Professional Documents
Culture Documents
A
Ordinary Differential
Equations (ODEs)
Section 1.1 p1
CHAPTER1
First-Order ODEs
Section 1.1 p2
1.1
1.1 Basic
Basic Concepts.
Concepts. Modeling
Modeling
Section 1.1 p3
1.1 Basic Concepts. Modeling
Section 1.1 p4
1.1 Basic Concepts. Modeling
Section 1.1 p5
1.1 Basic Concepts. Modeling
In
Inthis
thischapter
chapterwe
weshall
shallconsider
considerfirst-order
first-orderODEs.
ODEs.
Such
Suchequations
equationscontain
containonly
onlythe
thefirst
firstderivative
derivativey’y’and
andmay
may
contain
containyyand
andany
anygiven
givenfunctions
functionsof ofx.x.
Hence
Hencewe wecan
canwrite
writethem
themas as
(4)
(4) F(x,
F(x,y,y,y’)
y’)==00
or
oroften
oftenin
inthe
theform
form
y’y’==f f(x,
(x,y).
y).
This
Thisisiscalled
calledthe
theexplicit
explicitform,
form,inincontrast
contrastto
tothe
theimplicit
implicitform
form
(4).
(4).
For instance, the implicit ODE x −3y’ − 4y 2= 0 (where x ≠ 0)
−3 2
For instance, the implicit ODE x 3 y’2 − 4y = 0 (where x ≠ 0)
can
canbebewritten
writtenexplicitly
explicitlyas asy’y’==4x4xy 3 .2
y.
Section 1.1 p6
1.1 Basic Concepts. Modeling
•• A function
A function
yy==h(x)
h(x)
isiscalled
calledaasolution
solutionof ofaagiven
givenODEODE(4) (4)in
insome
someopen
open
interval
intervalaa<<xx<<bbififh(x)
h(x)isisdefined
definedand anddifferentiable
differentiable
throughout
throughoutthe theinterval
intervaland andisissuch
suchthat
thatthe
theequation
equation
becomes
becomesan anidentity
identityififyyand
andy’y’are
arereplaced
replacedwith
withhhand
and
h’,
h’,respectively.
respectively.
•• The
Thecurve
curve(the
(thegraph)
graph)ofofhhisiscalled
calledaasolution
solutioncurve.
curve.
•• Here,
Here,openopeninterval
intervalmeans
meansthatthatthetheendpoints
endpointsaaand andbbare
are
not
notregarded
regardedas aspoints
pointsbelonging
belongingto tothe
theinterval.
interval.
•• Also,
Also,aa<<xx<<bbincludes
includesinfinite
infiniteintervals
intervals−∞−∞<<xx<<b,b,aa<<xx<<
∞,∞∞<<xx<<∞∞(the
∞, (thereal
realline)
line)asasspecial
specialcases.
cases.
Section 1.1 p7
1.1 Basic Concepts. Modeling
EXAMPLE 2
Solution by Calculus. Solution Curves
The ODE y’ = dy/dx = cos x can be solved directly by integration on both
sides. Indeed, using calculus, we obtain y = ∫ cos x dx = sin x + c, where c
is an arbitrary constant. This is a family of solutions. Each value of c,
for instance, 2.75 or 0 or −8, gives one of these curves.
Figure 3 shows some of them, for c = −3, −2, −1, 0, 1, 2, 3, 4.
Section 1.1 p8
1.1 Basic Concepts. Modeling
EXAMPLE 3A
(A)Exponential Growth.
From calculus we know that y = ce0.2t has the derivative
dy
y 0.2 e 0.2 t 0.2 y
dx
Hence y is a solution of y’ = 0.2y (Fig. 4A). This ODE is of
the form y’ = ky.
With positive-constant k, it can model exponential growth,
for instance, of colonies of bacteria or populations of
animals. It also applies to humans for small populations in a
large country (e.g., the United States in early times) and is
then known as Malthus’s law.
Section 1.1 p9
1.1 Basic Concepts. Modeling
EXAMPLE 4A (continued)
(A) Exponential Growth.
y = ce0.2t
y = ce−0.2t
EXAMPLE 5
Radioactivity. Exponential Decay
y(t) = 0.5e−kt
Graphic
GraphicMethod
MethodofofDirection
DirectionFields.
