Professional Documents
Culture Documents
14 Complex Matrices (Unitary and Hermitian Matrices) .
14 Complex Matrices (Unitary and Hermitian Matrices) .
Note that if A is a matrix with real entries, then A* AT. To find the conjugate
transpose of a matrix, first calculate the complex conjugate of each entry and then take
the transpose of the matrix, as shown in the following example.
A 冤3 7i
2i
0
4i
冥
3 7i
2i 冤
0
4i 冥
3 7i 2i
A* AT 冤 0 4i 冥
Several properties of the conjugate transpose of a matrix are listed in the following
theorem. The proofs of these properties are straightforward and are left for you to
supply in Exercises 47–50.
UNITARY MATRICES
Recall that a real matrix A is orthogonal if and only if A1 AT. In the complex
system, matrices having the property that A1 A* are more useful, and such matrices
are called unitary.
A Unitary Matrix
冤0 4冥
1 4 0
4
冤
0 1冥
1 0
Because
冤0 1冥 I
1 0
AA* 2
Recall from Section 7.3 that a real matrix is orthogonal if and only if its row
(or column) vectors form an orthonormal set. For complex matrices, this property
characterizes matrices that are unitary. Note that a set of vectors
再v1, v2, . . . , vm冎
in C n (a complex Euclidean space) is called orthonormal when the statements below
are true.
1. 储vi储 1, i 1, 2, . . . , m
2. vi vj 0, i j
The proof of the next theorem is similar to the proof of Theorem 7.8 presented in
Section 7.3.
Show that the complex matrix A is unitary by showing that its set of row vectors forms
an orthonormal set in C 3.
冤 冥
1 1i 1
2 2 2
i i 1
A
冪3 冪3 冪3
5i 3i 4 3i
2冪15 2冪15 2冪15
SOLUTION
Let r1, r2, and r3 be defined as follows.
1 1i 1
冢 2, 冣 冢 冣
i i 1
r1 , , r2 , , ,
2 2 冪3 冪3 冪3
3 i 4 3i
冢2冪15, 2冪15, 冣
5i
r3
2冪15
Begin by showing that r1, r2 , and r3 are unit vectors.
1i
冤 冢2冣冢2冣 冢 冣冢1 2 i冣 冢 12冣冢 21冣冥 冤 4 42 14冥
1 1 1兾2 1 1兾2
储r1储 1
2
冤 冥
1 1 1 1兾2
3 3 3
1
3i 3i 4 3i 4 3i
冤冢2冪15冣冢2冪15冣 冢2冪15冣冢2冪15冣 冢 2冪15 冣冢 2冪15 冣冥
5i 5i 1兾2
储r3储
冤
60 60 60冥
25 10 25 1兾2
1
Then show that all pairs of distinct vectors are orthogonal.
REMARK
r2 r3 冢 冪3
i
冣冢2冪5i15冣 冢冪i 3冣冢23冪15i 冣 冢冪13冣冢42冪153i冣
Try showing that the column 5 1 3i 4 3i
vectors of A also form an 6冪5 6冪5 6冪5
orthonormal set in C 3. 0
So, 再r1, r2, r3冎 is an orthonormal set.
8.5 Unitary and Hermitian Matrices 423
HERMITIAN MATRICES
A real matrix is symmetric when it is equal to its own transpose. In the complex
system, the more useful type of matrix is one that is equal to its own conjugate
transpose. Such a matrix is called Hermitian after the French mathematician Charles
Hermite (1822–1901).
冤 冥
a1 a2i c1 c2i
b1 b2i d1 d2i
a a2i c1 c2i
1 冤
b1 b2i d1 d2i
. 冥
If A is Hermitian, then A A*. So, A must be of the form
a1 b1 b2i
A 冤b 1 b2i d1 冥
.
Similar results can be obtained for Hermitian matrices of order n n. In other words,
a square matrix A is Hermitian if and only if the following two conditions are met.
1. The entries on the main diagonal of A are real.
2. The entry aij in the ith row and the jth column is the complex conjugate of the
entry aji in the jth row and the ith column.
