Lecture 6

Numerical Methods
[MA-200]
The False Position Method (Regula-Falsi)
• Historically this method is known as “Regula Falsi” which is Latin and means: the
wrong rule. It is also known as method of false position

• The false position method is a numerical method to solve equations of the form
𝑓 𝑥 = 0.

• The aim is to find an approximation 𝑥𝑟 of a root r of the function 𝑓.

• Bisection algorithm cuts the interval into two equal segments, while the false
position method will try to generate a new interval in a faster way.
 Working procedure
i. f(x) is continuous on [a , b]

ii. If 𝑓 𝑎 . 𝑓(𝑏) < 0, then 𝑓(𝑥)=0 has solution in interval [a , b]

iii. Calculate
𝑎𝑓 𝑏 − 𝑏𝑓(𝑎)
𝑥1 =
𝑓 𝑏 − 𝑓(𝑎)
Find 𝑓 𝑥1 and observe the sign of the function
If 𝑓 𝑥1 < 0 then replace 𝑎 = 𝑥1
If 𝑓 𝑥1 > 0 then replace 𝑏 = 𝑥1
𝑥𝑖 −𝑥𝐸
iv. Repeat the Step iii until ≤ 𝜖, for given 𝜖.
𝑥𝐸
Example:
Find a root of the equation f(x)= 𝒙𝟐 -2 by using False Position Method in the interval [1, 2].
Solution:
ii.
To check f(x)= 𝑥 2 -2 is continuous on [1 , 2]
Since, f(x)= 𝑥 2 -2
𝑓 1 = 12 − 2 = −1
𝑓 2 = 22 − 2 = 2
So, f(x) is continuous on [1,2].
ii.
We have 𝑓(1) = −1 and 𝑓(2) = 2
𝑓 1 . 𝑓 2 = −1 2 = −2 ≤ 0
So, solution f(x)= 𝑥 2 -2=0 exist in interval [1 , 2]
Cont.
1st Iteration

𝑎𝑓 𝑏 − 𝑏𝑓(𝑎)
𝑥1 =
𝑓 𝑏 − 𝑓(𝑎)

(1)(2) − 2(−1)
𝑥1 = = 1.3333
2 − (−1)
𝑓 𝑥1 = 𝑓 1.3333 = 1.33332 − 2 = −0.2222 < 0

As, 𝑓 𝑥1 < 0 then replace 𝑎 = 𝑥1

We Get the modified interval

[1.3333, 2]
Cont.
2nd Iteration

𝑎𝑓 𝑏 − 𝑏𝑓(𝑎)
𝑥2 =
𝑓 𝑏 − 𝑓(𝑎)

(1.3333)(2) − 2(−0.2222)
𝑥2 = = 1.4000
2 − (−0.2222)
𝑓 𝑥2 = 𝑓 1.4000 = 1.4000 2 − 2 = −0.0401 < 0

As, 𝑓 𝑥2 < 0 then replace 𝑎 = 𝑥2

We Get the modified interval

[1.4, 2]
Cont.
3rd Iteration

𝑎𝑓 𝑏 − 𝑏𝑓(𝑎)
𝑥3 =
𝑓 𝑏 − 𝑓(𝑎)

(1.4)(2) − 2(−0.0401)
𝑥3 = = 1.4118
2 − (−0.0401)
𝑓 𝑥3 = 𝑓 1.4118 = 1.4118 2 − 2 = −0.0068 < 0

As, 𝑓 𝑥3 < 0 then replace 𝑎 = 𝑥3

We Get the modified interval

[1.4118, 2]
Cont.
4th Iteration

𝑎𝑓 𝑏 − 𝑏𝑓(𝑎)
𝑥4 =
𝑓 𝑏 − 𝑓(𝑎)

(1.4118)(2) − 2(−0.0068)
𝑥4 = = 1.4138
2 − (−0.0068)
𝑓 𝑥4 = 𝑓 1.4138 = 1.4138 2 − 2 = −0.0012 < 0

As, 𝑓 𝑥4 < 0 then replace 𝑎 = 𝑥4

We Get the modified interval

[1.4138, 2]
Cont.
5th Iteration

𝑎𝑓 𝑏 − 𝑏𝑓(𝑎)
𝑥5 =
𝑓 𝑏 − 𝑓(𝑎)

(1.4138)(2) − 2(−0.0012)
𝑥5 = = 1.4142
2 − (−0.0012)
𝑓 𝑥5 = 𝑓 1.4138 = 1.4138 2 − 2 = −0.0012 < 0