Fields.
Practical
PracticalExample
ExampleIllustrated
Illustratedin
inFig.
Fig.7.7.
We
Wecan canshow
showdirections
directionsof ofsolution
solutioncurves
curvesof ofaagiven
givenODE
ODE
••(1) by drawing short straight-line segments (lineal
(1) by drawing short straight-line segments (lineal
elements)
elements)in inthe
thexy-plane.
xy-plane.
This
Thisgives
givesaadirection
directionfield
field(or
(orslope
slopefield)
field)into
intowhich
whichyou
you
can
canthen
thenfit
fit(approximate)
(approximate)solution
solutioncurves.
curves.ThisThismay
mayreveal
reveal
typical
typicalproperties
propertiesof ofthe
thewhole
wholefamily
familyof ofsolutions.
solutions.
Figure
Figure77shows
showsaadirection
directionfield
fieldfor
forthe
theODE
ODE
••(2) y’y’==yy++xx
(2)
obtained
obtainedby byaaCAS
CAS(computer
(computeralgebra
algebrasystem)
system)andandsome
some
approximate
approximatesolution
solutioncurves
curvesfitted
fittedin.
in.
y’ = y + x
IfIfyou
youhave
haveno
noCAS,
CAS,first
firstdraw
drawaafew
fewlevel
levelcurves
curves
f(x,
f(x,y)y)==const
constof
off(x,
f(x,y),
y),
then
thenparallel
parallellineal
linealelements
elementsalong
alongeach
eachsuch
suchcurve
curve
(which
(whichisisalso
alsocalled
calledan isocline(等傾線
anisocline( 等傾線),),meaning
meaningaacurve
curve
of
ofequal inclination(傾斜角
equalinclination( 傾斜角)),
)),and
andfinally
finallydraw
draw
approximation
approximationcurves
curvesfit
fitto
tothe
thelineal
linealelements.
elements.
EXAMPLE 5
Mixing Problem
Mixing problems occur quite frequently in chemical industry. We
explain here how to solve the basic model involving a single tank.
The tank in Fig. 11 contains 1000 gal ( 加侖 )of water in which
initially 100 lb ( 磅 )of salt is dissolved. Brine runs in at a rate of
10 gal/min, and each gallon contains 5 lb of dissolved salt. The
mixture in the tank is kept uniform by stirring. Brine runs out at
10 gal/min. Find the amount of salt in the tank at any time t.
Extended Method:
Reduction to Separable Form
Certain
Certainnonseparable
nonseparableODEs ODEscan canbe
bemade
madeseparable
separablebyby
transformations
transformationsthat thatintroduce
introduceforforyyaanew
newunknown
unknown
function.
function.We Wediscuss
discussthis
thistechnique
techniqueforforaaclass
classof
ofODEs
ODEsof of
practical
practicalimportance,
importance,namely,
namely,forforequations
equations
y
(8) y' f .
(8) x
Here,
Here,f fisisany
any4(differentiable)
(differentiable)function
functionof ofy/x
y/xsuch
suchas
as
sin
sin(y/x),
(y/x),(y/x)
(y/x),4,and
andsosoon.
on.(Such
(SuchananODE
ODEisissometimes
sometimescalled
called
aahomogeneous
homogeneousODE, ODE,aatermtermwe weshall
shallnot
notuse
usebut
butreserve
reservefor
for
aamore
moreimportant
importantpurpose
purposein inSec.
Sec.1.5.)
1.5.)
Extended Method:
Reduction to Separable Form (continued)
The
Theform
formof
ofsuch
suchan
anODE
ODEsuggests
suggeststhat
thatwe
weset
sety/x
y/x==u;u;thus,
thus,
(9)
(9) yy==ux
ux andandby
byproduct
productdifferentiation
differentiation y’y’==u’x
u’x++u.u.
Substitution
Substitutioninto
intoy’y’==f(y/x)
f(y/x)then
thengives
givesu’x
u’x++uu==f f(u)
(u)
or
oru’x
u’x==f(u)
f(u)−−u.u.We
Weseeseethat
thatififf(u)
f(u)−−uu≠≠0,0,this
thiscan
canbe
be
separated:
separated: du dx
.
