Will you agree with me? Since waiting to be done
continuously, random variable in this case is called as CRV.
4. Results of any experiments accompanied by continuous changes at random over
c
a range of values may be classified as a continuous random variable.
vtu
Probability function/probability mass function f xi P X xi of a discrete
random variable:
f xi P (2 X 5) f x1 f x2 f x3 . . . f xn
ww
properties:
f xi 0
0 f xi 1
f xi 1
i
On the other hand, X is a continuous random variable, then its probability function will
be usually given or has a closed form, given as f ( x ) P( X x ) where x is defined over
a range of values., it is called as probability density function usually has some standard
m
form. This function too has the following properties:
f ( x) 0
0 f ( x) 1
o
f ( x) 1 .
s.c
To begin with we shall discuss in detail, discrete random variables and its distribution
functions. Consider a discrete random variable, X with the distribution function as
given below:
X xi x1
c x2 x3 . . . xn
vtu
f xi f x1 f x2 f x3 . . . f xn
Using this table, one can find probability of various events associated with X. For
example,
P xi X xj P X xi P X xi 1 up to + P X xj
w.
f xi f xi 1 f xi 2 + up to +f x j 1 +f x j
P xi X xj P X xi 1 P X xi 2 . . .+ P X xj 1
ww
f xi 1 f xi 2 + up to +f x j 1
P X xj 1 P X xj 1 1 P X x1 P X x2 up to +P X x j -1
0 F ( x) 1
when x , F ( x ) approaches 1
when x , F ( x) approaches 0
A brief note on Expectation, Variance, Standard Deviation of a Discrete Random
m
Variable:
i n
E(X ) xi f xi
i 1
o
i n
E X2 xi2 f xi
i 1
s.c
2
Var( X ) E X2 E( X )
ILLUSTRATIVE EXAMPLES:
Find (i) k; (ii) F (4) ; (iii) P ( X 5) ; (iv) P (2 X 5) (v) E(X) and (vi) Var (X).
Solution: To find the value of k, consider the sum of all the probabilities which equals
1
w.
X: 0 1 2 3 4 5 6
1 3 5 7 9 11 13
f xi :
49 49 49 49 49 49 49
21
P[2 X 5] P[ X 2] P[ X 3] P[ X 4] . Next to find E(X), consider
49
203
E( X ) xi f xi . To obtain Variance, it is necessary to compute
i 49
973
E X2 xi 2 f xi . Thus, Variance of X is obtained by using the
m
i 49
2
2 2 973 203
relation, Var( X ) E X E( X ) .
49 49
o
2. A random variable, X, has the following distribution function.
X: -2 -1 0 1 2 3
s.c
f xi : 0.1 k 0.2 2k 0.3 k
Next, P ( 2 X 2) P[ X 1] P[ X 0] P[ X 1] 0.5 .
2
Then E X 2 xi 2 f xi = 2.8. Var( X ) E X2 E( X ) = 2.8 - 0.64 = 2.16.
i
associated with X.
b b
P (a X b) f ( x ) dx , P (a X b) f ( x ) dx
a a
b x
m
P (a X b) f ( x ) dx , F ( x) P( X x) f ( x ) dx
a
E(X ) x f ( x ) dx , E X2 x 2 f ( x ) dx
o
2
Var( X ) E X2 E( X )
s.c
P (a X b) F (b) F (a )
dF ( x )
f ( x) , if the derivative exists
dx
SOME STANDARD DISTRIBUTIONS OF A DISCRETE RANDOM VARIABLE:
c
Binomial distribution function: Consider a random experiment having only two
vtu
outcomes, say success (S) and failure (F). Suppose that trial is conducted, say, n
number of times. One might be interested in knowing how many number of times
success was achieved. Let p denotes the probability of obtaining a success in a single
trial and q stands for the chance of getting a failure in one attempt implying that p + q =
1. If the experiment has the following characteristics;
w.
the probability of obtaining failure or success is same for each and every trial
trials are independent of one another
probability of having a success is a finite number, then
ww
We say that the problem is based on the binomial distribution. In a problem like this,
we define X as the random variable equals the number of successes obtained in n trials.
Then X takes the values 0, 1, 2, 3 . . . up to n. Therefore, one can view X as a
discrete random variable. Since number of ways of obtaining k successes in n trials
n n!
may be achieved in , therefore, binomial probability function may be
k k! n k !
n k nk
formulated as b(n, p, k ) p q .
k
Illustrative examples:
m
1. It is known that among the 10 telephone lines available in an office, the chance
that any telephone is busy at an instant of time is 0.2. Find the probability that (i)
exactly 3 lines are busy, (ii) What is the most probable number of busy lines and
o
compute its probability, and (iii) What is the probability that all the telephones are
busy?
s.c
Solution:
Here, the experiment about finding the number of busy telephone lines at an instant of
time. Let X denotes the number of telephones which are active at a point of time, as
c
there are n = 10 telephones available; clearly X takes the values right from 0 up to 10.
Let p denotes the chance of a telephone being busy, then it is given that p = 0.2, a finite
vtu
value. The chance that a telephone line is free is q = 0.8. Since a telephone line being
free or working is independent of one another, and since this value being same for each
and every telephone line, we consider that this problem is based on binomial
distribution. Therefore, the required probability mass function is
10
w.
