Systems of Differential Equations

If
Linear
i' *
I t.ae . and it one areas me

Given
qq.CH ,
Isijen ,
and fill ,
Hisn
functions defined
vector unknowns matrix and the hand side
function of ,
the
coefficient right
in an interval la ,
b) ,
the
system
term the system Is ) respectively
'

"
or
forcing of ,
.

Y
,
=
9,, Lt ) Y
,
t 9,24142 + . - .
t
9. nlt Yn ) + fit )

45 )
|
' '
Yz =
92 , It ) Y t 9224 ) Yz t
92nA Tn t
fact ) '

Fit) Ect )
I the
t
. -

and
goal
Alt ) both
.

,
In Qe ) and is
given
are
y +
=

itfacts
,

hi oh , 9mHz 9mA In
-

y
- - - -

It )
,
vector
'

all
functions satisfy
'
to
find I =
Fits that I - AH yet t Flt )
,

is called a
first order linear
system of differential equations .

for any
t in la ,
b ) .

rating are .

I .
Ein . ow ' i' *
I I. we Example
y
'
2-1 , t tinct )

{
k
-
t
order
t
linear
=

is
first system of
,
a
.

'
et htt )

=/
?)
' yes
=
y, + 3-12 t

Itt ) QEII 't Fit )

'
have and Cs ) becomes I = .
This is the
matrix notation
,
differential equations .

Using ,
we can
successively write
matrix
form of is ) .
 nth order linear
differential equation
iii.
into
Converting system
÷
:÷÷÷÷÷⇐i÷H:÷÷÷H
an a

:S
: t.it
:*
⇒

: :*:* The linear

differential equation
"
pn.tt ) In
"
't
'

y t
-

+ p at y potty
=
fat

=/! )
+

! I
. . .

=/;!) Tenn
,
ETH
If we set I ,
a and =
,
then the
system
linear
can be converted into a
first order system of differential equations

I )
' t cost )
Fits x-D
' ' '

simply becomes I =
RHI + .
Conversely , if acts =

3 est
and
by setting y
,
=
y ,
y
,
=
y ,
. . .

, Tn =
T

=/!!! )
→
'
Ect ) the system I
'

=
AHI 't FA can be written as
,
' '
linear
'

For instance to convert

y + pets y + gets y =
fit ) into or
system , we set
,

'
tet

It HiHi
iii. :
'
ty

ti:
costly
: I ,÷
Y +

I
=
-1

:
'

petty 't felt )

, , ' ' ' "
,
and It and
⇐ ⇐ y y follows -1 y T Tz y yay
-

y =
-

y
= =
=
. . - -
.

.eu , ,
,
,

. .
. . .

.
'
else
'
or
yz = -

gets -1 , -

play ,
+
felt ) .
Hence

the equations form of the I I' Fits

This is
system = acts t
" '

I I
=

:* I
face
, -12
-

is
'
= -

own ,
-

easy ,
+ fit ,
=

, .
1+1%1
 't ett et
similarly 't et
'
the
linear separately
' "
and
, y + pets y get , y ray =
fct ) can be
expressed as a
get y ,
= c
,
* I
,
y
,
=
ez 2
,
. . .

, yn
=
cu th .
Hence

}!
' "

system of differential equations by setting y y z y y =

y
) =/
=
-

constants
.

solution
, '
where
arbitrary
, , ,
is I # =
a
,
a
,
. . .

,
cu are .

' ' ' ' ' ' ' ' "

' '
It
follows y y
=
72
=
y y y y =
ray get ) y play ft I
-
= =
t
yes
= - - -

, ,
, , , z

'

yj Fits
'
solution
'
or else = -
ret ) y
,
-

get ) -12
-

petty ,
+ felt ) .
Hence This
suggests
that the
general of I = acts I +
should be

arbitrary

÷÷ iii.iii. iii.ii. iii. i

expected to
depend on n constants .

. .
.
.
.
. " .
Definition
'

solution I' F' as vector

function
'
The
general of the system I = a # + is a

)
It

I
Y ; a Cg Cn )
consider the linear
system of differential equations
.

first
.

order
.