Hermitian Matrices
冤 冥 冤 冥
3 2i 3i 1 2 3
c. 2 i 0 1i d. 2 0 1
3i 1 i 0 3 1 4
SOLUTION
a. This matrix is not Hermitian because it has an imaginary entry on its main diagonal.
b. This matrix is symmetric but not Hermitian because the entry in the first row and
second column is not the complex conjugate of the entry in the second row and first
column.
c. This matrix is Hermitian.
d. This matrix is Hermitian because all real symmetric matrices are Hermitian.
424 Chapter 8 Complex Vector Spaces
REMARK One of the most important characteristics of Hermitian matrices is that their
Note that this theorem implies
eigenvalues are real. This is formally stated in the next theorem.
that the eigenvalues of a real
symmetric matrix are real, as
stated in Theorem 7.7. THEOREM 8.10 The Eigenvalues of a Hermitian Matrix
If A is a Hermitian matrix, then its eigenvalues are real numbers.
PROOF
Let be an eigenvalue of A and let
a1 b1i
冤 冥
a bi
v 2 . 2
..
an bni
be its corresponding eigenvector. If both sides of the equation Av v are multiplied
by the row vector v*, then
v*Av v*共v兲 共v*v兲 共a12 b12 a22 b22 . . . an2 bn2兲.
Furthermore, because
共v*Av兲* v*A*共v*兲* v*Av
it follows that v*Av is a Hermitian 1 1 matrix. This implies that v*Av is a real
number, so is real.
To find the eigenvalues of complex matrices, follow the same procedure as for real
matrices.
冤 冥
3 2i 3i
A 2i 0 1i
3i 1 i 0
SOLUTION
ⱍ ⱍ
The characteristic polynomial of A is
3 2 i 3i
ⱍ I Aⱍ 2 i 1 i
3i 1 i
共 3兲共 2兲 共2 i兲关共2 i兲 共3i 3兲兴
2
3i 关共1 3i兲 3i兴
共 3 2 6兲 共5 9 3i兲 共3i 9 9兲
3 2
3 32 16 12
共 1兲共 6兲共 2兲.
So, the characteristic equation is 共 1兲共 6兲共 2兲 0, and the eigenvalues of A
are 1, 6, and 2.
8.5 Unitary and Hermitian Matrices 425
To find the eigenvectors of a complex matrix, use a procedure similar to that used
for a real matrix. For instance, in Example 5, to find the eigenvector corresponding to
the eigenvalue 1, substitute the value for into the equation
冤 冥冤 冥 冤 冥
3 2 i 3i v1 0
2 i 1 i v2 0
3i 1 i v3 0
to obtain
冤 冥冤 冥 冤 冥
4 2 i 3i v1 0
2 i 1 1 i v2 0 .
3i 1 i 1 v3 0
Solve this equation using Gauss-Jordan elimination, or a graphing utility or software
program, to obtain the eigenvector corresponding to 1 1, which is shown below.
TECHNOLOGY
1
Some graphing utilities and
software programs have built-in
programs for finding the
冤 冥
v1 1 2i
1
eigenvalues and corresponding
eigenvectors of complex
Eigenvectors for 2 6 and 3 2 can be found in a similar manner. They are
matrices. 1 21i 1 3i
冤 冥
6 9i
13
and
冤 冥
2 i , respectively.
5
LINEAR Quantum mechanics had its start in the early 20th century
ALGEBRA as scientists began to study subatomic particles and light.
APPLIED Collecting data on energy levels of atoms, and the rates of
transition between levels, they found that atoms could be
induced to more excited states by the absorption of light.
German physicist Werner Heisenburg (1901–1976) laid a
mathematical foundation for quantum mechanics using
matrices. Studying the dispersion of light, he used vectors to
represent energy levels of states and Hermitian matrices to
represent “observables” such as momentum, position, and
acceleration. He noticed that a measurement yields precisely
one real value and leaves the system in precisely one of a set
of mutually exclusive (orthogonal) states. So, the eigenvalues
are the possible values that can result from a measurement
of an observable, and the eigenvectors are the corresponding
states of the system following the measurement.