10
(i) To find the chance that 3 lines are busy i.e. P[X = 3] = b(10, 0.2,3) (0.2)3 (0.8)7
3
ww
(ii) With p = 0.2, most probable number of busy lines is n p 10 0.2 2. The
10
probability of this number equals b(10, 0.2, 2) (0.2) 2 (0.8)8 .
2
(iii) The chance that all the telephone lines are busy = (0.2)10 .
2. The chance that a bomb dropped from an airplane will strike a target is 0.4. 6
bombs are dropped from the airplane. Find the probability that (i) exactly 2
bombs strike the target? (ii) At least 1 strikes the target. (iii) None of the bombs
hits the target?
Solution: Here, the experiment about finding the number of bombs hitting a target.
m
Let X denotes the number of bombs hitting a target. As n = 6 bombs are dropped from
an airplane, clearly X takes the values right from 0 up to 6.
Let p denotes the chance that a bomb hits a target, then it is given that p = 0.4, a finite
o
value. The chance that a telephone line is free is q = 0.6. Since a bomb dropped from
airplane hitting a target or not is an independent event, and the probability of striking a
s.c
target is same for all the bombs dropped from the plane, therefore one may consider
that hat this problem is based on binomial distribution. Therefore, the required
10
probability mass function is b(10, 0.4, k ) (0.4) k (0.8)6 k .
k
c
(i) To find the chance that exactly 2 bombs hits a target,
vtu
10
i.e. P[X = 2] b(10, 0.4, 2) (0.4) 2 (0.8) 4
2
(ii) Next to find the chance of the event, namely, at least 1 bomb hitting the target; i.e.
P[ X 1] 1 P[ X 1] 1 P[ X 0] 1 (0.6)6 .
(iii) The chance that none of the bombs are going to hit the target is P[X=0] = (0.6) 6 .
w.
n k nk
probability mass function given by b(n, p, k ) p q . Consider the expectation of X,
k
namely,
k n
n k n k
E( X ) k pq
k 0 k
k n
n!
k pk q n k
k 0 k! n k !
k n
n(n 1)!
pp k 1q ( n 1 1 k)
k 0 (k 1)! n 1 1 k !
k n
(n 1)!
np p k 1q[( n 1 ( k 1)]
m
k 1 (k 1)![(n 1)! (k 1)!]
k n
n 1 k 1 [( n 1 ( k 1)]
np p q
k 2 k 1
o
np ( p q)n 1
s.c
np as p q 1
Thus, expected value of binomial distribution function is np .
To find variance of X, consider
k n
n k n
E X2 k2 pq k
c
k 0 k
k n
n k n
vtu
k
k (k 1 1) pq
k 0 k
k n k n
n! n k n
k (k 1) pk q n k
k p q k
k 0 k! n k ! k 0 k
k n
n(n 1)(n 2)!
p 2 p k 2 q[( n 2 ( k 2)]
E(X )
w.
n(n 1) p 2 ( p q) n 2
np . Since p +q = 1, it follows that
n(n 1) p 2 np
2
Therefore, Var( X ) E X2 - E( X )
2
n(n 1) p2 np np
n2 p 2 np 2 np n2 p 2
m
distributed random variable is Var( X ) npq .
o
A DISCUSSION ON POISSON DISTRIBUTION FUNCTION
s.c
This is a limiting case of the binomial distribution function. It is obtained by considering
that the number of trials conducted is large and the probability of achieving a success in
a single trial is very small i.e. here n is large and p is a small value. Therefore, Poisson
distribution may be derived on the assumption that n and p 0 . It is found that
Poisson distribution function is
e
c k
vtu
p( , k ) . Here, np and k 0, 1, 2, 3, . . . . .
k!
k 0 k 0
k k 1
e
k 0 ( k 1)!
k k 1 k k 1
e . But e , therefore it follows that for a
ww
k 1 (k 1)! k 1 (k 1)!
Poisson distribution function, E ( X ) . Next to find Variance of X, first consider
k
E X2 k 2 p( , k )
k 0
k k
e
k2
k 0 k!
k k
e
k (k 1 1)
k 0 k!
k k k k
e e
k (k 1) k
k 0 k! k 0 k!
k k
e
k (k 1) E( X )
m
k 0 k!
k k 2
2
e
k 0 (k 2)!
o
k k 2
2
e
(k 2)!
s.c
k 2
2
e e . Thus, E X 2 2
. Hence, Variance of the Poisson
2
distribution function is Var( X ) E X2 - E( X ) . The standard deviation is
Var( X ) c
Illustrative Examples:
vtu
1. It is known that the chance of an error in the transmission of a message
through a communication channel is 0.002. 1000 messages are sent
through the channel; find the probability that at least 3 messages will be
received incorrectly.
Solution: Here, the random experiment consists of finding an error in the transmission
w.
of a message. It is given that n = 1000 messages are sent, a very large number, if p
denote the probability of error in the transmission, we have p = 0.002, relatively a small
number, therefore, this problem may be viewed as Poisson oriented. Thus, average
ww
e 2 2k
is . p(2, k ) ,k 0, 1, 2, 3, . . . . Here, the problem is about finding the
k!
probability of the event, namely,
P( X 3) 1 P ( X 3) 1 {P[ X 0] P[ X 1] P[ X 2]}
k 2
e 2 2k
1
k 0 k!
