0% found this document useful (0 votes)
70 views13 pages

Indeterminate Coefficients

The document discusses the method of undetermined coefficients for solving non-homogeneous linear differential equations, providing a general solution format and examples of particular solutions. It explains the method's applicability to specific types of functions and includes examples demonstrating the process. Additionally, it covers the Euler-Cauchy equation of order n, detailing the substitution method for finding solutions and the resulting general solution format.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
70 views13 pages

Indeterminate Coefficients

The document discusses the method of undetermined coefficients for solving non-homogeneous linear differential equations, providing a general solution format and examples of particular solutions. It explains the method's applicability to specific types of functions and includes examples demonstrating the process. Additionally, it covers the Euler-Cauchy equation of order n, detailing the substitution method for finding solutions and the resulting general solution format.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Solution of Non-Homogeneous Equations

Method of Undetermined Coefficients


Prepared by M.C. Luis E. Castro-Solís

Considerthegeneralnon-homogeneouslinearequationofn-thorderwithcoefficients
constants.

(1) ay+ an-1yn-1 + ... + a1y' + a0y=f(x)

The general solution of (1) is of the form:

(2) y =
c y + yp

where andc,thecomplementaryfunctionisthegeneralsolutionoftheassociatedhomogeneousequation
(1)

(3) anyn+ an-1yn-1 + ... + a1y' + a0y = 0

yy,pitisasimpleparticularsolutionof(1)

The method of undetermined coefficients is a direct way to find andpwhen f(x) at (1) is
simple enough for us to make a well-founded assumption about the shape
generalofandpThe method of undetermined coefficients applies whenever the external function,
f(x)isalinearcombinationof(finite)productsoffunctionsofthetypesa)apolynomialin
x, b) an exponential function, exp(rx), or c) cos(kx), sin(kx), and also ypdoes not satisfy (3) —
as in the case of example 4—.

Example 1.- Let y'' + 3y' + 4y = 3x + 2, find a particular solution.

Solution.- Since f(x) = 3x + 2 we will assume yp= Ax + B, from where y'p=A,y”p= 0;

Substitutingintothedifferentialequation:

[0] + 3[A] + 4[Ax+B] = 3x + 2


4Ax + (3A + 4B) = 3x + 2
4A = 3Ç3A + 4B = 2
3/4

yp= ¾x - 1/16⌫
Example2.-Findaparticularsolutionofy''-4y=2e3x

Solution.- Given that f(x)= 2e3xitisreasonabletoassumeandp= Ae3x, where are you from?p= 3Ae3x,yp=
9Ae3x;

Substitutingintothedifferentialequation:

[9Ae3x] -4[Ae3x] = 2e3x


5Ae3xequals 2e3x
5A = 2
2/5

yp= 2/5 e3x⌫

Example 3.- Find a particular solution of 3y'' + y' - 2y = 2 cos x

Solution.- f(x) = 2cosx®yp=Acos(x)+Bsin(x),y'p-Asenx + Bcosx


yp-Acosx - Bsinx

Substitutingintothedifferentialequation:

3[-Acosx - Bsinx] + [-Asin x + Bcosx] - 2[Acosx + Bsinx] = 2cosx


(-5A + B)cosx + (-A - 5B)senx = 2cosx
2Ç - A - 5B = 0
-5/13

yp-5/13 cosx + 1/13 sinx⌫

The method of indeterminate coefficients is not as simple as it seems, as illustrated by the


nextexample.

Example 4.- Obtain a solution of y'' - 4y = 2e2x

Solution.- f(x) = 2e2x®yp Ae2x,and'p= 2Ae2x,andp=4Ae2xitwouldbeareasonableassumptionifnot


outside because it satisfies the associated homogeneous problem and” - 4y = 0, as will be seen in
continuation,whensubstitutinginthedifferentialequationitremains:

[4Ae2x] - 4[Ae2x] = 2e2x


Thisimpliesthat0=2e2xbut there is no x that satisfies the equation, therefore the assumption is
incorrect; and in fact we demonstrate thatAe2xsatisfies the homogeneous problem.