' ' , , , ,

I = I l tie .
,
ca
,
. .

,
en ) = that depends on n
arbitrary
, , , ,
, ,
. . .

,
, ,

÷,
y
=
"

| !!
, '
that be
satisfy I Fits
'
This
decoupled system easily constants and I
k
is a can a ca en It ) act ) Cts
for t in
-

any
'
. . .
=
+
, , ,

.
.

,
solved
by solving arch
of the equations our interval la , b) .
 Result ( Existence and
Definition Uniqueness )
'
I 't Fit ,
'

I a

{
=

consider the initial value

The problem that consists of finding a vector
function I A that satisfy problem I⇐ , =
Tj
Fit ,
' ' '

I I Fits
{ It
acts t
If and continuous and around initial
=

called 241 are at t to then the

initial value problem
-
-

is an ,

to , =
I; ( 3 t ) 3 (2)
3
- -

t
value
problem has a
unique
solution Itt ) that is
guaranteed to be
3 -

=/
4T

Example (9 ( 3 t2)
) the interval that contains
3C e
defined in
largest open
t to and where

eat )
- -
-
-

that Itt the solution the

Given )
( za t 3 q ,
is
general of
Fct )
+

both Alt ) and are continuous .

I 7)
'
's
{ %pg
-
'

I I
I ?I )
' =

system
'
value problem
-

I the initial
'
solve
I = ' ?
cost I
Yztlnft
,
Y ! It )
.

-1
Example
- -

, = , s e ,

{ YI
consider initial value
the
problem = get y
,
+
II ,
-12
+ It -61

we need to select the constants a Cz so that

,
5
Y (3) = 2 & Yz (3)
= -

at ,

ii.
y'd 1%4=1:I⇐Y÷÷ . I :: .
Hence
iii. Kitkat .

Find the
largest interval ca
,
b) where the solution
of the initial value
problem

is
guaranteed
to be
defined .
 Note
¥
I
cost )

)
-

t
Here QLH continuous in
f-f) U ( I 5) U ( ) and
though started with
first outer linear
general
= is a 5 to Even we the
, , ,

at
g.
¥ ,
'

Fits
'

I
'

that ) system I Acts

from will restrict ourselves

| /
now
hit = t on we

F- It )
,

continuous
=

le 6 ,
is in I - a
,
o
) U I 2
,
to ) .
The
largest open
Fits
.

'

to the case where = T and Act ) is a constant nxn matrix i. e.

, ,

interval that contains where continuous ( 5) The

t =3
,
alt ) is is
I ,
.

'

A-T constant
'

consider where A
will
only the
system I
we =
is a
,

largest open interval that contains t =3

,
where Fits is continuous is ( 2
,
to ) .

nxn matrix .

The interval that contains where both

largest open t =3 act ) and Flt ) are

continuous is La
b) =
'
(z 5 ) A ( 2 *a ) =
( 2,5 ) .
Hence the solution
, , ,

of the initial value problem is

guaranteed to be
defined in C 2,5 )

( 2,5 )

•
①
O O I I
① I

O '
12 I 2 3 4 5
 '
Result solutions Titi II Ties
'

Main
'

independent , ,
. . .

, of the system I =
AT .

All . . . 04h

:Ji inn)
matrix solutions let's look
If A is constant then the
general
To
find those
for them in the
form
.

= an nxn
,
,

=/! )
n ,
'
't

,
' '
'
J
'

AT I 's I column
solution
of the
system I = is
given by
e
,
where x is a
parameter and is a nonzero

' '

IIA I
'

Titi vector determined

=
9 Tilts + c
,
+ . . .
t Cn Tilt ) ,
both to be so that I = A

where
Ict II Tilts linearly indepeud solutions of the
system
÷:*
) at are

it:÷÷÷it÷÷÷÷÷l⇒e*l÷:t
. . . .

, , ,

That means .

'

• Ii ① = A
Ict , , II Ct KATIA ,
. . .