2 2
1 e 1 2 2 1 5e
2. A car hire –firm has two cars which it hires out on a day to day basis. The
number of demands for a car is known to be Poisson distributed with mean
1.5. Find the proportion of days on which (i) There is no demand for the car
and (ii) The demand is rejected.
m
Solution: Here, let us consider that random variable X as the number of persons or
demands for a car to be hired. Then X assumes the values 0, 1, 2, 3. … . . . It is given
that problem follows a Poisson distribution with mean, 1.5 . Thus, required probability
o
1.5
e (1.5) k
mass function may be written as p(1.5, k ) .
k!
s.c
(i) Solution to I problem consists of finding the probability of the event, namely
1.5
P[X = 0] = e .
(ii) The demand for a car will have to be rejected, when 3 or more persons approaches
c
the firm seeking a car on hire. Thus, to find the probability of the event P[ X
(1.5)2
3].
vtu
1.5
Hence, P[ X 3] 1 P{ X 3] 1 P[ X 0,1, 2] = e 1 1.5 .
2
Illustrative examples based on Continuous Random Variable and it‟s Probability
Density Function
o
1. Suppose that the error in the reaction temperature, in C, for a controlled
w.
x2
, 1 x 2
f ( x) 3
ww
0 elsewhere.
x
Solution: Consider F ( x ) P( X x) f (t )dt
x x
Case (i) x -1 F ( x) f ( t )dt 0dt 0
m
2 2
t2 t3 8 1
0 dt 0 1 . Therefore,
1
3 9 1
9
o
0, x 1
3
x 1
F ( x) , 1 x 2.
9
s.c
1, x 2.
x2
2kxe , for x 0
2. If the p.d.f of a R.V. X having is given by f ( x )
c 0, for x 0.
Find (a) the value of k and (b) distribution function F(X) for X.
vtu
x2
WKT 2kxe dx 1
0
ke t dt 1(put x 2 t)
0
w.
t
ke 0
1
(0 k ) 1 k 1
ww
x
F ( x) P( X x) f ( t )dt 0, if x 0
0 x
f ( t )dt f ( t )dt , if x 0
0
x
2 2
x2
0 2te t dt ( e z )0x (1 e ).
0
2
1 e x , for x 0
F ( x)
0, otherwise.
3. Find the C.D.F of the R.V. whose P.D.F is given by
x
, for 0<x 1
2
m
1
, for 1<x 2
f ( x) 2
3 x
, for 2<x 3
o
2
0, otherwise
s.c
x 0
Solution: Case (i) x 0 F ( x) f ( t )dt 0dt 0
x 0 x
t x2
Case (ii) 0 < x 1 F ( x) f ( t )dt 0dt dt
0
2 4
1 x
t 1 2x 1
0 dt dt
0
2 1
2 4
0 1 2 x
Case (iv) 2 < x 3 F ( x) f ( t )dt f ( t )dt f (t )dt f (t )dt
w.
0 1 2
6x x2 5
F ( x)
4
Case (v) for x >3, F(x) = 1. Therefore,
ww
0, if x 0
2
x
, if 0<x 1
4
2x 1
F ( x) , if 1<x 2
4
6 x x2 5
, if 2<x 3
4
1, if x>3
m
If X denotes the number of years, find the probability that the I.C. will work
properly
(a) less than 8 years
(b) beyond 8 years
(c) anywhere from 5 to 7 years
o
(d) Anywhere from 2 to 5 years.
s.c
x 0, for x 3
Solution: We have F ( x ) f ( t )dt 9
0 1 , for x 3.
x2
8
9
For (a): P ( x 8) f ( t )dt 1 2 0.8594
0
8
c
For Case (b): P(x > 8) = 1 – P(x 8) = 0.1406
vtu
2 2
For Case (c): P (5 x 7) = F (7) – F (5) = (1-9/7 ) – (1-9/5 ) = 0.1763
2
For Case (d): P (2 x 5) = F (5) – F (2) = (1-9/5 ) – (0) = 0.64
4
F ( x) c( x 1) , 1 x 3
1, x 3.
d
Solution: We know that f ( x ) [ F ( x )]
dx
0, x 1
f ( x) 4c( x 1)3 , 1 x 3
0, x 3.
4c( x 1)3 , 1 x 3
f ( x)
0, otherwise
3 3
4c ( x 1)3 dx 1 c ( x 1)4 1
1 1
1
m
16c 1 c
16
Using this, one can give the probability function just by substituting the value of
c above.
o
A discussion on some standard distribution functions of continuously distributed
s.c
random variable:
e x x >0,
The exponential distribution function is given by, f ( x)
ww
0, otherwise.
x
The probability distribution function may be written as F ( x) f ( x )dx which may be
1 e x , if 0 x
computed as F( x ) .
0, otherwise.
e x e x 1 1
m
x 1 - 0
2 2
0
Consider E X 2 x 2 f ( x )dx
o
x2 e x dx
0
s.c
x x x
x2
e
2x
e
2
e 2
= 2 3 2
0
2 1
Var ( X ) E X2 - E( X ) .
2
minutes.
