Numerical Analysis

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

Numerical Analysis

Analysis generally implies solving a problem through equations. The equations


must be reduced to an answer through procedures of algebra, calculus, differential
equations, partial differential equations etc.

Numerical Analysis is essentially analysis using only procedures of arithmetic, i.e.


addition, substraction, multiplication, division and comparision.

💡 Because the operations involved in numerical analysis are exactly what


computers can do, numerical analysis and computers are intimately related.

Goals we wish to achieve while using numerical analysis:

1. Accuracy: Amount and ability to control error

2. Efficiency: Number of steps taken to reach our solution

3. Stability: Small Deviations in output corresponding to small deviations in input


(otherwise it is unstable)

Notations

C(R) : Set of all functions ‘f’ such that f is continous on the entire real line R
C′ (R) : Set of all functions ‘f’’ such that f is continous on the entire real line R
C2 (R) : Set of all functions ‘f 2 ’ such that f is continous on the entire real line R
C∞ (R) : Set of all functions ‘f n ’ such that f is continous on the entire real line R

Numerical Analysis 1
Taylor’s Theorem with Lagrange form of Remainder

If f ∈ Cn [a, b] and if f n+1 exists in (a, b) then for point c and x ∈ [a, b] there exists a
ξ between c and x such that

f(x) = Σnk=0 k!1 f k (c)(x − c)k + En (x)

1
where En (x) = (n+1)!
f n+1 (ξ)(x − c)n+1

Mean Value Theorem

If f ∈ C[a, b] and f ′ exists in (a,b), then for any c and x ∈ [a, b] there exists a ξ
between c and x such that f(x) = f(c) + (x − c)f ′ (ξ)

Rolle’s Theorem

∈ C[a, b] and f ′ exists in (a,b), and if f(a) = f(b) = 0 then f ′ (ξ) = 0 for some
If f
ξ ∈ (a, b)

Error

X —> True Value

X ∗ —> Approximation

Numerical Analysis 2
Absolute Error = ∣X − X∗∣

∣x−x∗ ∣
Relative Error = ∣x∣

Kinds of Errors in Numerical Procedures

1. Error in Original Data: Problems depend on measurements of doubtful accuracy in


the real world.

2. Blunders: Human error is natural, we can make errors causing in overall blunder in
the result.

3. Truncation Error : Errors caused because of truncating the procedure or equation.

4. Propagated Error: Error caused by magnification of error in subsequent steps

5. Round off Error: Floating Point numbers are usually not exacly represented, but are
rounded off, which results in round off error.

Numerical Analysis 3

You might also like