(10)
(10) f (u) u x
Extended Method:
Reduction to Separable Form (continued)
We
Werecall
recallfrom
fromcalculus
calculusthat
thatififaafunction
functionu(x,
u(x,y)y)has
has
continuous
continuouspartial
partialderivatives,
derivatives,itsitsdifferential
differential(also
(alsocalled
called
its
itstotal
totaldifferential)
differential)isisu u
du dx dy.
x y
From
Fromthis
thisititfollows thatififu(x,
followsthat u(x,y)y)==cc==const,
const,du
du==0.0.
AAfirst-order
first-orderODE
ODEM(x,
M(x,y)y)++N(x,
N(x,y)y’
y)y’==0,0,written
writtenas
as
(use
(usedy
dy==y’dx
y’dxas
asin
inSec.
Sec.1.3)
1.3)
(1)
(1) M(x,
M(x,y)y)dx
dx++N(x,
N(x,y)y)dy
dy==00
isiscalled
calledan
anexact
exactdifferential
differentialequation
equationififthe
thedifferential
differential
form
formM(x,
M(x,y)y)dx
dx++N(x,
N(x,y)y)dy
dyisisexact,
exact,that
thatis,
is,this
thisform
formisisthe
the
differential
differential u u
du dx dy
(2)
(2) x y
ofofsome
somefunction
functionu(x,
u(x,y).
y).Then
Then(1) (1)can
canbe
bewritten
written
du
du==0.0.
ByByintegration
integrationwe weimmediately
immediatelyobtainobtainthe
thegeneral
generalsolution
solution
ofof(1)
(1)in
inthe
theform
form
(3)
(3) u(x,
u(x,y)y)==c.c.
Section 1.4 p35
1.4 Exact ODEs. Integrating Factors
This
Thisisiscalled
calledan animplicit
implicitsolution,
solution,in incontrast
contrastto toaasolution
solution
yy==h(x)
h(x)asasdefined
definedin inSec.
Sec.1.1,
1.1,which
whichisisalso
alsocalled
calledananexplicit
explicit
solution,
solution,forfordistinction.
distinction.Sometimes
Sometimesan animplicit
implicit solution
solution can
can
bebeconverted
convertedto toexplicit
explicitform.
form.(Do(Dothis
thisfor
forxx2++yy2==1.)
2 2
1.)IfIfthis
this
isisnot
notpossible,
possible,your
yourCASCASmaymaygraph
graphaafigure
figureofofthe
thecontour
contour
lines
lines(3)(3)of
ofthe
thefunction
functionu(x,
u(x,y)y)and
andhelp
helpyou youinin
understanding
understandingthe thesolution.
solution.
Comparing
Comparing(1) (1)and
and(2),
(2),we
weseeseethat
that(1)
(1)isisan
anexact
exactdifferential
differential
equation
equationififthere
thereisissome
somefunction
functionu(x,
u(x,y)y)such
suchthat
that
(4)
(4) (a) (a) u (b)
(b) u
M N.
From this we can
x derive a formulayfor checking whether (1)
From this we can derive a formula for checking whether (1)
isisexact
exactor ornot,
not,as asfollows.
follows.
Let
LetMMand
andNNbebecontinuous
continuousandandhave
havecontinuous
continuousfirst
first
partial
partialderivatives
derivativesininaaregion
regionininthe
thexy-plane
xy-planewhose
whose
boundary
boundaryisisaaclosed
closedcurve
curvewithout
withoutself-intersections.
self-intersections.Then
Then
by
bypartial
partialdifferentiation
differentiationofof(4)
(4)(see
(seeApp.
App.3.2
3.2for
fornotation),
notation),
M 2u N 2u
.
y y x x x y
By
Bythe
theassumption
assumptionof
ofcommunity
communitythe
thetwo
twosecond
secondpartial
partial
derivatives
derivativesare
areequal.
equal.Thus
Thus
M N
(5)
(5) .
y x
This
Thiscondition
conditionisisnot
notonly
onlynecessary
necessarybut butalso
alsosufficient
sufficientfor
for
(1)
(1)totobebean
anexact
exactdifferential
differentialequation.
equation.