Solution: Let X denotes the random variable equals number of minute’s conversation
may last. It is given that X is exponentially distributed with mean 3 minutes. Since for
1 1
an exponential distribution function, mean is known to be , so 2 or =0.5 . The
ww
0.5e 0.5 x , if x 0 ,
Probability density function can now be written as f ( x ) .
0, otherwise.
(i) To find the probability of the event, namely,
3
P[ X 3] 1 P[ X 3] 1 0.5e 0.5 x dx
0
4
(ii) To find the probability of the event, namely P [ X 4] 0.5e 0.5 x dx .
0
5
(iii) To find the probability of the event P[3 X 5] 0.5e 0.5 x dx .
3
m
2. in a town, the duration of a rain is exponentially distributed with mean equal to
5 minutes. What is the probability that (i) the rain will last not more than 10
minutes (ii) between 4 and 7 minutes and (iii) between 5 and 8 minutes?
o
Solution: An identical problem to the previous one. Thus, may be solved on
similar lines.
s.c
Discussion on Gaussian or Normal Distribution Function
Among all the distribution of a continuous random variable, the most popular and widely
c
used one is normal distribution function. Most of the work in correlation and regression
analysis, testing of hypothesis, has been done based on the assumption that problem
vtu
follows a normal distribution function or just everything normal. Also, this distribution is
extremely important in statistical applications because of the central limit theorem,
which states that under very general assumptions, the mean of a sample of n mutually
Independent random variables (having finite mean and variance) are normally
w.
distributed in the limit n . It has been observed that errors of measurement often
possess this distribution. Experience also shows that during the wear – out phase,
component life time follows a normal distribution. The purpose of today’s lecture is to
have a detailed discussion on the same.
ww
The normal density function has well known bell shaped curve which will be shown on
2
1 x
1
the board and it may be given as f ( x) e2 , - x
2
x
distribution function, namely, F ( x) P( X x) f ( x) dx has no closed form, evaluation
computed numerically and recorded in special table called normal distribution table.
m
However, It pertain to the standard normal distribution function by choosing and
z
1 t2
and their entries are values of the function, Fz ( z ) e 2
dt. Since the
o
2
s.c
z
1
that Fz ( z ) f (t ) dt =1 - F ( z ) .
z
2
X b)
c
Thus, tabulations are done for positive values of z only. From this it is clear that
P(a X ) 1 P( X a) 1 F (a)
Note: Let X be a normally distributed random variable taking a particular value, x, the
w.
x
corresponding value of the standardized variable is given by z . Hence,
x
F ( x) P( X x) F
z
.
ww
Here, the experiment consists of finding the life of electric bulbs of a particular make
(measured in hours) from a lot of 2000 bulbs. Let X denotes the random variable equals
the life of an electric bulb measured in hours. It is given that X follows normal
m
= 1 Fz 1.8333 1 0.9664 0.0336
Therefore, number of electrical bulbs with life expectancy more than 2150 hours is
o
0.0336 2000 67 .
1950 2040
s.c
Next to compute the probability of the event; P ( X 1950 hours)=F
z 60
Therefore, in a lot of 2000 bulbs, number of bulbs with life expectancy less than 1950
hours is 0.0668 * 2000 = 134 bulbs. c
Finally, to find the probability of the event, namely,
vtu
P (1920 X 2060) F (2060) F (1920)
F 0.3333 F 2
z z
w.
F 0.3333 1 F 2
z z
m
(i) To find the probability of the event P[ X 44.5] 1 F (44.5)
44.5 37.6
1 Fz
4.6
o
1 Fz 1.5
s.c
1 0.9332 0.0668 .
(ii) To find the probability of the event, P[ X 35.0] F (33.5)
35.0 37.6
Fz
4.6
c Fz 0.5652
vtu
1 Fz 0.5652
1 0.7123 0.2877 .
(iii) Consider the probability of the event P[30.0 X 40.0]
F (40) F (30)
w.
40 37.6 30 37.6
Fz Fz
4.6 4.6
Fz (0.5217) Fz ( 1.6522)
0.6985 1 0.9505 0.6490
ww
m
Consider P[ X 240] 1 P[ X 240]
1 F (240)
o
240 200
1 Fz
16
s.c
1 Fz 2.5
= 1 – 0.9938
= 0.00621
c
Next consider P[X > 240 | X > 210]
vtu
P[ X 240 and X > 210]
P[ X 210]
P[ X 240 ] 1 P[ X 240 ]
P[ X 210] 1 P[ X 210]
w.
240 200
1 Fz
16 1 Fz (2.5)
210 200 1 Fz (0.625)
1 Fz
16
ww
1 0.9939
0.2335
1 0.73401
o m
s.c
c
vtu
w.
ww
Unit VIII
Sampling Theory
Statistical Inference is a branch of Statistics which uses probability concepts to deal with
uncertainty in decision making. There are a number of situations where in we come
m
across problems involving decision making. For example, consider the problem of
buying 1 kilogram of rice, when we visit the shop, we do not check each and every rice
grains stored in a gunny bag; rather we put our hand inside the bag and collect a
o
sample of rice grains. Then analysis takes place. Based on this, we decide to buy or
not. Thus, the problem involves studying whole rice stored in a bag using only a sample
s.c
of rice grains.
