Download the App for Test series

Numerical Analysis Short Notes

𝑏−𝑎 𝑟
● ℎ = 𝑛
, 𝐸𝑓(𝑥) = 𝑓(𝑥 + ℎ), 𝐸 𝑓(𝑥) = 𝑓(𝑥 + 𝑟ℎ) =
1 1
−1 1/2 −1/2 1 −2
● 𝐷 = 𝐸 + 1, ∇ = 1 − 𝐸 , δ = 𝐸 −𝐸 ,μ = 2
2
(𝐸 + 𝐸 )
( )
𝑓 𝑥1 −𝑓(𝑥0)
( )
● 𝐷𝑥 𝑓 𝑥0 = 𝑓 𝑥0𝑥1 =
1
( ) (𝑥1−𝑥0)
;
( )
𝑓 𝑥1𝑥2 −𝑓(𝑥0𝑥1)
( )
𝑓 𝑥0, 𝑥1, 𝑥2 = 𝐷𝑥 𝑥 𝑓 𝑥0 =
1 2
( ) (𝑥2−𝑥0)
ℎ𝐷 1
● 𝐸 =𝑒 or 𝐷 = ℎ
log 𝑙𝑜𝑔 𝐸 =
1 1
ℎ
log 𝑙𝑜𝑔 (1 + ∆) =− ℎ
log 𝑙𝑜𝑔 (1 − ∇)
● Fundamental Theorem of difference calculus:- 𝑓(𝑥) is n degree
𝑡ℎ 𝑛+1
poly then 𝑛 difference is constant and 𝐷 𝑓(𝑥) = 0 ∀𝑥,
𝑛+1 𝑛+1 𝑛+1
∇ 𝑓(𝑥) = 0, 𝐷 𝑓(𝑥) = 0, 𝐷 𝑓(𝑥) = 0
𝑛 𝑛 𝑛 𝑛 𝑛
● 𝐷 𝑓(𝑥) = 𝑛! 𝑎0ℎ and ∇ 𝑓(𝑥) = 𝑛! 𝑎0ℎ , 𝐷 𝑓(𝑥) = 𝑎0

● Interpolation: Two types - 1. Equal interval 2. Unequal

interval
2
N.G.F → 𝑓(𝑥) = 𝑓 𝑥0 + 𝑢𝐷𝑦0 +( ) 𝑢(𝑢−1)
2!
𝐷 𝑦0 + …

𝑢(𝑢+1) 2
NGB → 𝑦𝑛 + 𝑢∇𝑦𝑛 + 2!
∇ 𝑦𝑛 + …

TestPlus - The Practice App

Whatsapp : 9352579499
 Download the App for Test series
𝑛 𝑥−𝑥𝑗
Lagrange → ∑ 𝐿𝑖(𝑥) = ∏ 𝑥𝑖−𝑥𝑗
𝑗=0 𝑗≠1

N.D.D.I →
( ) ( )( ) (
𝑓(𝑥) = 𝑓 𝑥0 + 𝑥 − 𝑥0 𝑓 𝑥0𝑥1 + 𝑥 − 𝑥0 𝑥 − 𝑥1 𝑓 𝑥0𝑥1𝑥2 + … )( )( )
● Sum of all Lagrange coefficients is 1.

● Numerical Differentiation :- 𝑢 = (𝑥 − 𝑥0)/𝑛

2 3 2
2 3 4
𝑓'(𝑥)≈
1
𝑛 ( )
[∆𝑓 𝑥0 +
2𝑢−1
2! ( )
∆ 𝑓 𝑥0 +
3𝑢 −6𝑢+2
6
∆ 𝑓 𝑥0 + ( ) 2𝑢 −9𝑢 +11𝑢−3
12
∆ 𝑓(𝑥0)
2
2 3 4
𝑓"(𝑥)≈
1
𝑛
2 ( )
[ ∆ 𝑓 𝑥0 + (𝑢 − 1)∆ 𝑓 𝑥0 + ( ) 6𝑢 −18𝑢+11
12 ( )
∆ 𝑓 𝑥0 + … ]

