Solutions: Vandermonde, Diagonal Matrices, Jordan Form, Isomorphism, Plane, Assignments of Linear Algebra

Solutions to problem set 9 of math 110: linear algebra. The solutions cover topics such as vandermonde determinant, similarity of diagonal matrices, jordan canonical form, isomorphism of vector spaces, plane in r3, and minimal polynomial. The document also includes proofs for some of the statements.

Typology: Assignments

Pre 2010

Uploaded on 10/01/2009

koofers-user-t0d-1
koofers-user-t0d-1 🇺🇸

8 documents

1 / 3

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Problem Set 9 (due November 12)
MATH 110: Linear Algebra
Each problem is worth 10 points.
PART 1
1. Curtis p. 161 10.
2. Curtis p. 226 7.
3. Curtis p. 243 6.
Solutions: 1. Vandermonde Determinant: Let Dbe the Vandermonde
determinant and let A=Qi<j(ζiζj). Let x=ζi. Then by expanding across
the ith row of the Vandermonde matrix in the determinant calculation, we
see that D=p(x) where pis a polynomial of degree n1. In fact, since ζj
is a root of pfor all j6=i(since if x=ζjwe have two rows of the matrix
equal and hence its determinant is 0) we have that D=c(xζ1· · · (xζn)
with the term (xζi) missing in the product. Since iis arbitrary in the
argument above, we see that D=kA where kis a polynomial in ζ1,...ζn.
But examining the coefficient of ζn1
1ζn2
2· · · 1 in D, we see that it is either
1 or 1, and therefore kmust be either 1 or 1.
2. Similarity of Diagonal matrices: Two similar matrices have the same
eigenvalues, and diagonal matrices have as their diagonal entries their eigen-
values. Using these two facts, we see that if the matrices are diagonal and
similar, then certainly their diagonals are rearrangements of each other (since
they must have the same elements, namely the eigenvalues). Conversely, if
the diagonal elements are rearrangements of each other, then the matrices are
similar because they represent the same linear transformation, with respect
to the same basis except with the basis elements rearranged.
3. We are assuming that Vis finite dimensional. Suppose that λv=λw.
Then for any linear functional f, we have that f(v) = f(w). Therefore
f(vw) = 0. Let B={v1, . . . , vn}be a basis for Vand C={fi}be the
dual basis of B. Then in particular, fi(vw) = 0 for every i, and so if we
write vw=Pn
i=1 civi, we have that ci= 0 for every i. Therefore vw= 0
and v=w. Thus we have that our map is 1 1, and it follows that it is an
isomorphism, since dim(V)) = dim(V). To see this, let D={µi}be the
dual basis of C, that is µi(fj) = 0 if j6=iand 1 if j=i. Notice that in fact,
1
pf3

Partial preview of the text

Download Solutions: Vandermonde, Diagonal Matrices, Jordan Form, Isomorphism, Plane and more Assignments Linear Algebra in PDF only on Docsity!

Problem Set 9 (due November 12) MATH 110: Linear Algebra

Each problem is worth 10 points. PART 1

  1. Curtis p. 161 10.
  2. Curtis p. 226 7.
  3. Curtis p. 243 6.

Solutions: 1. Vandermonde Determinant: Let D be the Vandermonde determinant and let A =

∏ i<j (ζi^ −ζj^ ). Let^ x^ =^ ζi. Then by expanding across the ith row of the Vandermonde matrix in the determinant calculation, we see that D = p(x) where p is a polynomial of degree n − 1. In fact, since ζj is a root of p for all j 6 = i (since if x = ζj we have two rows of the matrix equal and hence its determinant is 0) we have that D = c(x − ζ 1 · · · (x − ζn) with the term (x − ζi) missing in the product. Since i is arbitrary in the argument above, we see that D = kA where k is a polynomial in ζ 1 ,... ζn. But examining the coefficient of ζ 1 n −^1 ζ 2 n −^2 · · · 1 in D, we see that it is either 1 or −1, and therefore k must be either 1 or −1.

  1. Similarity of Diagonal matrices: Two similar matrices have the same eigenvalues, and diagonal matrices have as their diagonal entries their eigen- values. Using these two facts, we see that if the matrices are diagonal and similar, then certainly their diagonals are rearrangements of each other (since they must have the same elements, namely the eigenvalues). Conversely, if the diagonal elements are rearrangements of each other, then the matrices are similar because they represent the same linear transformation, with respect to the same basis except with the basis elements rearranged.
  2. We are assuming that V is finite dimensional. Suppose that λv = λw. Then for any linear functional f , we have that f (v) = f (w). Therefore f (v − w) = 0. Let B = {v 1 ,... , vn} be a basis for V and C = {fi} be the dual basis of B. Then in particular, fi(v − w) = 0 for every i, and so if we write v − w =

∑n i=1 civi, we have that^ ci^ = 0 for every^ i. Therefore^ v^ −^ w^ = 0 and v = w. Thus we have that our map is 1 − 1, and it follows that it is an isomorphism, since dim(V ∗)∗) = dim(V ). To see this, let D = {μi} be the dual basis of C, that is μi(fj ) = 0 if j 6 = i and 1 if j = i. Notice that in fact,

mui = λvi. This is because λvi (fj ) = fj (vi) which is 1 if i = j and 0 if i 6 = j. Then λv 1 ,... , λvn is a basis for (V ∗)∗. Therefore the spaces have the same dimension (since they have bases of the same size). PART 2 Problem 1(20) Let V be a real vector space of functions spanned by the set of real values functions {ex, xex, x^2 ex, e^2 x} and let T be the linear transformation T : V → V defined by T (f ) = f ′, the derivative of f. Find the Jordan canonical form of T. Solution: The characteristic polynomial is h(x) = (x − 1)^3 (x − 2) which is also the minimal polynomial (this requires a bit of checking). It follows, by computing the companion matrices, that the Jordan canonical form is

  

  .

Problem 2(10) Prove that if V is isomorphic to W then V ∗^ is isomorphic to W ∗. Is the converse true (prove or give a counterexample)? Solution: I meant for you to assume that V, W are finite dimensional, in which case the claim is true (and the converse), by the same arguments as number 3 in part 1. Problem 3 (10) a) Let T : R → R be a linear transformation. Show that T (x) = cx where c ∈ R is some constant. Proof: We will use the fact that L(R, R) is isomorphic to M 1 , 1 (R) (the 1 ×1 real matrices). Given this fact, the claim is clear since using the standard basis to represent T , we find that Ax = cx. b) Let T : R 2 → R be a linear transformation. Show that T (x, y) = c 1 x + c 2 y. Proof: Same as above, except now we find that L(R 2 , R) is isomorphic to M 1 , 2 (R). Again, using the standard basis, we find that Avt^ (now v is a 2 × 1 vector v = [xy]) is just c 1 x + c 2 y. c) Generalize parts a) and b) to a linear transformation of the form T : Rn → R.