This topic considers two different classes of problems
1. Hypothesis testing – we test a statement about the population parameter from
which the sample is drawn.
2.
c
Estimation – A statistic obtained from the sample collected is used to estimate the
vtu
population parameter.
First what is meant by hypothesis testing?
This means that testing of hypothetical statement about a parameter of population.
Conventional approach to testing:
The procedure involves the following:
w.
1. First we set up a definite statement about the population parameter which we call it
as null hypothesis, denoted by H 0 . According to Professor R. A. Fisher,
Null Hypothesis is the statement which is tested for possible rejection under the
ww
As we are studying population parameter based on some sample study, one can not do
the job with 100% accuracy since sample is drawn from the population and possible
sample may not represent the whole population. Therefore, usually we conduct
analysis at certain level of significance (lower than 100%. The possible choices include
99%, or 95% or 98% or 90%. Usually we conduct analysis at 99% or 95% level of
m
significance, denoted by the symbol . We test H 0 against H1 at certain level of
significance. The confidence with which a person rejects or accepts H 0 depends upon
o
the significance level adopted. It is usually expressed in percentage forms such as 5%
or 1% etc. Note that when is set as 5%, then probability of rejecting null hypothesis
s.c
when it is true is only 5%. It also means that when the hypothesis in question is
accepted at 5% level of significance, then statistician runs the risk of taking wrong
decisions, in the long run, is only 5%. The above is called II step of hypothesis testing.
Critical values or Fiducial limit values for a two tailed test:
Sl. No
c Level of
significance
Theoretical Value
vtu
1 1% 2.58
2 2% 2.33
3 5% 1.96
w.
Critical values or Fiducial limit values for a single tailed test (right and test)
Tabulated value 1% 5% 10%
Setting a test criterion: The third step in hypothesis testing procedure is to construct a
test criterion. This involves selecting an appropriate probability distribution for the
particular test i.e. a proper probability distribution function to be chosen. Some of the
distribution functions used are t, F, when the sample size is small (size lower than 30).
However, for large samples, normal distribution function is preferred. Next step is the
computation of statistic using the sample items drawn from the population. Usually,
samples are drawn from the population by a procedure called random, where in each
and every data of the population has the same chance of being included in the sample.
Then the computed value of the test criterion is compared with the tabular value; as
m
long the calculated value is lower then or equal to tabulated value, we accept the null
hypothesis, otherwise, we reject null hypothesis and accept the alternate hypothesis.
Decisions are valid only at the particular level significance of level adopted.
o
During the course of analysis, there are two types of errors bound to occur. These are
(i) Type – I error and (ii) Type – II error.
s.c
Type – I error: This error usually occurs in a situation, when the null hypothesis is
true, but we reject it i.e. rejection of a correct/true hypothesis constitute type I error.
Type – II error: Here, null hypothesis is actually false, but we accept it. Equivalently,
accepting a hypothesis which is wrong results in a type – II error. The probability of
c
committing a type – I error is denoted by where
vtu
= Probability of making type I error = Probability [Rejecting H 0 | H 0 is true]
On the other hand, type – II error is committed by not rejecting a hypothesis when it is
false. The probability of committing this error is denoted by . Note that
Critical region:
A region in a sample space S which amounts to Rejection of H 0 is termed as critical
region.
One tailed test and two tailed test:
ww
This depends upon the setting up of both null and alternative hypothesis.
A note on computed test criterion value:
of achieving a success in a single trial, q stands for the chance that there is a
m
failure in the trial and n refers to the size of the sample.
2. The magnitude of standard error gives an index of the precision of the
parameter.
o
s.c
ILLUSTRATIVE EXAMPLES
1. A coin is tossed 400 times and the head turned up 216 times. Test the
hypothesis that the coin is un– biased?
Solution: First we construct null and alternate hypotheses set up H 0 : The coin is not
c
a biased one. Set up H1 : Yes, the coin is biased. As the coin is assumed be fair and
it is tossed 400 times, clearly we must expect 200 times heads occurring and 200 times
vtu
tails. Thus, expected number of heads is 200. But the observed result is 216. There is
a difference of 16. Further, standard error is npq . With p = ½, q = ½ and n =
If we choose 5% , then the tabulated value for a two tailed test is 1.96. Since, the
calculated value is lower than the tabulated value; we accept the null hypothesis that
coin is un – biased.
2. A person throws a 10 dice 500 times and obtains 2560 times 4, 5, or 6. Can this
ww
The die is unfair. We consider that problem is based on a two – tailed test. Let us
choose level of significance as 5% then, the tabulated value is 1.96. Consider
Expected value - observed result
computing test criterion, zcal ; here, as the dice is tossed
standard error
by a person 5000 times, and on the basis that die is fair, then chance of getting any of
2500 - 2560
obtaining 4 or 5 or 6 is 2500. Hence, zcal 1.7 which is lower than 1.96.
35.36
Hence, we conclude that die is a fair one.
m
3. A sample of 1000 days is taken from meteorological records of a certain
district and 120 of them are found to be foggy. What are the probable limits to the
percentage of foggy days in the district?
o
Solution: Let p denote the probability that a day is foggy in nature in a district as
s.c
120
reported by meteorological records. Clearly, p 0.12 and q = 0.88. With n =
1000
1000, the probable limits to the percentage of foggy days is given by p 3 pqn .