● Numerical Integration:-
1. Trapezoidal formula:- point-2 degree-1
n=(b-a)

formula =
ℎ
2 ( )
[ 𝑦0 + 𝑦𝑛 + 2(𝑦1 + 𝑦2 + … + 𝑦𝑛−1)] (No error
if poly degree=1)

TestPlus - The Practice App

Whatsapp : 9352579499
 Download the App for Test series

1
2. Simpson 3
rule:- point-3 degree-2
n=(b-a)/2 (n is even)
formula =
ℎ
3 ( )
[ 𝑦0 + 𝑦𝑛 + 4 (𝑜𝑑𝑑 𝑡𝑒𝑟𝑚𝑠 ) + 2 (𝑒𝑣𝑒𝑛 𝑡𝑒𝑟𝑚𝑠)]
=
ℎ
3 ( ) ( )
[ 𝑦0 + 𝑦𝑛 + 4 𝑦1 + 𝑦3 + … + 𝑦𝑛−1 + 2 𝑦2 + … + 𝑦𝑛−2 ] ( )
(4th order error)
( No error if degree =3)

3
3. Simpson 8
rule :- point-4 degree-3
formula =
3
8 ( ) (
[ 𝑦0 + 𝑦𝑛 + 3 𝑦1 + 𝑦2 + 𝑦4 + 𝑦5 + … + 2(𝑦3 + 𝑦6)] ) (
No error if degree=3)
(𝑏−𝑎) 4 4
( 4th order error → − 80
ℎ 𝑓 (ξ))

4. Weddle’s rule :- point-7 degree-6

3
formula = 10
ℎ[𝑦0 + 5𝑦1 + 𝑦2 + 6𝑦3 + 𝑦4 + 5𝑦5 + 𝑦6] ( No
error if degree 5)
(𝑏−𝑎) 6 6
( error degree - 6) − 840
ℎ 𝑓 (ξ)(error)

● Numerical solution of equations: ( Algebraic and

Transcendental Method):-

Root finding problem :- IVT (intermediate)→ Bisection, Regula falsi

TestPlus - The Practice App

Whatsapp : 9352579499
 Download the App for Test series

FPT (fixed point) → Secant Method,

Newton Raphson Method

1. Bisection Method:- if 𝑓(𝑎). 𝑓(𝑏) < 0 ⇒next iteration will be

𝑓( ) =?
𝑎+𝑏
2
(𝑏−𝑎)𝑓(𝑏)
2. Regula falsi Method:- If 𝑓(𝑎). 𝑓(𝑏) < 0 ⇒ 𝑥1 = 𝑏 − 𝑓(𝑏)−𝑓(𝑎)
(𝑥𝑛−𝑥𝑛−1) 𝑓(𝑥𝑛)
3. Secant Method:- If 𝑥𝑛+1 = 𝑥𝑛 − 𝑓(𝑥𝑛))−𝑓(𝑥𝑛−1)

4. Newton Raphson Method:-

( )
𝑓 𝑥𝑛
𝑥𝑛+1 = 𝑥𝑛 − '
𝑓 (𝑥𝑛)

Method Order of C (asymptotic

Convergence constant)
Bisection 1 (linear) 1/2
Regula falsi 1 (linear) 0<c<1
Secant 5+1
=1.618 | 𝑓"(α) |
2 C= | ' |
| 2𝑓 (α) |
Newton Raphson 2 (quadratic) | 𝑓"(α) |
C= | ' |
| 2𝑓 (α) |

● Newton Raphson method converges more rapidly than secant

method.
1
2
< 𝐶 ≤ 1 then Bisection method is more better

TestPlus - The Practice App

Whatsapp : 9352579499
 Download the App for Test series
1
0< 𝐶 < 2
then Regula falsi is better

1
C= 2
both equivalent

TestPlus - The Practice App

Whatsapp : 9352579499