Beruflich Dokumente
Kultur Dokumente
:
:
(*)
:
:
Where the coefficients aijs, and gis are arbitrary functions of t. If every
term gi is constant zero, then the system is said to be homogeneous.
Otherwise, it is a nonhomogeneous system if even one of the gs is nonzero.
D-1 - 1
x a11
1
x 2 a 21
a
x3 = 31
: :
:
:
x a n1
n
x
a12
a 22
a32
:
:
an 2
... a1n x1 g1
... a 2 n x 2 g 2
... a3n x3 g 3
+
:
: : :
:
: : :
... a nn x n g n
x
1
x2
x3
x =
,
:
x n
x1
x
2
x3
x= ,
:
xn
g1
g
2
g3
g= .
:
g n
For a homogeneous system, g is the zero vector. Hence it has the form
x = Ax.
D-1 - 2
Examples:
(i) The mechanical vibration equation m u + u + k u = F(t) is equivalent to
x1 =
x2 =
x2
k
F (t )
x1 x2 +
m
m
m
(ii)
y 2y + 3y 4y = 0
is equivalent to
x1 = x2
x2 = x3
x3 = 4 x1 3 x2 + 2 x3
D-1 - 3
This process can be easily generalized. Given an n-th order linear equation
any(n) + an1 y(n1) + an2 y(n2) + + a2 y + a1 y + a0 y = g(t).
Make the substitutions: x1 = y, x2 = y, x3 = y, , xn = y(n1), and xn = y(n).
The first n 1 equations follow thusly. Lastly, substitute the xs into the
original equation to rewrite it into the n-th equation and obtain the system of
the form:
x1
x2
x3
:
:
xn1
xn
=
=
=
:
:
=
x2
x3
x4
:
:
xn
a0
a
a
a
g (t )
x1 1 x2 2 x3 ... n1 xn +
an
an
an
an
an
D-1 - 4
Exercises D-1.1:
1 3 Convert each linear equation into a system of first order equations.
1.
y 4y + 5y = 0
2.
y 5y + 9y = t cos 2t
y(4) + 3y y + 2y 6y = 11
3.
4. Rewrite the system you found in (a) Exercise 1, and (b) Exercise 2, into a
matrix-vector equation.
5. Convert the third order linear equation below into a system of 3 first
order equation using (a) the usual substitutions, and (b) substitutions in the
reverse order: x1 = y, x2 = y, x3 = y. Deduce the fact that there are multiple
ways to rewrite each n-th order linear equation into a linear system of n
equations.
y + 6y + y 2y = 0
Answers D-1.1:
1.
x1 = x2
x2 = 5x1 + 4x2
3.
4. (a)
5. (a)
x1 =
x2 =
x3 =
x4 =
2.
x1 = x2
x2 = x3
x3 = 9x1 + 5x3 + t cos 2t
x2
x3
x4
6x1 2x2 + x3 3x4 + 11
0 1
x =
x
5 4
x1 = x2
x2 = x3
x3 = 2x1 x2 6x3
(b)
(b)
0 1 0 0
0 0 1 0
x =
x +
9 0 5 t cos 2t
x1 = 6x1 x2 + 2x3
x2 = x1
x3 = x2
D-1 - 5
Review topics:
1. What is a matrix (pl. matrices)?
A matrix is any rectangular array of numbers (called entries). Each
entrys position is addressed by the row and column (in that order)
where it is located. For example, a52 represents the entry positioned at
the 5th row and the 2nd column of the matrix A.
D-1 - 6
1 0
I2 = 0 1 ,
1 0 0
0 1 0
, etc.
I3 =
0 0 1
x = any n 1 vector
a b e
c d g
f a e b f
=
h c g d h
D-1 - 7
a b ka kb
k
=
,
c d kc kd
a b
c d
e
g
f ae + bg
=
h ce + dg
af + bh
cf + dh
a b x ax + by
c d y = cx + dy .
D-1 - 8
a b
det
= ad bc
c d
Note: The determinant is a function whose domain is the set of all
square matrices of a certain size, and whose range is the set of all real
(or complex) numbers.
a b
c d ,
A =
1
ad bc
d b
c a .
D-1 - 9
Examples: Let A =
1 2
5 2
and B =
2 3
1 4 .
4 (3)
1 2 2 3 2 2
=
(i) 2A B = 2
4 4
5 2 1 4 10 (1)
0 1
= 11 0
1 2 2 3 2 + 2 3 8 4 11
1 4 = 10 2 15 + 8 = 8 7
(ii) AB = 5
2
2 3 1 2 2 15 4 6 13 10
= 1 + 20 2 + 8 = 19
BA = 1 4 5
2
10
(iii) det(A) = 2 (10) = 12,
det(B) = 8 3 = 5.