Let matrix A be a diagonal Hermitian matrix that represents
an observable. Then consider a physical system whose
state is represented by the column vector u. To measure the
value of the observable A in the system of state u, you can
find the product
冤 冥冤 冥
a11 0 0 u1
u*Au 关u1 u2 u3兴 0 a22 0 u2
0 0 a33 u3
共u1 u1兲a11 共u2 u2兲a22 共u3u3兲a33.
PROOF
To prove part 1, let v1 and v2 be two eigenvectors corresponding to the distinct (and
real) eigenvalues 1 and 2. Because Av1 1v1 and Av2 2v2, you have the
equations shown below for the matrix product 共Av1兲*v2.
共Av1兲*v2 v1*A*v2 v1*Av2 v1*2v2 2v1*v2
共Av1兲*v2 共1v1兲*v2 v1*1v2 1v1*v2
So,
2v1*v2 1v1*v2 0
共 2 1兲v1*v2 0
v1*v2 0 because 1 2
and this shows that v1 and v2 are orthogonal. Part 2 of Theorem 8.11 is often called the
Spectral Theorem, and its proof is left to you.
The eigenvectors of the Hermitian matrix shown in Example 5 are mutually orthogonal
because the eigenvalues are distinct. Verify this by calculating the Euclidean inner
products v1 v2, v1 v3, and v2 v3. For example,
v1 v2 共1兲共1 21i兲 共1 2i兲共6 9i兲 共1兲共13兲
共1兲共1 21i兲 共1 2i兲共6 9i兲 13
1 21i 6 9i 12i 18 13
0.
The other two inner products v1 v3 and v2 v3 can be shown to equal zero in a
similar manner.
冤 冥
3 2i 3i
A 2i 0 1i .
3i 1i 0
SOLUTION
The eigenvectors of A are shown after Example 5. Form the matrix P by normalizing
these three eigenvectors and using the results to create the columns of P.
储v1储 储共1, 1 2i, 1兲储 冪1 5 1 冪7
储v2储 储共1 21i, 6 9i, 13兲储 冪442 117 169 冪728
储v3储 储共1 3i, 2 i, 5兲储 冪10 5 25 冪40
So,
冤 冥
1 1 21i 1 3i
冪7 冪728 冪40
1 2i 6 9i 2 i
P .
冪7 冪728 冪40
1 13 5
冪7 冪728 冪40
Try computing the product P*AP for the matrices A and P in Example 7 to see that
1
冤 冥
0 0
P*AP 0 6 0
0 0 2
where 1, 6, and 2 are the eigenvalues of A.
You have seen that Hermitian matrices are unitarily diagonalizable. It turns out
that there is a larger class of matrices, called normal matrices, that are also unitarily
diagonalizable. A square complex matrix A is normal when it commutes with its
conjugate transpose: AA* A*A. The main theorem of normal matrices states that
a complex matrix A is normal if and only if it is unitarily diagonalizable. You are asked
to explore normal matrices further in Exercise 56.
The properties of complex matrices described in this section are comparable to the
properties of real matrices discussed in Chapter 7. The summary below indicates the
correspondence between unitary and Hermitian complex matrices when compared with
orthogonal and symmetric real matrices.
8.5 Exercises
冤 冥
Finding the Conjugate Transpose In Exercises 1– 4, i i i
determine the conjugate transpose of the matrix.
冪2 冪3 冪6
i 1 2i 2 i i i i
冤 冥 冤 冥
i 16. A
1. 2. 冪2 冪3 冪6
2 3i 1 1
i i
0
冤 冥 冤 冥
0 5 i 冪2i 2 i 冪3 冪6
3. 5i 6 4 4. 5 3i
冪2i 4 3 0 6i Row Vectors of a Unitary Matrix In Exercises 17–20,
(a) verify that A is unitary by showing that its rows are
Finding the Conjugate Transpose In Exercises 5 and orthonormal, and (b) determine the inverse of A.