Wewillassumey=Axe2xp, and'p= 2Axe2x+ Ae2x,yp= 4Axe2x+ 4Ae2x

Substitutingintothedifferentialequation:

4Axe2x+ 4Ae2x4[Axe2x] = 2e2x


4Ae2x= 2e2x
4A = 2
½

yp= ½e2x ⌫

Example5.-Findaparticularsolutionofy''+4y=3x3

Solution.- andp=Ax3+Bx2+ Cx + D, y'p= 3Ax2+ 2Bx + C, y”p= 6Ax + 2B

Substitutingintothedifferentialequation

[6Ax + 2B] + 4[Ax3+Bx2+ Cx + D] = 3x3


4Ax3+ 4Bx2(4C + 6A)x + (4D + 2B) = 3x3
4A = 3Ç4B = 0Ç6A + 4C = 0Ç2B + 4D = 0
3/4

yp= ¾x3- 9/8 x⌫

Example6.-Solvetheinitialvalueproblem

y" - 3y' + 2y = 3e-x-10cos3x


1
y'(0) = 2

Solution.-Wewillfirstsolvethecomplementaryfunction,thatis,theproblem
homogeneous.

y'' - 3y' + 2y = 0
m2- 3m + 2 = 0
(m-1)(m-2) = 0
m1=1, m2=2

yc = C1ex+ C2e2x
to be determined the values C1andC2solving now the forced problem, given that f(x)=3e-x
-10cos3x, we will assume (by linearity)

yp = Ae-x+ Bcos3x + Csin3x


y’p =-Ae-x- 3Bsin3x + 3Ccos3x
yp= Ae-x- 9Bcos3x - 9Csin3x

substitutinginthedifferentialequation:

Ae-x- 9Bcos3x - 9Csen3x] -3[-Ae-x- 3Bsin(3x) + 3Ccos(3x)


+ 2[Ae-x+ Bcos3x + Csen3x] = 3e-x-10cos3x
6Ae-x+ (-7B-9C)cos3x + (-7C+9B)sen3x = 3e-x-10cos3x
6A = 3Ç - 7B - 9C = -10Ç - 7C + 9B = 0
1/2

yp= ½e-x+ 7/13 cos3x + 9/13 sin3x

Thegeneralsolutionofthedifferentialequationisgivenbyy=y+yp c

y = C1ex+ C2e2x+ ½e-x+ 7/13 cos3x + 9/13 sin3x

deriving

y' = C1ex+ 2C2e2x- ½e-x- 21/13 sin(3x) + 27/13 cos(3x)

substitutingtheinitialconditionsinthetwopreviousexpressions

1
1 = C1+ C2+ 27/26

y'(0) = 2
2 = C1+ 2C2+ 41/26

solving for C1, C2:

c1=-½, C2= 6/13

The sought solution is:

y = -½ex+ 6/13 e2x+ ½e-x+ 7/13 cos(3x) + 9/13 sin(3x)⌫


Non-Homogeneous Linear Equations
Method of Undetermined Coefficients
Prepared by M.C. Luis E. Castro-Solís

Not all equations with variable coefficients can be solved analytically; for certain
forsufficientlysimpletypesofdifferentialequationsitispossibletoobtaintheirsolution;inthecase
Of second-order equations with one known solution, it is possible to acquire a second solution.
linearlyindependentthroughthemethodoforderreduction.

Themethodallows,givenasolutionofalinearordinarydifferentialequation
homogeneousofthe2ndorder,[Link]
canonicalform:

y" + p(x)y' + q(x)y = 0

andasolution

y1(x)

thatistosayand1(x)suchthaty1+p(x)y1' + q(x)y1= 0;Asecond solution is desired in the form of

y2(x)=u(x)y1(x)

Deriving:

y2='uy'1+yu’
1
y2=uy+2y11u+y1u”

Substitutingintothedifferentialequation

uy+
1 2y1u+y1[u]+p(x)[uy1'+yu'
1 ]+q(x)[u(x)y1(x)]=0

Reorganizing terms

u(x)[y1+p(x)y1' + q(x)y1]+u"y+
1 2u’y1' + p(x)u'y1= 0

fall
u" y +1 2u'y1’+ p(x)u’y1= 0

with the order gear reducer, w = u', w' = u''


y1w' + (2y1' + p(x) y1)w = 0

whatislinearinw,incanonicalform

w’+ [2y1'/y1+ p(x)]w = 0

with integrating factor r(x) = exp(∫P(x)dx)

r(x) = exp[2y]1'/y1+ p(x)]dx = exp(2lny1+op(x)dx) = y12expòp(x)dx

applyingittotheequationinw:

