Cliffs Quick Revie w
Differential Equations by Steven A . Ledu c
Series Editor Jerry Bobrow, Ph .D .
Cliffs Quick Re...
163 downloads
1850 Views
16MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Cliffs Quick Revie w
Differential Equations by Steven A . Ledu c
Series Editor Jerry Bobrow, Ph .D .
Cliffs Quick Revie w
Differential Equations by Steven A . Ledu c
Series Editor Jerry Bobrow, Ph .D .
CliffsQuickReviewTM Differential Equations Published by: Wiley Publishing, Inc . 909 Third Avenue New'York, NY 1002 2 www.wiley.co m Copyright 0 1995 Wiley Publishing, Inc ., New York, New York Library of Congress Control Number : ISBN : 0-8220-5320-9 Printed in the United States of Americ a 1098765 4 1 B/ST/QR/QS/I N No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by an y means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section s 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, 222 Rosewoo d Drive, Danvers, MA 01923, 978-750-8400, fax 978-750-4470 . Requests to the Publisher for permission should b e addressed to the Legal Department, Wiley Publishing, Inc ., 10475 Crosspoint Blvd ., Indianapolis, IN 46256 , 317-572-3447, fax 317-572-4447, or e-mail permcoordinator@wiley . cor n LIMIT OF LIABILITY/DISCLAIMER OF WARRANTY : THE PUBLISHER AND AUTHOR HAVE USE D THEIR BEST EFFORTS IN PREPARING THIS BOOK . THE PUBLISHER AND AUTHOR MAKE NO REPRESENTATIONS OR WARRANTIES WITH RESPECT TO THE ACCURACY OR COMPLETENESS OF TH E CONTENTS OF THIS BOOK AND SPECIFICALLY DISCLAIM ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE . THERE ARE NO WARRANTIES WHICH EXTEND BEYOND THE DESCRIPTIONS CONTAINED IN THIS PARAGRAPH . NO WARRANTY MAY B E CREATED OR EXTENDED BY SALES REPRESENTATIVES OR WRITTEN SALES MATERIALS . THE ACCU RACY AND COMPLETENESS OF THE INFORMATION PROVIDED HEREIN AND THE OPINION S STATED HEREIN ARE NOT GUARANTEED OR WARRANTED TO PRODUCE ANY PARTICULA R RESULTS, AND THE ADVICE AND STRATEGIES CONTAINED HEREIN MAY NOT BE SUITABLE FO R EVERY INDIVIDUAL . NEITHER THE PUBLISHER NOR AUTHOR SHALL BE LIABLE FOR ANY LOSS O F PROFIT OR ANY OTHER COMMERCIAL DAMAGES, INCLUDING BUT NOT LIMITED TO SPECIAL , INCIDENTAL, CONSEQUENTIAL, OR OTHER DAMAGES . FULFILLMENT OF EACH COUPON OFFE R IS THE RESPONSIBILITY OF THE OFFEROR . Trademarks : Wiley, the Wiley Publishing logo, Cliffs, CliffsNotes, the CliffsNotes logo, CliffsAP, CliffsComplete , CliffsTestPrep, CliffsQuickReview, CliffsNote-a-Day and all related logos and trade dress are registered trademarks o r trademarks of Wiley Publishing, Inc ., in the United States and other countries . All other trademarks are property o f their respective owners. Wiley Publishing, Inc ., is not associated with any product or vendor mentioned in this book . For general information on our other products and services or to obtain technical support, please contact our Custome r Care Department within the U .S . at 800-762-2974, outside the U .S . at 317-572-3993, or fax 317-572-4002 . Wiley also publishes its books in a variety of electronic formats . Some content that appears in print may not be avail able in electronic books .
CONTENTS
REVIEW AND INTRODUCTION Differentiation Partial Differentiation Integration Techniques of Indefinite Integration Integration by substitution Integration by parts Partial Integration Introduction to Differential Equations
FIRST-ORDER EQUATIONS Exact Equations Integrating Factors Separable Equations Homogeneous Equations Linear Equations Bernoulli's Equation
SECOND-ORDER EQUATIONS Linear Combinations and Linear Independence Linear Equations Homogeneous Equations Homogeneous Linear Equations with Constan t Coefficients The Method of Undetermined Coefficients Linear combinations of n functions Variation of Parameters The Cauchy-Euler Equidimensional Equation Reduction of Order Type 1 : Second-order equations with th e dependent variable missing Type 2 : Second-order nonlinear equations with th e independent variable missing
DIFFERENTIAL EQUATIONS
1 1 8 10 12 12 14 16 18
29 29 34 41 47 52 60
63 63 66 67 71 77 79 90 97 100 100 105
CONTENTS
Type 3 : Second-order homogeneous linear equation s where one (nonzero) solution is known
POWER SERIES Introduction to Power Series Taylor Series Power Series Solutions of Differential Equations First-order equations Second-order equations
THE LAPLACE TRANSFORM Linear Transformations The Laplace Transform Operator Using the Laplace Transform to Solv e Differential Equations
A FEW APPLICATIONS Applications of First-Order Equations Orthogonal trajectories Radioactive decay Newton's Law of Cooling Skydiving (part I) Applications of Second-Order Equations Skydiving (part II) Simple harmonic motion Damped oscillations Electric circuits and resonance
10 7
11 3 11 3 11 7 123 123 128
13 5 13 5 13 7 14 8
15 7 15 7 15 7 162 165 167 17 1 17 1 17 4 17 8 18 2
CLIFFS QUICK REVIEW
REVIEW AND INTRODUCTIO N
Before embarking on a study of differential equations, it i s essential to solidify your technical facility with differentiation an d integration . This section serves to review the definitions an d techniques of differentiation and integration and then to introduc e you to the study of differential equations .
Differentiatio n Given a function y = f (x), its derivative, denoted by y' or dy /dx , gives the instantaneous rate of change of f with respect to x . Geometrically, this gives the slope of the curve (that is, the slope o f the tangent line to the curve) y = f (x) . y = f (x )
tangent line to curv e atx=xo
■ Figure 1
■
The second derivative identifies the concavity of the curve y = f (x) . A portion of a differentiable curve y = f (x) from x = a to x = b is said to be concave up if the curve lies above its tangent line s between a and b, and concave down if it lies below its tangent lines .
DIFFERENTIAL EQUATIONS
REVIEW AND INTRODUCTIO N
A curve y = f (x) is concave up at those points x where the secon d derivative is positive, and concave down where the second derivative is negative . Points where the concavity changes are called inflection points and are located at those points x0 where f " (x0 ) = 0 but f"' (x 0) 0 . concave u p f " (x)> o curve abov e its tangent lines
■ Figure 2 ■ Table 1 below lists the most frequently used properties o f derivatives and, for later reference, the corresponding properties o f integrals .
CLIFFS QUICK REVIE W
REVIEW AND INTRODUCTIO N
Table 1 COMPUTATIONAL PROPERTIES O F DIFFERENTIATION AND INTEGRATIO N Integration
Differentiation Linearity :
Linearity :
d(ku) = k du
f k du = k f du = ku
d(u + v) = du +dv
(du + dz)) = u + v
Product rule :
Integration by parts :
d(uv) = u dv + v du
f u dv = uv — f v du
Quotient rule : u
d V
=
vdu — udv v2
Chain rule : (u v)' = (u.' v) v'
Integration by substitution : (u' . 0 • v'
u
o v
In addition to being familiar with the definitions and fundamen tal properties, you should, of course, be able to actually differentiat e a function . Although Table 2 below does not contain ever y differentiation formula, it will probably suffice to differentiat e almost every function you are likely to encounter in practice . Again , for later reference, integration formulas are listed alongside th e corresponding differentiation formulas . (Note : To avoid the repeti tion of writing "+ c" after every result in the right-hand column, th e arbitrary additive constant c has been omitted from each of th e integration formulas, as in Table 1 above . ) DIFFERENTIAL EQUATIONS
REVIEW AND INTRODUCTION
Table 2 DIFFERENTIATION AND INTEGRATION FORMULA S Differentials
Integral s
d(constant) = 0
ti n +
d(u") = nu"-' du
u"du
d (e") = e" d u
e" du = e"
n+1
1 d(ln u) = —d u u
f !du u
d(sin u) = cos u du
f cosudu=sin u
d (cos u) = –sin u du
d (cot u) = – csc 2 u du d (sec u) = sec u tan u d u d(csc u) = –csc u cot u du
f sec' u du = tan u f csc'- u du = —cot u f sec u du = In sec u + tan u csc u du = —ln l csc u + cot u j
du
du = arcsin u
d(arctan u) = + u du l -
, du = arctan u 1 + u-
4
—
in l u l
sin u du = —cos u
d(tan u) = sec u du
d(arcsin u) =
(nom—1 )
u2
CLIFFS QUICK REVIEW
REVIEW AND INTRODUCTIO N
Example 1 : Differentiate each of the following :
(b)
J' = xZe x
(c)
y = In x/x
(e)
y = \/x2
(f)
y = sin(x2 )
(g)
y = sin 2 x
(h)
tanx y = e
(i)
y = csc(sin
+
1
The solutions are as follows : (a)
y ' =6x— 5
(b)
Using the product rule, y' = x 2
(c)
By the quotient rule,
• e x + e x • 2x = xe v(x
+ 2)
1
x•x—lnx•1
y' =
1—ln x x2
All of the remaining parts use the chain rule (as embodied in the formulas in Table 2) . (d)
y' =4(x 3 +x— 1) 3 (3x2 + 1 )
DIFFERENTIAL EQUATIONS
REVIEW AN D INTRODUCTIO N
+ 1)1/2 ( e ) y = (x2
+ 1)-' /2 . 2x =
y' =
(f)
y' = 2x cos(x 2 )
(g)
y = (sinx) 2 = y' = 2 sinx cosx = sin 2x
(h)
y' = etan .r sect x
(i)
y' _ —csc(sin f) cot(sin
cos • 2\5c
x \/x '- + 1
■
Example 2 : What is the equation of the tangent line to the curv e y = e-' In x at the point (1, 0) ? The first step is to find the slope of the tangent line at x = 1 , which is the value of the derivative of y at this point : slope at point (1, 0) =
dy
dx
1
= e x . — + lnx•e k
x
= e x= l
Since the point-slope formula says that the straight line with slop e m which passes through the point (x0 , yo) has the equatio n Y — Yo =m (x—xo )
the equation of the desired tangent line is y = e(x — 1) .
Example 3 : Is the curve y = arcsin down at the point (4, 6) ?
■
concave up or is it concav e
Concavity is determined by the sign of the second derivative .
6
CLIFFS QUICK REVIEW
REVIEW AN D
INTRODUCTION
Since d(arcsin u) =
1 vl — u 2
du
the first derivative of y = arcsin V.- i s 1
1 v/-x-
\/1
1 (x — x-) - ' — 2 r—x2
Its second derivative is therefore 1 4 (x —
x 2 ) - 3/2 (1
2x — 1
2x)
4\/(x _ x- )
For x = 4'- , the denominator in the expression above for y" is positive (as it is for any x in the interval 0 < x < 1), but the numerator i s negative . Therefore, y"(4) < 0, and the curve is concave down at th e point (4, ) . ■
Example 4 : Consider the curve given implicitly by the equatio n 3x2y—y=x+
1
What is the slope of this curve at the point where it crosses the x axis ? To find the slope of a curve defined implicitly (as is the cas e here), the technique of implicit differentiation is used : Differentiat e both sides of the equation with respect to x ; then solve the resultin g equation for y' . 3xy-y3=x+ 1
(3x2y' + 6xy) — 3y y' = 1 y' (3x 2 —
3y2 )
= 1 — 6xy 1 -6xy
Y'
DIFFERENTIAL EQUATIONS
3(x2 _ Y )
REVIEW AN D INTRODUCTION
The curve crosses the x axis when y = 0, and the given equatio n clearly implies that x = -1 at y = 0 . From the expression directl y above, the slope of the curve at the point (—1, 0) i s 1
1 — 6xy
Y ' I(-1,0)
=
3(x2
y2 )
1u,
3
■
Partial Differentiatio n Given a function of two variables, f (x, y), the derivative with respec t to x only (treating y as a constant) is called the partial derivative off with respect to x and is denoted by either of /8x or f.. Similarly, th e derivative of f with respect to y only (treating x as a constant) i s called the partial derivative off with respect toy and is denoted b y either of /cry or f, . The second partial derivatives
off come
in four types : Notation
■
Differentiate f with respect tox twice . (That is, differentiate f with respect to x ; then differentiate the result with respec t to x again . )
a2f n►x- or j ,
■ Differentiate f with respect toy twice . (That is, differentiate f with respect toy ; then differentiate the result with respec t toy again . )
af ay
or f., .
Mixed partials : ■
8
First differentiate f with respect tox ; the n differentiate the result with respect toy .
of ay ax
or L.,.
CLIFFS QUICK REVIEW
REVIEW AND INTRODUCTIO N
■ First differentiate f with respect toy ; then differentiate the result with respect tox .
a2f axay
or
f .,_
For virtually all functions f (x, y) commonly encountered in practice , will be identical to f, .; that is, the order in which the derivative s are taken in the mixed partials is immaterial .
f,
Example 5 :
Iff (x, y) = 3x2y + 5x — 2y 2 +
1, find f., f, .,
an d
First, differentiating f with respect to x (while treating y as a constant) yields f,-=6xy+ 5 f
Next, differentiating constant) yields
with respect to y (while treating x as a f .=3x2—4y
The second partial derivative with respect to x ; therefore,
means the partial derivative o f
a
a
fit = (f,)-, =~x(f,-)=c~x(6xy+5)=6Y
The second partial derivative f, ., . means the partial derivative of f. with respect toy ; therefore , fn .= (f .),,=
The mixed partial toy ; therefore,
f
a
(f. ) =
Y
( 3x2—4y ) = - 4
means the partial derivative
off.
with respec t
(f,-), . = ~ (f = 1( 6xy +5) =6 x
DIFFERENTIAL EQUATIONS
9
REVIEW AN D INTRODUCTION
x
The mixed partial f, means the partial derivative off. with respec t to .x ; therefore, a
f«=(f3 = x (fr)= Note that
fa. = f
as expected .
a ax
3x'--4y)=6 x
■
Integratio n Indefinite integration means antidifferentiation ; that is, given a function f (x), determine the most general function F(x) whos e derivative is f (x) . The symbol for this operation is the integral sign , followed by the integrand (the function to be integrated) and a differential, such as dx, which specifies the variable of integration . On the other hand, the fundamental geometric interpretation o f definite integration is to compute an area . That is, given a functio n f (x) and an interval a < x < b in its domain, the definite integra l off from a to b gives the area bounded by the curve y = f (x), the x axis, and the vertical lines x = a and x = b . The symbol for thi s operation is the integral sign with limits of integration (a and b) , followed by the function and the differential which specifies th e variable of integration .
f,
r,
■ Figure 3 ■
10
CLIFFS QUICK REVIEW
REVIEW AND INTRODUCTIO N
From their definitions, you can see that the processes o f indefinite integration and definite integration are really very different . The indefinite integral of a function is the collection of function s which are its antiderivatives, whereas the definite integral of a function requires two limits of integration and gives a numerica l result equal to an area in the xy plane . However, the fact that bot h operations are called "integration " and are denoted by such simila r symbols suggests that there is a link between them . The Fundamental Theorem of Calculus says that differentiatio n (finding the slope of a curve) is the inverse operation of definit e integration (finding the area under a curve) . More explicitly, Part I of the Fundamental Theorem says that if a function is integrated (t o form a definite integral with a variable upper limit of integration) , and the result is then differentiated, the original function i s recovered ; that is, differentiation "undoes" integration . Part I I gives the connection between definite and indefinite integrals . I t says that a definite integral can be computed by first determining a n indefinite integral (so computing the area under a curve is done b y antidifferentiating) . The Fundamental Theorem of Calculus (Part I) : Iff is continuous, then d dx
f f (t) dt = f (x) .
The Fundamental Theorem of Calculus (Part II) : If f is continuous with antiderivative F, then Jf(x) dx = F(b) — F(a) .
DIFFERENTIAL EQUATIONS
REVIEW AND INTRODUCTIO N
Example 6 : Evaluate the integra l (x4 —3x2 +x— 1) dx Using the first integration formula in Table 2, every functio n whose derivative equals f (x) = x4 — 3x2 + x — 1 is given b y (x4 — 3x 2 +x— 1)dx= .k x5 —x3 +4x2 — x+ c where c is an arbitrary constant .
■
Techniques of Indefinite Integratio n
Integration by substitution. This section opens with integration by substitution, the most widely used integration technique, illustrate d by several examples . The idea is simple : Simplify an integral b y letting a single symbol (say the letter u) stand for some complicate d expression in the integrand . If the differential of u is left over in the integrand, the process will be a success .
Example 7 : Determine
f x,,/x2 + 1 dx Let u = x2 + 1 (this is the substitution) ; then du = 2x dx, and th e given integral is transformed int o
f ~x 2 + 1 xdx= f
. 'du =
which transforms back to 4 (x2 + 1) 3 / 2 + c .
u h/ 2 du =
• 72 u 3/2 + c
■
CLIFFS QUICK REVIEW
REVIEW AN D INTRODUCTION
Example 8 : Integrate sin 3 x cosx dx Let u = sin x; then du = cos x dx, and the given integral become s
f ll 3 du =
u 4 +c =
Example 9 : Evaluate
dx :
sin 4 x+ c
f tan x dx
First, rewrite tan x as sin x/cosx, then let u = cosx, du = –sin x
f tanxdx=
sm x cosx ~
f
duu
= —InIul +c = —ln~cosx~ + c ■
Example 10: Evaluate xe x- dx
Let u = x2 ; then du = 2x dx, and the integral is transformed int o fxe2dx=feh1du=eh1+c=e+c
■
Example 11 : Determine
f sect x tan x dx Let u = sec x; then du = sec x tan transformed int o
x dx,
and the integral i s
f sec t x tan x dx = f sec x • sec x tan x dx = f u du = u 2 + c = 2 sec t x + c
DIFFERENTIAL EQUATIONS
■
REVIEW AND INTRODUCTIO N
Integration by parts . The product rule for differentiation say s d (uv) = u dz► + v du . Integrating both sides of this equation give s uv = f u dv + f v du, or equivalentl y
f u dz► = uv — f v d u This is the formula for integration by parts . It is used to evaluate integrals whose integrand is the product of one function (u) and th e differential of another (dv) . Several examples follow .
Example 12: Integrate
f xe dx Compare this problem with Example 10 . A simple substitutio n made that integral trivial ; unfortunately, such a simple substitutio n would be useless here . This is a prime candidate for integration by parts, since the integrand is the product of a function (x) and th e differential (e x. dx) of another, and when the formula for integratio n by parts is used, the integral that is left is easier to evaluate (or, i n general, at least not more difficult to integrate) than the original . Let u = x and dz► = ex dx ; then v=ek
u=x du = dx
dv = et dx
and the formula for integration by parts yield s
5 u dv = uv — 5 v du 5 xe x dx = xe x — 5 e dx =xe x — e x, + c = e x (x
14
— 1)+c
■
CLIFFS QUICK REVIE W
REVIEW AND INTRODUCTIO N
Example 13 : Integrate
f x cos x dx Let u = x and dv = cos x dx ; the n u=x
v = sin x
du=dx
dzv=cosxdx
The formula for integration by parts give s
f
f u dv = uzv – f v d u x cosxdx =xsinx –
sin xdx
= x sin x + cos x + c
Example 14: Evaluate
f
Inxdx
Let u = inx and dzv = dx ; the n u=lnx
zv= x
1
du = – dx dv = dx x
and the formula for integration by parts yield s
f u dzv = uv – f v du
f lnXdx=xtnx— f
i x--dx
=xlnx –x + c
DIFFERENTIAL EQUATIONS
x
■
REVIEW AND INTRODUCTIO N
Partial Integratio n Suppose it is known that a given function f (x) is the derivative of some function F(x) ; how is F(x) found? The answer, of course, is to integrate f (x) . Now consider a related question : Suppose it is known that a given function f (x, y) is the partial derivative with respect to x of some function F(x, y) ; how is F(x, y) found? The answer is to integrate f(x, y) with respect to x, a process I refer to as partia l integration . Similarly, suppose it is known that a given function f(x, y) is the partial derivative with respect to y of some functio n F(x, y) ; how is F(x, y) found? Integrate f (x, y) with respect toy . Example 15 : Let M(x, y) = 2xy2 + x2 — y . It is known that M equal s f;_ for some function f(x, y) . Determine the most general suc h function f (x, y) . Since M(x, y) is the partial derivative with respect to x of som e function f(x, y), M must be partially integrated with respect to x t o recover f. This situation can be symbolized as follows : itv
f(x ,Y) f (x, y) `
f(
> M(x, y ) M(x, y)
)iL
Therefore, f(x,y)= fM(x,y)8x = (2xy 2 +x' — f(x,y)=x'v2+
16
y)ax
x3 —xy+tp(Y )
CLIFFS QUICK REVIEW
REVIEW AN D INTRODUCTION
Note carefully that the "constant" of integration here is an y (differentiable) function of y —denoted by i/r(y) since any suc h function would vanish upon partial differentiation with respect to x (just as any pure constant c would vanish upon ordinary differentia tion) . If the question had asked merely for a function f (x, y) fo r which f. = M, you could just take r(y) = 0 . ■ Example 16 : Let N(x, y) = sin x cos y — xy + 1 . It is known that N equals f,, for some function f (x, y) . Determine the most general suc h function f (x, y) . Since N(x, y) is the partial derivative with respect to y of som e function f (x, y), N must be partially integrated with respect to y t o recover f. This situation can be symbolized as follows :
> N(x, y )
f(x, y ) f (x, y)
f()i)y
N(x,y )
Therefore, f (x, y )
=
f N(x, y ) ay f (sinxcosy — xy + 1)ay
f (x, y) = sinx siny —
+ y + (x )
Note carefully that the "constant" of integration here is any (differentiable) function of x —denoted by 4(x) since any such function would vanish upon partial differentiation with respect toy . If the question had asked merely for a function f (x, y) for which f,. = N, you could just take (x) = 0 . ■
DIFFERENTIAL EQUATIONS
REVIEW AND INTRODUCTIO N
Introduction to Differential Equation s In high school, you studied algebraic equations lik e 3x-2(x-4)=5x, x 2 -8x+15=0,
and
16x-21= 4
The goal here was to solve the equation, which meant to find th e value (or values) of the variable that makes the equation true . Fo r example, x = 2 is the solution to the first equation because onl y when 2 is substituted for the variable x does the equation become a n identity (both sides of the equation are identical when and onl y whenx = 2) . In general, each type of algebraic equation had its own particular method of solution ; quadratic equations were solved by on e method, equations involving absolute values by another, and so on . In each case, an equation was presented (or arose from a wor d problem), and a certain method was employed to arrive at a solution, a method appropriate for the particular equation at hand . These same general ideas carry over to differential equations , which are equations involving derivatives . There are different type s of differential equations, and each type requires its own particula r solution method . The simplest differential equations are those o f the form y' = f (x) . For example, consider the differential equatio n dy dx = 2
x
It says that the derivative of some function y is equal to 2x . To solv e the equation means to determine the unknown (the function y ) which will turn the equation into an identity upon substitution . I n this case all that is needed to solve the equation is an integration :
CLIFFS QUICK REVIEW
18
REVIEW AND INTRODUCTION
Thus, the general solution of the differential equation y' = 2x is y = x' + c, where c is any arbitrary constant . Note that there are actually infinitely many particular solutions, such asy = x 2 + 1,y = x — 7, or y = x 2 + Tr, since any constant c may be chosen . Geometrically, the differential equation y' = 2x says that at each point (x, y) on some curve y = y(x), the slope is equal to 2x . Th e solution obtained for the differential equation shows that thi s property is satisfied by any member of the family of curvesy = x 2 + c (any only by such curves) ; see Figure 4 .