6, use a software program or graphing utility to find the
冤 冥 冤 冥
conjugate transpose of the matrix. 4 3 1 i 1 i
i
5 5 2 2
17. A 18. A
冤 冥
1i 0 1 i 3 4 1 1
2i 1 0 2i 5 5
i
冪2 冪2
5.
1i i 2 4i
i 2i 1 0
19. A 冤
1 冪3 i 1 冪3 i
2冪2 冪3 i 1 冪3 i 冥
冤 冥
2i 1 1 2i
冤 冥
0 2i 2i 1i 0 1 0
6.
i 2i i 1 1 i 1i
20. A 0
1 2i 4 0 2i 冪6 冪3
2 1
Non-Unitary Matrices In Exercises 7–10, explain why 0
冪6 冪3
the matrix is not unitary.
Identifying Hermitian Matrices In Exercises 21–26,
冤0 冥 冤i 冥
i 0 1 i
7. A 8. A determine whether the matrix is Hermitian.
0 1
冤i 0冥 冤0 冥
0 i i 0
冤 冥
1i 0 i 21. 22.
9. A 冪2 冪2
i
冤 冥 冤 冥
0 1 0 0 2i 1 0 i 1
23. 2 i i 0 24. 2 i i 0
冤 冥
1 1 1i
1 0 1 0 1 0
2 2 2
i 1 i
10. A 2i 3i
冪3
1
冪3
1
冪3
1i
25. 冤2 1i 2 3i 冥
冤 冥
2 2 2 1 冪2 i 5
26. 冪2 i 2 3i
Identifying Unitary Matrices In Exercises 11–16,
5 3i 6
determine whether A is unitary by calculating AA*.
1i 1i 1i 1i Finding Eigenvalues of a Hermitian Matrix In
11. A 冤1 i 1i 冥 12. A 冤1 i 1i 冥 Exercises 27–32, determine the eigenvalues of the matrix A.
冤i 0冥 冤i 3冥
冤 冥
i i 0 i 3 i
27. A 28. A
i
冤 冥 冪2 冪2
0
13. A 14. A
0 i i i 1i 2i
冤1 i 冥 冤2 i 冥
3 0
冪2 冪2
29. A 30. A
2 4
冤 冥
4 3
冤 冥
1 4 1i
5 5
15. A 31. A 0 i 3i
3 4
i i 0 0 2i
5 5
8.5 Exercises 429
冤 冥
i i Unitary Matrices In Exercises 45 and 46, use the result of
2 Exercise 44 to determine a, b, and c such that A is unitary.
冪2 冪2
i
冤 冥
32. A 2 0 6 3i
1 1 a 1 a
冪2
i
45. A
冪2 冤
b c 冥 46. A
冪2
冪45
0 2 b c
冪2
Proof In Exercises 47– 50, prove the formula, where A
Finding Eigenvectors of a Hermitian Matrix In and B are n ⴛ n complex matrices.
Exercises 33 – 36, determine the eigenvectors of the 47. 共A*兲* A 48. 共A B兲* A* B*
matrix in the indicated exercise.
49. 共kA兲* kA* 50. 共AB兲* B*A*
33. Exercise 27 34. Exercise 30
51. Proof Let A be a matrix such that A* A O. Prove
35. Exercise 31 36. Exercise 28 that iA is Hermitian.
Diagonalization of a Hermitian Matrix In Exercises 52. Show that det共A兲 det共A兲, where A is a 2 2 matrix.
37–41, find a unitary matrix P that diagonalizes the
Determinants In Exercises 53 and 54, assume that the
matrix A.
result of Exercise 52 is true for matrices of any size.
2i
冤i 0冥 冤2 i 冥
0 i 0 53. Show that det共A*兲 det共A兲.