As, 𝑓 𝑥5 < 0 then replace 𝑎 = 𝑥5

We Get the modified interval

[1.4142, 2]
Cont.
6th Iteration

𝑎𝑓 𝑏 − 𝑏𝑓(𝑎)
𝑥6 =
𝑓 𝑏 − 𝑓(𝑎)

(1.4142)(2) − 2(−0.0002)
𝑥6 = = 1.4143
2 − (−0.0002)
𝑓 𝑥6 = 𝑓 1.4143 = 1.4143 2 − 2 = 0.0001 > 0

As, 𝑓 𝑥6 > 0 then replace 𝑏 = 𝑥6

We Get the modified interval

[1.4142, 1.4143]
Relative Error:
Let 𝑥𝑖 is the approximate solution of the non-linear equation and 𝑥𝐸 is the exact
value of the function 𝑓(𝑥).
𝑥𝑖 − 𝑥𝐸
𝑅. 𝐸 =
𝑥𝐸
Since 𝑥𝐸 = 1.4142
𝑥1 −𝑥𝐸 1.3333−1.4142
i=1, 𝑥𝐸
= 1.4142
= 0.0572
𝑥2 −𝑥𝐸 1.4000−1.4142
i=2, 𝑥𝐸
= 1.4142
= 0.0100
𝑥3 −𝑥𝐸 1.4118−1.4142
i=3, 𝑥𝐸
= 1.4142
= 0.0017
𝑥4 −𝑥𝐸 1.4138−1.4142
i=4, 𝑥𝐸
= 1.4142
= 0.0003
𝑥5 −𝑥𝐸 1.4142−1.4142
i=5, 𝑥𝐸
= 1.4142
= 0 < 0.0001 = 10−4
𝑥6 −𝑥𝐸 1.4143−1.4142
i=6, 𝑥𝐸
= 1.4142
= 0.0001
 Homework

Exercise: 2.3
Q3(b, c), Q4(b), Q9, Q10, Q13(a), Q14(b)
Quiz Announcement

Wednesday 31st April 2021

Lecture 1 to 6
Timings 40 to 50 mins
Linear System of Equations
Special Matrices
General system of Linear Equations

System of equations for 𝑛 unknowns 𝑥1 , 𝑥2 , … , 𝑥𝑛 :

𝑎11 𝑥1 + 𝑎12 𝑥2 + … + 𝑎1𝑛 𝑥𝑛 = 𝑏1

𝑎21 𝑥1 + 𝑎22 𝑥2 + … + 𝑎2𝑛 𝑥𝑛 = 𝑏2
⋮ ⋮
𝑎𝑛1 𝑥1 + 𝑎𝑛2 𝑥2 + … + 𝑎𝑛𝑛 𝑥𝑛 = 𝑏𝑛
Equivalent matrix equation
Right-hand-side vector 𝑏
𝑎11 𝑎12 ⋯ 𝑎1𝑛 𝑥1 𝑏1
𝑎21 𝑎22 ⋯ 𝑎2𝑛 𝑥2 𝑏2
⋮ ⋮ ⋮ ⋮ = ⋮
𝑎𝑛1 𝑎𝑛2 ⋯ 𝑎𝑛𝑛 𝑥𝑛 𝑏𝑛

Coefficient matrix 𝐴
Column vector 𝑥
of unknowns

𝐴𝑥 = 𝑏
Example

3𝑥1 + 2𝑥2 = 18 3 2 𝑥1 18
ቊ =
−𝑥1 + 2𝑥2 = 2 −1 2 𝑥2 2
Graphical solution
𝑥2

Solution: 𝑥1 = 4, 𝑥2 = 3

𝑥1
 Solution Space

𝑥2 𝑥2 𝑥2

One solution Infinite number

of solutions

No solution

𝑥1 𝑥1 𝑥1
Gauss Elimination Method
• Gauss elimination aims to transform the system of linear equations in an
equivalent system that can be solved in very short time

• This is done by a systematic procedure replacing equations by new ones obtained

by multiplication and subtraction of equations from the original system
 Working Procedure
• Consider given a system of linear equations