Using the data available in this problem, one obtains the answer as
c
0.12 3 0.12 88 1000 . Equivalently, 8.91% to 15.07%.
4. A die was thrown 9000 times and a throw of 5 or 6 was obtained 3240 times.
vtu
On the assumption of random throwing, do the data indicate that die is biased?
(Model Question Paper Problem)
Solution: We set up the null hypothesis as H0 : Die is un - biased. Also,
assumption that distribution is normally distributed, the tabulated value is 1.96. The
chance of getting each of the 6 numbers is same and it equals to 1/6 therefore chance
of getting either 5 or 6 is 1/3. In a throw of 9000 times, getting the numbers either 5 or
ww
1
6 is ×9000 = 3000 . Now the difference in these two results is 240. With p = 1/3, q =
3
Difference
2/3, n = 9000, S.E . npq = 44.72. Now consider the test criterion zcal =
S.E.
240
5.367 which is again more than the tabulated value. Therefore, we reject null
44.72
hypothesis and accept the alternate that die is highly biased.
m
distribution of a statistic is approximately normal and (ii) values given by the samples
are sufficiently close to the population value and can be used in its place for calculating
o
standard error. When the standard deviation of population is known, then S.E (X) =
n
where denotes the standard deviation of population . When the standard deviation of the
s.c
p
population is unknown, then S.E (X) = where is the standard deviation of the
n
sample.
Fiducial limits of population mean are:
c
95% fiducial limits of population mean are X 1.96
vtu
n
x-
99% fiducial limits of population mean are X 2.58 . Further, test criterion zcal
n S.E.
ILLUSTRATIVE EXAMPLES
w.
1. A sample of 100 tyres is taken from a lot. The mean life of tyres is found to
be 39, 350 kilo meters with a standard deviation of 3, 260. Could the
sample come from a population with mean life of 40, 000 kilometers?
ww
Establish 99% confidence limits within which the mean life of tyres is
expected to lie.
Solution: First we shall set up null hypothesis, H0 : 40, 000 , alternate hypothesis as
H1 : 40, 000 . We consider that the problem follows a two tailed test and
x-
expression for finding test criterion, zcal . Here, =40, 000, x 39, 350 and
S.E.
3,260
3, 260 , n = 100. S.E. = 326 . Thus, z cal 1.994. As this value is slightly
n 100
greater than 1.96, we reject the null hypothesis and conclude that sample has not come
from a population of 40, 000 kilometers.
The 99% confidence limits within which population mean is expected to lie is given as
m
x 2.58×S.E. i.e. 39,350±2.58×326 = (38, 509, 40, 191) .
2. The mean life time of a sample of 400 fluorescent light bulbs produced by a
company is found to be 1, 570 hours with a standard deviation of 150
o
hours. Test the hypothesis that the mean life time of bulbs is 1600 hours
against the alternative hypothesis that it is greater than 1, 600 hours at 1%
s.c
and 5% level of significance.
Solution: First we shall set up null hypothesis, H0 : 1, 600 hours , alternate hypothesis
as H1 : 1, 600 hours . We consider that the problem follows a two tailed test and
chose 5% .
c
Then corresponding to this, tabulated value is 1.96.
x-
Consider the
vtu
expression for finding test criterion, zcal . Here, =1, 600, x 1, 570, n = 400 ,
S.E.
150 hours so that using all these values above, it can be seen that z cal 4.0 which is
really greater than 1.96. Hence, we have to reject null hypothesis and to accept the
alternate hypothesis.
w.
S1 n1 x1 1
S2 n2 x2 2
then standard error of difference between the means of two samples S1 and S2
2 2
2 Difference of sample means
is S.E = 1
and the test criterion is Zcal . The rest
n1 n2 Standard error
of the analysis is same as in the preceding sections.
When the two samples are drawn from the same population, then standard error is
m
1 1 Difference of sample means
S.E = and test criterion is Zcal .
n1 n2 Standard error
o
When the standard deviations are un – known, then standard deviations of the two
s12 s22
samples must be replaced. Thus, S.E = where s1 and s2 are standard
s.c
n1 n2
deviations of the two samples considered in the problem.
ILLUSTRATIVE EXAMPLES
1. Intelligence test on two groups of boys and girls gave the following data:
Data
c Mean Standard
deviation
Sample size
vtu
Boys 75 15 150
Girls 70 20 250
Is there a significant difference in the mean scores obtained by boys and girls?
the mean scores obtained by boys and girls. The alternate hypothesis is considered as
H1 : Yes, there is a significant difference in the mean scores obtained by boys and girls.
We choose level of significance as 5% so that tabulated value is 1.96.