[y12expòp(x)dx]w’ + [y12expòp(x)dx][2y11+ p(x)]w = 0

Itispossibletoidentifytheleftsideasd[wr]:

d[w r] = 0

integrating

wr=C

wy12expòp(x)dx = C

solving for w

w = (C/y12) exp[-òp(x)dx] = u’

Integratingagain

u = Cò{exp[-òp(x)dx]/y12}dx + K

with C=1 and K=0

u = ò{exp[-òp(x)dx]}/y12

andfinaly

y2=u(x)y1

is given by
2
y2=yò{exp[-òp(x)dx]/y
1 1 }dx

It can be demonstrated that the Wronskian, W(y1an,d2)¹0 and therefore and1y and y are linearly
2
independentsothegeneralsolutionofthedifferentialequationy''+p(x)y'+q(x)y=0
it would be

y(x) = C1y1+ C2y2

for a pair of constants C1andC2.

Example.-Considerthelinearhomogeneousordinarydifferentialequationwithvariablecoefficients:

x2y" - 5xy' + 9y = 0

Noting that the degree of the coefficients corresponds with the order of the derivative (form
equidimensional or Euler-Cauchy), it is possible that its solution is of the form

y1= x r

Deriving

r-1
y'1= rx

r-2
y1= r(r-1)x

SubstitutingintotheDifferentialEquation

x2[r(r-1)x r-2] - 5x[rxr-1 ] + 9[x]


r
= 0

r(r-1)x r - 5rx r+ 9x r= 0

As x¹0"rÎr Â

r(r-1) - 5r + 9 = 0

r2- 6r + 9 = 0

(r-3)2= 0

r1,2= 3 (repeated real roots)


Asolutionwillbe

y1=x3

Using the reduction of order formula, the second solution is:

3 6
y2=xò{exp[-ò(5/x)dx]/x }dx = x3òdx/x = x3lnx

Thegeneralsolutionis:

y = c1x3+ c2x3lnx = x (c1+3 c2lnx)


Non-Homogeneous Linear Equations
Euler-Cauchy Equation of Order n
Prepared by M.C. Luis E. Castro-Solís

TheequidimensionalorgeneralformoftheEuler-Cauchydifferentialequationofordernis

i (i)
S iaixy = 0 for i = 0, 1, .., n

or expanding the expression

anxny(n)+ an-1xn-1y(n-1) + ... + a2x2y+a1xy' + a0y = 0

r
It is possible to prove that the substitution y = x only produces terms of the same degree (degree r).
r
Consider y = x and its successive derivatives, if they exist:

y = xr
y' = rx r-1
y" = r(r-1)x r-2
y(3) = r(r-1)(r-2)x r-3

y(4) = r(r-1)(r-2)(r-3)x r-4

:
y(n-1)= r(r-1)(r-2)...(r-n+2)x r-n+1
y(n) = r(r-1)(r-2)...(r-n+1)x r-n

SubstitutingintothedifferentialequationofE/C:

anxn[r(r-1)(r-2)...(r-n+1)x r-n
] + a n-1
n-1x r(r-1)(r-2)...(r-n+2)x
r-n+1
]+
2 r-2 r-1 r
... + a2x [r(r-1)x ] + a1x[rx] + a0x = 0

Noticing that the products

r
xnxr-n =xn-1xr-n+1 = ... = x2xr-2 =xxr-1 Invalid input

falls f(r) = 0

anr(r-1)(r-2)...(r-n+1) + an-1r(r-1)(r-2)...(r-n+2)+...+ a2r(r-1)+a1r+ a0= 0


Thisexpressionisanalgebraicexpressionofdegreeninthevariabler;accordingtothefundamentaltheorem.
of algebra, has n roots, r1, r2, ..., rthese roots each provide a solution of the
r
linearly independent form, with which the n-parameter solution of the
Euler-Cauchy equation of order n.

y(x) = c1x r1+ c2x r2+ ... + cxn

Second-order Euler-Cauchy equation


Consider the equidimensional form of the second-order E-C equation.