■ Figure 4
■
Since these curves were obtained by solving a differential equatio n which either explicitly or implicitly involves taking an integral the y are sometimes referred to as integral curves of the differentia l equation (particularly when these solutions are graphed) . If on e particular solution or integral curve is desired, the differentia l equation is appended with one or more supplementary conditions . These additional conditions uniquely specify the value of th e arbitrary constant or constants in the general solution . For example , consider the proble m dy
dx=2x
and y=2when x= 0
DIFFERENTIAL EQUATIONS
19
REVIEW AND INTRODUCTIO N
The initial condition "y = 2 when x = 0 " is usually abbreviate d "y(0) = 2," which is read "y at 0 equals 2 ." The combination of a differential equation and an initial condition (also known as a constraint) is called an initial value problem (abbreviated IVP) . For differential equations involving higher derivatives, two o r more constraints may be present . If all constraints are given at th e same value of the independent variable, then the term IVP stil l applies . If, however, the constraints are given at different values o f the independent variable, the term boundary value problem (BVP ) is used instead . For example , this is an IVP : y" + 2y' — 3y = 0, y(0) = 1, y'(0) = 5 T
same i l
but this is a BVP :
y" + 2y' — 3y = 0,
y(0) = 1, y'(l) = 5 T
different T
To solve an IVP or BVP, first find the general solution of th e differential equation and then determine the value(s) of th e arbitrary constant(s) from the constraints .
Example 17 : Solve the IVP y' = 2x y(0) = 2 As discussed above, the general solution of this differentia l equation is the family y = x 2 + c . Since the constraint says that y must equal 2 when x is 0 , y(0)=2
[x'+c],-_,►=2
so the solution of this IVP is y = x2 + 2 .
0 2 +c=2
c= 2
■
CLIFFS QUICK REVIE W
20
REVIEW AN D INTRODUCTION
Example 18: Consider the differential equation y" + 2y' — 3y = 0 . Verify that y = c 1e v + c2e -=x (where c l and c, are arbitrary constants) is a solution . Given that every solution of this differentia l equation can be written in the form y = c 1e v + c,e -3x, solve the IVP y" + 2y' -3y= 0 y(0)= 1 y'(0) = 5
To verify that y = c i e x + c2e -3x is a solution of the differentia l equation, substitute . Since y' = c 1e x" — 3c,e -=
and
y" = c 1 e x +
once c 1e - + c,e -3' is substituted for y, the left-hand side of th e differential equation become s y"+2y' -3y= (c i e x +9c,e -3x )+2(c i ex- -3c,e - ' ) — 3(c 1 e x + c,e - 3x- ) = (c i e x
+
2c 1 e x 3c i ex ) — 6c,e -3x — 3c,e -3x + (9c,e )
=0
3
Now, to satisfy the conditions y(0) = 1 and y ' (0) = 5, the constants c l and c, must be chosen so that y(0)
1 = [c 1 e x
+
c,e-3"]x._1) = 1
c 1 + c2 = 1
an d y'(0)=5
[c 1 e' -3c,e-3x]x.=1,=5
c 1 -3c,= 5
Solving these two equations yields c 1 = 2 and c, = -1 . Thus, th e particular solution specified by the given IVP isy = 2e `" — e- 3x . n
DIFFERENTIAL EQUATIONS
REVIEW AN D INTRODUCTION
The order of a differential equation is the order of the highest derivative that appears in the equation . For example, y' = 2x is a first-order equation, y" + 2y' – 3y = 0 is a second-order equation , and y" – 7y ' + 6y = 12 is a third-order equation . Note that th e general solution of the first-order equation from Example 1 7 contained one arbitrary constant, and the general solution of th e second-order equation in Example 18 contained two arbitrar y constants . This phenomenon is not coincidental . In most cases, th e number of arbitrary constants in the general solution of a differentia l equation is the same as the order of the equation .
Example 19 : Solve the second-order differential equation y" = x + cos x. Integrating both sides of the equation will yield a differentia l equation for y' : y'
=
f y" (x + cos x) dx
=2x2 +sinx+c1 Integrating once more will give y : y =
y'
(2x-+sinx+c,)dx =x3 –cosx+c 1x+c2 where c 1 and c 2 are arbitrary constants . Note that there are tw o arbitrary constants in the general solution, which you shoul d typically expect for a second-order equation . ■
REVIEW AND INTRODUCTIO N
Example 20: For the following IVP, find the solution valid fo r x>0 : y"'=105x—
2 x'
+ 6
y(1)= 7 y'(1) = 3 7 y"(1)=7 3 The general solution of a third-order differential equatio n typically contains three arbitrary constants, so an IVP involving a third-order differential equation will necessarily have three constraint equations (as is the case here) . As in Examples 17 and 19, th e given differential equation is of the form y( 't'
=
f(x )
where y ( " ) denotes the nth derivative of the function y . Thes e differential equations are the easiest to solve, since all they requir e are n successive integrations . Note how the first-order differentia l equation in Example 17 was solved with one integration, and th e second-order equation in Example 19 was solved with two integra tions . The third-order differential equation given here will be solve d with three successive integrations . Here's the first : y"=
(105x' /2 — 2x -3
+6)dx
= 70x312 + x -2 + 6x + c , The value of this first arbitrary constant (c1) can be found by applying the condition y " (1) = 73 : y"(1) = 73
[70x 312 + x -2 + 6x + c,],._, = 73 X70+1+6+c,=73~c,=— 4
Thus,y" = 70x312 + x -2 + 6x — 4. DIFFERENTIAL EQUATIONS
REVIEW AND INTRODUCTIO N
Now, perform the second integration, which will yield y' : y' = f(7ox3/2 +x 2 + 6x - 4) v =28x /
2–x
-'
+
3x222 –4x+c ,
The value of this arbitrary constant (c 2 ) can be found by applying the constraint y' (1) = 37 : y'(1) = 37 = [28x 52 –x-' + 3x2 – 4x +
=37
28– 1 +3–4+c,=3 7 =11 Therefore,y' = 28x' ' 2 –x-' + 3x2 – 4x + 11 . Integrating once mor e will give the solution y : y = f (28x" – x - ' + 3x 2 – 4x + 11) dx =8x 71 2–lnx+x3 –2x2 + 1lx+c 3 The value of this arbitrary constant ( c 3 ) can be found by applyin g the condition y(1) = 7 : y(l) = 7
[8 x712 – In x + x 3 – 2x 2 + 1 lx + c3 ] .,._ 1
=
7
8–0+1–2+11+c3 = 7 c3 = - 1 1 Thus, the solution is y = 8x 71 – In x + x3 – 2x 2 2
+
11x – 11 .
A few technical notes about this example : ■ The given differential equation makes sense only for x > 0 (note the f and 2/x3 terms) . To respect this restriction, th e problem states the domain of the equation and its solutio n [that is, the set of values of the variable(s) where th e equation and solution are valid] asx > 0 . Always be aware o f the domain of the solution .
REVIEW AND INTRODUCTIO N
■ Although the integral of x-' is usually written in Ix', the absolute value sign is not needed here, since the domain o f the solution is x > 0, and 'xi = x for any x > 0 . ■ Contrast the methods used to evaluate the arbitrary constants in Examples 18 and 20 . In Example 18, the constraints were applied all at once at the end . In Example 20, however , the constants were evaluated one at a time as the solutio n progressed . Both methods are valid, and each particula r problem (and your preference) will suggest which t o use . ■
Example 21 : Find the differential equation for the family of curve s x2 + y 2 = c 2 (in the xy plane), where c is an arbitrary constant . This problem is a reversal of sorts . Typically, you're given a differential equation and asked to find its family of solutions . Here , on the other hand, the general solution is given, and an expressio n for its defining differential equation is desired . Differentiating bot h sides of the equation (with respect to x) give s x2
+
y-' = c ' dy
2x+2y-= 0 dy x+y
— d x= 0 dy
x
dx
y
This differential equation can also be expressed in another form , one that will arise quite often . By "cross multiplying," the differential equation directly above becomes ydy= –xdx
DIFFERENTIAL EQUATIONS
11
REVIEW AN D INTRODUCTION
which is then normally written with both differentials (the dx and the dy) together on one side : xdx+ydy= 0
Either y' = —x/y or x dx + y dy = 0 would be an acceptable way o f writing the differential equation that defines the given family (o f circles) x2 + y2 = c2 . n
Example 22: Verify that the equation y = ln(x/y) is an implicit solution of the IVP ydx—x(y+1)dy= 0 y(e) = 1
First note that it is not always possible to express a solution i n the form "y = some function of x ." Sometimes when a differentia l equation is solved, the solution is most naturally expressed with y' s (the dependent variable) on both sides of the equation, as in y = ln(x/y) . Such a solution is called an implicit solution, as opposed t o an explicit solution, which has y all by itself on one side of the equation and a function of x only on the right (as in y = x 2 + 2, fo r example) . Implicit solutions are perfectly acceptable (in some cases , necessary) as long as the equation actually defines y as a function o f x (even if an explicit formula for this function is not or cannot be found) . However, explicit solutions are preferable when available . Perhaps the simplest way to verify this implicit solution is t o follow the procedure of Example 21 : Find the differential equatio n for the solution y = ln(x/y) . To simplify the work, first rewrit e ln(x/y) as In x — In y : y = lnx — ln y dy 1 1 dy dx x y dx
26
CLIFFS QUICK REVIEW
REVIEW AN D INTRODUCTION
dx
1 + Y/
x
dy
1/x
dx 1 + (1/y ) dy
1 /x
xy
d x 1 + (1 /y) xy dy
y
dx x(y + 1 ) ydx-x(y+ 1)dy= 0
Therefore, the differential equation given in the statement of the problem is indeed correct . The initial condition is also satisfied , since 1 = ln(e/ 1) implies y(e) = 1 satisfiesy = ln(x/y) . ■
Example 23: Discuss the solution to each of the differential equations
(y')-+x-=0
and (y')--+y'= 0
The first differential equation has no solution, since no real valued function y = y (x) can satisfy (y' )2 = —x 2 (because squares o f real-valued functions can ' t be negative) . The second differential equation states that the sum of tw o squares is equal to 0, so both y' and y must be identically 0 . Thi s equation does have a solution, but it is only the constant functio n y = 0 . Note that this differential equation illustrates an exception t o the general rule stating that the number of arbitrary constants in th e general solution of a differential equation is the same as the order o f the equation . Although (y' ) 2 + y2 is a first-order equation, it s general solutiony = 0 contains no arbitrary constants at all . ■
DIFFERENTIAL EQUATIONS
REVIEW AND INTRODUCTIO N
One final note : Since there are two major categories of derivatives, ordinary derivatives lik e dy
dy
►► and, ch . = y
y'
j
and partial derivatives such a s f —ax
.fx.
and
a2f
ay y ax
.f.,,.
there are two major categories of differential equations . Ordinary differential equations (ODEs), like the equations in all of th e examples above, involve ordinary derivatives, while partial differen tial equations (PDEs), such a s 8f
82f
~t=k cox-
and
8 2w ax-
82w
+ , = 0 ay
involve partial derivatives . Only ordinary differential equations ar e examined in this book .
28
CLIFFS QUICK REVIEW
FIRST-ORDER EQUATIONS
From the Introduction, you know that a first-order differentia l equation is one containing a first but no higher derivative of th e unknown function . For virtually every such equation encountered i n practice, the general solution will contain one arbitrary constant , that is, one parameter, so a first-order IVP will contain one initia l condition . There is no general method that solves every first-orde r equation, but there are methods to solve particular types .
Exact Equation s Given a function f (x, y) of two variables, its total differential df i s defined by the equation df=~f dx+ f ~ dy Y
Example 1 : If f (x, y) = x-y
+ 6x - y 3,
then
df=(2xy+6)dx+(x--3y 2 )dy
n
The equation f(x, y) = c gives the family of integral curves (that is , the solutions) of the differential equatio n df = 0
Therefore, if a differential equation has the form af
af
dx+ —dy=O (* ) ax dy
for some function f (x, y), then it is automatically of the form
DIFFERENTIAL EQUATIONS
df = 0 ,
29
FIRST-ORDER EQUATIONS
so the general solution is immediately given by f (x, y) = c . In thi s case, 8f of dx + ~ dy ax Y is called an exact differential, and the differential equation (*) i s called an exact equation . To determine whether a given differentia l equation M(x, y) dx + N(x, y) dy = 0 is exact, use th e Test for Exactness: A
differential equation M dx + N dy = 0 is exact i f
and only if aM
aN
ay
ax
Example 2: Is the following differential equation exact ? ( y2— 2x ) dx +(2xy+ 1 ) dy = 0 The function that multiplies the differential dx is denote d M(x,y), so M(x,y) = y2 — 2x ; the function that multiplies th e differential dy is denoted N(x, y), so N(x, y) = 2xy + 1 . Sinc e aM
y
=2y
and
8N
~►x
=2y
the Test for Exactness says that the given differential equation i s indeed exact (since M,. = NO . This means that there exists a function f(x, y) such tha t af _
-x
30
M(x,y)=-2x
af
and - = N(x,y)=2xy+ 1 Y
CLIFFS QUICK REVIE W
FIRST-ORDER EQUATIONS
and once this function f is found, the general solution of th e differential equation is simply f(x,y)= c
(where c is an arbitrary constant) .
■
Once a differential equation M dx + N dy = 0 is determined t o be exact, the only task remaining is to find the function f (x, y) suc h that f, = M and f, = N . The method is simple : Integrate M wit h respect to x, integrate N with respect toy, and then "merge" the tw o resulting expressions to construct the desired function f.
Example 3 : Solve the exact differential equation of Example 2 :
(y 2 - 2x) dx + ( ay + 1)dy= 0 First, integrate M(x, y) = 2x with respect to x (and ignore the arbitrary "constant" of integration) : y2 –
f M (x , y ) ax=
( y22
2x)axxy 2
x2
Next, integrate N(x, y) = 2xy + 1 with respect toy (and again ignore the arbitrary "constant " of integration) :
f N (x, y) ay =
(2xy + 1)ay = xy 2 + y
Now, to "merge" these two expressions, write down each ter m exactly once, even if a particular term appears in both results . Here the two expressions contain the terms xy 2 , and y, s o –x2,
f(x,y)
DIFFERENTIAL EQUATIONS
xy2 -
x2 + y
I
FIRST-ORDER EQUATION S
(Note that the common term xy 2 is not written twice .) The genera l solution of the differential equation is f(x, y) = c, which in this cas e becomes xy 2 –x 2 +y=c
■
Example 4: Test the following equation for exactness and solve it i f it is exact : x(1 – siny)
dy = (cosx – cosy – y) dx
First, bring the dx term over to the left-hand side to write th e equation in standard form : (y+cosy–cosx)cGr+(x–xsiny)dy= 0 Therefore, M(x, y) = y + cos y – cos x, and N(x, y) = x – x sin y . Now, since aM ay
aN
= 1 – siny and
ax
= 1 – sin y
the Test for Exactness says that the differential equation is indeed exact (since M,. = NX ) . To construct the function f (x, y) such tha t fX = M and f,. = N, first integrate M with respect to x :
f M(x, y) ax =
(y + cosy – cos x) ax = xy + x cos y – sin x
Then integrate N with respect toy :
f N(x, y) ay =
(x – x sin y) ay = xy + x cos y
Writing all terms that appear in both these resulting expression s without repeating any common terms gives the desired function : f (x, y) = xy + x
cosy – sin x
CLIFFS QUICK REVIEW
FIRST-ORDER EQUATIONS
The general solution of the given differential equation is therefore xy+xcosy — sinx=c
■
Example 5 : Is the following equation exact ? (3xy -y 2 )dx+x(x
- y)dy = 0
Since aM aY
a d (3xy—y2) = x — 2y Y
but 0N
a (x2— Ay
ax ax
)= — y
it is clear that M,. ~ Na , so the Test for Exactness says that this equation is not exact . That is, there is no function f (x, y) whose derivative with respect to x is M(x, y) = 3xy — and which at th e same time has N(x, y) = x(x — y) as its derivative with respec t toy . ■ y2
Example 6: Solve the IV P (3x2y — 1) dx + (x 3 + 6y — y 2 ) dy = 0 y(0) = 3
The differential equation is exact becaus e aM _ a (3x 2y— 1)=3x 2 ay ay
and
Integrating M with respect to x give s
aN 0x
a ax
(x3
+6y—y 2 )=3x 2
f M(x,y)ax= f (3x 2y — i)ax=x ;y— x
DIFFERENTIAL EQUATIONS
FIRST-ORDE R EQUATION S
and integrating N with respect toy yield s
f N (x , y ) iy= f(x 3 + ~y —y 2 )3y =x 3y + 3y : —
y3
Therefore, the function f (x, y) whose total differential is th e left-hand side of the given differential equation i s f(x, y) = x-y - x + 3y- -y 3
and the general solution i s x 3y — x + 3y 2 —
y3= c
The particular solution specified by the IVP must have y = 3 whe n x = 0; this condition determines the value of the constant c : i)v=3 = C
[x 3y—x+3y 2 —
0—0+27—9=c
18= c
Thus, the solution of the IVP i s x 3y—x+3y 2 — y3= 18
■
Integrating Factor s If a differential equation of the for m M(x,y)dx + N(x,y)dy = 0 (* ) is not exact as written, then there exists a function µ(x, y) such tha t the equivalent equation obtained by multiplying both sides of (* ) by µ, (,uM) dx + (µN) dy = 0 is exact . Such a functionµ is called an integrating factor of th e original equation and is guaranteed to exist if the given differentia l equation actually has a solution . Integrating factors turn nonexac t
34
CLIFFS QUICK REVIEW
FIRST-ORDE R EQUATIONS
equations into exact ones. The question is, how do you find an integrating factor? Two special cases will be considered . Case 1 :
Consider the differential equation Mdx + N dy = 0 . If thi s equation is not exact, then M, . will not equal N,. ; that is, M,.— N,- 0 . However, if M, . — N,. N
is a function of x only, let it be denoted by (x) . The n µ(x )
=
will be an integrating factor of the given differentia l equation . Case 2:
Consider the differential equation Mdx + N dy = 0 . If thi s equation is not exact, then M, will not equal N, . ; that is , M,. — N,- 0 . However, if —M
is a function of y only, let it be denoted by 0(y) . The n µ(Y) = of o y) (n.
will be an integrating factor of the given differentia l equation .
Example 7 : The equatio n
(3xy—y 2)dx+x(x—y)dv= 0 is not exact, sinc e
M,.= a (3xy — y 2 ) = 3x — 2y but N,.=a(x'—xy)_2x — y Y
DIFFERENTIAL EQUATIONS
FIRST-ORDER EQUATIONS
(recall Example 5) . However, note tha t M, .—N,.
(3x — 2y) — (2x — y)
x—y
1
N
x(x—y)
x(x — y)
x
is a function of x alone . Therefore, by Case 1 , ef(1/ .v)ILV
= e In .v = x
will be an integrating factor of the differential equation . Multiplyin g both sides of the given equation by ,u, = x yield s (3x 2y —xy 2 )dx+(x 3 —x 2y )dy= 0 µM— M
µN
which is exact because aM aN a = 3x-'— may = ax Y Solving this equivalent exact equation by the method described i n the previous section, M is integrated with respect to x ,
f M ax = f(3x2y — xy 2 )
ax
= x 3y — x 2y 2
and Nis integrated with respect toy :
5 Nay= 5 (x3 —x-y)ay=x 3y—3x'y (with each "constant" of integration ignored, as usual) . Thes e calculations clearly give xy— ;xy 2 = c as the general solution of the differential equation .
■
Example 8 : The equatio n (x + y) sin y dx + (x sin y + cosy) dy = 0
36
CLIFFS QUICK REVIE W
FIRST-ORDER EQUATIONS
is not exact, sinc e M,. = (x + y) cosy + sin y
but
N,. = sin y
However, note that M, . — Na
_
(x + y) cosy + sin y — sin y
—(x + y) sin y
—M
cos y sin y
is a function ofy alone (Case 2) . Denote this function by iff(y) ; since cos y dy
y) dy = —
sin = — ln(sin y ) y
the given differential equation will hav e eh(' = e —In(sin-v)
= e in(sinr) - i
=
(sin y) — '
as an integrating factor . Multiplying the differential equatio n through by ,u, = (sin y)-' yields
cos y1
(x+y)cLr+ x+ S ~ ny ldy= U .Y~ µM= M
µN= N
which is exact because M,. = 1 = N,.
To solve this exact equation, integrate M with respect to x an d integrate N with respect toy, ignoring the "constant" of integratio n in each case :
f Max= f(x +y)3x = x 2 +xy y f Nay= f (x+ cos )ay=y+lnsiny I sin y
DIFFERENTIAL EQUATIONS
FIRST-ORDE R EQUATIONS
These integrations imply tha t +xy+lnIsinyl= c is the general solution of the differential equation .
■
Example 9: Solve the IV P (3&y + x) dx + e-' dy = 0 y(0) = 1 The given differential equation is not exact, sinc e M, . =
a y
(3e-5) + x) = 3e- v
but
N,
a (e-'-) = e-' . = ~x
However, note that M,. — N,. N
3e-'- e-
= 2
which can be interpreted to be, say, a function of x only ; that is, thi s last equation can be written as (x) = 2 . Case 1 then says tha t 0-0 ctv = e ft -r = e -, t. of
cr
will be an integrating factor . Multiplying both sides of the differential equation by µ(x) = e ' yields (3e 3-'y + xe 2') dx + (e''-') dy = 0
which is exact because M,. = 3e 3 ' = N,.
38
CLIFFS QUICK REVIE W
FIRST-ORDER EQUATIONS
Now, since
Max=
(3e 3 5' +xe'`)ax=e3`y +axe-' – e-`
and Nay =
f e3`. ay = e; 'y
(with the "constant" of integration suppressed in each calculation) , the general solution of the differential equation i s e 3`y+xe-
-e- = c
The value of the constant c is now determined by applying the initia l condition y(0) = 1 :
[e3`y + xe ;
– le?` "v =u,,=~ = C
=
c
Thus, the particular solution i s e3xy+xe' –
=
which can be expressed explicitly a s 3e_3' .+e-(1–2x ) Y
4
Example 10: Given that the nonexact differential equatio n
(5xy2 –2y)dx+(3xy–x)dy= 0 has an integrating factor of the form µ(x, y) = for some positive integers a and h, find the general solution of the equation . is a n Since there exist positive integers a and h such that integrating factor, multiplying the differential equation through b y
DIFFERENTIAL EQUATIONS
39
FIRST-ORDER EQUATIONS
this expression must yield an exact equation . That is , (5.x"+ ly b+2 - 2x"yh+ 1) dx + (3x"+ .?yh+ 1 - x"+ lyh) dy µM=M
0
(* )
µN= N
is exact for some a and b . Exactness of this equation mean s M, . = N,. 5(b + 2)x"+'y''+' – 2(b + 1)x"y'' = 3(a + 2)x"+'y''+' – (a + 1)x"y h By equating like terms in this last equation, it must be the case tha t 5(b+2)=3(a+2)
and
2(b+ 1)=a+ 1
The simultaneous solution of these equations is a = 3 and b = 1 . Thus the integrating factor x"y'' is x 3y, and the exact equatio n .Mdx+ Ndy=Dread s (5x 4y' - 2x 3y') dx + (3x5y' - x 4y) dy = 0 Now, since (5x 4y 3 – 2x3y2)ax=x5y3 –x4y 2
5 Max= and
5
Nay
=
5
(3x 5y' – x4y) ay = x 5y 3 – , x4y ,
(ignoring the "constant" of integration in each case), the genera l solution of the differential equation (*) and hence the original differential equation is clearly x'y' – ;x 4y , = c
40
■
CLIFFS QUICK REVIEW
FIRST-ORDER EQUATIONS
Separable Equation s Simply put, a differential equation is said to be separable if th e variables can be separated . That is, a separable equation is one tha t can be written in the form F(y) dy =. G (x) dx
Once this is done, all that is needed to solve the equation is t o integrate both sides . The method for solving separable equation s can therefore be summarized as follows : Separate the variables and integrate .