Difference of means
Consider zcal = . The standard error may be calculated as
ww
Standard Error
152 202 75 - 70
S.E = + =1.761 , The test criterion is zcal = 2.84 . As 2.84 is more than
150 250 1.761
1.96, we have to reject null hypothesis and to accept alternate hypothesis that there are
some significant difference in the mean marks scored by boys and girls.
m
50 50 14.14
calculated value is certainly lower than the two tabulated values, we accept the
hypothesis there is no significant difference in the make of the two bulbs produced by
the companies.
o
A discussion on tests of significance for small samples
s.c
So far the problem of testing a hypothesis about a population parameter was based on
the assumption that sample drawn from population is large in size (more than 30) and
the probability distribution is normally distributed. However, when the size of the
sample is small, (say < 30) tests considered above are not suitable because the
assumptions on which they are based generally do not hold good in the case of small
samples. IN particular, here one cannot assume that the problem follows a normal
c
distribution function and those values given by sample data are sufficiently close to the
population values and can be used in their place for the calculation of standard error.
vtu
Thus, it is a necessity to develop some alternative strategies to deal with problems
having sample size relatively small. Also, we do see a number of problems involving
small samples. With these in view, here, we will initiate a detailed discussion on the
same.
Here, too, the problem is about testing a statement about population parameter; i.e. in
ascertaining whether observed values could have arisen by sampling fluctuations from
some value given in advance. For example, if a sample of 15 gives a correlation
w.
coefficient of +0.4, we shall be interested not so much in the value of the correlation in
the parent population, but more generally this value could have come from an un –
correlated population, i.e. whether it is significant in the parent population. It is widely
accepted that when we work with small samples, estimates will vary from sample to
sample.
ww
Further, in the theory of small samples also, we begin study by making an assumption
that parent population is normally distributed unless otherwise stated. Strictly, whatever
the decision one takes in hypothesis testing problems is valid only for normal
populations.
Sir William Gosset and R. A. Fisher have contributed a lot to theory of small samples.
Sir W. Gosset published his findings in the year 1905 under the pen name “student”.
He gave a test popularly known as “t – test” and Fisher gave another test known as “z –
test”. These tests are based on “t distribution and “z – distribution”.
Gosset was employed by the Guinness and Son, Dublin bravery, Ireland which did not
permit employees to publish research work under their own names. So Gosset adopted
the pen name “student” and published his findings under this name. Thereafter, the t –
distribution commonly called student’s t – distribution or simply student’s distribution.
The t – distribution to be used in a situation when the sample drawn from a population is
m
of size lower than 30 and population standard deviation is un – known. The t – statistic,
i n 2
xi x
x i 1
t cal is defined as t cal n where S , x is the
o
n 1 12 S n 1
sample mean, n is the sample size, and x i are the data items.
s.c
The t – distribution function has been derived mathematically under the assumption of a
normally distributed population; it has the following form
1
2
t 2
f (t ) C 1 where C is a constant term and = n - 1 denotes the number of
c
degrees of freedom. As the p.d.f. of a t – distribution is not suitable for analytical
treatment. Therefore, the function is evaluated numerically for various values of t, and
vtu
for particular values of . The t – distribution table normally given in statistics text
books gives, over a range of values of , the probability values of exceeding by chance
value of t at different levels of significance. The t – distribution function has a different
value for each degree of freedom and when degrees of freedom approach a large
value, t – distribution is equivalent to normal distribution function.
The application of t – distribution includes (i) testing the significance of the mean of a
w.
random sample i.e. determining whether the mean of a sample drawn from drawn from
a normal population deviates significantly from a stated value (i.e. hypothetical value of
the populations mean) and (ii) testing whether difference between means of two
independent samples is significant or not i.e. ascertaining whether the two samples
comes from the same normal population? (iii) Testing difference between means of two
ww
• Note down the sample size, n and the number of degrees of freedom,
m
x
The test criterion is then calculated using the formula, t cal n
S
Later, the calculated value above is compared with tabulated value. As long as
the calculated value matches with the tabulated value, we as usual accept the
o
null hypothesis and on the other hand, when the calculated value becomes more
than tabulated value, we reject the null hypothesis and accept the alternate
s.c
hypothesis.
ILLUSTRATIVE EXAMPLES
1. The manufacturer of a certain make of electric bulbs claims that his bulbs have
a mean life of 25 months with a standard deviation of 5 months. Random
samples of 6 such bulbs have the following values: Life of bulbs in months: 24,
20, 30, 20, 20, and 18. Can you regard the producer‟s claim to valid at 1% level of
c
significance? (Given that t tab 4.032 corresponding to 5 ).
Solution: To solve the problem, we first set up the null hypothesis H0 : 25 months ,
vtu
alternate hypothesis may be treated as H0 : 25 months . To set up 1% , then
tabulated value corresponding to this level of significance is
th
t tab | 1% and =5 4.032 (4.032 value has been got by looking in the 5 row ) . The test criterion
i n 2
x xi x
is given by t cal n where S i 1 .
S n 1
w.
ww
Consider
2
xi x xi x xi x
24 1 1
26 3 9
30 7 49
20 23 -3 9
20 -3 9
m
18 -5 25
o
102 23 25
Thus, S 20.4 4.517 and t cal 6 1.084 . Since the calculated
5 4.517
s.c
value, 1.084 is lower than the tabulated value of 4.032; we accept the null hypothesis as
mean life of bulbs could be about 25 hours.
blood pressure readings before and after the injection of the drug. The alternate
hypothesis is H0 : before after i.e. the stimulus resulted in an increase in the blood
2
xi x xi x xi x
5 2 4
2 -1 1
8 5 25
-1 -4 16
3 0 0
m
0 -3 9
-2 -5 25
o
1 -2 4
3
5 2 4
s.c
0 -3 9
4 1 1
6 3 9
8
c 5 25
vtu
Total = 39 - Total = 132
i n 2
xi x
i 1 132 x
Consider S 11 3.317 . Therefore, t cal n may be
n 1 12 S
0 3
obtained as t cal 13 3.2614 . As the calculated value 3.2614 is more than the
w.