(iv) A =
1
2 (10)
1/ 6
2 2 1 2 2 1 / 6
=
=
5 1 12 5 1 5 / 12 1 / 12
D-1 - 10
D-1 - 11
a b
r
A rI =
c
d
1 0 a r
0 1 = c
b
d r .
D-1 - 12
b
a r
det
= (a r )(d r ) bc = r 2 (a + d )r + (ad bc) = 0
d r
c
It is a degree 2 polynomial equation of r, as you can see.
This polynomial on the left is called the characteristic polynomial of
the (original) matrix A, and the equation is the characteristic equation
of A. The root(s) of the characteristic polynomial are the eigenvalues
of A. Since any degree n polynomial always has n roots (real and/or
complex; not necessarily distinct), any n n matrix always has at least
one, and up to n different eigenvalues.
Once we have found the eigenvalue(s) of the given matrix, we put
each specific eigenvalue back into the linear system (A r I) x = 0 to
find the corresponding eigenvectors.
D-1 - 13
Examples:
A=
2 3
4 3
3
2 3
1 0 2 r
r
=
0 1 4
A rI =
3 r .
4 3
Its characteristic equation is
3
2 r
det
= (2 r )(3 r ) 12 = r 2 5r 6 = (r + 1)(r 6) = 0
3 r
4
The eigenvalues are, therefore, r = 1 and 6.
Next, we will substitute each of the 2 eigenvalues into the matrix
equation (A r I) x = 0.
For r = 1, the system of linear equations is
2 + 1 3
3 3
0
x
=
x
=
4 4
0 .
(A r I) x = (A + I) x =
3 + 1
4
Notice that the matrix equation represents a degenerated system of 2
linear equations. Both equations are constant multiples of the
equation x1 + x2 = 0. There is now only 1 equation for the 2
unknowns, therefore, there are infinitely many possible solutions.
This is always the case when solving for eigenvectors. Necessarily,
there are infinitely many eigenvectors corresponding to each
eigenvalue.
D-1 - 14
1
k1 = .
1
3
2 6
4 3
0
x
=
x
=
4 3
0 .
(A r I) x = (A 6 I) x =
3 6
4
Both equations in this second linear system are equivalent to
4 x1 3 x2 = 0. Its solutions are given by the relation 4 x1 = 3 x2.
Hence, the eigenvectors corresponding to r = 6 are all nonzero
multiples of
3
k2 = .
4
D-1 - 15
a 0
0 d , or
a b
0 d , or
a 0
c d .
Then the eigenvalues are just the main diagonal entries, r = a and d in all 3
examples above.
2. If A is any 2 2 matrix, then its characteristic equation is
a r
det
c
b
= r 2 (a + d )r + (ad bc) = 0
d r
r 2 Trace(A) r + det(A) = 0.
Note: For any square matrix A, Trace(A) = [sum of all entries on the main
diagonal (running from top-left to bottom-right)]. For a 2 2 matrix A,
Trace(A) = a + d.
D-1 - 16
Example:
2 3
4 3
3 3
.
which has roots r = 1 and 6. For r = 1, the matrix (A r I) is
4 4
Take the first row, (3, 3), which is a non-zero vector; put a minus sign to the
first entry to get (3, 3); then switch the entry, we now have k1 = (3, 3). It
is indeed an eigenvector, since it is a nonzero constant multiple of the vector
we found earlier.
On very rare occasions, both rows of the matrix (A r I) have all zero
entries. If so, the above algorithm will not be able to find an eigenvector.
Instead, under this circumstance any non-zero vector will be an eigenvector.
D-1 - 17
Exercises D-1.2:
Let C =
5 1
7
3
and
D=
0
2
2 1 .
Answers D-1.2:
1 1
25 3
,
(ii)
31 14
1
8 1
10 2
,
(ii)
2. (i)
3
1
8 3
1. (i)
3
s
s
5. (i) r1 = 2, k1 =
; r2 = 4, k 2 = ; s = any nonzero number
7 s
s
s
0
(ii) r1 = 2, k1 =
; r2 = 1, k 2 = ; s = any nonzero number
2 s / 3
s
4. (i)
7/8
D-1 - 18
x1 = a x1 + b x2
x2 = c x1 + d x2
It has the alternate matrix-vector representation
x =
a b
c d x.