Example 11 : Solve the equation 2y dy (x 2 + 1) dx . Since this equation is already expressed in "separated" form , just integrate : 2ydy=(x 2 +1)dx
f 2y dy = (x' + 1) dx y- - 'x 3 +x+c
■
Example 12 : Solve the equatio n x dx + secx siny dy = 0 This equation is separable, since the variables can be separated : sec x sin y dy = —x dx sin ydy = —xcosxdx The integral of the left-hand side of this last equation is simpl y fsinydy= —cos y
DIFFERENTIAL EQUATIONS
41
FIRST-ORDER EQUATION S
and the integral of the right-hand side is evaluated using integratio n by parts : —x cos x dx = — x cos x dx _ —[x sin x — sin xdx ] (xsinx+cosx+c ) The solution of the differential equation is therefor e cos y = x sin x + cosx + c
■
Example 13 : Solve the IVP dy x(ex- + 2 ) 6y 2 dx = y(0) = 1 The equation can be rewritten as follows : 6y 2 dy=x(e x- +2)dx Integrating both sides yield s
f 6y 2 dy = x(ex" + 2) dx 2Y 3 = ex' + x 2 + c Since the initial condition states that y = 1 at x = 0, the parameter c can be evaluated : [ 2y3 ]y=i = [4e
+x2+c]~._2=4+c
c= 4
2 The solution of the IVP is therefore 2)2 3 = 4e-t + x-- + 4 o r 4y 3
42
= e x' + 2x 2
+3
■
CLIFFS QUICK REVIEW
FIRST-ORDER EQUATIONS
Example 14: Find all solutions of the differential equatio n (x 2 -1)y 3 dx+x 2 dy=0 . Separating the variables and then integrating both sides give s x 2 dy = —(x' . — 1)y 3 dx dy 1—x 2 = x, dx Y
(t)
fy-3dy=f(x-2_ 1) dx —'2y
= — x -' — x — c
1 1 =— +x+ c x 2yAlthough the problem seems finished, there is another solution o f the given differential equation that is not described by the family ;y -2 = x-' + x + c . In the separation step marked (t), both side s were divided by y 3 . This operation prevented y = 0 from bein g derived as a solution (since division by zero is forbidden) . It just s o happens, however, that y = 0 is a solution of the given differentia l equation, as you can easily check (note : y = 0 dy = 0) . Thus, the complete solution of this equation must includ e 1 1 both the family 2y- , = + x + c and x the constant function y = 0 The lesson is clear : If both sides of a separable differential equation are divided b y some function f(y) (that is, a function of the dependen t variable) during the separation process, then a valid solutio n may be lost . As a final step, you must check whether the constan t function y = yo [where f(yo) = 0] is indeed a solution of the give n differential equation . If it is, and if the family of solutions foun d DIFFERENTIAL EQUATIONS
43
FIRST-ORDE R EQUATIONS
by integrating both sides of the separated equation does no t include this constant function, then this additional solution mus t ■ be separately stated to complete the problem .
Example 15 : Solve the equation dy = (1 +
e-x)(Y2
–
Separating the variables give s dy
y 2 -1=(1+e-x)dx
(t )
(To achieve this separated form, note that both sides of the origina l equation were divided by y 2 – 1 . Thus, the constant functions y = 1 and y = -1 may be lost as possible solutions ; this will have to be checked later .) Integrating both sides of the separated equatio n yields
f
f
dy z— 1
+ y _ 1 ) dY =
f
(1 +
f+
e-A)
cLr
e -x )dy
– 1nfy+1~+1nIy– 1I =x – e -x + c In
y– 1 =2(x–e-x+c ) y+ 1
Now, both constant functions y = 1 and y = -1 are solutions of the original differential equation (as you can check by simply notin g that y = ± 1 dy/dx = 0), and neither is described by the famil y above . Thus the complete set of solutions of the given differentia l
44
CLIFFS QUICK REVIEW
FIRST-ORDER EQUATIONS
equation includes both the family In
y -1 y+ 1
=2(x—e ' + c )
and the constant functions y = ± 1
Example 16: Solve the differential equation xydx —
■
2 + 1)dy=0 .
Separate the variables , (x' + 1)dy =xydx .dy
xdx
y x- + 1
(t )
and integrate both sides : dy
x dx
y
x' + 1
Inly~ =
In(x'+1)+c '
2lnIy1 =ln(x 2 + 1)+In c In (y ' ) = In [c(x 2 + 1) ] y'=c(x 2 +1 )
Note that in the separation step (t), both sides were divided by y ; thus, the solution y = 0 may have been lost . Direct substitution o f the constant function y = 0 into the original differential equatio n shows that it is indeed a solution . However, the family y 2 = c(x 2 + 1 ) already includes the function y = 0 (take c = 0), so it need not b e separately mentioned . ■
DIFFERENTIAL EQUATIONS
45
FIRST-ORDE R EQUATIONS
Example 17 : Find the curve r = r(0) in polar coordinates that solve s the IVP dr d8 –r
tan 0
r(Tr) = 2
The given equation is separable, since it can be expressed in th e separated form dr — = –tan 8dO (t ) r
Now integrate both sides : dr
–sin Bd 9 COS
B
In Irl = In cos 91 + c ' 1nIrl =lnlcosB1 +lnlc ' InIrl = In c cos 0 1 r=ccos B
Since the solution curve is to pass through the point with pola r coordinates (r, 0) _ (2, Tr) , 2=ccosTr
c = -2
The solution of the IVP is therefor e r = -2 cos 0
This is a circle of diameter 2, tangent to they axis at the origin ; se e Figure 5 . Note : In the separation step (t), both sides were divide d by r (which is the dependent variable here) . However, even thoug h r = 0 formally satisfies the differential equation, it clearly does no t satisfy the initial condition r(-rr) = 2 . ■
CLIFFS QUICK REVIE W
46
FIRST-ORDER EQUATION S
(r,9)=(2,n )
■ Figure 5
■
Homogeneous Equation s A function equation
is said to be homogeneous of degree n if th e
f(x, y)
f(2x , zy) =
z`f(x , y )
holds for all x, y, and z (for which both sides are defined) .
Example 18: The function degree 2, sinc e f(
zx
, zy )
=
f(x, y) = x 2
+
y2
is homogeneous of
(zx ) 2 + (zy ) 2 = z2(x2 + y2) = z 2f(x, y )
•
Example 19 : The function f(x, y) = /x8 — 3x 2y h is homogeneous o f degree 4, since f(zv ,
zy ) = V ( ) 8 - 3 ( ) 2 (zy ) 6 = Vz 8 W - 3x2y6 ) =
DIFFERENTIAL EQUATIONS
— 3x 2y 6 = z 4f(x, y)
■
47
FIRST-ORDER EQUATIONS
Example 20 : The function f (x, y) = 2x + y is homogeneous of degree 1, sinc e f(zx , zy ) =2 (zx) + (zy ) =z 2x + y) = z' f (x, y)
■
Example 21 : The function f (x, y) = x 3 — y 2 is not homogeneous, since z ,y 2 f (zx, 2y ) _ (zx )3 — (z') 2 = z3x3 —
which does not equal z' f (x, y) for any n .
■
Example 22: The function f (x, y) = x 3 sin (y/x) is homogeneous o f degree 3, since zY f (zx, zy) = (zx) 3 sin ! = z-; x 3 sin x = z 3f (x, y)
■
A first-order differential equatio n M(x, y) dx + N(x, y) dy = 0 is said to be homogeneous if M(x, y) and N(x, y) are both homogeneous functions of the same degree .
Example 23 : The differential equatio n (x 2 — y 2 ) dx + xy dy = 0 is homogeneous because both M(x, y) = x 2 — y 2 and N(x, y) = xy are ■ homogeneous functions of the same degree (namely, 2) .
CLIFFS QUICK REVIE W 48
FIRST-ORDER EQUATION S
The method for solving homogeneous equations follows from thi s fact : The substitution y = xv (and therefore dy = xdv + vdx) transforms a homogeneous equation into a separable one .
Example 24 : Solve the equation
(x' — y 2 )
dx + xy dy = 0 .
This equation is homogeneous, as observed in Example 23 . Thus to solve it, make the substitutionsy = XI) and dy = x dv + v dx : [x 2 — (xv) 2]
dx + [x(xv)](x dv + v dx) = 0
(x- - x 'zv--)
dx + x 3z, dz~ + x-z►- dx = 0 x-dx+xvdv= 0 dx + xv dv = 0
This final equation is now separable (which was the intention) . Proceeding with the solution, dx
v dv = - x
dx x
,,dv _ -1-x 2 =
—In
+c'
Therefore, the solution of the separable equation involving x and v can be written C
x
DIFFERENTIAL EQUATIONS
49
FIRST-ORDER EQUATIONS
To give the solution of the original differential equation (whic h involved the variables x and y), simply note that v _ yx
y = xv
Replacing v by y/x in the solution above gives the final result :
l (12 – (n
2
c x
y2
= 2x 2 In
c
x
This is the general solution of the original differential equation . ■
Example 25 : Solve the IVP 2(x + 2y) dx + (y – x) dy = 0 y(1) = 0 Since the function s M(x, y) = 2(x + 2y)
and
N(x, y) = y – x
are both homogeneous of degree 1, the differential equation i s homogeneous . The substitutions y = xv and dy = x dv + v dx transform the equation into 2(x + 2xv) dx +
(xv –
x)(x dz) + v dx) = 0
which simplifies as follows : 2x dx + 4xzvdx +x2 vdzv –x 2 dv +xzv 2 dx – xvdx = 0 (2x + 3xv + xz) 2 ) dx + x(2 + 3v
+ z v 2)
(2 + 3v +
(x2z) – x2)
dv = 0
dx + x2(v – 1) dv = 0
z, 2 )
dx + x(v - 1) dzv = 0
CLIFFS QUICK REVIE W 50
FIRST-ORDER EQUATION S
The equation is now separable . Separating the variables and inte grating gives x(v — 1)dv= —(2+3v +v 2 )dx v — 1 ) + 3v + 2 dv
dx x
(I )
The integral of the left-hand side is evaluated after performing a partial fraction decomposition : v— 1 v 2 +3v+2 Therefore,
f
3 -2 1 + v+ 2 v+
v— 1 (v+ 1)(v+2)
f (
v— 1 v2+3v+2dv
)
-2 3 dz v v + 1 +v+2
= -2lnJz~+ 11 +31niv+2 1 =ln(v+ 1) - 22 (v+2)3 1 The right-hand side of (t) immediately integrates to cLr x
In+c ' =1nIcx - ' l
Therefore, the solution to the separable differential equation (t) i s (v + 1) -2 (v + 2) 3 = cx - ' Now, replacing v by y/x give s y
—+ 1 x
DIFFERENTIAL EQUATIONS
-2
y
(— X
3
+
2 =cx - '
FIRST-ORDE R EQUATION S
as the general solution of the given differential equation . Applying the initial conditiony(1) = 0 determines the value of the constant c : I
0
1-z I0
)3
Thus, the particular solution of the IVP i s 2
y x
1)_
IX
-+-
2 _ 8C- I
which can be simplified to (2x + y )3 = 8(x + y )2 as you can check . Technical note : In the separation step (t), both sides wer e divided by (v + 1)(v + 2), and v = -1 and v = -2 were lost a s solutions . These need not be considered, however, because eve n though the equivalent functions y = –x and y = -2x do indee d satisfy the given differential equation, they are inconsistent with th e ■ initial condition .
Linear Equation s A first-order differential equation is said to be linear if it can b e expressed in the form y' + P(x)y = Q(x ) where P and Q are functions of x . The method for solving suc h equations is similar to the one used to solve nonexact equations . There, the nonexact equation was multiplied by an integratin g factor, which then made it easy to solve (because the equation became exact) .
CLIFFS QUICK REVIEW
FIRST-ORDER EQUATIONS
To solve a first-order linear equation, first rewrite it (if necessary) in the standard form above ; then multiply both sides by th e integrating factor µ4x) = e 11
d
The resulting equation,
dy +µPy
=
µQ
(*)
is then easy to solve, not because it's exact, but because th e left-hand side collapses :
dy + µPy = dy +ef P c!x Py µ ms
k
dy
d
+y(e5Pdx
dy
)
d ,u,
dx + Y dx
Therefore, equation (*) become s d
Iv (AY)
= /I Q
making it susceptible to an integration, which gives the solution : µY =
(µQ) dx
Do not memorize this equation for the solution ; memorize the step s needed to get there .
DIFFERENTIAL EQUATIONS
FIRST-ORDER EQUATIONS
Example 26 : Solve the differential equatio n y'+2xy= x
The equation is already expressed in standard form, with P(x ) 2x and Q(x) = x . Multiplying both sides by µ(x ) = ef"v =
e
f2 (i =
transforms the given differential equation int o ex' y' + 2xe-" - y = xe-' d dx (ex Y) = xe'
Notice how the left-hand side collapses into (µy) ' ; as shown above , this will always happen . Integrating both sides gives the solution : ex - y
= fxe U dx
ex)? =+ c y= ;+ce -~ -
■
Example 27 : Solve the IVP 1 y' + —y = sin x x Y(n) = 1
Note that the differential equation is already in standard form . Since P(x) = 1/x, the integrating factor i s µ(x ) = efP`x = eft'
54
iv)( =
e in _ x
CLIFFS QUICK REVIE W
FIRST-ORDER EQUATIONS
Multiplying both sides of the standard-form differential equation b y µ = x gives xy' +y =x sin x
(xy)' = x sin x Again, note how the left-hand side automatically collapses int o (µy)' . Integrating both sides yields the general solution : xy = fxsinx v xy = –x cosx + sin x + c Applying the initial condition y(Tr) = 1 determines the constant c : Tr . 1 = –Tr cos Tr+sinTr+c = c= 0 Thus the desired particular solution i s xy = –x cosx + sin x or, since x cannot equal zero (note the coefficient P(x) = 1/x in th e given differential equation) , y=
sin x – cosx x
■
Example 28 : Solve the linear differential equation dy x- = -2y cbc First, rewrite the equation in standard form : dy
DIFFERENTIAL EQUATIONS
2 + x~Y =O
(* )
FIRST-ORDE R EQUATION S
Since the integrating factor here i s µ(x ) =
e fPd
=
e f(2
)th =
e_ 2 /
multiply both sides of the standard-form equation (*) by dy
e -2/x
+
2 .x 2
µ
= e-2/x,
e -2/x y = 0
collapse the left-hand side, d
~ (e -2/x y) = 0
and integrate : e -?iay— c
Thus the general solution of the differential equation can b e expressed explicitly as y = ce 2/x-
n
Example 29 : Find the general solution of each of the followin g equations : (a)
dy
dy
(b)
4 —xy= 0 4
• — x y =x 4
Both equations are linear equations in standard form, wit h -4/x . Since
P(x)
f PdY = f - -x ctx = 4
- 4 In = In (x -4 )
the integrating factor will b e = ,u(x) = efPdx = eln (x-4) x -4 CLIFFS QUICK REVIEW
56
FIRST-ORDER EQUATIONS
for both equations . Multiplying through by ,u, = x - 4 yield s d a _.
d
dx
(x- 4y) = 0
for equation (a )
(x -4y)
for equation (b )
=1
Integrating each of these resulting equations gives the general solutions : x -4y = c
y = cx 4
x -4y = x + c y = cx 4 + x 5
for equation (a ) for equation (b)
■
Example 30: Sketch the integral curve of (1 +x2 )y' =x ( 1 —y ) which passes through the origin . The first step is to rewrite the differential equation in standar d form : (1+x2)y'+xy=x .
x
x
Y'+l+x2y1+x2
(* )
Sinc e
f
Pdx=
x 2dx=ln(1 +x 2 ) +x
the integrating factor i s µ(x)
2 = e fPdx = e(1/2) In (1+x 2 ) = e ln (1+x ) 1/' =
DIFFERENTIAL EQUATIONS
(1
+x2)1/ 2
FIRST-ORDER EQUATIONS
Multiplying both sides of the standard-form equation (*) by (1 + x 22 ) ' /2 gives x
(1+x2)'/2y'+
µ =
x (1 + x2 )' / 2
(1 + x 2 )' /2y
As usual, the left-hand side collapses into (µy)' , d
[(1 + x2 )" 2y]
=
x 1+x2 'i 2
and an integration gives the general solution : +x2)1/2y
=
f x(1 +x 2 )
(1 + x 2 1 /2y
=
(1 +
(1
2)
2 dx
x2)1/2 +
c
y = 1 +c(1 +x 2 - ' /2 2)
2
To find the particular curve of this family that passes through th e origin, substitute (x, y) = (0, 0) and evaluate the constant c : 0= 1+c(1+0) -1/ ma c = — 1 2
Therefore, the desired integral curve i s y
1
1 ~l1+x 2
which is sketched in Figure 6 .
Example 31 : An object moves along the x axis in such a way that it s position at time t > 0 is governed by the linear differential equatio n dx + ~ - t_)x = t dt(t If the object was at position x = 2 at time t = 1, where will it be a t time t = 3?
CLIFFS QUICK REVIE W
58
FIRST-ORDE R EQUATION S
■ Figure 6
■
Rather than having x as the independent variable and y as th e dependent one, in this problem t is the independent variable and x i s the dependent one . Thus, the solution will not be of the form "y = some function of x" but will instead be "x = some function of t . " The equation is in the standard form for a first-order linea r equation, with P = t — t - ' and Q = t' . Since
f Pdt= f(t - t
1 )dt = . t 2 -
ln t
the integrating factor i s µ( t ) =
efPdt = e(t 2 /2)-In t = e t ,l2e -In t = et 2 /2eln t -1 = t -1et 2 / 2
Multiplying both sides of the differential equation by this integrating factor transforms it into dx
t-"2/2 dt +
t_2)et _,
(1
,/2
x = tet -
As usual, the left-hand side automatically collapses , d dt [t
, - 'e t -1 x1 = te' 2
DIFFERENTIAL EQUATIONS
59
FIRST-ORDE R EQUATIONS
and an integration yields the general solution : t - 'e t2/2x =
f te` 2/2 dt
=e t2/ 2-+ c x = t(1 + ce
- 1 '/2 )
Now, since the condition "x = 2 at t = 1 " is given, this is actually a n IVP, and the constant c can be evaluated : 2 = 1 • (1 + ce - ' 12 )
=
ce -1/2
=
1
c = e' / 2
Thus, the position x of the object as a function of time t is given b y the equation x = 41
+ e(' -' 2 )/2]
and therefore, the position at time t = 3 i s x I,_ 3
which is approximately 3 .055 .
= 3(1 + e -4 ) ■
Bernoulli's Equatio n The differential equatio n y ' + P (x )y = Q(x) y " is known as Bernoulli's equation . If n = 0, Bernoulli's equatio n reduces immediately to the standard form first-order linear equation : y' + P(x)y = Q(x ) If n = 1, the equation can also be written as a linear equation : y' + P(x) y = Q(x )y y' + [P (x ) — Q(x )] y = 0 CLIFFS QUICK REVIE W
60
FIRST-ORDE R EQUATION S
However, if n is not 0 or 1, then Bernoulli's equation is not linear . Nevertheless, it can be transformed into a linear equation (an d therefore solved by the method of the preceding section) by firs t multiplying through by n y -
y
,
-n
y'
+
P(x)y' -n
=
Q(x)
and then introducing the substitutions =
w w'
y'
-n
( 1
—
n )Y
-ny
The equation above then become s 1 1—n
w' + P(x)w = Q(x )
which is linear in w (since n 1) . Example 32 : Solve the equatio n Y ' + xY = xY 3
Note that this fits the form of the Bernoulli equation with n 3 . Therefore, the first step in solving it is to multiply through by y - n = Y -3 : y _3y' +x,-2 =x
(* )
Now for the substitutions ; the equation s w=y l-3 =y - 2
transform (*) int o
or, in standard form,
DIFFERENTIAL EQUATIONS
61
FIRST-ORDER EQUATIONS
Notice that the substitutions were successful in transforming th e Bernoulli equation into a linear equation (just as they wer e designed to be) . To solve the resulting linear equation, first determine the integrating factor : = e 2 µ(x) = e 1 = el- a& -x
Multiplying (* *) through the µ = e x` yields -
e -2w ' – 2xe -x`w = – 2xe -x ` (e' 2w)' = – 2xe -x ` And an integration gives e -x`w = e -x`w =
7
– 2xe -x` d.x e _x2
+ c
w = 1 + cex2 The final step is simply to undo the substitution w = solution to the original differential equation is therefore 1 y2
= 1 +ce
x
z
y- 2 .
Th e
■
CLIFFS QUICK REVIE W 62
SECOND-ORDER EQUATION S
Second-order differential equations involve the second derivativ e of the unknown function (and, quite possibly, the first derivative a s well) but no derivatives of higher order. For nearly every second order equation encountered in practice, the general solution wil l contain two arbitrary constants, so a second-order IVP must includ e two initial conditions .
Linear Combinations and Linear Independenc e Given two functions y 1 (x) andy Z (x), any expression of the form C LYI + c 2Y2 where c l and c 2 are constants, is called a linear combination of y l and y 2 . For example, if y 1 = ex and y 2 = x2 , the n 5ex + 7x 2 ,
ex – 2x 2 ,
and
–ex + x2
are all particular linear combinations of y l and y 2 . So the idea of a linear combination of two functions is this : Multiply the functions b y whatever constants you wish ; then add the products .
Example 1 : Is y = 2x a linear combination of the functions y, = x and y 2 = x 2? Any expression that can be written in the for m c lx+c 2x2
is a linear combination of x and x2 . Since y = 2x fits this form by taking c 1 = 2 and c2 = 0, y = 2x is indeed a linear combination of x and x2 . ■
DIFFERENTIAL EQUATIONS
63
SECOND-ORDER EQUATION S
Example 2 : Consider the three functions y, = sin x, y 2 = cosx, an d = sin(x + 1) . Show that y 3 is a linear combination of y, and y 2 .
y3
The addition formula for the sine function say s sin(x + 1) = cos 1 sin x + sin 1 cos x Note that this fits the form of a linear combination of sin x and cos x , c 1 sinx+c,cos x by taking c l = cos 1 and c, = sin 1 .