• Find Augmented Matrix for the given equation

• Transform Augmented Matrix into Upper triangular/ Echelon form

• Find Equations corresponding to Upper Triangular Matrix/ Echelon Matrix

• Using back substitutions, find solutions of given system of equations

General idea
𝑎11 𝑎12 𝑎13 ⋯ 𝑎1𝑛 𝑥1 𝑏1
𝑎11 𝑎12 ⋯ 𝑎1𝑛 𝑥1 𝑏1 ′ ′
Forward 0 𝑎22 𝑎23 ⋯ 𝑎′ 2𝑛 𝑥2 𝑏 ′
𝑎21 𝑎22 ⋯ 𝑎2𝑛 𝑥2 𝑏2 ′ ′
2
⋮ = elimination 0 0 𝑎33 ⋯ 𝑎3𝑛 𝑥3 = 𝑏3′
⋮ ⋮ ⋮ ⋮
𝑥𝑛 ⋮ ⋮ ⋮ ⋮ ⋮ ⋮
𝑎𝑛1 𝑎𝑛2 ⋯ 𝑎𝑛𝑛 𝑏𝑛 𝑥𝑛
0 0 0 ⋯ 𝑎′ 𝑛𝑛 𝑏𝑛′

𝑏1 − σ𝑛𝑖=2 𝑎1𝑖
′
∙ 𝑥𝑖
𝑎11 𝑎12 𝑎13 ⋯ 𝑎1𝑛 𝑥1 𝑏1 𝑎11
′ ′
0 𝑎22 𝑎23 ⋯ 𝑎′ 2𝑛 𝑥2 𝑏 ′
2 Back
𝑥1 ⋮
0 0 ′
𝑎33 ⋯ ′
𝑎3𝑛 𝑥3 = 𝑏3′ substitution ⋮ 𝑏 ′
− 𝑎 ′
∙ 𝑥𝑛
= 𝑛−1 𝑛−1,𝑛
⋮ ⋮ ⋮ ⋮ ⋮ ⋮ 𝑥𝑛−1 ′
𝑥𝑛 𝑎𝑛−1,𝑛−1
′ 𝑥𝑛
0 0 0 ⋯ 𝑎𝑛𝑛 𝑏𝑛′ 𝑏𝑛′
′
𝑎𝑛𝑛
Example with two equations
• Consider
𝑥1 + 𝑥2 = 3
ቊ
3𝑥1 − 4𝑥2 = 2

• Subtracting three times the first equation from the second yields

𝑥1 + 𝑥2 = 3
ቊ
−7𝑥2 = −7
Example of a two equations
• Solve
𝑥1 + 𝑥2 = 3
ቊ
−7𝑥2 = −7

• We first solve the last equation:

𝑥2 = 1
• And then the first one:

𝑥1 + 1 = 3 ⇒ 𝑥1 = 2
Remark
• There are various ways to form an upper triangular system

𝑟2 ← 𝑅2 − 3𝑅1
𝑥1 + 𝑥2 = 3 𝑥1 + 𝑥2 = 3
ቊ ቊ
3𝑥1 − 4𝑥2 = 2 −7𝑥2 = −7
Used in Gauss elimination

𝑥1 + 𝑥2 = 3 𝑟2 ← 3𝑅1 − 𝑅2 𝑥1 + 𝑥2 = 3
ቊ ቊ
3𝑥1 − 4𝑥2 = 2 7𝑥2 = 7
 Augmented matrix notation
• A common notation is:

1 1 | 3 𝑟2 ← 𝑅2 − 3𝑅1 1 1 | 3
Pivot element 3 −4 | 2 0 −7 | −7

We want to get a 0 here

• Solving leads to
2
𝑥=
1
Example with three equations
• Consider
−3𝑥1 + 2𝑥2 − 𝑥3 = −1
ቐ 6𝑥1 − 6𝑥2 + 7𝑥3 = −7
3𝑥1 − 4𝑥2 + 4𝑥3 = −6
• Augmented matrix notation

−3 2 −1 | −1
6 −6 7 | −7
3 −4 4 | −6
 Example with three equations
• Solving:
−3 2 −1 | −1 𝑟2 ← 𝑅2 − −2 𝑅1 −3 2 −1 | −1
6 −6 7 | −7 𝑟3 ← 𝑅3 − −1 𝑅1 0 −2 5 | −9
3 −4 4 | −6 0 −2 3 | −7

−3 2 −1 | −1 −3 2 −1 | −1
𝑟3 ← 𝑅3 − 𝑅2
0 −2 5 | −9 0 −2 5 | −9
0 −2 3 | −7 0 0 −2 | 2
Example with three equations
• Back substitution

−3 2 −1 | −1 2
Back substitution
0 −2 5 | −9 𝑥= 2
0 0 −2 | 2 −1
Gauss Jordan Elimination
Working Procedure

• Find Augmented Matrix for the given equation.

• Transform augmented Matrix into Normal form/ Reduced Echelon form.

• Find Equations corresponding to Normal form.

• Find solutions of given system of equations.

Example
Use Gauss-Jordan elimination to solve the system

The augmented matrix of the system is

Cont.
End of Section