LINEAR
ALGEBRA
6.1 Defining Matrices
What is a matrix?
◦ A matrix is a rectangular array of numbers. Each number is called an element.
◦ They are defined in terms of the number of rows and columns.
◦ The matrices above are
◦ If a matrix has the same number of rows as columns it is called a square matrix.
◦ A matrix with 1 row (and several columns) or 1 column (and several rows), like the last two above, it
is called a vector.
◦ We can transpose a matrix by swapping its row and columns e.g.
◦ A matrix is just a number but is often called a scalar.
◦ Matrices are usually identified with a capital letter.
LINEAR
ALGEBRA
6.2 Simple Matrix Arithmetic
Scalar multiplication
◦ Any matrix can be easily multiplied ◦ Likewise, each term can be divided
by a scalar. by a common factor.
◦ Each element being multiplied by
the scalar.
◦ Let , then
Addition and
Subtraction
◦ Addition and subtraction can only be
performed when the matrices are
conformable that is in this case have
identical dimensions.
LINEAR
ALGEBRA
6.3 Matrix Multiplication
Matrix multiplication is confusing
◦ Matrices can be multiplied when they are conformable and this time it means that the number of
columns in the first matrix must equal the number of rows in the second.
◦ The number of rows in the first and the number of columns in the second giving the dimensions of the result
◦ Thus square matrices with the same dimensions can be multiplied but non-square matrices with the same dimensions
cannot.
◦ This also means that when we multiply matrices, the order matters; (generally).
◦ So, we can multiply
◦ a by another matrix (result is )
◦ a by a matrix (result is )
◦ but not a by another matrix
◦ or a 2 by a 4 matrix
◦ But we can multiply a by a (result is )
◦ What makes matters worse, the method of multiplication is far from obvious and demands considerable
practice.
◦ But, learn to think “row by column, row by column…”
Row by column
◦ We start with a pair of matrices ◦ Or a by a
◦ NB is conformable
◦ Other combinations are unusual
◦ (
◦ We commonly multiply a square
matrix by a vector (here
LINEAR
ALGEBRA
6.4 Inverses & equations
The idea of an inverse
◦ Consider the trivial equation this is easily ◦ We can generalise this, so for
solved by dividing both sides by 4.
◦ However, another way of looking at the
problem is, instead, to multiply both sides
by
◦ We can use different notation since
◦ Multiplying by is exactly the same idea as ◦ So
dividing by 4.
◦ Of course, and 3 so
◦ We don’t need to write the number 1, so
our solution is and, checking ◦ is called the (multiplicative) inverse of
Matrix equations and the identity
matrix
◦ We will soon be solving matrix ◦ The identity matrix does the same thing
equations which are written in the when we multiply a matrix by it, that 1
does, when we multiply a number by it.
form
◦ Nothing.
◦ They are solved in the same way
◦ The identity matrix is always a square
matrix and the smallest,
◦ represents the identity matrix which
is the matrix equivalent of 1. ◦ In general, an identity matrix has 1s on
the ”leading” diagonal and zeros
◦ So everywhere else.
Determinants
◦ The determinant is a scalar (simple ◦ For example, if then
number) which is unique to each matrix
(square matrices only).
◦ Or for,
◦ Finding determinants is part of the
process of finding inverse matrices.
◦ Finding determinants for and matrices ◦ In general, for
is reasonably manageable.
◦ It’s much more work for larger matrices.
◦ Finally, consider,
◦ In fact, finding the determinant of a is
pretty straightforward
◦ It is just the difference between the ◦ A matrix with a zero determinant is
products of the elements of the leading called a singular matrix (and usually
and trailing diagonals. causes problems)
Finding an inverse
◦ The inverse of a matrix is always ◦ For the general matrix the adjoint is
◦ The adjoint matrix is a manipulation ◦ That is swap the elements on the
of the original matrix. In the case of a leading diagonal and change the
the manipulation is straightforward. signs of the elements on the trailing
◦ Not so for a and very troublesome for diagonal.
anything bigger. ◦ So, for
◦ We only consider matrices
◦ And therefore is the inverse of .
Another example
◦ Just checking
◦ N.B singular matrices do not have
◦ So, , readers can check that inverses
◦ Remember their determinants are 0
and we cannot divide by zero.
Simultaneous Linear Equations
◦ The motivation for this is to explain how ◦ So,
we can use matrices to solve
simultaneous linear equations.
◦ represents the system of equations above,
that is,
◦ This is a case of two equations in two ◦ But as we have seen the solution to this is
unknowns but the same principles apply
to any number of equations, as long as we
can set up a square matrix. ◦ So
◦ Consider the multiplication where ◦ And the solution is
◦ And ◦ So
LINEAR
ALGEBRA
6.5 Eigenvalues & eigenvectors
The characteristic equation
◦ An eigenvector is a vector which when multiplied by a matrix yields a vector which is a multiple of itself;
the multiple being the eigenvalue.
◦ For example,
◦ So, we say that 2 is an eigenvalue and an eigenvector of the matrix
◦ Any square matrix has up to eigenvalues and eigenvectors. We are only concerned with matrices which
have up to two eigenvalues and eigenvectors.
◦ In general, we say, for matrix
◦ Where represents the eigenvalues and the eigenvectors.
◦ We re-write slightly as and then as to get the characteristic equation
◦ This sounds more complicated than it is in practice.
The eigenvalues
◦ If we consider , its characteristic equation is
◦ The eigenvector of 2 we noted earlier.
◦ These are a little more difficult to find but we try to find the eigenvectors corresponding to each
eigenvalue in turn.
The eigenvectors
◦ For ◦ For
◦ being the eigenvector ◦ being the eigenvector
corresponding to corresponding to
◦ So, from each row ◦ So, from each row
◦ In both cases, the values is 3 times ◦ In both cases, the yvalues is 2 times
the value, so one possible the xvalue, so one possible
eigenvector is , another is etc. eigenvector is , another is etc.