D-1 - 19
A related note, (from linear algebra,) we know that eigenvectors that each
corresponds to a different eigenvalue are always linearly independent from
each others. Consequently, if r1 and r2 are two different eigenvalues, then
their respective eigenvectors k1 anf k2, and therefore the corresponding
solutions, are always linearly independent.
D-1 - 20
Case I
If the coefficient matrix A has two distinct real eigenvalues r1 and r2, and
their respective eigenvectors are k1 and k2. Then the 2 2 system x = Ax
has a general solution
x = C1 k1 e r1 t + C2 k2 e r2 t .
Example:
x =
2 3
4 3 x.
1
k1 = ,
1
3
k2 = .
4
1 t
3 6 t
x = C1 e + C2 e
1
4
D-1 - 21
Example:
x =
3 2
2 2 x,
1
x(0) = 1
3 + 1 2
4 2
0
x
=
x
=
2 1
0 .
(A r I) x = (A + I) x =
2 + 1
2
Solving the bottom equation of the system: 2x1 x2 = 0, we get the
relation x2 = 2 x1. Hence,
1
k1 = ,
2
For r = 2, the system is
2
3 2
1 2
0
x
=
x
=
2 4
0 .
(A r I) x = (A 2 I) x =
2 2
2
Solving the first equation of the system: x1 2x2 = 0, we get the
relation x1 = 2x2. Hence,
2
k2 = .
1
D-1 - 22
1
x = C1 e t + C2
2
2 2 t
1 e .
1
x(0) = C1 e 0 + C2
2
2 0 C1 + 2C2 1
1 e = 2C + C = 1 .
2
1
That is
C1 + 2C2
2C1 + C2
= 1
= 1 .
1 t
2 2 t e t + 2e 2 t
x= e + e =
t
2t .
2
1
2e + e
D-1 - 23
Case II
A little detail: Similar to what we have done before, first there was the
complex-valued general solution in the form
x = C1 k1 e( + i ) t + C2 k2 e( i ) t .
We filter out the imaginary parts by carefully choosing two sets of
coefficients to obtain two corresponding real-valued solutions that are also
linearly independent:
u = e t (a cos( t ) b sin( t ) )
v = e t (a sin( t ) + b cos( t ) )
D-1 - 24
Example:
x =
2 5
1 2 x
5
2 i
0
x
=
0 .
(A r I) x =
2 i
1
Solving the first equation of the system: (2 i) x1 5x2 = 0, we get
the relation (2 i) x1 = 5x2. Hence,
5 5 0
k=
= 2 + 1 i = a + bi
2
0
0
x = C1 e0t cos(t ) sin(t ) + C2 e0 t sin(t ) + cos(t )
1
1
2
5 cos(t )
5 sin(t )
+ C2
= C1
2
cos(
t
)
+
sin(
t
)
2
sin(
t
)
cos(
t
)
D-1 - 25
Example:
x =
1 6
3
5 x,
0
x(0) = 2 .
6
1 ( 2 + 3i )
3 3i 6 0
x
=
3
= 0 .
(A r I) x =
3
5
(
2
+
3
i
)
3
3
i
Solving the second equation of the system: 3x1 + (3 3i) x2 = 0, we
get the relation x1 = (1 + i) x2. Hence,
1 + i 1 1
k=
= 1 + 0 i = a + bi
1
1
1
x = C1 e 2t cos(3t ) sin(3t ) + C2 e 2 t sin(3t ) + cos(3t )
0
0
1
cos(3t ) sin(3t )
2 t cos(3t ) sin(3t )
+ C2 e
= C1 e
cos(
3
t
)
sin(
3
t
)
2t
D-1 - 26
1
1
x(0) = C1 e 0 cos(0) sin(0) + C2 e 0 sin(0) + cos(0)
0
0
1
1
1 C + C2 0
= C1 + C2 = 1
=
1
0 C1 2
Therefore, C1 = 2 and C2 = 2. Consequently, the particular solution is
cos(3t ) sin(3t )
cos(3t ) sin(3t )
+ 2e 2t
x = 2e 2t
cos(3t )
sin(3t )
4 sin(3t )
2t
=e
2 cos(3t ) + 2 sin(3t )
D-1 - 27
Case III
Suppose the coefficient matrix A has a repeated real eigenvalues r, there are
2 sub-cases.
(i) If r has two linearly independent eigenvectors k1 and k2. Then the 2 2
system x = Ax has a general solution
x = C1 k1 e rt + C2 k2 e rt.
Note: For 2 2 matrices, this possibility only occurs when the coefficient
matrix A is a scalar multiple of the identity matrix. That is, A has the form
1 0 0
= 0 ,
0
1
Example:
x =
2 0
0 2 x.