■
Example 3: Can the function y = x 3 be written as a linea r combination of the functions y l = x and y, = x2 ? If the answer were yes, then there would be constants c l and c 2 such that the equation x 3 =cox+c,x2
(* )
holds true for all values of x . Lettingx = 1 in this equation give s 1 = c~ +
c2
and lettingx = -1 gives -1=-c l +c , Adding these last two equations gives 0 = 2c 2, so c, = 0 . And since c 2 = 0, c, must equal 1 . Thus, the general linear combination (* ) reduces to x3 = x which clearly does not hold for all values of x . Therefore, it is not possible to write y = x3 as a linear combination of y l = x an d Y2 = x2 . ■
64
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATIONS
One more . definition : Two functions y, and y, are said to be linearly independent if neither function is a constant multiple of the other . For example, the functions y i = x 3 and y2 = 5x3 are not linearly independent (they're linearly dependent), since y2 is clearly a constant multiple ofy, . Checking that two functions are dependen t is easy ; checking that they're independent takes a little more work .
Example 4 : Are the functions y 1 (x) = sin x and y2 (x) = cos x linearl y independent ? If they weren't, then y 1 would be a constant multiple ofy, ; that is , the equation sinx=ccos x would hold for some constant c and for all x . But substituting x = 7r/2, for example, yields the absurd statement 1 = 0 . Therefore, the above equation cannot be true : y 1 = sin x is not a constant multipl e of y 2 cos x ; thus, these functions are indeed linearly indepen dent . ■
Example 5 : Are the functions y l = et- and y, = x linearly independent ? If they weren't, then y 1 would be a constant multiple of y2 ; that is, the equation = cx
would hold for some constant c and for all x . But this cannot happen, since substituting x = 0, for example, yields the absurd statement 1 = 0 . Therefore, y, = ex is not a constant multiple ofy, = x ; these two functions are linearly independent . ■
DIFFERENTIAL EQUATIONS
65
SECOND-ORDE R EQUATIONS
Example 6 : Are the functions y l = xee' and y 2 independent ?
=
ex linearl y
A hasty conclusion might be to say no because y l is a multiple o f But y, is not a constant multiple of y2 , so these functions truly ar e independent . (You may find it instructive to prove they're independent by the same kind of argument used in the previous tw o examples .) ■
y2 .
Linear Equation s Recall that the order of a differential equation is the order of th e highest derivative appearing in the equation . Thus, a second-orde r differential equation is one that involves the second derivative o f the unknown function but no higher derivatives . A second-order linear differential equation is one that can b e written in the form a(x)y" + b(x)y ' + c(x)y = d(x)
where a (x) is not identically zero . [For if a (x) were identically zero , then the equation really wouldn't contain a second-derivative term , so it wouldn't be a second-order equation .] If a(x) 0, then bot h sides of the equation can be divided through by a(x) and th e resulting equation written in the form y" + p(x)y' + q(x)y = r(x) It is a fact that as long as the functions p, q, and r are continuous o n some interval, then the equation will indeed have a solution (on that interval), which will in general contain two arbitrary constants (a s you should expect for the general solution of a second-orde r differential equation) . What will this solution look like? There is n o explicit formula that will give the solution in all cases, only variou s methods that work depending on the properties of the coefficient CLIFFS QUICK REVIE W
66
SECOND-ORDE R EQUATIONS
functions p, q, and r . But there is something definitive and very important that can be said about second-order linear equations .
Homogeneous Equation s There are two definitions of the term "homogeneous differentia l equation ." One definition (already encountered) calls a first-orde r equation of the form M(x, y) dx + N(x, y) dy = 0
homogeneous if M and N are both homogeneous functions of th e same degree . The second definition and the one which you'll se e much more often states that a differential equation (of any order ) is homogeneous if once all the terms involving the unknow n function are collected together on one side of the equation, th e other side is identically zero . For example , y " — 2y' + y = 0
is homogeneous
y" — 2y' + y = x
is no t
but
For the rest of this book, the term " homogeneous" will refer only t o this latter definition . The nonhomogeneous equatio n a(x)y" + b(x)y' + c(x)y = d(x)
(* )
can be turned into a homogeneous one simply by replacing th e right-hand side by 0 : a(x)y" + b(x)y' + c(x)y = 0
(** )
Equation (* *) is called the homogeneous equation corresponding t o the nonhomogeneous equation (*) . There is an important connec -
DIFFERENTIAL EQUATIONS
67
SECOND-ORDER EQUATIONS
tion between the solution of a nonhomogeneous linear equatio n and the solution of its corresponding homogeneous equation . Th e two principal results of this section are given below :
Theorem A . If y, (x) and y,(x) are linearly independent solutions o f the linear homogeneous equation (* * ), then every solution is a linea r combination of y, and Y, . That is, the general solution of the linea r homogeneous equation is Y = c 1YI + c ?Y2
Theorem B . If y(x) is any particular solution of the linear nonhomogeneous equation (*), and if y1, (x) is the general solution of th e corresponding homogeneous equation, then the general solution o f the linear nonhomogeneous equation i s Y = Y,, + y
That is , general solution of linear _ general solution of correspondin g homogeneous equatio n nonhomogeneous equation — + particular solution of give n nonhomogeneous equatio n [Note : The general solution of the corresponding homogeneou s equation, which has been denoted here by y,,, is sometimes calle d the complementary function of the nonhomogeneous equation (* ) . ] Theorem A can be generalized to homogeneous linear equations o f any order, while Theorem B as written holds true for linea r equations of any order . You can see Theorem B illustrated for a first-order equation by looking back at Example 29 on page 56 . Theorems A and B are perhaps the most important theoretical fact s about linear differential equations definitely worth memorizing .
CLIFFS QUICK REVIEW
68
SECOND-ORDE R EQUATIONS
Example 7: The differential equatio n y" – 2y' +y= 0 is satisfied by the function s y i = ek
and y, = xe'
Verify that any linear combination of y l and y, is also a solution o f this equation . What is its general solution ? Every linear combination of y i = ev and y, = xee' looks like this : y = c sex + c 2xev
for some constants c 1 and c, . To verify that this satisfies th e differential equation, just substitute . If y = c,et + c-axe', the n y ' =c~eA +c,e t (x+1 ) y"=c,ex
+c2
(x+2 )
Substituting these expressions into the left-hand side of the give n differential equation give s y" – 2y ' + y = [c i ev + c,e' (x + 2) ] – 2 [c i e' + c,e' (x + 1)] + [c 1 e' + c xe' ] = c l (e' – 2e' + e') + c,[e'(x + 2) – 2e'(x + 1) + xe' ]
=0
3
Thus, any linear combination of y, = e'- and y 2 = xe' does indee d satisfy the differential equation . Now, since y, = e' and y2 = xev are linearly independent (see Example 6), Theorem A says that th e general solution of the equation i s y = c 1 e' + c,xe
DIFFERENTIAL EQUATIONS
■
69
SECOND-ORDE R EQUATION S
Example 8 : Verify that y = 4x — 5 satisfies the equatio n y " +5y ' + 4y = 16x . Then, given that y, = e_a and y 2 = e-4x are solutions of th e corresponding homogeneous equation, write the general solution of the given nonhomogeneous equation .
First, to verify that y = 4x — 5 is a particular solution of th e nonhomogeneous equation, just substitute . If y = 4x — 5, then y' = 4 and y" = 0, so the left-hand side of the equation become s y"+5y' + 4y = 0 + 5(4) + 4(4x — 5 ) 3
= 16x
Now, since the functions y, = e_ x and y 2 = e-4x are linearly independent (because neither is a constant multiple of the other) , Theorem A says that the general solution of the correspondin g homogeneous equation i s yh
= c,e - .r
+
c2e
-4'
Theorem B then say s y=c,e -x
+c 2e -4v
+4x
— 5
is the general solution of the given nonhomogeneous equation .
■
Example 9: Verify that both y, = sin x and y 2 = cos x satisfy th e homogeneous differential equation y" + y = 0 . What then is th e general solution of the nonhomogeneous equation y" + y = x? If)), = sin x, then y ;' = —sin x, soy;' + y, does indeed equal zero . Similarly, if y2 = cos x, then y' = —cos x, so y + y2 is also zero, as desired . Since y, = sin x and y2 = cos x are linearly independent (se e
CLIFFS QUICK REVIE W 70
SECOND-ORDER EQUATIONS
Example 4), Theorem A says that the general solution of th e homogeneous equation y" + y = 0 i s yh = C 1
Sln x
+ C2
cos X
Now, to solve the given nonhomogeneous equation, all that i s needed is any particular solution . By inspection, you can see tha t y = x satisfies y" + y = x . Therefore, according to Theorem B, th e general solution of this nonhomogeneous equation i s y= c l sin x+ c 2 cos x+ x
■
Homogeneous Linear Equations with Constant Coefficient s The general second-order homogeneous linear differential equation has the form a(x)y"
+ b(x)y' + c(x)y
=0
[a(x) $ 0 ]
If a(x), b(x), and c(x) are actually constants, a(x) = a 0, b(x) = b , c(x) = c, then the equation becomes simpl y ay" + by' + cy = 0 This is the general second-order homogeneous linear equation wit h constant coefficients . Theorem A above says that the general solution of this equatio n is the general linear combination of any two linearly independen t solutions . So how are these two linearly independent solution s found? The following example will illustrate the fundamental idea .
DIFFERENTIAL EQUATIONS
SECOND-ORDE R EQUATIONS
Example 10: Solve the differential equationy " — y ' — 2y = 0 . The trick is to substitute y = e mx (m a constant) into th e equation ; you will see shortly why this approach works . If y = e"-' , then y' = me"'x and y " = m 2e"'-, so the differential equation become s m 2emx — me' — 2e"'x = 0 The term emx can be factored out and immediately canceled (sinc e e mx never equals zero) : emx(m2—m — 2) = 0 m 2 —m—2= 0 This quadratic polynomial equation can be solved by factoring : m 2 —m—2= 0 (m + 1)(m — 2) = 0 m=-1, 2 Now, recall that the solution began by writing y = e mu . Since th e values of m have now been found (m = -1, m = 2), bot h y = e -x
and y = e
are solutions . Since these functions are linearly independen t (neither is a constant multiple of the other), Theorem A says tha t the general linear combinatio n y
= c ie -x
+ c2e 2x
is the general solution of the differential equation .
■
Note carefully that the solution of the homogeneous differentia l equation ay" + by' + cy = 0
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATIONS
depends entirely on the roots of the auxiliary polynomial equatio n that results from substituting y = e' and then canceling out the e'm term . Once the roots of this auxiliary polynomial equation ar e found, you can immediately write down the general solution of th e given differential equation . Also note that a second-order linea r homogeneous differential equation with constant coefficients wil l always give rise to a second-degree auxiliary polynomial equation , that is, to a quadratic polynomial equation . The roots of any quadratic equatio n (a 0 )
am t +bm+c=0
are given by the famous quadratic formul a m=
–b ± \/b 2 –4a c 2a
The quantity under the square root sign, b 2 – 4ac, is called th e discriminant of the equation, and its sign determines the nature o f the roots . There are exactly three cases to consider . Case 1 : The discriminant is positive . In this case, the roots are real and distinct . If the two roots ar e denoted m 1 and m 2 , then the general solution of the differentia l equation is y= C
1 em
1 x+
c ge
m
2x
Case 2: The discriminant is zero . In this case, the roots are real and identical ; that is, th e polynomial equation has a double (repeated) root . If this doubl e root is denoted simply by m, then the general solution of th e differential equation is y=c l e'
DIFFERENTIAL EQUATIONS
+c 2xe"
SECOND-ORDE R EQUATIONS
Case 3: The discriminant is negative .
In this case, the roots are distinct conjugate complex numbers , r ± si . The general solution of the differential equation is the n y =e'=` (c~ cossx+c2sinsx )
So here's the process : Given a second-order homogeneou s linear differential equation with constant coefficients (a 0) , ay" + by' +cy= 0
immediately write down the corresponding auxiliary quadrati c polynomial equation am t +bm+c= 0
(found by simply replacingy" by m2, y' by m, and y by 1) . Determin e the roots of this quadratic equation, and then, depending o n whether the roots fall into Case 1, Case 2, or Case 3, write th e general solution of the differential equation according to the for m given for that Case .
Example 11 : Solve the differential equationy" + 3y' — 10y = 0 . The auxiliary polynomial equation i s m 2 +3m—10= 0 whose roots are real and distinct : m 2 + 3m — 10 = 0 (m + 5)(m — 2) = 0 m=—5, 2
This problem falls into Case 1, so the general solution of th e differential equation is y = c 1e -
74
5~
. +
c 2e2z-
■
CLIFFS QUICK REVIE W
SECOND-ORDER EQUATION S
Example 12 : Give the general solution of the differential equatio n y" —2y ' +y=0 . The auxiliary polynomial equation i s m 2 — 2m + 1 = 0 which has a double root : 2m+1= 0 (m 1) 2 = 0 m=
1
This problem falls into Case 2, so the general solution of th e differential equation is y = c+ c-ixe as you verified earlier in Example 7 .
■
Example 13 : Solve the differential equation y" + 6y ' + 25y = 0 . The auxiliary quadratic equation i s 2 +6m+25= 0
m22
which has distinct conjugate complex roots : m
-6± J6 2 —4 . 1 . 25 -6±~—64 2 .1 = 2
-6±8 i 2 = -3 ± 4i
This problem falls into Case 3, so the general solution of th e differential equation i s y =e -3x (c 1 cos 4x+c,sin4x )
DIFFERENTIAL EQUATIONS
SECOND-ORDER EQUATION S
Example 14 : Solve the IV P
First, rewrite the differential equation in standard form : 2y"+y' –y= 0
Next, form the auxiliary polynomial equatio n 2m22+m – 1 = 0
and determine its roots : 2m 2 + m – 1 = 0 (2m – 1)(m + 1) = 0 m= ;,– 1
Since the roots are real and distinct, this problem falls into Case 1 , and the general solution of the differential equation is therefor e y
=
c1e 2 +
c2e-a
(* )
All that remains is to use the two given initial conditions to determine the values of the constants c 1 and c2 : [c1ev/2 + c,e"] x= o = 1 =
[c1e2
– c
These two equations for c 1 and to yield 2c1
2e -A ID
c2
=o = 2
c1 + 2 – c2
1 =
2
can be solved by first adding them
=3c 1
= 2
then substituting c 1 = 2 back into either equation to find c, = -1 .
76
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATIONS
From (*), the solution of the given IVP is therefor e y=2e`/?_e
■
The Method of Undetermined Coefficient s In order to give the complete solution of a nonhomogeneous linea r differential equation, Theorem B says that a particular solutio n must be added to the general solution of the correspondin g homogeneous equation . Now that the method for obtaining th e general solution of a homogeneous equation (with constant coefficients) has been discussed, it is time to turn to the problem o f determining a particular solution of the original, nonhomogeneou s equation . If the nonhomogeneous term d(x) in the general second-orde r nonhomogeneous differential equatio n a(x)y" + b(x)y' + c(x)y = d(x)
(* )
is of a certain special type, then the method of undetermine d coefficients can be used to obtain a particular solution . The special functions that can be handled by this method are those that have a finite family of derivatives, that is, functions with the property tha t all their derivatives can be written in terms of just a finite number o f other functions. For example, consider the function d = sin x . Its derivatives ar e d' = cos x,
d" = —sin x,
d"' = —cos x,
d o "') = sin x ,
and the cycle repeats . Notice that all derivatives of d can be writte n in terms of a finite number of functions. [In this case, they are sin x and cosx, and the set {sinx, cosx} is called the family (of derivatives ) of d = sin x .] This is the criterion that describes those nonhomogeneous terms d(x) that make equation (*) susceptible to the metho d of undetermined coefficients : d must have a finite family. DIFFERENTIAL EQUATIONS
SECOND-ORDER EQUATION S
Here's an example of a function that does not have a finit e family of derivatives : d = tan x. Its first four derivatives ar e d'
= sec'' x,
d"' = 2 d O)
d" = 2
sec' x tan x ,
sec 4 x + 4 sec'x tan '- x ,
= 16 sec 4 x tan x + 8 sec' x tan' x
Notice that the nth derivative (n 1) contains a term involvin g tan''-' x, so as higher and higher derivatives are taken, each one wil l contain a higher and higher power of tan x, so there is no way that al l derivatives can be written in terms of a finite number of functions . The method of undetermined coefficients could not be applied i f the nonhomogeneous term in (*) were d = tan x . So just what are the functions d(x) whose derivative families are finite? See Table 3 below . Table 3 NONZERO FUNCTIONS WITH A FINITE FAMILY OF DERIVATIVE S Function
Family
k (k : a constant ) x" (n : a nonnegative integer ) ek `
Ixn ,xn-1 ,
,x, 1 1
Ieki }
sin kx
{sin kx, cos kx }
cos kx
{sinkx,coskx }
a finite product of any of the preceding types
all products of th e individual family members
CLIFFS QUICK REVIEW
78
SECOND-ORDE R EQUATIONS
Example 15 : If d(x) = 5x 2, then its family is {x2, x, 1 } . Note that any numerical coefficients (such as the 5 in this case) are ignored whe n determining a function's family . ■
Example 16 : Since the function d(x) = x sin 2x is the product of x and sin 2x, the family of d(x) would consist of all products of th e family members of the functions x and sin 2x . That is , family of x sin 2x = {x, 1} • {sin 2x, cos 2x } = {x sin 2x, x cos 2x, sin 2x, cos 2x}
■
Linear combinations of n functions . A linear combination of tw o functions y, and y2 was defined to be any expression of the form c j y i + c,y
2
where c l and c2 are constants . In general, a linear combination of n functions y,, y,, . . . , y,, is any expression of the for m d y +c I
1
2 y2
+
. .
.
+c„ y„
where c 1 , . . . , c,, are constants . Using this terminology, the nonhomogeneous terms d(x) which the method of undetermined coefficients is designed to handle are those for which every derivative ca n be written as a linear combination of the members of a given finit e family of functions . The central idea of the method of undetermined coefficients i s this : Form the most general linear combination of the functions i n the family of the nonhomogeneous term d(x), substitute thi s expression into the given nonhomogeneous differential equation , and solve for the coefficients of the linear combination .
DIFFERENTIAL EQUATIONS
79
SECOND-ORDER EQUATION S
Example 17 : Find a particular solution (and the complete solution ) of the differential equatio n y" + 3y' — 10y=5x2 As noted in Example 15, the family of d = 5x2 is {x 2 , x, 11 ; therefore, the. most general linear combination of the functions i n the family is y = Ax2 + Bx + C (where A, B, and C are th e undetermined coefficients) . Substituting this into the given differential equation give s (2A)+3(2Ax+B)— 10(Ax 2
+Bx+C) = 5x2
Now, combining like terms yield s (— 10A )x2
+
(6A — 1OB)x+(2A+3B— 10C)=5x2
In order for this last equation to be an identity, the coefficients o f like powers of x on both sides of the equation must be equated . Tha t is, A, B, and C must be chosen so tha t -10A = 5 6A — lOB= 0 2A +3B- IOC= 0 The first equation immediately gives A = - 4. Substituting this int o the second equation gives B = - 4 and finally, substituting both o f these values into the last equation yields C = Therefore, a particular solution of the given differential equation i s y =
-4x2 -
1(►
x
- ly
tO(►
According to Theorem B, then, combining thisy with the result o f Example 11 gives the complete solution of the nonhomogeneou s differential equation : y = c 1 e- 5-' + c2e 2x — 4- x2 — T}i x — ~<<~. ■
CLIFFS QUICK REVIEW
80
SECOND-ORDER EQUATION S
Example 18 : Find a particular solution (and the complete solution ) of the differential equatio n y" — 2y' + y = sin x Since the family of d = sin x is {sin x, cos x), the most genera l linear combination of the functions in the family is = A sin x + B cos x (where A and B are the undetermined coefficients) . Substituting this into the given differential equation give s (—A sin x — B cos x) — 2(A cos x — B sin x) + (A sin x + B cos x) = sin x Now, combining like terms and simplifying yield s (2B) sinx + (—2A) cosx = sin x In order for this last equation to be an identity, the coefficients A and B must be chosen so that 2B = 1
-2A = 0 These equations immediately imply A = 0 and B = A particula r solution of the given differential equation is therefor e y = ;cos x According to Theorem B, combining this y with the result o f Example 12 yields the complete solution of the given nonhomogeneous differential equation :y = c 1 ex + c2xe v + cosx . ■
DIFFERENTIAL EQUATIONS
81
SECOND-ORDER EQUATION S
Example 19 : Find a particular solution (and the complete solution ) of the differential equation y"+6y' +25y=8e -'x
Since the family of d = 8e-'x is just {e-'x }, the most general linear combination of the functions in the family is simply y = Ae - 'x (where A is the undetermined coefficient) . Substituting this into the given differential equation give s (49Ae -') + 6(– 7Ae - 'x ) + 25 (Ae - 'x) = 8e- 7x
Simplifying yields 32Ae-7x = 8e -7x
In order for this last equation to be an identity, the coefficient A must be chosen so that 32A = 8 which immediately gives A = A particular solution of the give n differential equation is therefor e Y = 4 e -7x
and then, according to Theorem B, combining y with the result o f Example 13 gives the complete solution of the nonhomogeneou s ■ differential equation : y = e_3x. (c 1 cos 4x + c 2 sin 4x) + e- 7-v.
