Professional Documents
Culture Documents
18.03 Problem Set 1: Part II Solutions
18.03 Problem Set 1: Part II Solutions
03ProblemSet1: PartIISolutions
PartIpoints:0.9,1.6,2.4,3.4.
0. (T5Feb) (a) [4] The growth ratek(t) hasunits years
1
(sothat k(t)x(t)thasthe
sameunitsasx(t)). Thevariablethasunitsyears,sotheaaddedtoitmusthavethe
sameunits,andk
0
musthaveunitsyearsinorderfortheunitsofthefractiontowork
out.
(b) [3] x(t+t) x(t) +k(t)x(t)t,so x=k
0
x/(a+t)
2
.
(c)[4] Separate: dx/x = k
0
(a+t)
2
dt. Integrate: ln |x| +c
1
= k
0
(a+t)
1
+c
2
.
Amalgamate constants and exponentiate: x = e
c
e
k
0
/(a+t)
. Eliminate the absolute
c
| |
value: x=Ce
k
0
/(a+t)
,whereC=e . Reintroducethesolutionwelostbydividingby
xintherststep: allowC=0. Sothegeneralsolutionisx=Ce
k
0
/(a+t)
. (Notethat
theexponentk
0
/(a+t) isdimensionless,asanexponentmustbe.)
(d) [4] Whentgetsverylarge,theexponentgetsveryneartozero,sothereisanite
limitingpopulation: x
=C. Thusx(t) =x
e
k
0
/(a+t)
. Taket=0inthesolution:
k
0
/a k
0
/a
x
0
=x
e ,orx
=e x
0
.
1. (W6Feb) (a) [3] (Thepicturelacksthe m= 1isocline,sorry.)
(b)[3] The separatrix is the elbow curve that other integral curves snuggle up to to
the left and separate from as things move to the right. [Here are some facts about
solutionstothisdierentialequation,someofwhicharecontainedinlaterpartsofthis
problem: Everysolution has astartingtime(if I thinkofxastime)beforewhichit
doesnotexisteveryintegralcurveisasymptotictoaverticalstraightlinetoitsleft.
Everysolutionabovetheseparatrixhasanendingtimeafterwhichitdoesnotexist
itis asymptotic to a dierent vertical straightline to itsright. Every solutionbelow
theseparatrixsurvives forever, and becomesasymptotic (fromabove)to the graphof
y=
x. Thesearealsoexactlythesolutionswithacriticalpointeachhasonelocal
maximum. Theseparatrixitselfisasolution,averyspecialone: itistheonlysolution
whichlivesforeverbuthasnomaximumandisnotasympotictoy=
x.Insteaditis
asymototictoy=
x.]
(c) [3] Thepointoftangencyoccurswherethetangentlinetothem=1isoclinehas
slope1. Them=1isoclineisthelocusofx=y
2
+1. Dierentiateimplicitly (using
the chain rule)1= 2yy
,soify
=1theny=1/2. Theequationthengivesx= 5/4,
so (a, b) = (5/4, 1/2).
(d)[3]f(x) =
x,orindeedf(x) =
x+mforanym. Therestoftheproblem
dependsonusingf(x) =
x,though,asnotedonthewebsite. Forlargex,y(x)>
x, but if m < 0 there will be solutions of the type we are considering such that
y(x) <
x+mforsomex >a.
(i)Followthe solutionfromthe x = aline. Suppose thatitdoescross the m = 1
isocline, andsuppose that therst crossing is at a point (x, y). Sinceitstartedbelow
them=1isocline,itmustbecrossingitfrombelow. Thismeansthattheslopeofthe
solutionmustbegreaterthantheslopeoftheisoclineatthecrossingpoint. Theslope
ofthe isoclineat thecrossing point is greater than1,sotheslopeofthesolutionatthe
crossing pointmustbe greaterthan1too.Butwhenasolutioncrossesthe1isocline,
itmustcrossitwithslopeexactly1. Soitcantcrossitatall.
It will be clearer if I dont try to follow (ii) and (iii) (as explained on the website).
(i) shows that solutions y(x) such that y(a) < b are trapped below the 1 isocline.
NowIwanttosaythatifasolutionisabovethelowerbranchofthenullclineatsome
point (x, y(x)),thenitstaysabovethenullclineforevermore. Theslopeofthenullcline
isnegative; therstcrossing must befromabove; so atthecrossing theslopeof the
solutionmustbeevenmorenegative;butwhenitcrossesthenullclineitmusthaveslope
zero.
FinallyIclaim that if (a, y(a)) isbelowthenullcline,thentheintegralcurvemusteven-
tuallycrossthenullcline. Untilitdoes,itisintheregionwherethedirectioneldhas
positiveslopes. Soitisincreasing. Butthelowerbranchofthenullclineisisfalling;so
thetwocurvesmusteventuallyintersect.
Onecanalsoarguethatifthesolutionnevercrossesthenullclinethen,afterrisingat
thestartitmust eventually starttofall, since thenullcline does. So itmust reacha
maximumsomewhereinbetween. Butallmaximaofsolutionsoccuralongthenullcline.
(e) [3]Criticalpointsofsolutionsofy
=y
2
xoccurwheny
=0,thatis,alongthe
null-cline: so ifthere is one at (c, d) thenc=d
2
.
(f)[3] Yes. Here are two arguments: All solutions are increasing for small values of
x. Fromwhatwecansee,thesolutionswithcriticalpointsaretheoneswhichbecome
asymototictothegraphofy=
It
q| = e e .
Eliminatetheabsolutevalueandreintroducethelostsolution: Ix q= . Solvefor
x: x= (q/I) +Ce
It
(wherethis C istheearlieronedividedbyI).
(c)[5] Constant trust value means x = 0, which says Ix = q or x = q/I. So with
q=12,000dollars/yearandI= 0.05,x=$240,000. (IfScroogesocksaway morethan
this,thenthetrust fundcould payout the $1000/monthandstill grow.Butthiswouldnt
beScrooge.)
(d) [5] We want tond the constantof integration whichmakesx(T) = 0, whereT= 20:
0 =x(T) = (q/I) +Ce
IT
,orC=(q/I)e
IT
. Thusx= (q/I)(1 e
IT
e
It
). Nowwe
cansett= 0tondtherequiredinitialvalueofthetrust:x(0) = (q/I)(1 e
IT
). With
T= 20andI= 0.05,1 e
IT
= 1 e
1
0.63212056.Thustheinitial fundingisabout
63%ofwhatitwasin(c): x(0) ($240,000)(.63212056) $151,708.93.
n(n 1)h
2
. Withh= 1/n thisgivesourestimate
MIT OpenCourseWare
http://ocw.mit.edu
18.03 Differential Equations
Spring 2010
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.