1
k1 = ,
0
0
k2 = .
1
1
x = C1 e 2t + C2
0
2008, 2012 Zachary S Tseng
0 2 t
1 e .
D-1 - 28
x = C1 k e rt + C2 (k t e rt + e rt).
Where the second vector is any solution of the nonhomogeneous linear
system of algebraic equations
(A r I ) = k .
Example:
x =
1 4
4 7 x,
2
x(0) = 1 .
4
1 + 3
4 4
0
x
=
x
=
4 4
0 .
(A r I) x =
7 + 3
4
Both equations of the system are 4x1 4x2 = 0, we get the same
relation x1 = x2. Hence, there is only one linearly independent
eigenvector:
1
k= .
1
D-1 - 29
4 4
4 4
1
= 1 .
2
1 / 4
Choose 2 = 0, we get = 0 .
A general solution is, therefore,
1 3t 1 4 3t
1 3 t
x = C1 e + C2 t e + e
1
0
1
1
12t 2 3t
1
1 4
x = e 3 t 12 t e 3t + e 3t =
e
1
0
1
12t + 1
D-1 - 30
x = Ax.
First find the two eigenvalues, r, and their respective corresponding
eigenvectors, k, of the coefficient matrix A. Depending on the eigenvalues
and eigenvectors, the general solution is:
I. Two distinct real eigenvalues r1 and r2:
x = C1 k1 e r1 t + C2 k2 e r2 t .
II. Two complex conjugate eigenvalues i, where + i has as an
eigenvector k = a + b i:
x = C1 k1 e rt + C2 k2 e rt.
(ii) When only one linearly independent eigenvector exist
x = C1 k e rt + C2 (k t e rt + e rt).
Note: Solve the system (A r I) = k to find the vector .
D-1 - 31
Exercises D-1.3:
1. Rewrite the following second order linear equation into a system of two
equations.
y + 5y 6y = 0
Then: (a) show that both the given equation and the new system have the
same characteristic equation. (b) Find the systems general solution.
2 7 Find the general solution of each system below.
2
2.
x =
x.
5 10
4.
8 4
x =
x.
1 4
6.
2 1
x =
x.
1 2
3 6
x.
3 3
3.
x =
5.
3 2
x =
x.
1 5
7.
2 5
x =
x.
1 4
2
1 1
9.
4 0
x =
x,
0 4
5
x(3) = .
2
10.
1
x =
2
4
x,
3
0
x(1) = .
3
11.
x =
2
8
x,
6
8
x(0) = .
0
12.
1
x =
1
1
x,
1
6
x(0) = .
4
D-1 - 32
9
x,
3
3
x(55) = .
5
3
x,
1
1
x(20) = .
1
13.
x =
1
14.
x =
2
15. For each of the initial value problems #8 through #14, how does the
solution behave as t ?
16. Find the general solution of the system below, and determine the
possible values of and such that the initial value problem has a solution
that tends to the zero vector as t .
5 1
x =
x
,
x
(0)
=
.
3
7
Answers D-1.3:
1 6 t
(b) x = C1 e + C 2
6
1
+ C 2 e 5 t
1
1. (a) r2 + 5r 6 = 0,
1 t
1 e
7 3 t
2. x = C1 e
5
cos( 3t ) + sin( 3t )
cos( 3t ) + sin( 3t )
+ C2
3. x = C1
cos( 3t )
sin( 3t )
6t
6t
6t
4. x = C1 e + C2 t e + e
1
1
0
D-1 - 33
4 t cos( t ) sin( t )
4 t cos( t ) + sin( t )
+
C
e
5. x = C1 e
2
sin( t )
cos( t )
1 t
1 3t
6. x = C1 e + C 2 e
1
1
5 t
7. x = C1 e + C 2
1
1 3t
1 e
4 cos t 2 sin t
x
=
e
8.
2 cos t 4 sin t
2e5t 5 2e t +1
10. x = 5t 5
t +1
2e + e
5 + e 2t
12. x =
2t
5 e
5e 3t 60 6e 4t 80
14. x =
3t 60
+ 4e 4t 80
5e
t
5e 4t + 12
9. x =
4 t + 12
2e
4e 2t + 4e10t
11. x =
2t
10 t
2e + 2e
9 6e 6t +330
13. x =
6 t +330
3 + 2e
0
x(t ) = . For #10, 11, 12, and 14, the limits do not
15. For #8 and 9, tlim
0
exist, as x(t) moves infinitely far away from the origin. For #13,
9
lim x(t ) =
t
3
1 2t
1 4t
x
=
C
e
+
C
16.