,
Fiat =
Aino ,
for any
t

'
" 't
combination (e T ) AT
Till linear other
Equating the heads
-

and two to
• . No is a
of the
Tjcti ,
i. e.
,
A
y = A = e .

't
ett AT
'

AT 's
'

( Ice ) Ties
' '

Inch )
'
J T
'
del to for any t
I A I ⇒ x e ⇐ a
.
= =
,
. . .
, , ,

" ' '

I solution I'
of
'

'
'
Hence
for e to be a =
AT ,
I must be an
eigenvalue
this
'

AT amount
Based on result
, solving I = to
finding n
linearly '

of I must be associated
eigenvector of
A and an A with 7 .
 '
'
It Ir
'

solution multiple thereof I associated

' '
solution
Below owe the
steps to
follow
to
find the
general of I' =
AI ,
Take or a as a
of =
AT ,

where matrix real

A is a constant nxn .
with the
eigenvalue X=r .

the
1 .
Find all the
eigenvalues of the matrix A
by finding
roots
of 3 . For
every pair of complex conjugate eigenvalues a
Ifi , find the

=/! )
?!!!
X Qin
Qi 042
I'
-

associated
- - -

system
,
the
eigenvectors by solving
the characteristic equation del CA
TIKO ⇒
/ 1=0
-
-

,,
-
@lat Bi ) ) I
, t 04202 t . - -
t
I
, = o

)
% U + "
-
&(" " B " " + - - -
+ % = °

⇐
' "
that )V→=O→
To
simplify ,
we will assume It the roots a. are
simple .
KTBDI Xx )

2 .
For
every
real
eigenvalue x -
-
r
, find the associated
eigenvectors
I
=/! ) dnt

'
Vi t ane Iz t . . - .
+ Cann
Hpi
-

) )Y=o

solution I 's T
(Q , ,
- r ) I
,
+ 042 Iz t
- - . + Qin In = O The should be k
Hsi )
-
where k is our
arbitrary constant

by solving the
system CARI
.
) FIT # ,
) Girl
:
.
, +1922-402 t . - -
+ aan In =o

and
Iim . to is a
fixed solution
of ⇐* , .
Take
Ref Etsitty
, i )
a "
U
'
+ 92 U2 + ' - -

by
Hann -
-

- o

Tmf EM
'
" "
Jr
'

The solution should be J , k

,
where and
Thin ) .
or a
multiple thereof as thetwo solutions

'

I
'
k is an
arbitrary constant and to is a
fixed solution
of 1*3 .
associated with the
eigenvalues at Bi & a -
Bi .
4 .
txt hen
finished with all the
eigenvalues A
,
72
,
. .

,
in
you
should have
Example

system
'

'
y =
24 , -1472 '

/ =/ :! ]
,
the I'
'
solutions
Iie
'

AT
'
real
n
Ii Ctl
, IIA ,
. .

,
, of the system I = .
solve ⇐ I
, =3
, + ,
,

They are
automatically linearly independent and the
general
I 3)
, 2 4
I
find the
eigenvalue of They the roots of delta XI )
We A o
are
-
-
-

. . -

,
solution the
of system is

=/ 43
}
t

1=4 ⇒
-

del -

CAXI
-
) -
I 3 -
X ) - 12 =
I - 57 -
6 = Atl ) I x -
6)
.
,

II Inc
'

I Ct ) =
a CH +
Cz Lt ) +
. . .
+ Cu t )

Hence the
del LA XI ) ⇒ ( Atl )C X G) o Hence
eigenvalues 7 I 72=6
-
o
. -
= = =
-

,
.

2 . We
find the associated
eigenvectors .

•
Eigenvectors associated with 2 , = - i

=/Yg)
'

I such then
If is an
eigenvector ,

Iz
14 4502
3 I 4 O

I
=

! f
'

)V→= J
' ,
I 34+402=0 ⇐ =

( A ItX.
-

-
= ⇐ CAt I ⇐
⇐
so ,
are ,
=o

I ;)I I ?)
'
-
"
Hence I = = -
re
.
-
 '

eigenvector I
[I ) the solution Example
'
the and
selecting U2 =3
,
we
get =
y,
=
2-1
+342

{
,

solve the initial value

problem '

tf I I
it '
y 242
sy
e'
-

I
= t
I
'
-

I
,
ai = = .

y
,
= - 2
, -120=3
associated with start the solution then the initial conditions
•
Eigenvectors 72=6 We by finding general ,
we
impose .