IfIf(1)
(1)isisexact,
exact,the
thefunction
functionu(x,
u(x,y)y)can
canbe
befound
foundbybyinspection
inspection
ororininthe
thefollowing
followingsystematic
systematicway.
way.From
From(4a)
(4a)we
wehave
havebyby
integration
integrationwithwithrespect
respectto
toxx
(6)
(6) u M dx k( y);
in
inthis
thisintegration,
integration,yyisisto
tobe
beregarded
regardedas asaaconstant,
constant,and
andk(y)
k(y)
plays
playsthe
therole
roleof
ofaa“constant”
“constant”of ofintegration.
integration.
To
Todetermine
determinek(y),
k(y),we
wederive
derive∂u/∂y
∂u/∂yfrom
from(6),
(6),use
use(4b)
(4b)to
toget
get
dk/dy,
dk/dy,and
andintegrate
integratedk/dy
dk/dytotoget
getk.k.
We
Wemultiply
multiplyaagiven
givennonexact
nonexactequation,
equation,
(12)
(12) P(x,
P(x,y)y)dx
dx++Q(x,
Q(x,y)y)dy
dy==0,0,
by
byaafunction
functionFFthat,
that,in
ingeneral,
general,will
willbe
beaafunction
functionof
ofboth
bothxx
and
andy.y.The
Theresult
resultwas
wasan anequation
equation
(13)
(13) FP
FPdx
dx++FQ
FQdydy==00
that
thatisisexact,
exact,so
sowe
wecan
cansolve
solveititas
asjust
justdiscussed.
discussed.Such
Suchaa
function
functionisisthen
thencalled
calledan
anintegrating
integratingfactor
factorof
of(12).
(12).
Let
LetFF==F(x).
F(x).Then
ThenFFy y==0,0,and
andFFx x==F’F’==dF/dx,
dF/dx,so
sothat
that(15)
(15)
becomes
becomes
FP
FPy ==F’Q
F’Q++FQ
FQx. .
y x
Dividing
Dividingby
byFQ
FQand
andreshuffling
reshufflingterms,
terms,we
we have
have
1 dF 1 P Q
(16) R, where R .
(16) F dx Q y x
Theorem 1
Integrating
IntegratingFactor
FactorF(x)
F(x)
IfIf(12)
(12)isissuch
suchthat
thatthe
theright
rightside
sideRRofof(16)
(16)depends
dependsonly
onlyon
onx,x,then
then
(12)
(12)has
hasan anintegrating
integratingfactor
factorFF==F(x),
F(x),which
whichisisobtained
obtainedby
by
integrating
integrating(16)
(16)and
andtaking
takingexponents
exponentson onboth
bothsides.
sides.
(17) F( x) exp R( x)dx.
(17)
Note:
(12) P(x, y) dx + Q(x, y) dy = 0,
1 dF 1 P Q
(16) R, where R .
F dx Q y x
Similarly,
Similarly,ififF*
F*==F*(y),
F*(y),then
theninstead
insteadof
of(16)
(16)we
weget
get
1 dF * 1 Q P
(18) F * dx R*,
(18) where R* .
P x y
Theorem 2
Integrating
IntegratingFactor
FactorF*(x)
F*(x)
IfIf(12)
(12)isissuch
suchthat
thatthe
theright
rightside
sideR*R*ofof(18)
(18)depends
dependsonly
onlyonony,y,
then
then(12)
(12)hashasan
anintegrating
integratingfactor
factorF*
F*==F*(x),
F*(x),which
whichisisobtained
obtained
from
from(18)
(18)andandtaking
takingexponents
exponentson onboth
bothsides.
sides.
(19) F *( y ) exp R*( y)dy
(19)
Note:
(12) P(x, y) dx + Q(x, y) dy = 0,
1 dF * 1 Q P
(18) R*, where R* .
F * dx P x y
29
F*(y) = e–y
29
y(0) = –1
u(0, – 1) = 1+ 0 + e = 3.72
ex + xy + e–y = 1 + e = 3.72
1.5
1.5 Linear
Linear ODEs.
ODEs. Bernoulli
Bernoulli
Equation.