a2x2y+a1xy' + a0y = 0

Writingincanonicalform

x2y”+Poxy' + Qoy = 0

with Po=(a1/a2) and Qo= (a0/a2)

r
Assumingthatthesubstitutiony=xanditsderivatives satisfytheequation,substituting

y = xr

r-1
y' = rx

r-2
y'' = r(r-1)x

fall

x2r(r-1)x r-1 + P oxrx


r-1
+ Qr ox = 0

r
dividing by x¹0"rÎ Â

f(r) = r(r-1) + Por + Qo= 0

o
f(r) = r2+ (Po - 1)r + Qo= 0

with roots

r1,2= ½(-(Po-1)±[(Po-1)2-4Qo] ½)
Therearethreecasesdependingonwhetherthevalueofthediscriminantargumentisgreaterthan,equalto,orlessthan.
zero.

Distinct real roots and = c1x + c2r1x r2

ï
If (Po-1) -4Qoi = 0, Repeated real roots and = c1x r+ c2x rl n x
2

ï
Complex conjugate roots
y = c1xacos(blnx) + c2xasen(blnx)

Case I (distinct roots) is obtained directly from the substitution we started with. Let's see.
in more detail cases II and III.

Case II. Repeated real roots

r
Ifthevalue(Po-1)2-4Qo= 0®r1= r2= r = -½(P0-1)®y1=c1x

We will use 'order reduction' to calculate a 2nd solution from this. First, let's remember
that the formula of the order reduction method

2
y2=yò[exp(-òp(x)dx)/y
1 1 ]dx

we derive it from the normalized form y” + p(x)y’+ q(x)y = 0; before applying it we must
identify p(x); based on the canonical form we have been working with

x2y”+Poxy' + Qoy = 0

whichisnormalizedto

y" + (Po/x)y’+ (Qo/x2y = 0

it is seen that p(x) = Po/x, from where

r 2r r 2r 2r r -Po
y2=xòexp(-òPo/xdx)/x dx = xòexp(-Poln(x)/x dx = x^o x / x dx

r -Po2[-½(Po-1)] r
y2=xòx /x dx = xòx-Po+Po-1 dx = xòdx/x =r xlnx r

[Link]

y = c1x r+ c2x lrn x


Note - In general, in a C-E equation of order n, if roots of multiplicity m are present, the
the corresponding solution would be:
y(x) = x(c1+r c2lnx+c3(lnx)2+ ... + cm(lnx)m-1)

Case III.- Conjugate complex roots

Ifthevalue(Po-1)2-4Qo< 0®r1,2= a±bi®y = c1xa+bi + c2xa-bi

Theobviousidentityandalnx=xait will be useful in this case, since:

xa±by= e(a±bi)lnx= ealnx ±iblnx


e = ealnx[cos(blnx)±isen(blnx)]

thislastpartwithEuler'sformula,e±ikx= coskx ± isenkx. The general solution is given by:

a
y = ealnx[C1cos(blnx)+C2sen(blnx)] = x [C1cos(blnx)+C2sen(blnx)]

2
Example 1.- Find the general solution of 2x y'' + xy' - 15y = 0

Solution.- The characteristic equation will be:

2r(r-1) + r - 15 = 0
2r2- r - 15 = 0
(2r + 5)(r - 3) = 0
r1-5/2, r2= 3

y(x) = C1x-5/2 x+ C2x3⌫

Example 2.- Find the general solution of x y'' + 7xy' + 13y = 02

Solution.- The characteristic equation will be:

r(r-1) + 7r + 13 = 0
r^2 + 6r + 13 = 0
r=-3±2i

3
y(x) = [C1cos(2lnx) + C2sen(2lnx)]/x⌫

Example3.-Findthegeneralsolutionofxy+6x2y'' + 7xy'3 +(3)y = 0


Solution.- The characteristic equation is:

r(r-1)(r-2) + 6r(r-1) + 7r + 1 = 0
r3+ 3r2+ 3r + 1 = 0
(r+1)3= 0

r = r1= r2= r3-1

Arootofmultiplicity3requiresasecond-degreepolynomial.

y(x) = x (c1+ c2lnx+c3(lnx)2) ⌫


-1

You might also like