=/Yg) FIT
=LI} )
'
matrix
=P
coefficient
'

find eigenvalues of
I such then CA ist the the
If is
eigenvector ) CAGI ) 1 we A
-

our ⇐
-

.
, .

46 Ig
I 62=0
{
4 O

equation
t
-

solution
I of
=

character 0kt
-

have
,
,

They the CA XI ) We
-

⇐ ⇐
re
⇐ { I
,
-
V2 = o ⇐ re
,
= vz are , a -
=o .

3N Iz re = o
3 =o
,
-
-
, ,

I I !
2
del LA II ) K ) 12 X ) (3) C- 3) X 2) 9 It
follows
-
-

=
= - x - - - -
t
2
. .

Hence I 's
;) =/ !! ]
fu =
y
I!) .

selecting us = I
,
we
get the
eigenvector
'
diet (A XI ) 2+9=0 2=-9
.

= o ⇐ ( X -

2) ⇒ ( X 2) - = i ) ⇐ X = 2 I 32

II
2T
et
't

=/ ! ) and the
JI
solution I; = = e
Ii) '

Hence the
eigenvalues are 7 ,
= 2+3 i & 72=2 - 32 .

solution associated
general given by find eigenvectors
3 .
The is 2 .
We the

I to Ties it
I;! I
"
associated
I I ;] I :)
with i
-

a a a eigenvectors 7 2+3
-

= t =
a t c. e = • ,
=

'

Et IT
=/! ) INIT (INIT
'
and ebt & ttczet If such then (A 4th
)
so Lt ) 49 Cz ygct ) 3g e- is an
eigenvector I A
-

t ⇐
-
=
y
= -
,
.

,
 equivalent associated
find
That's the need
to
system • Notice there was no to the
eigenvectors with the

'

I:* ÷
'

I :i: Ii: conjugate eigenvalue

air si
⇐
⇐ 72=2 - .

IIfu; ) =/ ! ) I !)
i
initial select constants
-

problem
-

Hence i leads to the to solve the value need to the

u
selecting us Now we
-
- =
,
.

, ,

!
fws
f-! )
( 2h43 " )
)
"

fee
" '

a c
,
in
Iit , = so that I co , =

y¥=Re( E
""
I!)
' " ,
I

Ll
eigenvector = and the two solutions
i I) and -
C
, sink t ) t Cz cos t)
)
which leads to the
system
E
" " "

Iie Im ( I iI)

ly
]
'
we have

y! =/?;
,

fee
=

)
)
.

-20434+3^434
)
9=-2
I Hence Ice , =

Et tee: I I :;
III I )
" " Cz 3
eat
=

=
E .
2 sink t ) +3 cos t) ) ,

i i i
@t )

.eu/fwsxty+ifnn3ty
- - sin costs )
t

'
24

I
y +342
Et
I ]
=
cos t) " sin ( 3D

{
, ,

It
follows that
Itt ) =
.

singe , , yjct ) = e
wspt
and the Notice the ivp y , =
-

zy ,
+
zyz
could have been solved
using Laplace
@ ) ,
2
42
=
Y @) = -

, ,

solution
general is

!{
transform follows
fee
90434+4 sin est) ) as :

)
""

I eatGt) y
w
est
) I
sin
TEH =
g + ↳ =

- sin Cst ) cos

-
G sink t ) t Cz costst )
 ' '
Take conditions
Laplace transform of
f
the both and do solve
• y ,
=
2-1
,
t 3T
,
-1g =-3
-4+2-12
to
get How we the
system 237,14¥
when no initial are
prescribed ?