Example 20 : Find the solution of the IV P y" – y' – 6y = – ex + 12x
y(0) = 1 y'(0) _ -2
CLIFFS QUICK REVIEW
82
SECOND-ORDER EQUATION S
The first step is to obtain the general solution of the correspond ing homogeneous equatio n y" — y' — 6y= 0 Since the auxiliary polynomial equation has distinct real roots , m 2 —m—6= 0
(m + 2)(m — 3) = 0 m = -2, 3 the general solution of the corresponding homogeneous equation i s Yh =
c 1 e-
+ c 2 e3x•
Now, since the nonhomogeneous term d(x) is a (finite) sum of functions from Table 3, the family of d(x) is the union of the familie s of the individual functions . That is, since the family of —e'" is {e'"} , and the family of 12x is (x, 11 , the family of —e`' + 12x = [the family of —e'' ) U [the family of 12x] = {e''} U {x, 1} = { e', x, 1 ) The most general linear combination of the functions in the famil y of d = —ex + 12x is therefore = Ae' + Bx + C (where A, B, and C are the undetermined coefficients) . Substituting this into the give n differential equation give s (Aex)—(Ae+B)—6(Ae+Bx+C)= —ex + 12x
Combining like terms and simplifying yield s (—bA)ex
+
(— 6B)x + (—B — 6C) =
+ 12x
DIFFERENTIAL EQUATIONS 83
SECOND-ORDE R EQUATIONS
In order for this last equation to be an identity, the coefficientsA, B , and C must be chosen so that - 6A=— 1 - 6B = 1 2 —B —
6C = 0
The first two equations immediately give A = 4 and B = - 2 , whereupon the third implies C = . A particular solution of th e given differential equation is therefor e Y = , e'—2x + According to Theorem B, then, combining this y with y,, gives th e complete solution of the nonhomogeneous differential equation : y = + c,e3r + -,,ea — 2x + . Now, to apply the initial conditions and evaluate the parameters c, and c, : y(0) = 1 = [c ) e -?~+ c,ev + y'(0) = -2
— 2x +
+ 3c,e 3x +
=1
c, + c, + = 1
— 2]X=O = - 2
-2c ) + 3c, — = - 2 Solving these last two equations yields c, = and c, = ~, . Therefore , the desired solution of the IVP i s y= (e' + 1)+yex +e')—2x
■
Now that the basic process of the method of undetermine d coefficients has been illustrated, it is time to mention that it isn' t always this straightforward . A problem arises if a member of a family of the nonhomogeneous term happens to be a solution of the corresponding homogeneous equation . In this case, that family must be modifie d before the general linear combination can be substituted into th e original nonhomogeneous differential equation to solve for th e undetermined coefficients . The specific modification procedure wil l be introduced through the following alteration of Example 20 . CLIFFS QUICK REVIE W
84
SECOND-ORDER EQUATION S
Example 21 : Find the complete solution of the differential equation Y"
Y' -6y= 10e
The general solution of the corresponding homogeneous equation was obtained in Example 20 : y1,
= c l e -'-~ + c,e k
Note carefully that the family {e3` } of the nonhomogeneous term d = 10e3; contains a solution of the corresponding homogeneou s equation (take c, = 0 and c, = 1 in the expression for yh ) . Th e "offending" family is modified as follows : Multiply each member of the family byx and try again . multiply eac h family membe r by x
{ e ~~ }
> {
Since the modified family no longer contains a solution of th e corresponding homogeneous equation, the method of undetermined coefficients can now proceed . (If xe 3- had been again a solution of the corresponding homogeneous equation, you woul d perform the modification procedure once more : Multiply each member of the family by x and try again .) Therefore, substituting = Axe = into the given nonhomogeneous differential equatio n yields (9Axe x + 6Ae x ) — (3Axe- + Ae 3~") — 6(Axe= ) = 10e = 5 Ae" = 10e = A= 2 This calculation implies that y = 2xe= is a particular solution of th e nonhomogeneous equation, so combining this with yh gives th e complete solution : y = c 1e _,x + c,e" + 2xe 3x
■
DIFFERENTIAL EQUATIONS
85
SECOND-ORDE R EQUATIONS
Example 22 : Find the complete solution of the differential equation y"–2y' =62–3e
2
First, obtain the general solution of the corresponding homogeneous equation y"–2y' = 0 Since the auxiliary polynomial equation has distinct real roots , m 2 – 2m = 0 m(m — 2) = 0 m=0, 2 the general solution of the corresponding homogeneous equation i s yh
= c i e~u+ c,eZ = c l + c,e '
The family for the 6x 2 term is {x2 , x, 11, and the family for the – 3ek term is simply {ex/2 } . This latter family does not contain a solution o f the corresponding homogeneous equation, but the family x, 1 } does (it contains the constant function 1, which matches yh whe n c l = 1 and c2 = 0) . This entire family (not just the "offending " member) must therefore be modified : "2
{x2,
multiply eac h family membe r by x
{x2,x, 1 }
> {x3 ,
x2,
x}
The family that will be used to construct the linear combination y i s now the union x l U frx/2} _ {x3 , x 2 , x, e- 2 } {x3, x2,
86
CLIFFS QUICK REVIE W
SECOND-ORDER EQUATIONS
This implies that y = Ax3 + Bx2 + Cx + Deg /2 (where A, B, C, and D are the undetermined coefficients) should be substituted into th e given nonhomogeneous differential equation . Doing so yield s (6Ax+2B+4De al ')–2(3Ax2
+2Bx+C+ 4Deg/2 ) =6x 22 –3e
/2
which after combining like terms read s ( – 6A)x2 +(6A – 4B)x+(2B–2C)+(–4De 2 )=6x2 –3e-12 In order for this last equation to be an identity, the coefficients A, B , C, and D must be chosen so tha t -6A = 6 6A – 4B = 0 2B – 2C = 0 4D
= -3
These equations determine the values of the coefficients : A = -1 , B = C = –z, and D = 4 . Therefore, a particular solution of th e given differential equation i s y= —x3 –4x--4x+4e4 According to Theorem B, then, combining this y with y', gives th e complete solution of the nonhomogeneous differential equation : y=c 1 + c 2e 2x — x 3 — 4x2 — 4x + 4e" . ■
DIFFERENTIAL EQUATIONS
87
SECOND-ORDE R EQUATIONS
Example 23 : Find the complete solution of the equatio n y" + 4y = x sin 2x + 8 First, obtain the general solution of the corresponding homogeneous equation y" + 4y = 0 Since the auxiliary polynomial equation has distinct conjugat e complex roots, m2 + 4 = 0 m = ±2i the general solution of the corresponding homogeneous equation i s yh
=c 1 cos2x+c2 sin 2x
Example 16 above showed that th e family of x sin 2x = {x, 1} • {sin 2x, cos 2x } = {x sin 2x, x cos 2x, sin 2x, cos 2x } Note that this family contains sin 2x and cos 2x, which are solution s of the corresponding homogeneous equation . Therefore, this entire family must be modified :
{x sin 2x, x cos 2x, sin 2x, cos 2x}
multiply eac h family membe r by x
{x 2 sin 2x, x 2 cos 2x, x sin 2x, x cos 2x } None of the members of this family are solutions of the correspond ing homogeneous equation, so the solution can now proceed a s usual . Since the family of the constant term is simply Ill, the famil y used to construct y is the union
CLIFFS QUICK REVIE W
88
SECOND-ORDE R EQUATIONS
{x 2
sin 2x, x 2 cos 2x, x sin 2x, x cos 2x} U {1 } = {x' sin 2x, x 2 cos 2x, x sin 2x, x cos 2x, 1 1
This implies tha t .-j; = Ax' sin 2x + Bx' cos 2x + Cx sin 2x + Dx cos 2x + E (where A, B, C, D, and E are the undetermined coefficients ) should be substituted into the given nonhomogeneous differentia l equation y" + 4y = x sin 2x + 8 . Doing so yield s [–8Bx+(2A -4D)]sin2x+ [8Ax + (2B + 4C)] cos 2x + 4E = x sin 2x + 8 In order for this last equation to be an identity, A, B, C, must be chosen so that
D,
and E
-8B = 1
2A–4D= 0 8A = 0 2B + 4C = 0
4E = 8 These equations determine the coefficients : A 0, B = –,, C = -116, D = 0, and E = 2 . Therefore, a particular solution of the give n differential equation i s y= --x'cos 2x+x sin2x+ 2 According to Theorem B, then, combining this y with y,, gives th e complete solution of the nonhomogeneous differential equation : ■ y = c 1 cos2x+c2sin2x–4-x'cos2x+ xsin2x+2 .
DIFFERENTIAL EQUATIONS
89
SECOND-ORDER EQUATIONS
Variation of Parameter s For the differential equatio n a(x)y " + b(x)y' + c(x)y = d(x)
(* )
the method of undetermined coefficients works only when th e coefficients a, b, and c are constants and the right-hand term d(x) i s of a special form . If these restrictions do not apply to a given nonhomogeneous linear differential equation, then a more powerful method of determining a particular solution is needed : th e method known as variation of parameters . The first step is to obtain the general solution of the correspond ing homogeneous equation, which will have the for m Yh =
C
IY1
+ c 2Y 2
where y, and Y 2 are known functions . The next step is to vary th e parameters ; that is, to replace the constants c, and c, by (as ye t unknown) functions 7' 1(x) and v 2(x) to obtain the form of a particular solution y of the given nonhomogeneous equation : Y
= z ► , (x )Y
1+
v 2(x )Y ,
The goal is to determine these functions v and v 2 . Then, since the functions y, and Y2 are already known, the expression above for y yields a particular solution of the nonhomogeneous equation . Combining y with y,, then gives the general solution of the non homogeneous differential equation, as guaranteed by Theorem B . Since there are two unknowns to be determined, v, and v 2, tw o equations or conditions are required to obtain a solution . One o f these conditions will naturally be satisfying the given differentia l equation . But another condition will be imposed first. Sincey will b e substituted into equation (*), its derivatives must be evaluated . Th e first derivative of y- i s Y'
90
=
7 '1Yi + 7 'iYi +
+ 7 ''L;Y 2 CLIFFS QUICK REVIEW
SECOND-ORDER EQUATION S
Now, to simplify the rest of the process and to produce the firs t condition on v l and v 2 set +
V zY2
=
0
This will always be the first condition in determining v 1 and v, ; th e second condition will be the satisfaction of the given differentia l equation (*) .
Example 24: Give the general solution of the differential equation y" + y = tan x. Since the nonhomogeneous right-hand term, d = tan x, is not of the special form that the method of undetermined coefficients can handle, variation of parameters is required . The first step is t o obtain the general solution of the corresponding homogeneou s equation, y" + y = O . The auxiliary polynomial equation i s m2 + 1 = 0 whose roots are the distinct conjugate complex numbers m = ±i = 0 ± li . The general solution of the homogeneous equation is therefore y = c I sinx+c,cos x
Now, vary the parameters c l and c, to obtai n y = v sin x +
v2
cos x
Differentiation yield s = v, cos x + v I sin x – v, sin x + v ; cos x Next, remember the first condition to be imposed on v l and v ;Y 1
DIFFERENTIAL EQUATIONS
v2 :
+ v;y, = 0
91
SECOND-ORDER EQUATIONS
that is, I) ; sin x + v; cos x = 0
(1 )
This reduces the expression for y' t o y ' = v1 cos x — II2 sin x
so, then,
= — v1 sin x + vj cos x — v2
cos
x—
z)4
sin x
Substitution into the given nonhomogeneous equation y" + y = tan x yield s (— v 1 sin x + v ; cos x — v2 cos x — v ; sin x) + (v 1 sin x + v2 cosx ) = tan x v;
cos
x — v2 sin x = tan x (2 )
Therefore, the two conditions on vl and v 2 are v ; sin x + v ; cos x = 0 v cos x — v sin x = tan x
(Z )
To solve these two equations for v ; and v, first multiply the firs t equation by sin x; then multiply the second equation by cos x : v sin 2 x + v ; cos x sin x = 0 v ; co s 2 x - v2 sin x cos x = sin x Adding these equations yield s v ; (sin e x + cos2 x) = sin x
= sin x
Substituting v ; = sin x back into equation (1) [or equation (2)] the n gives z~2 =
92
—tan x sin x
CLIFFS QUICK REVIEW
SECOND-ORDER EQUATIONS
Now, integrate to find v> and integration in each case) :
V2
(and ignore the constant o f
= v;
=f
sin xdx
= –cos x and v2
=
cos x
dx
cos t x – 1 cos x
dx
= (cos x – sec x) dx = sin x — In Isecx + tan x Therefore, a particular solution of the given nonhomogeneou s differential equation i s y= V I
sin x + V 2 cos x
– cos x
sin x + [sin x – In ~ sec x + tan x ] cos x
= –cosx In ( sec x + tan x l Combining this with the general solution of the correspondin g homogeneous equation gives the general solution of the nonhomogeneous equation : y = c, sin x +
DIFFERENTIAL EQUATIONS
C2
cos x — cos x In I sec x + tan x~
■
93
SECOND-ORDE R EQUATIONS
In general, when the method of variation of parameters i s applied to the second-order nonhomogeneous linear differentia l equation a(x)y" + b(x)y' + c(x)y = d(x ) with y = v i (x)y, + v 2 (x)y2 (where y h = c, y, + c2y2 is the genera l solution of the corresponding homogeneous equation), the tw o conditions on v, and V2 will always b e v iYi + v 2' y 2 = 0 a (x )[v i y i + v 1Yfl = d(x)
(1) (2)
So after obtaining the general solution of the correspondin g homogeneous equation (yh = + c,y,) and varying the parameters by writingy = v,y, + v,y,, go directly to equations (1) and (2 ) above and solve for v ; and v .
Example 25 : Give the general solution of the differential equation y" — 2y' + y = e~ In x Because of the In x term, the right-hand side is not one of th e special forms that the method of undetermined coefficients ca n handle ; variation of parameters is required . The first step, obtainin g the general solution of the corresponding homogeneous equation , y" — 2y' + y = 0, was done in Example 12 : yh
=c,y, +c,y, =c l ev +c,xe '
Varying the parameters gives the particular solutio n Y
=z 'iYi +zv 2y; =v i e' +v,xe4
and the system of equations (1) and (2) become s v>e l
+ 7~2xe'
0
v eA+v ;ex (x+ 1) =eIn x
94
(1 ) (2 )
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATIONS
Cancel out the common factor of ex in both equations ; then subtrac t the resulting equations to obtai n v4 = In x Substituting this back into either equation (1) or (2) determine s v ; = –x In x Now, integrate (by parts, in both these cases) to obtain v l and from v ; and v 2' :
f xlnx=
1 1 (lnx)~x-) — I
f (!)(!
V2
)
=1x 2 (1 – 2 Inx ) v2
= f v l; = f Inx dx =
(ln x)(x)
1 (x)xdx =xlnx— 1
Therefore, a particular solution i s Y
+
=
7 '2Y 2
= v > eA + v,xe x
=4x 2 (1 – 21nx)e~ +x(lnx – 1)xe x = ~'-x 2ex (21n x – 3 ) Consequently, the general solution of the given nonhomogeneou s equation is y=
+ c 2xex +
DIFFERENTIAL EQUATIONS
x 2ea (21n x – 3)
■
95
SECOND-ORDER EQUATION S
Example 26 : Give the general solution of the following differential equation, given that y i = x and y 2 = x are solutions of its corresponding homogeneous equation :
3
2y " –3xy' + 3y = 12x 4 Since the functionsy 1 = x and = x 3 are linearly independent , x
y2
Theorem A says that the general solution of the correspondin g homogeneous equation i s yh
= Y I +c,y,=c 1 x+c,x
3
Varying the parameters c l and c2 gives the form of a particular solution of the given nonhomogeneous equation : .-Y- =
v 1YI + 1'2Y 2 = v 1 x + v,x 3
where the functions v i and V 2 are as yet undetermined . The two conditions on v> and V2 which follow from the method of variation o f parameters are v ;y 1 + v,y, = 0
(1 )
a (x) [v i y i + v ; y ;] = d(x) (2 ) which in this case (y, = x, y 2 = x 3 , a = x 2, d = 12x4 ) becom e v ;x + v ;x 3 = 0 x 2 [v j + v4 3x22 ] = 12x 4 (2) Solving this system for v ; and v ; yields v ; = – 6x
2
and v ; = 6
from which follow v,
v
96
(–6x 2) dx = - 2x
= 7, ; =
2=
f
v;=
f
3
6dx=6x
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATION S
Therefore, the particular solution obtained i s Y = z 'IY1 = V lx = (—
+ z '2Y 2
+ v,x 3
2e)x
+
(6 )x '
Y = 4x 4 and the general solution of the given nonhomogeneous equation i s y=c l x+c,x 3 +4x 4
■
The Cauchy-Euler Equidimensional Equatio n The second-order homogeneous Cauchy-Euler equidimensiona l equation has the form ax 2y "
+
bxy' + cy = 0
where a, b, and c are constants (and a 0) . The quickest way t o solve this linear equation is to substitute y = x"t and solve for m . I f y = x m , then y' = mxm-1 and y" = m (m — 1)x m- 2 so substitution into the differential equation yield s am(m —
1)m
+ bmxm + cx"' = 0
x m [am(m — 1) + bm + c] = 0 am(m — 1) + bm + c = 0 am' + (b — a)m + c = 0 (* )
DIFFERENTIAL EQUATIONS
97
SECOND-ORDE R EQUATIONS
Just as in the case of solving second-order linear homogeneou s equations with constant coefficients (by first setting y = e lm'. an d then solving the resulting auxiliary quadratic equation for m), thi s process of solving the equidimensional equation also yields a n auxiliary quadratic polynomial equation . The question here is, ho w isy = x m to be interpreted to give two linearly independent solution s (and thus the general solution) in each of the three cases for th e roots of the resulting quadratic equation ? Case 1 : The roots of (*) are real and distinct. If the two roots are denoted m 1 and m,, then the general solution of the second-order homogeneous equidimensional differential equation in this case i s Y
=
C 1 xm'
+
C Zx" ,
Case 2: The roots of (*) are real and identical. If the double (repeated) root is denoted simply by m, then th e general solution (for x > 0) of the homogeneous equidimensiona l differential equation in this case i s y = clx m +c,x"'ln x
Case 3 : The roots of (*) are distinct conjugate complex numbers . If the roots are denoted r ± si, then the general solution of th e homogeneous equidimensional differential equation in this case is y = x'"[c 1 cos(s In x) + c, sin(s In x) ]
Example 27 : Give the general solution of the equidimensiona l equation x 2y"—3xy' + 3y = 0
98
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATIONS
Substitution ofy = x m results i n x 2 m(m — 1).x"' -2 — 3x•mxm- '
+
3x"' = 0
x"'[m(m— 1)—3m+3]= 0 m(m — 1) — 3m + 3 = 0 m 2 — 4m + 3 = 0 (m — 1)(m — 3) = 0 m = 1, 3 Since the roots of the resulting quadratic equation are real an d distinct (Case 1), both y = x' = x and y = x 3 are solutions an d linearly independent (see Example 17), and the general solution o f this homogeneous equation i s y =cox+c,x 3 (Compare this result with the statement of Example 26 .)
■
Example 28 : For the following equidimensional equation, give th e general solution which is valid in the domain x > 0 : x 2y " —3xy ' + 4y = 0 Substitution ofy = x"' results i n x2 . m (m — 1)x"n — 3x • mx,m — 1 + 4x "' = 0 x"'[m(m— 1) — 3m + 4] = 0 m(m — 1) — 3m + 4 = 0 m 2 —4m+4= 0 (m—2) 2 = 0 m = 2, 2
DIFFERENTIAL EQUATIONS
99
SECOND-ORDE R EQUATIONS
Since the roots of the resulting quadratic equation are real an d identical (Case 2), both y = x 2 and y = x 22 In x are (linearl y independent) solutions, so the general solution (valid for x > 0) o f this homogeneous equation i s y=c l x '-+c,x 2 ln x
■
If the general solution of a nonhomogeneous equidimensional equation is desired, first use the method above to obtain the general solution of the corresponding homogeneous equation ; then apply variation of parameters (see Example 26) .
Reduction of Orde r Some second-order equations can be reduced to first-order equations, rendering them susceptible to the simple methods of solvin g equations of the first order . Three particular types of such secon d order equations will be discussed in this section . Type 1 : Second-order equations with the dependent variabl e missin g Type 2 : Second-order nonlinear equations with the independen t variable missing Type 3 : Second-order homogeneous linear equations where on e (nonzero) solution is know n
Type 1 : Second-order equations with the dependent variable missing. Examples of such equations includ e y" + y' = x
and xy" — 2y' = 12x 2
The defining characteristic is this : The dependent variable, y, doe s not explicitly appear in the equation . This type of second-orde r
100
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATIONS
equation is easily reduced to a first-order equation by the transformation = w
Y
This substitution obviously implies y" = w', and the origina l equation becomes a first-order equation for w . Solve for th e function w ; then integrate it to recovery .
Example 29: Solve the differential equation y" + y' = x . Since the dependent variable y .is missing, let y' = w and y" = w' . These substitutions transform the given second-order equation int o the first-order equation w'+w= x
which is in standard form . Applying the method for solving suc h equations, the integrating factor is first determined , µ=e fPd'- = e f ` x = e x and then used to multiply both sides of the equation, yieldin g + exw = xex d dx (e
xw) = xex e xw
f xex. dx = xe x — f ex dx
=
= xe x — (ea + c l )
Therefore, w=x — 1 —c 1 e
DIFFERENTIAL EQUATIONS
101
SECOND-ORDER EON PONS
Now, to give the solution y of the original second-order equation , integrate : y' = w may=
w=
(x - 1 —c l e -x )dx
This gives y=2x 2
—x+c l e -x +c 2
Referring to Theorem B, note that this solution implies that y = c l e -x + c 2 is the general solution of the corresponding homogeneou s equation and that y = 2x 2 – x is a particular solution of th e nonhomogeneous equation . (This particular differential equatio n could also have been solved by applying the method for solvin g second-order linear equations with constant coefficients .) ■
Example 30 : Solve the differential equatio n xy" – 2y' = 10x 4
Again, the dependent variable y is missing from this second order equation, so its order will be reduced by making th e substitutions y' = w and y" = w' : xw' — 2w = 10x 4
which can be written in standard form 2 w' -- w= lox = x The integrating factor here i s
µ
= efPdx = ef(-2/x)dx = e-21nx _ eln(x - `' ) = x - 2
which is used to multiply both sides of the equation, yieldin g
102
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATIONS
x -2w' – 2x -3w = 10x d dx (x
-2w) = 1Qx
x-2w=
f 10xdx
x - 2w =5x 2 +c ; w=5x 4 +c ;x Integrating w gives y : y' =w 'y= =
w +
-(5x4
= x5 +
c ;x 2 ) dx
cx 3
+c 2
Letting c l = 6, , the general solution can be writte n y=x5 +c lx 3 +c 2
n
Example 31 : Sketch the solution of the IV P y"+ (Y ' ) 2= 0 y(0) = 1 y' (0) = 1 Although this equation is nonlinear [because of the term (y') 2 ; neither y nor any of its derivatives are allowed to be raised to an y power (other than 1) in a linear equation], the substitutions y ' = w and y" = w' will still reduce this to a first-order equation, since th e variable y does not explicitly appear . The differential equation i s transformed into w' + w 2 = 0
DIFFERENTIAL EQUATIONS 103
SECOND-ORDER EQUATIONS
which is separable : dw
—w -
dw = dx wdw = d — x w1 — =x+c i w w =
1 x + c1
Since y' = w, integrating gives y
w
1 dx= ln(x+c I ) + c , x+c
Now apply the initial conditions to determine the constants c 1 an d c2: y(0) = 1 y' ( 0) = 1
[ln(x + c i ) + c,]a.
= 1 = In c 1 + c, = 1
[(x + c l) -' ]X_(► = 1 = c i = 1
Because c, = 1, the first condition then implies c_, = 1 also . Thus th e solution of this IVP (at least forx > -1) i s y=ln(x+ 1) + 1 whose graph is shown in Figure 7 .
104
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATIONS
■ Figure 7
■
Type 2 : Second-order nonlinear equations with the independen t variable missing . Here's an example of such an equation : y 2y" - (y' ) 3 = 0
The defining characteristic is this : The independent variable, x, doe s not explicitly appear in the equation . The method for reducing the order of these second-orde r equations begins with the same substitution as for Type 1 equations , namely, replacingy' by w . But instead of simply writingy" as w', th e trick here is to express y" in terms of a first derivative with respect to y . This is accomplished using the chain rule : d dy
y"
dz. dx
dw =
dx
dw dy =
dydx
dw = dy Y
Therefore, dw Y "=w d Y
This substitution, along with y' = w, will reduce a Type 2 equatio n to a first-order equation for w . Once w is determined, integrate to find y .