37. A 38. A
4
54. Prove that if A is unitary, then ⱍdet共A兲ⱍ 1.
冤 冥
i i
2 55. (a) Prove that every Hermitian matrix A can be written as
冪2 冪2
the sum A B iC, where B is a real symmetric
i
39. A 2 0 matrix and C is real and skew-symmetric.
冪2
(b) Use part (a) to write the matrix
i
0 2
1i
冪2
冤1 i 冥
2
A
3
4 2 2i
40. A 冤
2 2i 6 冥 as the sum A B iC, where B is a real symmetric
matrix and C is real and skew-symmetric.
冤 冥
1 0 0
(c) Prove that every n n complex matrix A can be
41. A 0 1 1 i
written as A B iC, where B and C are
0 1 i 0
Hermitian.
(d) Use part (c) to write the complex matrix
42. Consider the following matrix.
冤2 i 冥
i 2
冤 冥
2 3 i 4 i A
1 2i
A 3i 1 1i
4i 1i 3 as the sum A B iC, where B and C are
Hermitian.
(a) Is A unitary? Explain.
56. (a) Prove that every Hermitian matrix is normal.
(b) Is A Hermitian? Explain.
(b) Prove that every unitary matrix is normal.
(c) Are the row vectors of A orthonormal? Explain.
(c) Find a 2 2 matrix that is Hermitian, but not unitary.
(d) The eigenvalues of A are distinct. Is it possible
to determine the inner products of the pairs of (d) Find a 2 2 matrix that is unitary, but not
eigenvectors by inspection? If so, state the Hermitian.
value(s). If not, explain why not. (e) Find a 2 2 matrix that is normal, but neither
(e) Is A unitarily diagonalizable? Explain. Hermitian nor unitary.
(f) Find the eigenvalues and corresponding eigenvectors
of your matrix in part (e).
43. Show that A In is unitary by computing AA*.
(g) Show that the complex matrix
44. Let z be a complex number with modulus 1. Show that
冤0 i冥
the matrix A is unitary. i 1
冤 冥
1 z z
A
冪2 iz iz is not diagonalizable. Is this matrix normal?
430 Chapter 8 Complex Vector Spaces
8 Review Exercises
Operations with Complex Numbers In Exercises Converting to Polar Form In Exercises 31– 36,
1–6, perform the operation. determine the polar form of the complex number.
1. Find u z: u 2 4i, z 4i 31. 4 4i 32. 2 2i
2. Find u z: u 4, z 8i 33. 冪3 i 34. 1 冪3i
3. Find uz: u 4 2i, z 4 2i 35. 7 4i 36. 3 2i
4. Find uz: u 2i, z 1 2i
Converting to Standard Form In Exercises 37– 42,
u find the standard form of the complex number.
5. Find : u 6 2i, z 3 3i
z
u
6. Find : u 7 i, z i
冤 冢 6 冣 i sin冢 6 冣冥
37. 5 cos
z
Finding Zeros In Exercises 7–10, use the given zero to
冤 冢 3 冣 i sin冢 3 冣冥
38. 2 cos
Vector Operations in C n
冤 冥
In Exercises 55 – 58, find the 9 2i 2
indicated vector using u ⴝ 冇4i, 2 ⴙ i冈, v ⴝ 冇3, ⴚi 冈, and 76. 2 i 0 1 i
w ⴝ 冇3 ⴚ i, 4 ⴙ i冈. 2 1 i 3
55. 7u v 56. 3iw 共4 i兲v
57. iu iv iw 58. 共3 2i兲u 共2i兲w Finding Eigenvalues and Eigenvectors In Exercises
77 and 78, find the eigenvalues and corresponding
Finding the Euclidean Norm In Exercises 59 and 60, eigenvectors of the matrix.
determine the Euclidean norm of the vector.
冤 冥
2 0 i
2i
冤 冥
4
59. v 共3 5i, 2i兲 60. v 共3i, 1 5i, 3 2i兲 77. 78. 0 3 0
2i 0
i 0 2
Finding the Euclidean Distance In Exercises 61 and
62, find the Euclidean distance between the vectors. 79. Proof Prove that if A is an invertible matrix, then A*
61. v 共2 i, i兲, u 共i, 2 i兲 is also invertible.
62. v 共2 i, 1 2i, 3i兲, u 共4 2i, 3 2i, 4兲 80. Determine all complex numbers z such that z z.