S L } y
} Y @) = 2 LIT ) t
32372 } (s 2) L3 -1 } 3 LI -123 2 ,

{
-1
- -
= = -

{
-

, ,
, , ,

Answer constants
S L }
-121 Yz @ I
3
LIT , } +2 LI -123 3 237,3 t Is 2) LIT } -12
=3 : Assume y @) k and y @) he where k and K2 are
arbitrary
-
- = -
= = = .

, ,
, , , ,

linear
This is a 2×2 (
algebraic ) system in L } y
,
} and Ls y
, }
•
Taking the Laplace transform of both equations we
get
⇒}
,

• •
solve the linear system .
The determinant
of the
system is A = Is = 2549
Cs-

SLIT }

+3k
ki
Y
-

y = 2
LI -43 t 4h37 } ( S 2) LIT , }
-
-
4 LI -123 = ki

{
, , ,

Using
Cramer 's rule to solve
for Lh y, } & LIY ,
} ,
we
get { Ll 3h37 3h27
⇒

3 LIT } (s 3) 227 he
s -123 Yzco ) } } }
-

t
-
= =
t
-

, , , ,

: it : it 'dI:
"

it ; 3.1
-

earn owl
mas
-

=
-

= .

=/
S -
2 -
a
the
• The determinant of system is A
→ s .
,
/ = Cs -

2) Is -3 ) -
12 = 5-5 s .
6 = ( Stl ) I S -
G)

's rule
• • • Take the inverse Laplace transform and solve
for y
,
A and
get ) Using
Cramer
,
we
get
'
ki he "

?I:}I
's
2)
Et I Etc
'
" k

¥173
'
I
I!}! I K2# I
' '
-

L'
III Lay
I
-

)
'

I lets I I 2¥ +3£ set ninety MY and

}
-

yes = e # =
3N = 2W +3
I =
-
. -
-
= = =
, ,
,
,
K2 ( Sti ) ( S -
6)

Et £13
' '
Et I Et (
-

bit ) = 2-
I 3£
?
-23¥ ,
let ) =
I Ibo
}{ } ,
A =

Iyo +2¥ ) ,
A = 3 costsHt 2 sine t)
) Next
,
we
find the
partial faction decomposition of each rational
function .
 t4k2=A
4kt
take
3k
kits
k ( 3)
have
s
have
-

A
!
We ,
(s ( ) In
6) B
summary
3) t Sti we
⇒
.
-

= t ,

( Sti ) ( s -
G) s+ , s

4k K2
-74 ke
t 3k
4k -14kg = -7A A = ( 4k -4kg )/7 y Lt ) =
'
e- +
,
ebt
I
,

{
-

, ,
leads to
setting s= - I then 5=6
3h , + a he = 7 B
⇐
B = 4h17
kit&

3h22
t
Yzct ) -73k e-
4k¥34 ebt
,
-
462 -1462
=
+
,
4k -462 3k -14k
Hence Lf y } Cs ) =
7
+
7
=
, i
+
, I

,
Stl S 6 7 Stl 7 S -
6
ke hi 3kt k2
that have
-

letting
It &
follows Cz
-

c = = we
,
4k 4k2 7 ,
t 3k -14kg Gt
-

Thus
-
,

y.CH
,
=
e + e
t
7 7
Y Ctl
=
4g e- + czebt

{
-

K2 +3k t

similarly have 's 2) '

2) +3k e- czebt
-

we kzcs ALS 6) TBC ) Ct ) 39

sea s.BG ⇒ Stl
Yz
a
= +
-

=
-

, + ,

( Sti ) ( S -
G ) ,

3h22 -13k 7 A A= k213 3k ' " 7

{ {
-

,
- -

⇐ which solution obtained

Setting s= -
I then s=6 leads to
an ,
,→z ,
, ,
13=(44+34)/7
is the same
general we
using
the
eigenvalue -

3k 4kg -13k eigenvector method

3kg
.
-
, ,

3k2 3kt 4h22 -13k ,

Lf -1234 ) 7 7 I I
Hence
-

=
= +

Stl S -
6 7 Stl 7 S -
6

3k t
ebt
Thus -73k 4ketfk
' ,
e-
yzct ) = +