Equation. Population
Population Dynamics
Dynamics
AAfirst-order
first-orderODE
ODEisissaid
saidto
tobe
belinear
linearififititcan
canbe
bebrought
brought
into
intothe
theform
form
(1)
(1) y’y’++p(x)y
p(x)y==r(x),
r(x),
by
byalgebra,
algebra,and
andnonlinear
nonlinearififititcannot
cannotbe
bebrought
broughtinto
intothis
this
form.
form.
Homogeneous
HomogeneousLinear LinearODE.
ODE.
We
Wewant
wantto tosolve
solve(1)
(1)on
onsome
someinterval
intervalaa<<xx<<b,b,call
callititJ,J,and
and
we
webegin
beginwithwiththe
thesimpler
simplerspecial
specialcase
casethat
thatr(x)
r(x)isiszero
zerofor for
all
allxxin
inJ.J.(This
(Thisisissometimes
sometimeswritten
writtenr(x)
r(x)≡≡0.)
0.)Then
Thenthe theODE ODE
(1)
(1)becomes
becomes
(2)
(2) y’y’++p(x)y
p(x)y==00
and
andisiscalled
calledhomogeneous.
homogeneous.
The
Thegeneral
generalsolution
solutionof ofthe
thehomogeneous
homogeneousODE ODE(2)(2)isis
(3)
(3) y(x) = ce −∫p(x)dx (c = ±e *
−∫p(x)dx
y(x) = ce (c = ±ec*when
c
whenyy>/<
>/<0);
0);
here
herewe
wemay
mayalso
alsochoose
choosecc==00andandobtain
obtainthe
thetrivial
trivial
solution
solutiony(x)
y(x)==00for
forall
allxxin
inthat
thatinterval.
interval.
34
34
1.5 Linear ODEs. Bernoulli Equation. Population Dynamics.
Reduction to Linear Form. Bernoulli Equation
Numerous
Numerousapplications
applicationscan canbe bemodeled
modeledby byODEs
ODEsthat
thatare
are
nonlinear
nonlinearbut butcan
canbe betransformed
transformedto tolinear
linearODEs.
ODEs.OneOneofofthe
the
most
mostuseful
usefulones
onesof ofthese
theseisisthe theBernoulli
Bernoulliequation
equation
(9)
(9) y’ y’ + p(x)y = g(x)ya (a(aany
+ p(x)y = g(x)y a
anyreal
realnumber).
number).
IfIfaa==00or
oraa==1,1,Equation
Equation(9) (9)isislinear.
linear.Otherwise
Otherwiseititisis
nonlinear.
nonlinear.Then
Thenwe weset
set 1−a
u(x)
u(x)==[y(x)]
[y(x)]1−a. .
WeWedifferentiate
differentiatethisthisand
and−asubstitute
substitutey’y’ from
from (9),
(9),obtaining
obtaining
u’u’==(1(1−−a)y
a)y−ay’y’==(1(1−−a)y a)y−a(gy
−a
(gya−−py).
a
py).
Simplification
Simplificationgivesgives
u’u’==(1(1− −a)(g
a)(g −− pypy
1−a
1−a),),
where
where y1−a==uuon
y1−a
onthe
theright,
right,so sothat
thatwe wegetgetthe
thelinear
linearODE
ODE
(10)
(10) u’u’++(1(1−−a)pu
a)pu==(1 (1−−a)g. a)g.
Section 1.5 p58
1.5 Linear ODEs. Bernoulli Equation. Population Dynamics.
Example 4
Logistic Equation
Logistic Equation
Solve the following Bernoulli equation, known as the logistic equation (or
Solve the following Bernoulli equation, known as the logistic equation (or
Verhulst equation)
Verhulst equation)
(11) y’ = Ay − By2 2
(11) y’ = Ay − By
Solution. Write (11) in the form (9), that is,
Solution. Write (11) in the form2 (9), that is,
y’ − Ay = − By 2
y’ − Ay = − By
to see that a = 2 so that u = y1−a1−a = y−1−1. Differentiate this u and substitute y’ from
to see that a = 2 so that u2 = y −2= y . Differentiate this
−1 u and substitute y’ from
(11), u’ = −y y’ = −y (Ay − By 2
) = B − Ay .
