Computational Physics
• In this course we will learn many of the
important numerical techniques used in
computational physics and demonstrate them
with the help of computer programs.
• The programming language I’ll use is
MATLAB
• You may use any other programming language
you are more familiar with for assignments etc.
The major numerical techniques we will learn
and use are the following.
1. Finding roots of a function of one and more
variables.
kL
Example : k tan k
2 2
2
Solve for k
Newton–Raphson method in one dimension
and extension to higher dimensions.
2 . Interpolation and Curve fitting
Example : Interpolating/curve fitting experimental
data.
Polynomial and Spline interpolation, Best fit
by minimizing RMS difference
3. Numerical integration
Example : Cornu’s spiral
t
u 2 t
u 2
x ( t ) cos du y( t ) cos du
0 2 0 2
Euler method, Simpson Method, Gaussian
quadrature etc.
4. Numerical solutions of ordinary differential
equations with initial as well as boundary
conditions.
Example : Solving the Airy’s equation
y xy 0 y(0) 0 ; y(L) 0
Euler Method, Runge-Kutta methods of
different order, Shooting method.
5. Numerical solutions of PDE with boundary
conditions
Example : The wave, the heat and Laplace
equations.
Finite Difference and Finite Element methods.
6. Monte Carlo techniques
Examples : Brownian motion, Diffusion, Ising
Model, Monte-Carlo integration
The Metropolis Algorithm,
7. Numerical Quantum Mechanics
Example : Finding energy eigenvalues
and eigenfunctions in different setting,
time evolution of wavefunctions etc.
Beginning With MATLAB
1. MATLAB : An Introduction with Applications
by
Amos Gilat
Basic Operations With Matrices
MATLAB is the short form for
Matrix Laboratory
Not
Mathematics Laboratory
Every variable is a matrix. A single number
is a 1 x 1 matrix. An 1-d array of length N is
either a 1 x N matrix (row vector) or N x 1
matrix (column vector).
In MATLAB one does not need to declare the
data/variable type. Any alphanumeric string,
beginning with a alphabet can be used as a
real, integer or complex variable.
The arithmetic operations : + ; - ; * ; / ; ^
X^3 is x raised to the power 3
Built-in MATLAB Functions
sqrt(x) nthroot(x,n) exp(x) log(x)
Log10(x) factorial(x) abs(x) sin(x)
sind(x) cos(x) cosd(x) tan(x) tand(x)
Cot(x) cotd(x) asin(x) acos(x) tanh(x)
round(x) : round(17/5) = 3 round(18/5)=4
fix(x) : fix(13/5)=2
ceil(x) : ceil(13/5)=3
rem(x,y) : rem(13,5)=3
rand : generates a random number
between 0 and 1
randi(n) : generates a random integer b/w 1
and n
rand(m,n) : generates a random 2 x 3 matrix
with each element a random number b/w 0
and 1
randi(m,n,p) : generates a m x n matrix
with each element a random number b/w 1
and p
Creating a Matrix
A = [2.5 -3.2 1.8]
A = [1 2 3;4 5 6;7 8 9]
A=2:3:17 A=linspace(x,y,n)
Matrix Operations
A*B A multiplied by B. Dimensions must match.
inv(A) : Inverse of A
det(A) : determinant of A
trace(A) : Trace of A
A= ones(m,n) : m x n matrix with all ones
A = zeros(m,n) : m x n matrix with all zeros
A = eye(n) : n x n identity matrix
sin(A) cos(A) exp(A) atan(A) etc. create
matrices where the operation (sin, cos, exp
etc. are applied to each element of the
matrix)
A^n : matrix A multiplied with itself n times.
A.^n : each element of the matrix is raised
to the nth power.
If A and B are of the same dimension, then
A./B is element by element division of A with
those of B
The “for” loop
for k=1:n
--------
--------
--------
end
Let us numerically prove the following.
If A is a square matrix, then
det(exp(A)) = exp(trace(A))