DIFFERENTIAL EQUATIONS
105
SECOND-ORDER EQUATIONS
Example 32 : Solve the differential equatio n y 22y"_ ( y' ) 3= 0
The substitutions y' = w and y" = w(dw/dy) transform this second-order equation for y into the following first-order equatio n for w : y2
dw w d –w'= 0 y
~dw
wy- d
y
– w-= 0
Therefore, w = 0 or
dw y2–w2=
dy
0
The statement w = 0 means y' = 0, and thus y = c is a solution fo r any constant c . The second statement is a separable equation, an d its solution proceeds as follows :
y
dw dy dw w2
= w2
dy – y 2
dw 1dy w2 –
y
1
1
W
106
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATIONS
Now, since w = dy/dx, this last result become s dx 1 dy =
+Cl
y
(y - ' + c,) dy = dx (y - ' + c j ) dy = dx
which gives the general solution, expressed implicitly as follows : lnIyl +cry=x+c 2 Therefore, the complete solution of the given differential equatio n is y=c
or
In I y~+cry=x+c,
■
Type 3 : Second-order homogeneous linear equations where one (nonzero) solution is known . Sometimes it is possible to determin e a solution of a second-order differential equation by inspection , which usually amounts to successful trial and error with a fe w particularly simple functions . For example, you might discover tha t the simple function y = x is a solution of the equation x 2 y " — xy' + y = 0 or that y = ex satisfies the equatio n .xy—(x+l)y'+y= 0 Of course, trial and error is not the best way to solve an equation , but if you are lucky (or practiced) enough to actually discover a solution by inspection, you should be rewarded .
DIFFERENTIAL EQUATIONS
107
SECOND-ORDE R EQUATIONS
If one (nonzero) solution of a homogeneous second-orde r equation is known, there is a straightforward process for determining a second, linearly independent solution, which can then b e combined with the first one to give the general solution . Let y , denote the function you know is a solution . Then let y = y 1 v(x) , where v is a function (as yet unknown) . Substitute y = y i z' into th e differential equation and derive a second-order equation for v . Thi s will turn out to be a Type 1 equation for v (because the dependen t variable, v, will not explicitly appear) . Use the technique describe d earlier to solve for the function v ; then substitute into the expressiony = y i v to give the desired second solution .
Example 33 : Give the general solution of the differential equatio n x2y" —xy' +y= 0 As mentioned above, it is easy to discover the simple solutio n y = x. Denoting this known solution by y l , substitute y = y l v = xv into the given differential equation and solve for v . If y = xv, the n the derivatives are y' =xv' + v y " =xv " +2v ' Substitution into the differential equation yield s x 2 (xv" + 2v') — x(xv' + v) + xv = 0 x 3 v" + x-'v ' = 0 71
108
'
- v = 0 x
+ -
CLIFFS QUICK REVIEW
SECOND-ORDE R EQUATIONS
Note that this resulting equation is a Type I equation for v (becaus e the dependent variable, v, does not explicitly appear) . So, letting v' = w and v " = w', this second-order equation for v becomes th e following first-order equation for w : W' +
— x
W
= 0
(*)
The integrating factor for this standard first-order linear equation i s
µ = e 1'
=
ef(1 /x)dx
=
'elnx = x
and multiplying both sides of (*) by ,u, = x give s xw' +w= 0 d dx (xw)
0
xw= c w=cx - ' Ignore the constant c and integrate to recover v : v'=w~v=
fw=
x - 'dx=lnIx '
Multiply this by y, to obtain the desired second solution , y, = y,v = x In I x The general solution of the original equation is any linear combination of y 1 = x and y 2 = x In lx l : y = c i x + c2x In 'x i This agrees with the general solution that would be found if thi s problem were attacked using the method for solving an equidimensional equation . ■
DIFFERENTIAL EQUATIONS 109
SECOND-ORDER EQUATIONS
Example 34 : Determine the general solution of the followin g differential equation, given that it is satisfied by the function y = ex : xy" — (x + 1)y' + y = 0 Denoting the known solution by y 1, substitute y = y l z) = exv int o the differential equation . With y = exv, the derivatives are y' = ex v ' + exv y ►► = e z
x ,"
+
2ex v' + exv
Substitution into the given differential equation yield s x(exv " + 2ex v ' + e xv) — (x + 1) (e xv ' + exv) + exv = 0
which simplifies to the following Type 1 second-order equation fo r v: xv " + (x — 1)v' = 0 Letting v ' = w, then rewriting the equation in standard form, yield s w' +
x— 1 w = 0 (* ) x
The integrating factor in this case is e fPdx
e
fIl-(l/x)]dx
e x-inx- =eXe -inx =
ex
x Multiplying both sides of (*) by p . = ex /x yield s d ex dx x w
w = cxe -x
110
CLIFFS QUICK REVIE W
SECOND-ORDER EQUATIONS
Ignore the constant c and integrate to recover v : v =
= fxedx = —xe -x —
— e' dx
= —xe -x — e ' e -x(x + 1 ) Multiply this by y 1 to obtain the desired second solution , Y2=y1v
= ex
• [_ e_x (x
+ 1)] = —(x + 1 )
The general solution of the original equation is any linear combination ofy, and y 2 :
DIFFERENTIAL EQUATIONS
POWER SERIES
It often happens that a differential equation cannot be solved i n terms of elementary functions (that is, in closed form in terms o f polynomials, rational functions, ex , sin x, cos x, In x, etc .) . A power series solution is all that is available . Such an expression is nevertheless an entirely valid solution, and in fact, many specifi c power series that arise from solving particular differential equation s have been extensively studied and hold prominent places in mathematics and physics .
Introduction to Power Serie s A power series in x about the point xo is an expression of the for m co + c 1(x
—
x0) +
c2(x — x0)2 +
where the coefficients c,, are constants . This is concisely written using summation notation as follows :
n= 0
cn (x — xo ) n
Attention will be restricted to x 0 = 0 ; such series are simply called power series in x : co + c 1 x + c 2 x2 +
=
n= 0
c n.x' n
A series is useful only if it converges (that is, if it approaches a finit e limiting sum), so the natural question is, for what values of x will a given power series converge? Every power series in x falls into on e of three categories :
DIFFERENTIAL EQUATIONS
POWER SERIE S
Category 1 : The power series converges only forx = 0 . Category 2 : The power series converges for G R and diverges (that is, fails to converge) for Ix' > R (where R i s some positive number) . Category 3 : The power series converges for all x. Since power series that converge only for x = 0 are essentiall y useless, only those power series that fall into Category 2 or Category 3 will be discussed here . The ratio test says that the power serie s
E Cnxf
n= 0
will converge if lim n-~x
n+ l Cn+l x C~ n
<1
(* )
and diverge if this limit is greater than 1 . But (*) is equivalent t o < lim n—x
Cn cn+ 1
so the positive number R mentioned in the definition of a Category 2 power series is this limit : R= lim
Cn
n->x
If this limit is co, then the power series converges for (xi < oo which means for all xand the power series belongs to Category 3 . R i s called the radius of convergence of the power series, and the set o f all x for which a real power series converges is always an interval , called its interval of convergence .
114
CLIFFS QUICK REVIEW
POWE R SERIES
Example 1 : Find the radius and interval of convergence for each o f these power series : 2"
x
(a ) ,1=
x
0n!
,1 =1
x
n3
3
(c)
,1
—
X"
,1=1 n
[Recall that n ! ("n factorial") denotes the product of the positiv e integers from 1 to n . For example, 4! = 1 • 2 • 3 • 4 = 24 . By definition, 0! is set equal to 1 . ] (a) In this power series, c,, = 2"/n !, so the ratio test says R = li m
it —~x
cil C r1 +
2'' (n +
1
I ) !
2,1+1
= li m n! • 11
n+ 1
= lim
2
=
x
Therefore, this series converges for all x . (b) The radius of convergence of the power series in (b) i s R = li m PI —~ x
C il
C"+ 1
li m it — x
n3
3 n+ 1
3" • (n +1) 3 = lim
n3
(n + l)
3,1+ 1
3"
=1 .3= 3
Since R = 3, the power series converges for Ix ~ < 3 and diverges for (x > 3 . For a power series with a finite interval of convergence, the question of convergence at the endpoints of the interval must b e examined separately. It may happen that the power series converge s at neither endpoint, at only one, or at both . The power serie s
±a
i1 =
1
il
Xf l
converges at neither the endpoint x = 3 nor x = -3 because the individual terms of both resulting serie s
DIFFERENTIAL EQUATIONS
POWER SERIES
1n
x
3
E
and
n=1
(-1)"n
3
it= 1
clearly do not approach 0 as n -> oc . (For any series to converge, it i s necessary that the individual terms go to 0 .) Therefore, the interva l of convergence of the power series in (b) is the open interval -3 < x < 3. (c) The radius of convergence of this power series is R = lim
„-max
c,, c„ + 1
1
n+1
n
1
= lim - • ►t-~x
= lim
,r->x
n+ 1
n
= 1
Since R = 1, the series
converges for Ix' < 1 and diverges for Ix' > 1 . Since this power series has a finite interval of convergence, the question of convergence at the endpoints of the interval must be examined separately . At the endpoint x = -1, the power series become s
I (-o
n= 1
nn
which converges, since it is an alternating series whose terms go t o 0. However, at the endpointx = 1, the power series become s
n=1 n
which is known to diverge (it is the harmonic series) . Therefore, th e interval of convergence of the power serie s x
1 _x”
n=1 n
is the half-open interval -1 < x < 1 .
116
■ CLIFFS QUICK REVIE W
POWER SERIE S
Taylor Serie s A Category 2 or Category 3 power series in x defines a function f by setting f (x) = co + c 1 x + c2 x 2 + c3x3 + c 4 x4
+ .. .=
c,,xn n=( )
(0)
for any x in the series' interval of convergence . The power series expansion for f (x) can be differentiated term by term, and the resulting series is a valid representation of f'(x) i n the same interval : f '(x) = c l + 2c 2 x + 3c 1x2
+
4c4 x3
+•••=
E nc„x" -'
,,= 1
(1 )
Differentiating again gives f " (x) = 2c2
+
6c3x+ 12c 4 x2
+ ••• =
n= 2
n(n —
1)c„x" -2
(2 )
and so on . Substitutin g x = 0 in equation (0) yields co = f (0) , x = 0 in equation (1) yields c l = f '(0) , f"(0 ) x = 0 in equation (2) yields c 2 = 2 , and in general, substitutingx = 0 in the power series expansion fo r the nth derivative off yields c„ _
DIFFERENTIAL EQUATIONS
f (0 0 )
n!
POWE R SERIES
These are called the Taylor coefficients off, and the resulting powe r series 'f(")(0) Yi
n=
o n .!
is called the Taylor series of the function
f.
Given a function f, its Taylor coefficients can be computed by th e simple formula above, and the question arises, does the Taylo r series off actually converge to f (x)? If it does, that is, i f f(x)=E f rt=0
(YI)(0) n.
x
Y1
for all x in some neighborhood of (interval around) 0, then th e function f is said to be analytic (at 0) . [More generally, if you form the Taylor series off about a pointx = xo , x f (") (xo) n=o
n•
Y (x — xo) '
and if this series actually converges to f (x) for all x in some neighborhood of xo, then f is said to be analytic at xo .] Polynomials are analytic everywhere, and rational functions (quotients of polynomials) are analytic at all points where the denominator is not zero . Furthermore, the familiar transcendental (that is, nonalgebraic ) functions e x, sin x, and cos x are also analytic everywhere . The Taylo r series in Table 4 are encountered so frequently that they are wort h memorizing . For a general power series (like the ones in Example 1), it i s usually not possible to express it in closed form in terms of familia r functions .
118
CLIFFS QUICK REVIEW
POWE R SERIES
Table 4 TAYLOR SERIE S f(n)o
f(x)
1
x
x2
x3
Ix"
(for Ix'
"=()
n=()
(for all x )
i n.
x5
,.2n+ 1
+ — ••• = I(—1) " 3! 5 ! (2n + 1) ! ll=O 2
4
cos x = 1 — ~+ i 2.
< 1)
xn
x3
ex= 1+x+ i+ i+•• — 2. 3.
sinx=x —
n!
ll=O
1 — x = 1 +x +x 2 +x 3 + . . . =
) x"
4.
. .=E(—1)" x 2i t (2n) . n=( )
(for all x )
(for all x)
Example 2 : Use Table 4 to find the Taylor series expansion of eac h of the following functions : (a)
1 1 — x2
(d)
(b)
1 1 —
(e)
x cosx
x
2
(c) ln(1 + x) (f) sinx cos x
(g) arctan x
DIFFERENTIAL EQUATIONS
119
POWE R SERIES
(a) Replacing x by x 1/(1 — x) give s
2
in the Taylor series expansion o f
1 = 1 +x2+(x2)2+(x2)3+ ••• 1 —x
2
=1 +x +x 2
4
= 1 x 211
for
for
1x 2 I
< 1
+
+ x6
Ix'
< 1
Ix 2 I
<1.
11= U
since
Ixi
< 1 is equivalent to
2
(b) Differentiating 1/(1 — x) gives 1/(1 — x) , so differentiating the Taylor series expansion of 1/(1 — x) term by term will giv e the series expansion of 1/(1 — x)2 : 1
d
x)2
(
1
1 —
d
= — Ix" = x ) dv n=()
E—
x
d
X 1l)
1r=()
=
nxil 11 =
1
for
Ix'
< 1
(c) First, replacing x by —x in the Taylor series expansion of 1/(1 — x) gives the expansion of 1/(1 + x) : 1 1 1+x 1--(—x)
1 +(—x)+(—x)2+(—x)3+•••
= E (—1)'1 x"
for
Ix'
ford—xi < 1
< 1
11 =( )
Now, since integrating 1/(1 + x) yields ln(1 + x), integrating th e Taylor series for 1/(1 + x) term by term gives the expansion fo r ln(1 + x), valid for Ix' < 1 :
120
CLIFFS QUICK REVIE W
POWER SERIE S
n(1 +x)=
x
E ( — 1)„ x n
1 l +x dx=
„=O
=
x
f(_1)n x n
rt=O
=
E(— 1 )n r,=0
xn+ l
n +
Technical note : Integrating 11(1 + x) yields ln(1 + x) + c (where c is some arbitrary constant), so strictly speaking, the equation abov e should have been written xn + 1
ln(1 + x) + c
= E (—1)'' n + 1 ,t
=0
However, substitutingx = 0 into this equation shows that c = 0, s o the expansion given above for ln(1 + x) is indeed correct . (d) Replacing x by —x 2 in the Taylor series expansion of & yields the desired result :
_
x (—x2)”
n=0
(e) gives
'
n•
X
= n=0 E (— 1 ) n
x 2" n.
Multiplying each term of the Taylor series for cos x by x
± (_1)h1 ,
~ x ~xcos= ,, = o (2n) !
(f) One way to find the series expansion for sin x cos x is to multiply the expansions of sin x and cos x. A faster way, however , involves recalling the trigonometric identity sin 2x = 2 sin x cos x and then replacing x by 2x in the series expansion of sin x :
DIFFERENTIAL EQUATIONS
POWER SERIES
1
x
(2x?' ' (2n+1) .
2n=O
(_ )n 22n
n=O
(2n+1)!
x2n +
(g) Since arctan x is the integral of 1/(1 + x 2 ), integrate th e series expansion of 1 /(1 + x 2 ) term by term : arctan x
= f 1 + x z dx
for IxZ l < 1
= ,r =(, (–1)n
x 2n+ 1 2n + 1
for (x ( < 1
Recall the technical note accompanying part (c) above (which also involved the term-by-term integration of a power series) . Th e integral of 141 + is actually arctan x + c, and the equatio n above should read x2)
arctan x + c = I (–1)n n=0
x2n+ 1 2n + 1
However, substituting x = 0 into this equation shows that c = 0, s o the expansion given above for arctan x is indeed correct . ■
CLIFFS QUICK REVIEW
POWER SERIE S
Power Series Solutions of Differential Equation s First-order equations. The validity of term-by-term differentiatio n of a power series within its interval of convergence implies tha t first-order differential equations may be solved by assuming a solution of the form y=
c xn rt
n =( )
substituting this into the equation, and then determining th e coefficients cn .
Example 3 : Find a power series solution of the for m y=cnxn n=( )
for the differential equation y ' –xy= 0
Substituting c nx n
y= n =0
into the differential equation yields x
x
E nCnxn—1 — n= 0 Cnx n+1
=0
n=1
Now, write out the first few terms of each series , (c 1 +2c 2 x+3c 3 x 2 +4c 4x 3 +•••)–(cox +c l x 2 +c 2 x 3 +•••)= 0 and combine like terms : c l + (2c 2 – c 0)x + (3c 3 – c 1 )x 2
DIFFERENTIAL EQUATIONS
+
(4c 4 – c 2 )x 3
+ • • • =
0
POWER SERIES
Since the pattern is clear, this last equation may be written a s
E2 ( nC r — C,,_2)x 11— I = 0
+
c
n=
In order for this equation to hold true for all x, every coefficient on the left-hand side must be zero. This means c 1 = 0, and for all n > 2 , nC„ —
C,,_2
= 0
This last equation defines the recurrence relation that holds for th e coefficients of the power series solution : c „- 2 n
c„ =
for n
2
Since there is no constraint on co, co is an arbitrary constant, and it i s already known that c, = 0 . The recurrence relation above says c, = -co and c3 = c l , which equals 0 (because c 1 does) . In fact, it is eas y to see that every coefficient c,, with n odd will be zero . As for c 4, th e recurrence relation says c4
=
c2 _
4
co 2 .4
and so on . Since all c„ with n odd equal 0, the desired power serie s solution is therefor e y=co+c2x2+C4x4+cx 6 + . . . =co+ co x2+ co x4+
2.4
2
=co 1 +
co 2 . 4 .6
1 1 1 2 x2+ 2 4 x4+ 2 4
+• •
6 x6+
. ..
x
= co
1 2n E nx n=0 n n. !2
124
CLIFFS QUICK REVIEW
POWER SERIE S
Note that the general solution contains one parameter (co), as expected for a first-order differential equation . This power series i s unusual in that it is possible to express it in terms of an elementary function . Observe : co
1
1
1 -— X 2 n.
x 2n = co
n=O n!2
n=O
,
It is easy to check that y = coe`' 2/2 is indeed the solution of the give n differential equation, y ' = xy . Remember : Most power series canno t be expressed in terms of familiar, elementary functions, so the fina l answer would be left in the form of a power series . ■
Example 4: Find a power series expansion for the solution of th e IVP y' = x + y y(0) = 1
Substituting y=
cx „ n= O
into the differential equation yield s x
E nc„x" -1
n=1
x
= x + ± c„x " n= O
or, collecting all the terms on one side , x
E
n=1
DIFFERENTIAL EQUATIONS
x
nC,,x" -1 — x —
n= O
C,1X"
= 0
POWER SERIE S
Writing out the first few terms of the series yield s (c 1 + 2c,x+3c3 x 2 +4c4 x 3
+ —
.
•• )
x — (co+c 1 x+c,x-+c3 x 3
+ •••) =
0
or, upon combining like terms , (c 1 — c 11)
+
(2c, — I — c 1 )x + (3c3 — c,)x 2
+
(4c4 — c3)x 3
+ • . . =
0
Now that the pattern is clear, this last equation can be writte n (c 1 —co)+(2c,— 1 —c1)x+(nc,l—1)x,l-1 = 0 19= 3
In order for this equation to hold true for all x, every coefficient on th e left-hand side must be zero. This mean s c 1 -c„=0,
2c,— 1 —c 1 = 0,
and
nc,, —
1
= 0 forn3
(* )
The last equation defines the recurrence relation that determine s the coefficients of the power series solution : c,, = c ,,-1 n
forn
3
The first equation in (*) says c 1 = co, and the second equation say s = ; (1 + co) . Next, the recurrence relation says
c, = 1(1 + c 1)
c3
c4
c2 =3 c3 4
4(1 + co)
1 +
3
=2 3
co
1+c0 2. 3 . 4
and so on . Collecting all these results, the desired power serie s solution is therefore y = c 0 +c 1 x+c 2x 2 +c 3x 3 +c4x 4
126
+
• .
CLIFFS QUICK REVIEW
POWER SERIE S
=co+cox+
1+ co 2 l+ co 1+ co 4 + 2 x + 2 3 x + 2 3 4x
= co+cox+ (1 +co)
1
x = c(1 + C pX +
(1 +
1
2 x2+ 2 3x-
1 +2 .3 .4x4+ •• •
1
X„ Co) E n =2 n .
Now, the initial condition is applied to evaluate the parameter co : x
y(0) = 1
i
co + cox + (1 + co) E i x" n=2
n.
=co= 1
Therefore, the power series expansion for the solution of the give n IVP is x
y=1+x+
2
E ix " n.
n=2
If desired, it is possible to express this in terms of elementar y functions . Since x
2 x" = 2 x0 + 2 E 0. 1. n=0 n .
X1
x 2
+ n=2 E n . Xn
equation ** may be written x 2 : y=1+x+ fix" n=2 n .
2 2 2 =1+x+ E, ix" – x 0 – i 1. n=0 n . 0. = 1 + x + (2e x – 2 – 2x )
2ext— x — 1 which does indeed satisfy the given IVP, as you can readily verify . ■ DIFFERENTIAL EQUATIONS
POWER SERIE S
Second-order equations . The process of finding power series solutions of homogeneous second-order linear differential equations i s more subtle than for first-order equations . Any homogeneou s second-order linear differential equation may be written in the form y" + p (x )y ' + q (x )y = 0 If both coefficient functions p and q are analytic at x0 , then x0 i s called an ordinary point of the differential equation . On the othe r hand, if even one of these functions fails to be analytic at x0 , then x0 is called a singular point . Since the method for finding a solutio n that is a power series in x0 is considerably more complicated if x0 is a singular point, attention here will be restricted to power serie s solutions at ordinary points .
Example 5 : Find
a power series solution in x for the IV P v"—xy'+y= 0 y(0) = 2 y'(0) = 3
Substituting y =
n=( )
c,1 x
,1
into the differential equation yield s x
,1
=2
x
n (n –
–
x
nc11x,1 + /1=1
c,,x" = 0 n=( )
(* )
The solution may now proceed as in the examples above, writing ou t the first few terms of the series, collecting like terms, and the n determining the constraints on the coefficients from the emergin g pattern . Here's another method .
128
CLIFFS QUICK REVIEW
POWE R SERIES
The first step is to re-index the series so that each one involve s x" . In the present case, only the first series must be subjected to thi s procedure . Replacing n by n + 2 in this series yield s x
x
n ( n – 1 )c,1x ,1
,1
=
=2
,1+2=2
( n + 2)[(n + 2) –
11
=E
c,,+,x(,1+2)_ 2
(n + 2)(n + 1)c„ + 2 x"
,1=( 1
Therefore, equation (*) become s x n=()
(n + 2)(n +
1)c,,+2x,1 –
x
x
"=1
,1= 0
E nc,,x" + E c„x,1 =
0
The next step is to rewrite the left-hand side in terms of a single summation . The index n ranges from 0 to 0c in the first and thir d series, but only from 1 to cc in the second . Since the common range of all the series is therefore 1 to cc, the single summation which wil l help replace the left-hand side will range from 1 to cc . Consequently , it is necessary to first write (* *) a s [2c2
+E "=1
x
(n + 2) (n + 1)c, + ,
x
–±nc„x"+ [co + ,1=
1
c,,x" = 0 ,1= 1
and then combine the series into a single summation : (2c2 + co) +
[(n + 2)(n + 1)c„+ 2 – (n – 1)c„]x" = 0
In order for this equation to hold true for all x, every coefficient on th e left-hand side must be zero. This means 2c, + co = 0, and for n 1 , the following recurrence relation holds : c"+'
DIFFERENTIAL EQUATIONS
n– 1 (n + 1)(n + 2) c "
129
POWER SERIE S
Since there is no restriction on co or c,, these will be arbitrary, an d the equation 2c2 + co = 0 implies c 2 = -1-co. For the coefficient s from c 3 on, the recurrence relation is needed : C
1— 1 3 = c c' '+' —( 1 + 1)(1 + 2 )
C
4
C
5
c `'
C2+'
2—1 C2 (2 + 1)(2 + 2)
3+2
3—1 - 3+1 3+2 c3
=C
c4+
4—1 c4 (4 + 1)(4 + 2)
'
c -7 =c5+2= (5
0 1 - CO 3 . 4 2 2 4 .50
0
3 — co 5 .6 2 3 - 4
5—1 4 7 = 1)(5 + 2c5—6 •0 0
The pattern here isn't too difficult to discern : c,, = 0 for all odd n > 3, and for all even n > 4, n— 3 c „ _ — n, c o This recurrence relation can be restated as follows : for all n c
=
2,
2n—3 3—2 n o (2n)! co — (2n)! c
The desired power series solution is therefore y=c„+c ix+c,x 2 +c 3 x 3 +cox` +c 5 x 5 +cox') + co, CO, 3c, 0 ~~ =c„+c,x— 2 x-+0— 4i x +0— x +•• • 6!