81. Proof Prove that if z is a zero of a polynomial
Finding Complex Inner Products In Exercises 63 – 66,
equation with real coefficients, then the conjugate of z
use the inner product 具u, v典 ⴝ u1v1 ⴙ 2u2v2 to find 具u, v典.
is also a zero.
63. u 共i, 3i兲 and v 共2i, 2i兲
82. (a) Find the determinant of the Hermitian matrix
64. u 共i, 2 i兲 and v 共5 i, i兲
冤 冥
3 2i 3i
65. u 共1 i, 1 2i兲 and v 共2 i, 1 2i兲 2i 0 1i .
66. u 共2 2i, 1兲 and v 共3 4i, 2兲 3i 1i 0
Finding the Conjugate Transpose In Exercises 67–70, (b) Proof Prove that the determinant of any
determine the conjugate transpose of the matrix. Hermitian matrix is real.
1 4i 3i 83. Proof Let A and B be Hermitian matrices. Prove that
67. A 冤 3i 2i 冥 AB BA if and only if AB is Hermitian.
84. Proof Let u be a unit vector in C n. Define
2i 2i
68. A 冤
2 2i 冥
H I 2uu*. Prove that H is an n n Hermitian and
1 2i
unitary matrix.
冥
2i 3 2i
冤
5
85. Proof Use mathematical induction to prove
69. A 2 2i 3 2i i
DeMoivre’s Theorem.
3i 2i 1 2i
86. Proof Show that if z1 z2 and z1z2 are both nonzero
冤 冥
2 1i i real numbers, then z1 and z2 are both real numbers.
70. A i 2 2i 0
87. Proof Prove that if z and w are complex numbers,
1 1i 2i then
Identifying Unitary Matrices In Exercises 71–74, ⱍz wⱍ
ⱍzⱍ ⱍwⱍ.
determine whether the matrix is unitary. 88. Proof Prove that for all vectors u and v in a complex
冤 冥 冤 冥
i 1 2i 1i inner product space,
冪2 冪2 4 4 具u, v典 14 关 储u v储2 储u v储2 i储u iv储2
71. 72.
i 1 i 冪2 i储u iv储2兴.
冪2 冪2 冪3 冪3
True or False? In Exercises 89 and 90, determine
冤 冥
1 0 1
冪2 冪2
whether each statement is true or false. If a statement is
true, give a reason or cite an appropriate statement from
冤 冥
1 0
73. 74. 0 i 0 the text. If a statement is false, provide an example that
i i
1i 1 i shows the statement is not true in all cases or cite an
0
2 2 appropriate statement from the text.
89. A square complex matrix A is called normal if it
Identifying Hermitian Matrices In Exercises 75 and
commutes with its conjugate transpose so that
76, determine whether the matrix is Hermitian.
AA* A*A.
冤 冥
1 1 i 2 i
90. A square complex matrix A is called Hermitian if
75. 1 i 3 i
A A*.
2i i 4
432 Chapter 8 Complex Vector Spaces
8 Project
A 冤0.8
0.6 0.8
0.6 冥
are complex, the same principles apply, and you can obtain complex exponential
solutions.
2. Convert the complex solutions to real solutions by observing that if a bi
is a (complex) eigenvalue of A with (complex) eigenvector v, then the real and
imaginary parts of etv form a linearly independent pair of (real) solutions. Use
the formula ei cos i sin .
3. Use the initial conditions to find the explicit form of the (real) solutions of the
original equations.
4. If you have access to a graphing utility or software program, graph the solutions
obtained in part 3 over the domain 0
t
3. At what moment are the two
populations equal?
5. Interpret the solution in terms of the long-term population trend for the two
species. Does one species ultimately disappear? Why or why not? Contrast this
solution to that obtained for the model in Chapter 7.
6. If you have access to a graphing utility or software program that can numerically
solve differential equations, use it to graph the solutions of the original system
of equations. Does this numerical approximation appear to be accurate?
Supri Suharjoto/Shutterstock.com