(11), u’ = −y2y’ = −y−2(Ay − By2) = B − Ay−1.
The last term is − Ay−1−1= − Au. Hence we have obtained the linear ODE
The last term is − Ay = − Au. Hence we have obtained the linear ODE
u’ + Au = B.
u’ + Au = B.
The general solution is [by (4)]
The general solution is−At [by (4)]
u = ce −At+ B/A.
u = ce + B/A.
Since u = 1/y, this gives the general solution of (11),
Since u = 1/y, this gives the general solution of (11),
1 1
(12) y At (Fig. 21)
(12) u ce B / A (Fig. 21)
Directly from (11) we see that y = 0 (y(t) = 0 for all t) is also a solution.
Directly from (11) we see that y = 0 (y(t) = 0 for all t) is also a solution.
Section 1.5 p59
1.5 Linear ODEs. Bernoulli Equation. Population Dynamics.
Example 4 (continued)
We
Wesee seethat
thatin
inthe
thelogistic
logisticequation
equation(11) (11)the
theindependent
independent
variable
variablettdoesdoesnotnotoccur
occurexplicitly.
explicitly.An AnODE ODEy’y’==f(t,f(t,y)y)in
in
which
whichttdoes doesnot
notoccur
occurexplicitly
explicitlyisisof ofthe
theform
form
(13)
(13) y’y’==f(y)
f(y)
and
andisiscalled
calledan anautonomous
autonomousODE. ODE.ThusThusthe thelogistic
logistic
equation
equation(11) (11)isisautonomous.
autonomous.
Equation
Equation(13) (13)has
hasconstant
constantsolutions,
solutions,called
calledequilibrium
equilibrium
solutions
solutionsor orequilibrium
equilibriumpoints.points.These
Theseare aredetermined
determinedby by
the
thezeros
zerosofoff(y),
f(y),because
becausef(y)
f(y)==00gives
givesy’y’==00byby(13);
(13);hence
hence yy
==const.
const.These
Thesezeros
zerosare
areknown
knownas ascritical
criticalpoints
pointsof of(13).
(13).An An
equilibrium
equilibriumsolution
solutionisiscalled
calledstable
stableififsolutions
solutionsclosecloseto toitit
for
forsome
somettremain
remainclose
closeto toititfor
forall
allfurther
furthert.t.ItItisiscalled
called
unstable
unstableififsolutions
solutionsinitially
initiallyclose
closetotoititdo
donot
notremain
remainclose close
to
toititas
asttincreases.
increases.
Section 1.5 p61
1.6
1.6 Orthogonal
Orthogonal Trajectories.
Trajectories.
Optional
Optional
An
Animportant
importanttype typeofofproblem
problemin inphysics
physicsororgeometry
geometryisistoto
find
findaafamily
familyof ofcurves
curvesthat
thatintersect
intersectaagiven
givenfamily
familyof of
curves
curvesatatright
rightangles.
angles.TheThenew
newcurves
curvesarearecalled
called
orthogonal
orthogonaltrajectories
trajectoriesof ofthe
thegiven
givencurves
curves(and
(and
conversely).
conversely).Examples
Examplesare arecurves
curvesof ofequal
equaltemperature
temperature
(isotherms)
(isotherms)and andcurves
curvesof ofheat
heatflow,
flow,curves
curvesof ofequal
equal
altitude
altitude(contour
(contourlines)
lines)ononaamap
mapand andcurves
curvesof ofsteepest
steepest
descent
descenton onthat
thatmap,
map,curves
curvesofofequal
equalpotential
potential
(equipotential
(equipotentialcurves,
curves,curves
curvesofofequal
equalvoltage-the
voltage-theellipses
ellipses
in
inFig.
Fig.24,
24,next
nextslide)
slide)and
andcurves
curvesof ofelectric
electricforce
force(the
(the
parabolas
parabolasin inFig.
Fig.24).
24).
Here
Herethe
theangle
angleof ofintersection
intersectionbetween
betweentwo twocurves
curvesisis
defined
definedto tobebethe
theangle
anglebetween
betweenthe thetangents
tangentsof ofthe
thecurves
curves
atatthe
theintersection
intersectionpoint.
point.Orthogonal
Orthogonalisisanother
anotherwordwordforfor
perpendicular.
perpendicular.