130
CLIFFS QUICK REVIE W
POWER SERIE S
1 1 3 =c(► 1 _ 2 x2 _4xa_6x~~ 1 =c(► 1 __ x 2
2
+ E 3 . — 2n x
—
. . . + c1x
z,T
n=? (2n) .
As expected for a second-order differential equation, the genera l solution contains two parameters (co and c l ), which will be determined by the initial conditions . Since y(0) = 2, it is clear that co = 2 , and then, since y ' (0) = 3, the value of c 1 must be 3 . The solution o f the given IVP is therefore x2
x 2(3 — 2n) (2n)! x " ,T=2
+E
■
Example 6: Find a power series solution in x for the differentia l equation (x 2 + 1)y"+y' —x-2y= 0 Substituting
y=Ec, T x,T TT
=O
into the given equation yield s x
x
(x 2 + 1) E n(n — 1)c„x" ,T=?
+ E nc„x" -1 ,T=1
x
E c„x" +2 = 0
—
n=( )
or x
x
E n(n — 1)c„x" + E n(n — 1)c„x" + Enc, lx" - 1
n=2
n=2
x
E c nx" +2 = 0 "=( ) DIFFERENTIAL EQUATIONS
(X )
POWER SERIE S
Now, all series but the first must be re-indexed so that each involve s x" : x n=2
=
n(n — 1)c„x" -2
=
n+2= 2
n= O
(n + 2) (n + 1)c„ + 2x "
x
/t = 1
ncrtx"-' =
x
it+ l = 1
x
1x(n+ l )— l =
(n + 1 )c,, +
n= O
x
n=()
c,,xft
+2
=
1]c„+2x(n+2)- 2
(n + 2)[(n + 2) —
(n +
1)cn+ l x "
x
n—2=0
C n—2x
(n—2)+2 n= 2
Cn_2 x
n
Therefore, equation (*) become s x
n=2
x
n(n -1)c„x"
+E
n=O
(n + 2)(n + 1)c,, +2 x" +
n= O
(n + 1)c,, + l x"
x
c,, _,x /1 = 0
—
(** )
n= 2
The next step is to rewrite the left-hand side in terms of a single summation . The index n ranges from 0 to oc in the second and third series, but only from 2 to oc in the first and fourth . Since the common range of all the series is therefore 2 to 00, the single summation which will help replace the left-hand side will range from 2 to 0c . It i s therefore necessary to first write (**) a s x
n=2
x
n(n — 1)c„x" + 2c, + 6c 3x
+ n=E2 (n + 2)(n +
+ [c l + 2c 2x +
E (n + 1)c,, +i x"
,1=2
1)c,, +2 x " x —
n= 2
c„_ 2x11 = 0
CLIFFS QUICK REVIEW
POWE R SERIE S
and then combine the series into a single summation : (c 1 +2c,)+(2c 2 +6c 3 ) x + E [n(n – 1)c„ + (n + 2)(n + 1)c,, +2 + (n + 1)c„+1 –
2] x"
= 0
n =2
Again, in order for this equation to hold true for all x, every coefficient on the left-hand side must be zero . This means c 1 + 2c, = 0, 2c2 + 6c3 = 0, and for n 2, the following recurrence relatio n holds : c" _ 2 – n(n – 1)c„ – (n + (n + 1)(n + 2 )
c"+2
Since there is no restriction on co or c 1 , these will be arbitrary; th e equation c 1 + 2c 2 = 0 implies c 2 = – 2 c 1, and the equation 2c 2 + 6c 3 = 0 implies c 3 =- -C2 = - 4( — C 1 ) = c 1 . For the coefficient s from C4 on, the recurrence relation is needed : C2_2 – C4 = C 2+2 =
2(2 – 1)C2 – (2 + 1 )C 2+ 1 (2 + 1)(2 + 2)
co–2c 2 –3c 3
3.4 C3_2 – C5 c3+2
ci
co–2(–2c 1 )–3(, c 3.4
3(3 – 1)C 3 – (3 + (3 + 1)(3 + 2 )
6c3 — 4c4
4. 5 = — U) C 11 — 12O C1
DIFFERENTIAL EQUATIONS
cl
4C11 + 24 0 1
1)C 3+ 1
6(40
4(T12-co + 24 0 1 )
4 .5
POWE R SERIES
Co = C4+2 =
c4 _, – 4(4 – 1)c 4 – (4 + 1 )c 4+ 1 (4 + 1)(4+ 2 )
C,–12c 4 –5c 5 5 . 6 —z c l — 12(lcO + ~4 c 1 ) – 5 . 6 -
To CO
720C 1
The desired power series solution is therefore
y=c () +c 1x+c 2 x 2
+C 3 x 3 +C4x 4
= co + c ix + ( - 1- c 1) x2 +(
CO
= CO(1 + 12 x4
+c 5x 5 +cox ('+ . . .
+ (c1)x3 +
12OCl )x5 + (-1C1 f~()x5
3f~()xf~ -
(I2 c () + Ac l ) x4 2 OCI )x`'
+
. . .l
1x5 - 23x(,+C(x - 21x2+ 61 x 3 +. 271x 4 - 12O 1 72( 1
.)
Determining a specific pattern to these coefficients would be a tedious exercise (note how complicated the recurrence relation is) , so the final answer is simply left in this form . ■
134
CLIFFS QUICK REVIE W
THE LAPLACE TRANSFORM
If you are faced with an IVP that involves a linear differentia l equation with constant coefficients, you can proceed by the metho d of undetermined coefficients or variation of parameters and the n apply the initial conditions to evaluate the constants . However , what if the nonhomogeneous right-hand term is discontinuous ? There exists a method for solving such problems that can also b e used to solve less frightening IVP's (that is, ones that do not involve discontinuous terms) and even some equations whose coefficient s are not constants . One of the features of this alternative method fo r solving IVP's is that the values of the parameters are not found afte r the general solution has been obtained . Instead, the initial conditions are incorporated right into the initial stages of the solution, so when the final step is completed, the arbitrary constants hav e already been evaluated .
Linear Transformation s A function is usually introduced as a rule which acts on a number t o produce a unique numerical result . That is, a function accepts a number as input and produces a number as output . For instance , consider the function defined by the equation f (x) = x2 ; it specifies a particular operation to be performed on any given value of x . When this function acts on the number 3, for example, it gives the result 9 , a fact which can be symbolized as follows : 3
H9
However, functions are not restricted to acting only on numbers t o produce other numbers . A function can also act on a function t o produce another function . These "superfunctions" are often referred to as operators or transformations . Therefore, an operato r accepts a function as input and produces a function as output . DIFFERENTIAL EQUATIONS
THE LAPLAC E TRANSFORM
A familiar example is the differentiation operator, v
D:
f '(x)
f(x)
This operator acts on a (differentiable) function to produce anothe r function : namely, the derivative of the input function . For example , 1)
x3
1)
3x'
sin x
1)
H cos x,
e`
H e`
Another well-known example is the integration operator, which act s on an (integrable) function to produce another function : it s integral . Since an operator is a function, it must produce one an d only one output for each input ; therefore, it makes sense here to consider an integration operator I of the for m f .
f
f(t)
dt
[The letter t, a dummy variable, is chosen simply to distinguish i t from x, the upper limit of integration . It ' s usually considered ba d form to write, for example, Jc~ f (x) dx . ] To illustrate, if f (x) = x 2 , the n I [ f l = ~► f (t) dt = £ t 2 dt=x 3
that is, x2 -
Both of these operators, differentiation D and integration I, enjoy an important property known as linearity . Any operator or transfor-
CLIFFS QUICK REVIEW
THE LAPLACE TRANSFORM
mation T is said to be linear if both of the following conditions , T[c .f]=c .T[f ] T[f +g] = T[f] + T[g]
always hold for all constants c and all admissible functions f and g. Sinc e the derivative of a constant multiple of a function is equal to th e constant times the derivative of the function , and since the derivative of the sum of two functions is the sum of th e derivatives , differentiation is indeed linear . Furthermore, since both of th e previous statements remain true when "derivative" is replaced b y "integral, " integration is also a linear transformation .
The Laplace Transform Operato r A particular kind of integral transformation is known as the Laplac e transformation, denoted by L . The definition of this operator i s L [f(x)] =
f
e-l' f(x) dx
The result called the Laplace transform off will be a function o f p, so in general, f (x )
DIFFERENTIAL EQUATIONS
F( p)
THE LAPLAC E TRANSFOR M
Example 1 : Find the Laplace transform of the function f(x) = x. By definition, x
L [x]
= fe 1 xdx
Integrating by parts yield s L[x]
= fePx xdx x x — _ e -pxp c~
x —
0
1 x = 0 + — ~~ e-rx. dx
p 1 p
1 --e —na p
1 --
edx
p
(for p > 0)
X
0
1 p2 Therefore, the function F( p) = 1 /p2 is the Laplace transform of th e function f (x) = x . [Technical note : The convergence of the imprope r integral here depends on p being positive, since only then wil l (x/p)e-P and e-rx approach a finite limit (namely 0) as x Therefore, the Laplace transform of f(x) = x is defined only fo r p>0 .] ■ In general, it can be shown that for any nonnegative integer n , the Laplace transform of f(x) = x" is
F(p)
n!
=
n+ 1
P
Like the operators D and Iindeed, like all operators the Laplace transform operator L acts on a function to produce anothe r
138
CLIFFS QUICK REVIE W
THE LAPLACE TRANSFORM
function . Furthermore, sinc e L[cf] =
f
x e ra 'cf(x)dx=c
x
f e'
fix) dx=cL[f ]
and I,[f + g ]
=
f
e-~'x [f(x ) + g (x )] dx [e'f(x) + e ,'g (x )] dx
x
e Px f(x) dx +
x
e ,» g(x) dx
=L[f] +L[g]
the Laplace transform operator L is also linear . [Technical note : Just as not all functions have derivatives o r integrals, not all functions have Laplace transforms . For a function f to have a Laplace transform, it is sufficient that f (x) be continuou s (or at least piecewise continuous) for x 0 and of exponentia l order (which means that for some constants c and X, the inequalit y ~f(x)I < ce '
holds for all x) . Any bounded function (that is, any function f tha t always satisfies ~ f (x) < M for some M > 0) is automatically o f exponential order (just take c = M and A = 0 in the definin g inequality) . Therefore, sin kx and cos kx each have a Laplace transform, since they are continuous and bounded functions . Furthermore, any function of the form eke , as well as any polynomial , is continuous and, although unbounded, is of exponential order an d therefore has a Laplace transform . In short, most of the function s you are likely to encounter in practice will have Laplace trans forms . ]
DIFFERENTIAL EQUATIONS
139
THE LAPLAC E TRANSFOR M
Example 2 : Find the Laplace transform of the function f(x) = x3 — 4x + 2 . Recall from the first statement following Example 1 that th e Laplace transform of f(x) = x" is F(p) = n!/p" + ' . Therefore, sinc e the Laplace transform operator L is linear, L[x3 — 4x+ 2] =L[x 3 ] + L[—4x] +L[2 ] = L[x3 ] 4L [x' ] + 2L [x"] 3!
p
1!
3+1
— 4
p
6 4
F( p
0!
1+I
+ 2
2
p
~~+ 1
■
Example 3 : Determine the Laplace transform off (x) = e l' . Apply the definition and perform the integration : x
L [e ll =
e. _," e " dx dx
(1
1 p—k
"
In order for this improper integral to converge, the coefficien t (p — k) in the exponential must be positive (recall the technica l note in Example 1) . Thus, for p > k, the calculation yields 1 L [e t ] _ p — k
140
■
CLIFFS QUICK REVIE W
THE LAPLAC E TRANSFORM
Example 4 : Find the Laplace transform of f (x) = sin kx. By definition, L [sin kx]
_
f
e -r'- sin kx dx
This integral is evaluated by performing integration by parts twice , as follows : 1 e -1"- sinkxdx = - - e -P"- sinkx p
1 - - e -~'a kcos kx dx p
1 - ak = - - e P sin kx + - e -P-' cos kx dx p p 1 _ - - e -/"- sin kx + p k 1 – - - e -'"- cos kx + p p
1
f p-e-I)x
(-k sin kx) dx
1 k -e -1"- sinkx - ;e -1"- coskx p P-
k-
e sin kx dx
so kp-
1 k e -,"- sin kx dx = -p- e -' sin kx -P-- e - P" cos kx
5 e -"" sin kx dx = 5 e -'"- sin kx dx = DIFFERENTIAL EQUATIONS
1 k - p- e -P"- sin kx - e -/" cos kx P-
- e -'"- (p sin kx + k cos kx ) p- + k -
141
THE LAPLACE TRANSFOR M
Therefore, L [sin kx]
=f
e —px
sin kx dx =
—CPI (psinkx+kcoskx) " p2
+
jo
k2
k = p
Z +k 2
for p > 0 . By a similar calculation, it can be shown tha t L [cos kx]
= p, p
k2
■
Example 5 : Determine the Laplace transform of the functio n f (x)
1 for0 <x < 2 0 for x > 2
pictured in Figure 8 :
1 0 2 ■ Figure 8
142
■
CLIFFS QUICK REVIEW
THE LA PLA C E TRANSFORM
This is an example of a step function . It is not continuous, but i t is piecewise continuous, and since it is bounded, it is certainly of exponential order . Therefore, it has a Laplace transform . o •
• •
L [ f (x)] =
e
Px
f (x) dx
+
fo2 e_Px f(x) dx 2 0
e -Px . 1dx+ e-Px
f
e_Px f(x ) dx
x
-Px 0dx 2 e
dx
1 – e -2P P
Table 5 (page 144) assembles the Laplace transforms of a few o f the most frequently encountered functions, as well as some of th e important properties of the Laplace transform operator L .
Example 6 : Use Table 5 to find the Laplace transform of f (x) = sin 2 x . Invoking the trigonometric identit y sin e x = 2(1
cos 2x )
linearity of L implies L[sin z x] =
[1 – cos 2x] =
DIFFERENTIAL EQUATIONS
11
p
2(p p z2 + +44
2 p(p2+4)
■
143
THE LAPLAC E TRANSFORM
Table 5 LAPLACE TRANSFORM S f (x )
F(p )
1
1
p n!
xn
pn 1
ekx
sin
p— k k
kx
cos kx
p2
+ k2
p
p2
+k
2
Linearity : L [c, f (x) + czg(x)] = c i L [ f (x)] + c 2L[g(x) ]
Shifting formula : L [e kxf (x)] = F(p — k),
where L [ f (x)] = F(p )
Laplace transform of derivatives :
144
L [Y'J =
PL IYJ
L [Y"l =
P ZL IYJ
— Y(o) — PY(o ) — Y '( o)
CLIFFS QUICK REVIEW
THE LAPLAC E TRANSFORM
Example 7 : Use Table 5 to find the Laplace transform of g(x) = 5.' .
x3e
The presence of the factor e 5-v suggests using the shifting formul a with k = 5 . Since 3! L [f (x)] = L[x ] =p + 3+1
1 =
6 a = F(p)
p4
the shifting formula says that the Laplace transform of f (x)e 5 i x 3e 5 i is equal to F(p — 5) . In other words, the Laplace transform o f x 3e 5 i is equal to the Laplace transform of x 3 with the argument p shifted top — 5 : -
=
-
L [x3e 5 i ] -
6 =
( p—5 )
4
■
Example 8 : Use Table 5 to find the Laplace transform of f (x) = e- 1. sinx — 3 . First, since L [sin x] = 1/(p 2 + 1), the shifting formula (wit h k = -2) says L[e -'-'sin
x]
=
(p+2) 2 + l
Now, because L [3] = 3 . L[l] = 3/p, linearity implie s L[e - '' sinx — 3] =
DIFFERENTIAL EQUATIONS
1 3 - (p+2)-+1 p
145
THE LAPLAC E TRANSFORM
Example 9: Use Table 5 to find a continuous function whose Laplace transform is F(p) = 12/p 5 . This example introduces the idea of the inverse Laplac e transform operator, L - I . The operator L -' will "un-do" the actio n of L . Symbolically, f (x) - F ( p )
-1
f (x)
F(p)
If you think of the operator L as changing f (x) into F(p), then th e operator L-' just changes F(p) back into f (x) . Like L, the invers e operator L -' is linear. More formally, the result of applying L -' to a function F(p) is to recover the continuous function f (x) whose Laplace transform is th e given F(p) . [This situation should remind you of the operators D and I (which are, basically, inverses of one another) . Each wil l un-do the action of the other in the sense that if, say, I changes f (x ) into F(x), then D will change F(x) back into f (x) . In other words , D = I-', so if you apply I and then D, you're back where yo u started . ] Using Table 5 (reading it from right to left) , 12 -1
p
146
s
24
-1
p
s
_
_1
24 p
4! =
= 4L -1 p
:4 4
CLIFFS QUICK REVIE W
THE LAPLACE TRANSFORM
Example 10: Find the continuous function whose Laplace trans form isF(p) = 1/(p 2 – 1) . By partial fraction decomposition , 1 p2 _
– p+ l + p– 1
1
Therefore, by linearity of L -' , L - 1 ~. [p 2
—1
= 1'
[p+ij
_4 e -X
+L - 1 [
Lp-1 =
Example 11 : Determine
+
2(ex – e')
■
12
L -'
First, note that p has been shifted to p + 2 = p – (–2) . Therefore, sinc e L - '
2 12
p + 9
=L - ' p2+ 4 .3
=4 .L-' 9
3
p +
32
=4sin3x
the shifting formula (with k = -2) implies L-'
12 -2x sin 3x 2 [(P + 2) + 9] = 4e
Example 12: Evaluate
p –
L -1 p2
■
1 6p+25
Although p 2 – 6p + 25 cannot be factored over the integers, it can be expressed as the sum of two squares : p2 —
6p + 25 =
DIFFERENTIAL EQUATIONS
(p 2 –
6p + 9) + 16=(p–3) 2 +42
147
THE LAPLACE TRANSFORM
Therefore, L -'
p — -6p +25
~
L_
p — (p
-3) 2 +42
(p—3)+ 2 + — 3) 3)2
.L
=L -,
p 3 L (p—3)--+4- +
=
L-1
p
3
(p—3)2+4- +
4
~L
( p—3 ) 2+4'
= e 3x cos4x + 4e 3x sin 4x =eNcos4x+ ;sin4x)
■
Using the Laplace Transform to Solve Differential Equation s In this section, you will learn how to use the Laplace transfor m operator to solve (first- and second-order) differential equation s with constant coefficients . In particular, the differential equation s must be IVP's with the initial condition(s) specified at x = 0 . The method is simple to describe . Given an IVP, apply th e Laplace transform operator to both sides of the differential equation . This will transform the differential equation into an algebraic equation whose unknown, F(p), is the Laplace transform of th e desired solution . Once you solve this algebraic equation for F(p) , take the inverse Laplace transform of both sides ; the result is th e solution to the original IVP .
148
CLIFFS QUICK REVIEW
THE LAPLAC E TRANSFOR M
Before this process is undertaken, it is necessary to see what th e Laplace transform operator does to y' and y " . Integration by parts yields e
L [Y'] =
y , (x) dx
I.
x
= e y(x) –
= –y(0) +p
–pe -Tx y(x) dx
e P'Y(x ) dx
( p > 0)
= –y(0) +pL[y ]
so L [Y ') = PL [Yl —y(o )
Replacingy byy' in this result gives the Laplace transform of y" : L [Y " I = pL LY ' J – Y ' ( 0 ) = p [ pL [Y] – Y(0 )] – Y ' ( 0 ) L [Y"l = p2L LYE – pY( 0 ) – Y ' ( 0 )
Example 13 : Use the Laplace transform operator to solve the IV P y 2y=e 3 , y(0) = - 5 Apply the operator L to both sides of the differential equation ; then use linearity, the initial condition, and Table 5 to solve fo r L [Y] :
DIFFERENTIAL EQUATIONS
149
THE LAPLACE TRANSFORM
L[y'] — 2L[y] = pL[y] — y(0) — 2L[y] = PL [Y] + 5 — 2L [Y]
p— 3
(P — 2)L[y] =
_ 3 - 5 p
-sn L [Y] — LP —
+ i6
2 )(P — 3 )
Therefore, y =L-'
—Sp + 1 6 (P — 2 )( p — 3 )
By partial fraction decomposition , -5p+16 ( p—2 )( p—3 )
-6
1
p - 2 +p - 3
so -6 1 y=L-2+ 3 = - 6e 2 +e ` _ +p _ p
is the solution of the IVP .
Usually when faced with an IVP, you first find the genera l solution of the differential equation and then use the initia l condition(s) to evaluate the constant(s) . By contrast, the Laplac e transform method uses the initial conditions at the beginning of th e solution so that the result obtained in the final step by taking th e inverse Laplace transform automatically has the constants evaluated .
