Mathematics for Economics
Lecture 3
ASEEM BHAGWAT
Eigenvalues and Eigenvectors
ASEEM BHAGWAT
• Suppose that λ is an eigenvalue of the n × n matrix A, and that v is an
eigenvector corresponding to the eigenvalue λ
• Then Av = λv = λIv, where I is the n × n identity matrix, so that
(A−λI)v = 0
• Since v ∈ Rn is non-zero, it follows that we must have
det (A−λI) = 0
• Solving this equation gives us the eigenvalues of A
ASEEM BHAGWAT
• The set {v ∈ Rn : (A−λI)v = 0} is the nullspace of the matrix (A−λI)
• The polynomial det (A−λI) is called the characteristic polynomial of
the matrix A
• For any root λ of the polynomial, the space described is called the
eigenspace corresponding to the eigenvalue λ
ASEEM BHAGWAT
Example
ASEEM BHAGWAT
• The characteristic polynomial equation is λ2 − 8λ + 12 = 0
• The eigenspace and eigenvector corresponding to λ = 2 is:
• The eigenspace and eigenvector corresponding to λ = 6 is:
ASEEM BHAGWAT
Example
ASEEM BHAGWAT
• (λ + 1)(λ − 2)(λ − 5) = 0
ASEEM BHAGWAT
ASEEM BHAGWAT
ASEEM BHAGWAT
ASEEM BHAGWAT
• Note that all three eigenspaces are lines through the origin
• Note that the three eigenvectors are linearly independent, and so
they form a basis for R3
ASEEM BHAGWAT
Example
ASEEM BHAGWAT
• (λ − 2)3 = 0
ASEEM BHAGWAT
ASEEM BHAGWAT
• Note that (A – 2I) has rank 1
• So, the corresponding eigenspace is of rank 2, that is a plane through
the origin
• So, we can only find two linearly independent eigenvectors of A
• So, R3 does not have a basis consisting of the eigenvectors of A
ASEEM BHAGWAT
Eigenvalues of Powers
• Note the following:
A2v = A(Av) = A(λv) = λ(Av) = λ(λv) = λ2v
• Thus, λ2 is an eigenvalue of A2, with the eigenvector v
• In fact, we can even say that λk is an eigenvalue of Ak, with the
eigenvector v
ASEEM BHAGWAT
Properties of Eigenvalues
• Product of eigenvalues equals the determinant
• For triangular matrices, diagonal elements are the eigenvalues
• Sum of the eigenvalues equals the sum of the diagonal elements, also
called trace of the matrix
• The trace of the matrix satisfies the following properties:
• tr(A) + tr(B) = tr(A + B)
• tr(kA) = k x tr(A)
• tr(AB) = tr(BA)
• tr(In) = n
ASEEM BHAGWAT
Eigenvalues of Products
• Homework: Find the flaw in the argument:
If λ is an eigenvalue of A, and β is an eigenvalue of B, then the product
is an eigenvalue of AB, because of the following reason:
ASEEM BHAGWAT
Diagonalization
• Consider the previous example
• We know that the eigenvalues are λ1 = 2 and λ2 = 6, with eigenvectors
ASEEM BHAGWAT
• We know that the vectors form a basis for R2, and so every element u
in R2 can be written in a unique form as
u = c1v1 + c2v2, and
Au = 2c1v1 + 6c2v2
• We can rewrite this as:
ASEEM BHAGWAT
• Now, if we write P and D as:
• Thus, we get u = Pc and Au = PDc, and so that Apc = PDc
• This implies (AP − PD)c = 0 for all c in R2, giving us AP = PD
• Since P is invertible, we get P−1AP = D
• Note that this follows from A being a basis for R2
ASEEM BHAGWAT
ASEEM BHAGWAT
• Suppose that A is an n × n matrix, with entries in R
• We say that A is diagonalizable if there exists an invertible matrix P,
with entries in R, such that P−1AP is a diagonal matrix, with entries in
R
• It follows from our Proposition that an n × n matrix A with entries in R
is diagonalizable if its eigenvectors form a basis for Rn
• We also establish the following result:
ASEEM BHAGWAT
ASEEM BHAGWAT
• In the process discussed above, we have assumed that all n roots of
the characteristic polynomial are real
• If not, we can permit the matrix to have complex eigenvalues, and
diagonalize the matrix in Cn
• Example (solve for Homework):
ASEEM BHAGWAT
Homework
ASEEM BHAGWAT
Application of Diagonalization
• Suppose we are given the matrix
• Now suppose we want to find A98
ASEEM BHAGWAT
• First, we diagonalize it as P−1AP = D
• This gives us PDP-1 = A
ASEEM BHAGWAT
• Finally, we calculate A98 as follows:
ASEEM BHAGWAT
