Você está na página 1de 28

Computational Methods in

Engineering (ME6000)
Sourav Rakshit
MDS-204
srakshit@iitm.ac.in
Contents
Interpolation

Numerical differentiation and integration

Numerical solutions of ODE

Numerical solutions of PDE


Evaluation: (50 marks)
Q1 7.5
Q2 7.5
Computer assignments 15
Final 20

Attendance requirement: >85%

Computer assignment grading based on Viva


Learning outcomes
At the end of the course:

You will know/revise some of the most widely used numerical


algorithms in engineering

You will not be timid of writing computer programs (unless you copy
all computer assignments from your friends without understanding
and get 0 in viva)
Interpolation
Table

Analytic function
Applications of interpolating functions
Interpolation

Differentiation

Integration
Approximating functions
Polynomials
Trigonometric
Exponential

Properties
Fitting polynomial to discrete data

Exact fit Approximate fit

Equally spaced and unequally spaced data (notes)


Properties of polynomials

Suitability of approximating polynomials -> Weierstrass approxm. theorem


Taylor series polynomial

Remainder
Taylor polynomial: error
Horners algorithm
Direct fit polynomials

Ill-conditioned
Difficult
to solve
Lagrange polynomials
Hermite polynomials
Nevilles method

Divided difference formula Data


1st Divided difference - Linear

2nd Divided difference - Quadratic 3rd Divided difference - Cubic


Table for Nevilles algorithm

Degree 3 polynomial
Divided difference tables and polynomials

Data
Same table as given by Nevilles algorithm
Difference tables and difference polynomials
Forward difference

Backward difference

Central difference
Difference tables

Forward difference table Backward difference table Central difference table


Newton forward difference polynomial
Newton backward difference polynomial
Approximate fitting: Least squares
Error when both
Error when X and Y are not fixed
Y is fixed
To get a simple function (need not pass through
all points) such that the sum of square of errors
is minimum
Error when
X is fixed

data
Best fit criteria
Least squares
Higher degree polynomial approximation using least squares

Você também pode gostar