1
2
zero as t provided that C1 = 0, which can be achieved whenever the
initial condition is such that = (i.e., + = 0, including the case = =
0).
D-1 - 34
Example:
x =
1 4
4 7 x,
2
x(0) = 1 .
Before we start, let us rewrite the problem into the explicit form of
individual linear equations:
x1 = x1 4 x2
x2 = 4 x1 7 x2
x1(0) = 2
x2(0) = 1
(1)
(2)
(s 1)L{x1} + 4L{x2} = 2
(1*)
4L{x1} + (s + 7)L{x2} = 1
(2*)
D-1 - 35
(1**)
(2**)
[(s 1) (s + 7) (16)]L{x2} = s 9
(s2 + 6s + 9)L{x2} = s 9
Therefore,
s 9
1
12
=
L{x2} = ( s + 3) 2 s + 3 ( s + 3) 2
(3)
(4)
[(s 1) (s + 7) + 16]L{x1} = 2s 18
(s2 + 6s + 9)L{x1} = 2s 18
2s 18
12
L{x1} = ( s + 3) 2 = s + 3 ( s + 3) 2
D-1 - 36
Therefore,
x1 12 te 3 t 2 e 3 t 12 t 2 3 t
x= =
=
e
3t
3t
x
12
t
+
1
12
te
+
e
2
Which agree with the solution we have found earlier using the
eigenvector method.
The method above can also, without any modification, be used to solve
nonhomogeneous systems of linear differential equations. It gives us a way
to solve nonhomogeneous linear systems without having to learn a separate
technique. In addition, it also allows us to tackle linear systems with
discontinuous forcing functions, if necessary.
Example:
x =
4 2
3
1 x +
t
2t 1 ,
3
x(0) = 5 .
x1 = 4x1 2x2 t
x2 = 3 x1 + x2 + 2t 1
x1(0) = 3
x2(0) = 5
1
s2
2 1
s2 s
(5)
(6)
Simplify:
D-1 - 37
1
3s 2 1
(s + 4)L{x1} + 2L{x2} = 2 + 3 =
s
s2
(5*)
2 1
5s 2 s + 2
3L{x1} + (s 1)L{x2} = 2 5 =
s
s
s2
(6*)
Multiplying eq. (5*) by s 1 and eq. (6*) by 2, then subtract the latter
from the former. We eliminate L{x2}, to find L{x1}.
3s 3 + 7 s 2 + s 3
(s + 3s + 2)L{x1} =
s2
2
Therefore,
3s 3 + 7 s 2 + s 3
1
3
11
L{x1} = s 2 ( s + 1)( s + 2) = 4( s + 2) 2s 2 + 4s
x1 =
1 2t 3 11
e t+
4
2
4
5s 3 12s 2 2s + 5
(s + 3s + 2)L{x2} =
s2
2
Therefore,
5s 3 12s 2 2s + 5
1
5 19
L{x2} = s 2 ( s + 1)(s + 2) = 4( s + 2) + 2s 2 4s
1 2t 5 19
e + t
x2 =
4
2
4
D-1 - 38
Finally,
11
1 2t 3
e
t
+
1 e 2 t 6 t + 11
x1 4
2
4
x= =
= 2 t
.
1
5
19
2
t
4
e
10
t
19
2 e + t
2
4
4
Exercise D-1.4:
Use Laplace transforms to solve each nonhomogeneous linear system.
5e 3t
10
1 3
1.
x =
x+
,
x
(0)
=
3.
3t
2 4
6e
2.
2 1
1
x =
x
+
8 ,
1 2
1
x(0) = .
1
3.
3 0
4 sin 2t
x =
x+
,
5 2
cos 2t
2
x(0) = .
1
4.
2 8
x =
x +
1 4
0
x(0) = .
4
3t + 1
6t 2 ,
D-1 - 39
Answers D-1.4:
13 3 t 51 2 t 3 5 t
10 e + 5 e 2 e
x=
1.
17 3 t 17 2 t
5t
e +
e + 3e
5
5
e 2 t cos t + 4 e 2 t sin t + 2
2. x =
2t
2t
4
e
cos
t
+
e
sin
t
18 3 t 8
12
e
+
cos
2
t
+
sin
2
t
13
13
13
x
=
3.
18 3 t 117 2 t
7
113
e
e
cos 2 t +
sin 2 t
52
52
52
13
7
55
55 6 t
2
e
t
t
+
12
2
12
x
=
55 6 t 3 2 1
4.
41
e t + t+
4
4
24
24
D-1 - 40