Linear Algebra Problem Set 8: Eigenvalues and Characteristic Polynomials, Assignments of Linear Algebra

Solutions to problem set 8 in math 110: linear algebra. The problems deal with eigenvalues, minimal polynomial, and characteristic polynomial. Topics include showing that matrices with equal eigenvalues have identical eigenvectors and traces, finding the exponential form of a matrix given its eigenvalues, and proving that two matrices with the same determinant and trace may or may not have the same characteristic polynomial depending on their size.

Typology: Assignments

Pre 2010

Uploaded on 10/01/2009

koofers-user-jpm
koofers-user-jpm 🇺🇸

10 documents

1 / 3

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Problem Set 8 (due Friday October 29)
MATH 110: Linear Algebra
Each problem is worth 10 points.
PART 1
1. Curtis p. 201 6.
2. Curtis p. 215 9.
1: Notice that if T2= 1 then the minimal polynomial of T m(x) divides
(x1)(x+ 1). There are 3 cases: m(x) = x+ 1, m(x) = x1, m(x) =
(x+ 1)(x1). In each case we can apply the triangular form theorem to find
that V+and Vare exactly the nullspaces appearing in the direct sum.
2: a) It suffices to show that AB and BA have the same eigenvalues.
Suppose λis an eigenvalue of AB. Then ABx =λx so BABx =λBx so Bx
is an eigenvector of BA with eigenvalue λ. Similarly, if µis an eigenvalue
of BA then it is also an eigenvalue of AB. Therefore AB and BA have the
same eigenvalues. So their traces, which are the sum of their eigenvalues, are
equal. b) A=C1BC or equivalently CA =BC for some C. Now
trace(B) = trace(BCC 1) = trace(C AC 1) = trace(AC C1) = trace(A).
c) We think of trace as a linear transformation T r :MnF. We are in-
terested in the nullity (or dimensional of the nullspace of T r). This can be
computed from the rank and nullity theorem to be n2rwhere ris the
dimension of the range of T r, which is 1. d) Matrices of the form AB BA
have trace 0 from part a) and therefore the subspace generated by them is
a subspace of the nullspace of T r. Call this subspace Z. Let A, B simulta-
neously range over all matrices with only one 1 and all the rest zero, where
the 1 does not lie on the diagonal. Looking at AB BA, we get matrices
that are either 0, or else have only one 1 in some position not on the di-
agonal, or else 1,1 consecutively somewhere on the diagonal. There are
n2n+n1 = n21 such matrices, and they are clearly independent,
therefore Zhas dimension at least n21. It follows that Zis the nullspace
of T r.
Note: part d) should have read: “is generated by matrices of the form
AB BA where A, B Mn(F)”.
1
pf3

Partial preview of the text

Download Linear Algebra Problem Set 8: Eigenvalues and Characteristic Polynomials and more Assignments Linear Algebra in PDF only on Docsity!

Problem Set 8 (due Friday October 29) MATH 110: Linear Algebra

Each problem is worth 10 points. PART 1

  1. Curtis p. 201 6.
  2. Curtis p. 215 9.

1: Notice that if T 2 = 1 then the minimal polynomial of T m(x) divides (x − 1)(x + 1). There are 3 cases: m(x) = x + 1, m(x) = x − 1 , m(x) = (x + 1)(x − 1). In each case we can apply the triangular form theorem to find that V+ and V− are exactly the nullspaces appearing in the direct sum. 2: a) It suffices to show that AB and BA have the same eigenvalues. Suppose λ is an eigenvalue of AB. Then ABx = λx so BABx = λBx so Bx is an eigenvector of BA with eigenvalue λ. Similarly, if μ is an eigenvalue of BA then it is also an eigenvalue of AB. Therefore AB and BA have the same eigenvalues. So their traces, which are the sum of their eigenvalues, are equal. b) A = C−^1 BC or equivalently CA = BC for some C. Now

trace(B) = trace(BCC−^1 ) = trace(CAC−^1 ) = trace(ACC−^1 ) = trace(A).

c) We think of trace as a linear transformation T r : Mn → F. We are in- terested in the nullity (or dimensional of the nullspace of T r). This can be computed from the rank and nullity theorem to be n^2 − r where r is the dimension of the range of T r, which is 1. d) Matrices of the form AB − BA have trace 0 from part a) and therefore the subspace generated by them is a subspace of the nullspace of T r. Call this subspace Z. Let A, B simulta- neously range over all matrices with only one 1 and all the rest zero, where the 1 does not lie on the diagonal. Looking at AB − BA, we get matrices that are either 0, or else have only one 1 in some position not on the di- agonal, or else 1, −1 consecutively somewhere on the diagonal. There are n^2 − n + n − 1 = n^2 − 1 such matrices, and they are clearly independent, therefore Z has dimension at least n^2 − 1. It follows that Z is the nullspace of T r. Note: part d) should have read: “is generated by matrices of the form AB − BA where A, B ∈ Mn(F )”.

PART 2

Problem 1(20) a) Prove that if A is an n × n matrix with all its eigenvalues equal to λ, then

exA^ = eλx

n∑− 1

k=

xk k!

(A − λI)k.

Proof: Simply plug into Putzer’s method (optional problem). The poly-

nomials Pk(A) are just (A − λI)k. Clearly rk(x) = x

k k! e

λx (^) (check by differen-

tiating that it satisfies the recurrence). b) A 3 × 3 matrix A has all its eigenvalues equal to λ. Show that

exA^ =

eλx((λ^2 x^2 − 2 λx + 2)I + (− 2 λx^2 + 2x)A + x^2 A^2 ).

Solution: Plug into part a) with n = 3. Problem 2(15) Curtis p. 226 2,3. Solutions in book. Problem 3 (10) Let A and B be n × n matrices with detA = detB and trA = trB. Prove that A and B have the same characteristic polynomial if n = 2, but that this need not be the case if n > 2. Solution: If n = 2 then the characteristic polynomial of a matrix looks like x^2 + ax + b where a is the negative of the trace and b is the determinant. So if n = 2 then the conditions on A and B ensure that the characteristic polynomial is the same. When n > 2 the only two of the coefficients have to be the same, but the actual characteristic polynomials don’t have to be. For example, consider the 3 × 3 diagonal matrices diag(2, 2 , 8) and diag(4, 2 , 4) (diag means the entries appear in the diagonal and all other entries are 0). It is easy to check these matrices have different characteristic polynomials but the same determinant and trace. Problem 4 (10) Suppose that the minimal polynomial of a linear transformation T : V → V satisfies m(x) = (x − α 1 )e 1 · · · (x − αs)es

as in the Triangular Form Theorem. Find the rational canonical form of T.