Section 1.6 p63
1.6 Orthogonal Trajectories.
(continued)
(continued)
In
Inmany
manycases
casesorthogonal
orthogonaltrajectories
trajectoriescan
canbe befound
foundusing
using
ODEs.
ODEs.In Ingeneral,
general,ififwe
weconsider
considerG(x,
G(x,y,y,c)c)==00to
tobe
beaagiven
given
family
familyofofcurves
curvesin
inthethexy-plane,
xy-plane,then
theneach
eachvalue
valueofofccgives
givesaa
particular
particularcurve.
curve.Since
Sinceccisisone
oneparameter,
parameter,such suchaafamily
familyisis
called
calledaaone-parameter
one-parameterfamilyfamilyofofcurves.
curves.
Problem
Problemof ofExistence
Existence
Under
Underwhat
whatconditions
conditionsdoes
doesan
aninitial
initialvalue
valueproblem
problemofofthe
theform
form
(1)
(1)have
haveatatleast
leastone
onesolution
solution(hence
(henceone
oneororseveral
severalsolutions)?
solutions)?
Problem
Problemof ofUniqueness
Uniqueness
Under
Underwhat
whatconditions
conditionsdoes
doesthat
thatproblem
problemhave
haveatatmost
mostone
one
solution
solution(hence
(henceexcluding
excludingthe
thecase
casethat
thatitithas
hasmore
morethan
thanone
one
solution)?
solution)?
Theorem 1
Existence
ExistenceTheorem
Theorem
Let
Letthe
theright
rightside
sidef(x,
f(x,y)y)ofofthe
theODE
ODEin
inthe
theinitial
initialvalue
valueproblem
problem
(1)
(1) y’y’==f(x,
f(x,y),
y), y(x
y(x0)0)==yy0 0
be
becontinuous
continuousatatall
allpoints
points(x,
(x,y)y)ininsome
somerectangle
rectangle
R:
R:|x
|x−−xx0|<|< a,a, | |yy−−yy0 | |<<bb (Fig.
(Fig.26)
26)
0 0
and
andbounded
boundedin
inR;
R;that
thatis,is,there
thereisisaanumber
numberKKsuch suchthat
that
(2)
(2) | |f(x, y)|≤≤KK
f(x,y)| for
forall
all(x,
(x,y)y)ininR.
R.
Then
Thenthe theinitial
initialvalue
valueproblem
problem(1) (1)has
hasatatleast
leastone
onesolution
solutiony(x).
y(x).
This
Thissolution
solutionexists
existsatatleast
leastfor
forall
allxxininthe
thesubinterval
subinterval|x |x−−xx0|<|<αα
0
ofofthe
theinterval
interval|x |x−−xx0|< |< a;a;here,
here,ααisisthe
thesmaller
smaller ofofthe
thetwo
two
0
numbersaaand
numbers andb/K.b/K.
Section 1.7 p67
1.7 Existence and Uniqueness of Solutions for Initial Value Problems
Theorem 1 (continued)
Theorem 2
Uniqueness
UniquenessTheorem
Theorem
Let
Letf fand
andits
itspartial
partialderivative
derivativefyfy==∂f/∂y
∂f/∂ybe
becontinuous
continuousfor
forall
all(x,
(x,
y)y)in
inthe
therectangle
rectangleRR(Fig.
(Fig.26)
26)and
andbounded,
bounded,say,
say,
(3)
(3)(a)
(a)|f|f(x, y)|≤≤K,
(x,y)| K, (b)
(b)|f|fy(x,
y (x,y)|
y)|≤≤M
M for
forall
all(x,
(x,y)y)ininR.
R.
Then
Thenthe
theinitial
initialvalue
valueproblem
problem(1) (1)has
hasatatmost
mostone
onesolution
solutiony(x).
y(x).
Thus,
Thus,bybyTheorem
Theorem1,1,the theproblem
problemhas hasprecisely
preciselyone
onesolution.
solution.This
This
solution
solutionexists
existsatatleast
leastfor
forall
allxxininthat
thatsubinterval
subinterval|x |<α.α.
|x−−xx0|<
0