Intro - Linear Algebra

IIT Kanpur July 201
20 followers
1 Video View·Dec 2, 2023
Prof. Arbind Kumar Lal, Department of Mathematics and Statistics, IIT-Kanpur
So, this slide is about the introduction to the course and the aim of the course. So,
the idea of the course is that, we have been solving a system of linear equations, alright,
in two equations in two unknowns, three equations in three unknowns till our school days. We
also learn determinant, what are inverse matrices and so on. But what we would like to do here
is that, we would like to give it a proper framework, alright. So, in place of solving
a system of equation in two variables, two unknowns, two equations and so on, we would
like to look at a general setup, what is called A X is equal to B, alright.
So, what we would like is that, our matrix A will be m cross n, X is n cross 1 and this
is a vector of unknowns, of unknowns and B is also given to us, it is of size m cross
1, fine. So, A is given, B is given, we want to find X as that A X is equal to B, fine.
Now, so to understand this, we need to know, what is a matrix. So, we need to know, need
to know, what is a matrix and then we are multiplying, so we are multiplying A with
X in some sense. So, we should know, what is called matrix multiplication, multiplication,
we need to know this. We also further need to know that, when I look at the system, say
3 X is equal to 5, then we know that, X is equal to 5 by 3 is a solution. So, in some
sense, what we are saying is that, X is equal to 3 inverse time 5 or 3 inverse means, reciprocal
of 3, reciprocal of 3 or what is called inverse of 3, fine.
So, in a similar sense, we need to talk of, what is called inverse of a matrix. So, it
turns out that, we can talk of inverse of a matrix being unique and making our work
easy is, when A is an n cross n matrix, alright. So, when A is a n cross n matrix that is,
it is a square matrix, we can talk of inverse of the matrix, fine. So, once we understand
these things, we proceed with the solution set. So, in the process, what we learn is,
the first thing that we learn is, what is called a matrix A being written as L times
U, where L is a n cross n matrix, U is also an n cross n matrix, but then there are some
special properties, what are called, L is called a lower triangular matrix and U is
called an upper triangular matrix. And so, what we are trying to say is, for
example L, the matrix looks like, lower I am saying, so it will look like 1 0 0 0 minus
1 0 2 4 5. So, the upper part is nothing but, it consists of 0 here and U will have, there
is a separate thing here, which will look like here 2 5 7 0 minus 1 minus 5 0 0 2, alright.
So, here the 0 will be here as such, alright. So, this is what we are going to look at.
So, the first thing that we are going to look at, when I solve a system, try to write the
matrix A as L U, fine. The next we would like to look at is, what is called a rank factorization
of A. So, A is equal to b times c. So, for me, A is going to be an m cross n matrix and
there is a notion of what is called rank of A. So, rank of A is r, if I say, alright,
it has some connotation, some definition, so let us not worry about it. So, rank of
A is r, then I can write matrix A as a matrix of size m cross r and c of r cross n and it
turns out that, this rank of B is same as rank of C is r. So, this is the two things
that we would like to know from a system of equations. Then, we would like to increase
our difficulty level and see, can we do it better, in the sense that, these things give
us a solution, which are unique, exact and things like that. Most of the questions that
we come across, does not have a solution or sometimes, if it has, it is very difficult
to find those solutions. So, we would like to look at approximate solutions and they
are the notions of what are called factorizing A. So, this was our two, third is writing
A as q times r is important. So, again m is this, m cross n, q will be something like
m cross r and r is less than equal to m, this is r cross n again, but the important thing
here is that, q has the property that, q transpose q, which is an r cross r matrix is identity.
So, in some sense, q behaves as a orthogonal matrix. So, this is the third task that we
are going to have. The fourth task that we are going to have in this course is, what
are called Eigen values, Eigen vectors, trying to understand Eigen values and Eigen vectors.
So, this idea leads us to what is called writing a matrix A, which is a square matrix
as say S D S inverse, where S is a non-singular matrix and D is a diagonal matrix. In general,
this is not possible for every matrix A, but what is possible is that, we can write as
this as say U T U transpose, where U is what is called a unitary matrix and T is upper
triangular. So, this is possible for every matrix U, there is no problem as such here.
So, given any matrix A, I can write like this, what is called a Schur's theorem.
And then, the last thing that we would like to do in this course is, given that we have
matrix A, which is m cross n, we would like to write it as U sigma V transpose, U and
V are unitary matrices, the sizes are meaningful and this sigma is called the similar values
of A, which are related with what are called Eigen values of A transpose or A transpose
A, whatever you want to say. So, they are related with this. So, that is the aim of
the course. The aim of the course is first to understand matrices, matrix multiplication,
invertibility, go on to understand A equal to L U, A equal to B C, A equal to Q R and
then, the two related with Eigen values, Eigen vectors. That is the aim of the course.
Thank you.