3.317
tabulated values of 3.055 and 2.179, we accept the alternate hypothesis that after the
drug is given to patients, there is an increase in the blood pressure level.
3. the life time of electric bulbs for a random sample of 10 from a large
consignment gave the following data: 4.2, 4.6, 3.9, 4.1, 5.2, 3.8, 3.9, 4.3, 4.4, 5.6 (in
ww
„000 hours). Can we accept the hypothesis that the average life time of bulbs is 4,
000 hours?
2
xi x xi x xi x
m
5.2 0.8 0.64
4.4
3.8 -0.6 0.36
o
3.9 -0.5 0.25
s.c
4.4 0.0 0.0
5.6 1.2 1.44
4.4 4.0
t cal 10 = 2.148. As the computed value is lower than the tabulated value of
0.589
w.
2.262, we conclude that mean life of time bulbs is about 4, 000 hours.
2
A discussion on test and Goodness of Fit
ww
Recently, we have discussed t – distribution function (i.e. t – test). The study was
based on the assumption that the samples were drawn from normally distributed
populations, or, more accurately that the sample means were normally distributed.
Since test required such an assumption about population parameters. For this reason,
A test of this kind is called parametric test. There are situations in which it may not be
possible to make any rigid assumption about the distribution of population from which
one has to draw a sample.
Thus, there is a need to develop some non – parametric tests which does not require
any assumptions about the population parameters.
2
With this in view, now we shall consider a discussion on distribution which does
not require any assumption with regard to the population. The test criterion
2
Oi Ei
2 i
corresponding to this distribution may be given as where
Ei
RT CT
Ei : Oi : Observed values , Ei : Expected values .
m
N
When Expected values are not given, one can calculate these by using the following
RT CT
relation; Ei : . Here, RT means the row total for the cell containing the row, CT
o
N
is for the column total for the cell containing columns, and N represent the total number
of observations in the problem.
s.c
2
The calculated value (i.e. test criterion value or calculated value) is compared
2
with the tabular value of value for given degree of freedom at a certain prefixed
level of significance. Whenever the calculated value is lower than the tabular value, we
continue to accept the fact that there is not much significant difference between
c
expected and observed results.
On the other hand, if the calculated value is found to be more than the value suggested
in the table, then we have to conclude that there is a significant difference between
vtu
observed and expected frequencies.
For the usage of test, the following conditions must checked before employing the test.
These are:
ww
ILLUSTRATIVE EXAMPLES
DEPT. OF MATHS/SJBIT Page 129
ENGG. MATHEMATICS-IV 10MAT41
1. From the data given below about the treatment of 250 patients suffering from a
disease, state whether new treatment is superior to the conventional test.
m
New one 140 30 170
Conventional 60 20 80
o
Total 200 50 280
s.c
Solution: We set up null hypothesis as there is no significance in results due to the
two procedures adopted. The alternate hypothesis may be assumed as there could be
some difference in the results. Set up level of significance as
2 2 2
112 100 71 50 32 10
+ 5% then tabulated value is
100 50 10
2
| 0.05, 1 3.841 .
c
vtu
RT CT
Consider finding expected values given by the formula, Expectation(AB)
N
where RT means that the row total for the row containing the cell, CT means that the
total for the column containing the cell and N, total number of frequencies. Keeping
these in view, we find that expected frequencies are
w.
136 34 170
ww
A 64 16 80
200 50 250
2 2
Oi Ei Oi Ei Oi Ei Oi Ei
Ei
60 64 -4 16 0.250
30 34 -4 16 0.471
20 16 4 16 1.000
m
Total 1.839
o
As the calculated value 1.839 is lower than the tabulated value | 0.05, 1 3.841 , we
accept the null hypothesis, namely, that there is not much significant difference between
the two procedures.
No. of heads
Frequency
c 0
6
1
s.c
2. A set of five similar coins is tossed 320 times and the result is
27
2
72
3
112
4
71
5
32
vtu
Test the hypothesis that the data follow a binomial distribution function.
Solution: We shall set up the null hypothesis that data actually follows a binomial
distribution. Then alternate hypothesis is, namely, data does not follow binomial
distribution. Next, to set up a suitable level of significance, 5% , with n = 6, degrees
2
of freedom is 5. Therefore, the tabulated value is | 0.05, 5 11.07 . Before
w.
proceeding to finding test criterion, first we compute the various expected frequencies.
As the data is set to be following binomial distribution, clearly probability density function
n k nk
is b n, p, k p q . Here, n 320, p 0.5, q 0.5 , and k takes the values
k
ww
The expected values are: Ei : 10, 50, 100, 100, 50, 10 . Consider
2 2 2
2 6 10 27 50 72 100
|cal +
10 50 100
2 2 2
112 100 71 50 32 10
+ = 78.68. As the calculated
100 50 10
m
value is very much higher than the tabulated value of 3.841, we reject the null
hypothesis and accept the alternate hypothesis that data does not follow the binomial
distribution.
o
c s.c
vtu
w.
ww