150
CLIFFS QUICK REVIEW
THE LAPLACE TRANSFOR M
Example 14 : Use Laplace transforms to solv e y" -3y' -4y= -16x y(0) = - 4 y'(o) _ - 5 Apply the operator L to both sides of the differential equation ; then use linearity, the initial conditions, and Table 5 to solve fo r L [y]:
L[y"] – 3L[y'] – 4L [Y] = L[–1(ix ]
= L[–lfix ] 6 [p2L[y] + 4p + s1 - 3{pL[y] + 4] - 4L[y} = – 1 P– 16 ( p -3p -4)L [ y ] +4p -7= P16
––4p+ 7 L [ y ]=
p--3p– 4
But the partial fraction decomposition of this expression for L[y] i s 16
––4p+ 7 -3p–4
16
– –, -4p+7 p2 p
-16–4p 3 +7p 2
-3p–4 p' p2 ( p+ 1 )( p–4 )
-3
+
4
DIFFERENTIAL EQUATIONS
1
-2
;+ p+l + p– 4
p p2
THE LAPLAC E TRANSFOR M
Therefore, y=L-'
3
4 + +
p p2
1 2 + p+l p– 4
which yields x
y= -3+4x+e_ —2e 4x
■
Example 15: Use Laplace transforms to determine the solution of the IVP y"—2y' + 5y = 0 y(0) = - 1 y' (0) = 7 Apply the operator L to both sides of the differential equation ; then use linearity, the initial conditions, and Table 5 to solve fo r L [y] :
L[ y° - Zy' + Sy] = L[0]
[p2L[y] — PY( 0 ) — y '(0)] — 2 [PL [Y] — Y( 0 )] + SL [Y] = 0
[ p 2L[y] + p — 7] — 2[pL[y] + 1 ] + SL [Y] = 0 (p 2 -2p+5)L[y]+p–9=0
–p+ 9 L[Y]—p~—2p+ 5
CLIFFS QUICK REVIEW
THE LAPLAC E TRANSFORM
Now , –p+9
–p+ 9
p 2 –2p+5 (p 2 –2p+1)+ 4 –p+ 9
(p - 1) z + 4 —(p—1)+ 8 (P — 1 ) z + 2 2 —(p—1)
4 . 2
(P —1 ) 2 + 22
(P —1 ) 2+22
so
( P –1 ) 2
+
LP_1
22
2 +22 ]
= — e x cos 2x + 4ex sin 2x or more simply, y = ex (4 sin 2x – cos 2x)
Example 16: Use the fact that if f (x) = L - '[F(p)], then for any positive constant k, L' fe-knF (P))
0 for x < k =
f (x – k) for x
to solve and sketch the solution of the IV P y' –
y – cr (x )
y(0) = 0
DIFFERENTIAL EQUATIONS
k
THE LAPLAC E TRANSFORM
where v- is the step function 0 for0x< 2 1 forx 2 shown in Figure 9 :
1
2 ■ Figure 9 ■ As mentioned earlier, the Laplace transform method is particularly well-suited to solving IVP's that involve discontinuous functions such as the step function r shown above . As usual, begin by taking the Laplace transform of both sides o f the differential equation : L [y' — y] = L
Mx)] (* )
Since y(0) = 0, the left-hand side of (*) reduces t o L [Y' — Y~ = L [Y'l — L [yl = [P L IY l — Y( o )) — L [Yl = ( p — 1 )L [Y l
Using the definition of L, the right-hand side of (*) is no w evaluated : L [cr(x) ] =
154
x
1
e -PXQ'(x) dx = i e -px dx = - - e -px~ P
e -2n
x _ 2
P
CLIFFS QUICK REVIEW
THE LAPLAC E TRANSFORM
Therefore, the transformed equation (*) read s e (P - 1 )L [Y] =
P
so e -2P y=L -
But e -2P
p ( p - 1)
-1 e
e p ( p - 1)
p+P- 1
so -,
1
-2P
y L
e
- L -,
1
2
e- P
P
P
Now, since L -' [1 (p - 1)] = e-', the formula given in the statemen t of the problem says 1
L
e
-1
0 Hex-2 rr' p -1
forx< 2 forx
2
and since L -1 [1 p] = 1, applying the formula given in the statemen t of the problem again yield s 11
e L
2P
p
forx < 2 = 101 for x 2
Alternatively, simply notice that e- 2n L[o-(x)] =
1 L -1 e-2P -
P
DIFFERENTIAL EQUATIONS
p
0 forx < 2 _ cr(x) _
1 forx >_ 2
THE LAPLACE TRANSFORM
Substituting these results into (**) gives the solution of the IVP : 0 Y =
forx < 2 forx 2
0 forx < 2 1 forx >— 2
which becomes 0
forx < 2 e-`" -2 — 1 for x >_ 2 This function is sketched in Figure 10 :
1
2 ■ Figure 10 ■
156
CLIFFS QUICK REVIEW
A FEW APPLICATIONS
It can be argued that the greatest importance of studying differential equations lies in their applications to science . Since the process of predicting and describing the changes in a physical system is perhaps the raison d'etre of science, and because the rates at whic h changes occur are expressed by derivatives, it is not surprising tha t so many scientific laws are formulated as differential equations . It is not possible here to list the tremendous number of practica l applications of differential equations, but this book would b e incomplete if there weren't some acknowledgment of their rol e outside pure mathematics .
Applications of First-Order Equation s Orthogonal trajectories . The term orthogonal means perpendicular, and trajectory means path or curve. Orthogonal trajectories, there fore, are two families of curves that always intersect perpendicularly. A pair of intersecting curves will be perpendicular if th e product of their slopes is -1, that is, if the slope of one is the negative reciprocal of the slope of the other . Since the slope of a curve is given by the derivative, two familes of curves f1 (x,y, c) = 0 and f,(x, y, c) = 0 (where c is a parameter) will be orthogona l wherever they intersect if
DIFFERENTIAL EQUATIONS
df,
-1
di
df,/dx
A FEW APPLICATIONS
Example 1 : The electrostatic field created by a positive poin t charge is pictured as a collection of straight lines which radiate awa y from the charge (Figure 11) . Using the fact that the equipotentials (surfaces of constant electric potential) are orthogonal to th e electric field lines, determine the geometry of the equipotentials o f a point charge .
■ Figure l I
■
If the origin of an xy coordinate system is placed at the charge , then the electric field lines can be described by the family y = cx
The first step in determining the orthogonal trajectories is to obtai n an expression for the slope of the curves in this family that does not involve the parameter c . In the present case , y cx dx
c
x (
)
The differential equation describing the orthogonal trajectories i s therefore
158
CLIFFS QUICK REVIEW
A FE W APPLICATIONS
dv x d
since the right-hand side of (**) is the negative reciprocal of th e right-hand side of (*) . Because this equation is separable, th e solution can proceed as follows : dy
x
dx
y
ydy=—xdx
f y dy = f —x dx 0y-=—x-+ c x 2 +y 2 =c -
where c 2 = 2c ' . The equipotential lines (that is, the intersection of the equipotentia l surfaces with any plane containing the charge) are therefore th e family of circles x 2 + y 2 = c 2 centered at the origin . The equipotential and electric field lines for a point charge are shown in Figure 12 . electric field lines
equipotentia l lines ■ Figure 12
DIFFERENTIAL EQUATIONS
■
159
A FE W APPLICATIONS
Example 2 : Determine the orthogonal trajectories of the family of circles x 2 + (y -- c) 2 = c 2 tangent to thex axis at the origin . The first step is to determine an expression for the slope of th e curves in this family that does not involve the parameter c . By implicit differentiation, x2+ ( y—c ) 2=c 2 x2 + y 2 = 2cy dy dy 2x + 2y = 2c dx dx dy
x
dx c
Y
To eliminate c, note that x 2 + (y c )' c -
x ' + y' = 2cy
c
x2 + y-
2y
The expression above for dy/dx may now be written in the form dy
x
ay — y- (~ )
x
dx c — y —x- +
-
y2
—
Y
Therefore, the differential equation describing the orthogona l trajectories is dy y-' —x dx
ay
since the right-hand side of (**) is the negative reciprocal of th e right-hand side of (* ) .
160
CLIFFS QUICK REVIEW
A FE W APPL I PP/ICA PONS
If equation r'') is written in the for m (y 22–x2)dx—2xydy= 0
note that it is not exact (since M, . = 2y but N. = - 2y) . However , because M,. -- Nx N
2
2y– (–2y) -2xy
x
is a function of x alone, the differential equation ha s
µ = e f (-2/x)d =
e-21n_,-
=
eln(a- `)
= x -2
as an integrating factor . After multiplying through by µ = x- 2 , the differential equation describing the desired family of orthogona l trajectories become s (x -2 y 2
–
1) dx – 2x - 'y dy = 0
which is now exact (because M,. = 2x - 22y = Ni ) . Sinc e Max =
(x -2y 2 –
1)ax= –x - ' y 2 – x
and
f
N ay
=
f(–'y)ay =
–
x - ' y -2
the solution of the differential equation i s – x = -2c (The reason the constant was written as -2c rather than as c will b e apparent in the calculation below .) With a little algebra, th e equation for this family may be rewritten : –x - ' y 2 – x = -2 c x - ' y 2 +x=2c y2+x-2 =2cx (x—c ) 2+y2=c 2
DIFFERENTIAL EQUATIONS
161
A FE W APPLICATIONS
This shows that the orthogonal trajectories of the circles tangent to the x axis at the origin are the circles tangent to the y axis at th e origin! See Figure 13 .
■ Figure 13
■
Radioactive decay. Some nuclei are energetically unstable and ca n spontaneously transform into more stable forms by various processes known collectively as radioactive decay . The rate at which a particular radioactive sample will decay depends on the identity of the sample . Tables have been compiled which list the half-lives o f various radioisotopes . The half-life is the amount of time require d for one-half the nuclei in a sample of the isotope to decay; therefore, the shorter the half-life, the more rapid the decay rate . The rate at which a sample decays is proportional to the amoun t of the sample present . Therefore, if x(t) denotes the amount of a radioactive substance present at time t, the n dx dt
—kx
(k > o)
CLIFFS QUICK REVIEW
162
A FE W APPLICATIONS
(The rate dx/dt is negative, since x is decreasing .) The positive constant k is called the rate constant for the particular radioisotope . The solution of this separable first-order equation i s x = xoe -k'
(* )
where x0 denotes the amount of substance present at time t = 0 . Th e graph of this equation (Figure 14) is known as the exponential deca y curve :
■ Figure 14
■
The relationship between the half-life (denoted T1,2) and the rat e constant k can easily be found . Since, by definition, x = - x0 at t = T1,2, (*) becomes xo = xoe -kT' ; -
– kT 1 /2 =
In
2
=
k=
In 2 T1, 2
Because the half-life and rate constant are inversely proportional , the shorter the half-life, the greater the rate constant, and , consequently, the more rapid the decay .
DIFFERENTIAL EQUATIONS
163
A FE WAPPLICATIONS
Radiocarbon dating is a process used by anthropologists an d archaeologists to estimate the age of organic matter (such as wood or bone) . The vast majority of carbon on earth is nonradioactiv e carbon-12 ( 12 C) . However, cosmic rays cause the formation o f carbon-14 ( 14 C), a radioactive isotope of carbon which become s incorporated into living plants (and therefore into animals) through the intake of radioactive carbon dioxide ( 14CO2 ) . When the plant o r animal dies, it ceases its intake of carbon-14, and the amoun t present at the time of death begins to decrease (since the 14C decay s and is not replenished) . Since the half-life of 14C is known to be 5730 years, by measuring the concentration of 14C in a sample, its age ca n be determined .
Example 3 : A fragment of bone is discovered to contain 20% of th e usual 14 C concentration . Estimate the age of the bone . The relative amount of 14C in the bone has decreased to 20% o f its original value (that is, the value when the animal was alive) . Thus, the problem is to calculate the value of t at whichx(t) = 0 .20x , (wherex = the amount of 14 C present) . Sinc e In 2 T 1/2
k
In 2 573 0
the exponential decay equation (*) say s 0.20x0 = x0 e—[(In2)/57301 t In (0 .20) = —
In 2 573 0 t
t = 5730 t
164
—in (0 .20 ) In 2
13,300 years
■
CLIFFS QUICK REVIE W
A FE W APPLICATIONS
Newton's Law of Cooling . When a hot object is placed in a coo l room, the object dissipates heat to the surroundings, and it s temperature decreases . Newton's Law of Cooling states that the rat e at which the object's temperature decreases is proportional to the difference between the temperature of the object and the ambien t temperature . At the beginning of the cooling process, the differenc e between these temperatures is greatest, so this is when the rate o f temperature decrease is greatest . However, as the object cools, th e temperature difference gets smaller, and the cooling rate decreases ; thus, the object cools more and more slowly as time passes . T o formulate this process mathematically, let T(t) denote the temperature of the object at time t and let T, denote the (essentially constant) temperature of the surroundings . Newton's Law o f Cooling then says dT dt —k(T —
T.)
(k > 0 )
Since T,. < T (that is, since the room is cooler than the object), T decreases, so the rate of change of its temperature, dT/dt, i s necessarily negative . The solution of this separable differentia l equation proceeds as follows :
fT—T dT
ln(T —
—k dt S
T) =
— kt + c '
T — T,. = T
=
T.+ce -k '
(* )
DIFFERENTIAL EQUATIONS
165
A FE W APPLICATION S
Example 4: A cup of coffee (temperature = 190°F) is placed in a room whose temperature is 70°F . After five minutes, the temperature of the coffee has dropped to 160°F . How many more minutes must elapse before the temperature of the coffee is 130°F ? Assuming that the coffee obeys Newton's Law of Cooling, it s temperature T as a function of time is given by equation (*) wit h T,=70 : T(t) = 70 + ce -k t
Because T(0) be evaluated :
190, the value of the constant of integration (c) ca n T(0) = 190 = 190=70+c = c= 12 0
Furthermore, since information about the cooling rate is provided (T = 160 at time t = 5 minutes), the cooling constant k can be determined : 160 = T(5 ) 160 = 70 + 120e - Sk e -5k =
4
-5k=1n 4
k=5ln 4 Therefore, the temperature of the coffee t minutes after it is place d in the room is T(t) = 70 + 120e-1''5)1°
166
43)1 '
CLIFFS QUICK REVIEW
A FE W APPLICATION S
Now, setting T = 130 and solving for t yield s 130 = 70 + 120e—[(1 /5)In(4/3)] t e —[(1 /5)ln(4/3)]t = 1
(—i In 43)t = I n t= t
2
5 In 2 In 12 minute s
This is the total amount of time after the coffee is initially placed i n the room for its temperature to drop to 130°F . Therefore, afte r waiting five minutes for the coffee to cool from 190°F to 160°F, it i s necessary to then wait an additional seven minutes for it to coo l down to 130°F . ■
Skydiving (part I) . Once a sky diver jumps from an airplane, ther e are two forces that determine her motion : the pull of the earth' s gravity and the opposing force of air resistance . At high speeds, th e strength of the air resistance force (the drag force) can be expresse d as kz v 2 , where v is the speed with which the sky diver descends and k is a proportionality constant determined by such factors as th e diver's cross-sectional area and the viscosity of the air . Once th e parachute opens, the descent speed decreases greatly, and th e strength of the air resistance force is given by Kv . Newton's Second Law states that if a net force Fnet acts on a n object of mass m, the object will experience an acceleration a give n by the simple equation Fnet
DIFFERENTIAL EQUATIONS
= ma
167
A FE W APPLICATIONS
Since the acceleration is the time derivative of the velocity, this la w can be expressed in the form dzv Fnet = m dt
(* )
In the case of a sky diver initially falling without a parachute, th e drag force is Fdrag = kv 2, and the equation of motion (*) become s mg — kz v'- =
m
dzv dt
or more simply, dv bz, 2 dt = g — where b = k/m . [The letter g denotes the value of the gravitationa l acceleration, and mg is the force due to gravity acting on the mass m (that is, mg is its weight) . Near the surface of the earth, g i s approximately 9.8 meters per second 2 .] Once the sky diver's descent speed reaches v = \/g/b = \/mg/k, the equation above says dv/dt = 0 ; that is, v stays constant . This occurs when the speed is great enough for the force of air resistance to balance the weight of the sky diver ; the net force and (consequently) the acceleration drop to zero . Thi s constant descent velocity is known as the terminal velocity . For a sky diver falling in the spread-eagle position without a parachute, th e value of the proportionality constant k in the drag equation Fdrag = kzv 2 is approximately 4 kg/m. Therefore, if the sky diver has a total mass of 70 kg (which corresponds to a weight of about 150 pounds) , her terminal velocity is Z1 terminal (no parachute)
=
\/mg k =
/(7o)(9 .8) l
or approximately 120 miles per hour .
168
52 m/ s
Z
CLIFFS QUICK REVIE W
A FE W APPLICATIONS
Once the parachute opens, the air resistance force become s F air resist = Kv, and the equation of motion (*) become s mg — Ky
=m
dv
dt
or more simply, dv By dt = g — where B = Kim . Once the parachutist's descent speed slows to v = g/B = mg/K, the equation above says dv/dt = 0 ; that is, v stays constant . This occurs when the speed is low enough for the weigh t of the sky diver to balance the force of air resistance ; the net force and (consequently) the acceleration reach zero . Again, this constant descent velocity is known as the terminal velocity. For a sk y diver falling with a parachute, the value of the proportionalit y constant K in the equation Fairresist = Ky is approximately 110 kg/s . Therefore, if the sky diver has a total mass of 70 kg, the termina l velocity (with the parachute open) is onl y vterminal (with parachute)
mg
(70)(9 .8 )
K
11 0
6 .2 m/ s
which is about 14 miles per hour . Since it is safer to hit the groun d while falling at a rate of 14 miles per hour rather than at 120 mile s per hour, sky divers use parachutes .
DIFFERENTIAL EQUATIONS
169
A FE W APPLICATIONS
Example 5 : After a free-falling sky diver of mass m reaches a constant velocity of v 1 , her parachute opens, and the resulting ai r resistance force has strength Kv . Derive an equation for the spee d of the sky diver t seconds after the parachute opens . As stated above, once the parachute opens, the equation of motion is dv g – By dt –
where B = Kim . The parameter that will arise from the solution o f this first-order differential equation will be determined by the initia l condition v(0) = v 1 (since the sky diver's velocity is z ► 1 at the momen t the parachute opens, and the "clock" is reset to t = 0 at thi s instant) . This separable equation is solved as follows : dv – g – Bv
f dt
1 –Bln(g–Bv)=t+c " ln(g – By) = –Bt + c ' g – By = ce -B '
Now, since v(0) = v, g – Bz), = c, the desired equation for th e sky diver's speed t seconds after the parachute opens is g–B7 = (g–Bv>)e-'' By = g – (g – Bv, )e-'3 ' B (1 - - v1 e - "' g
v
170
mg K
K vl i + (_ mg
)e-(K/m) t
CLIFFS QUICK REVIEW
A FE W APPLICATIONS
Note that as time passes (that is, as t increases), the term e -(KIm)t goes to zero, so (as expected) the parachutist's speed v slows t o mg/K, which is the terminal speed with the parachute open . ■
Applications of Second-Order Equation s Skydiving (part II) . The principal quantities used to describe th e motion of an object are position (s), velocity (v), and acceleratio n (a) . Since velocity is the time derivative of the position, an d acceleration is the time derivative of the velocity, acceleration is th e second time derivative of the position . Therefore, the position function s(t) for a moving object can be determined by writin g Newton's Second Law, Fne , = ma, in the for m F'net = m
d '-s
di"
and solving this second-order differential equation for s . [You may see the derivative with respect to time represented b y For example, s ("s dot") denotes the first derivative of s with respect to t, ands ("s double dot") denotes the second derivative o f s with respect to t . The dot notation is used only for derivatives with respect to time . ] a dot.
DIFFERENTIAL EQUATIONS
A FE W APPLICATIONS
Example 6: A sky diver (mass m) falls long enough without a parachute (so the drag force has strength kv 2 ) to reach her firs t terminal velocity (denoted v l ) . When her parachute opens, the ai r resistance force has strength Kv . At what minimum altitude must her parachute open so that she slows to within 1% of her new (muc h lower) terminal velocity (v 2 ) by the time she hits the ground ? Let y denote the vertical distance measured downward from the point at which her parachute opens (which will be designated tim e t = 0) . Then Newton's Second Law (Fnet = ma) becomes mg – Kv = ma, or, since v = y and a = y , mg – Ky=my This situation is therefore described by the IV P y
K
+m y= g y(0) = 0 ,y(0) = v l
The differential equation is second-order linear with constan t coefficients, and its corresponding homogeneous equation i s y+By= 0 where B = Kim . The auxiliary polynomial equation, r2 + Br = 0, has r = 0 and r = –B as roots . Since these are real and distinct, th e general solution of the corresponding homogeneous equation i s Yh = cl + c2e-Bt
= cl + c 2e -(K/m)t
The given nonhomogeneous equation has y = (mg/K)t as a particular solution, so its general solution i s y=c l +c 2e -(K/m)t
+ mg t
K
()
CLIFFS QUICK REVIEW
A FE W APPLICATIONS
Now, to apply the initial conditions and evaluate the parameters c , and c2 : c2e-(K/m)t+
y(0)=0
cl +
)) ( 0 ) = v 1
K [_c 2_ e_ 1m )t +
mg K t t- o
m
K t= o
+c2= 0
=0
c1
= z, l
–c2
K
m+
mg K = vi
These two equations imply cl
m
= K
Vi –
mg K
and
c2
m
= – K z> > –
mg K
Once these values are substituted into (*), the complete solution t o the IVP can be written a s
=
m
g k)( 1 mgL [ (v'
J
1
The derivative of this expression gives the velocity of the sky diver t seconds after the parachute opens : v
mg y
K
1
Note that equation (***) is equivalent to the result of Example 5 (a s it should be) . The question asks for the minimum altitude at which the sky diver's parachute must be open in order to land at a velocity o f (1 .01)x 2 . Therefore, set v equal to (1 .01)v 2 in equation (* * *) an d solve for t ; then substitute the result into (**) to find the desire d altitude . Omitting the messy details, once the expression in (* * *) i s set equal to (1 .01)v 2, the value of t is found to b e
DIFFERENTIAL EQUATIONS
A FE W APPLICATIONS
v 1K - 1 m mg t =KIn 1 .01 v 2K 1 mg
v1K m mg K I n 0 .0 1
and substituting this result into (* *) yield s v 1K mg m mg v~ y — K K In 0 .01 + g
m K
0 .0 1 v iK 1 mg I
1
To evaluate the numerical answer, the following values from th e discussion in Skydiving (part I) are used : mass of sky diver : m = 70 k g gravitational acceleration :
g = 9 .8 m/s 2
air resistance proportionality constant : K = 110 kg/ s terminal velocity without parachute : v 1 = terminal velocity with parachute :
v2
/mg k
52 m/ s
mg = K 6 .2 m/ s
These substitutions give a descent time t [the time interval betwee n the parachute opening to the point where a speed of (1 .01)z 2 i s attained] of approximately 4 .2 seconds, and a minimum altitude a t which the parachute must be opened of y 55 meters (a littl e higher than 180 feet) . ■ Simple harmonic motion . Consider a spring fastened to a wall, wit h a block attached to its free end at rest on an essentially frictionles s horizontal table . The block can be set into motion by pulling o r pushing it from its original position and then letting go, or b y striking it (that is, by giving the block a nonzero initial velocity) . Th e
174
CLIFFS QUICK REVIEW
A FE W APPLICATION S
force exerted by the spring keeps the block oscillating on th e tabletop . This is the prototypical example of simple harmoni c motion . The force exerted by a spring is given by Hooke's Law; this state s that if a spring is stretched or compressed a distance x from it s natural length, then it exerts a force given by the equatio n F = —kx
The positive constant k is known as the spring constant and is directl y related to the spring's stiffness : The stiffer the spring, the larger th e value of k . The minus sign implies that when the spring is stretche d (so that x is positive), the spring pulls back (because F is negative) , and conversely, when the spring is compressed (so that x i s negative), the spring pushes outward (because F is positive) . Therefore, the spring is said to exert a restoring force, since it always tries to restore the block to its equilibrium position (the positio n where the spring is neither stretched nor compressed) . The restoring force here is proportional to the displacement (F = —kx « x) , and it is for this reason that the resulting periodic (regularl y repeating) motion is called simple harmonic. Newton 's Second Law can be applied to this spring-bloc k system . Once the block is set into motion, the only horizontal forc e that acts on it is the restoring force of the spring . Therefore, th e equation F1et = ma becomes —kx = mz, o r k z+—x= 0 m This is a homogeneous second-order linear equation with constan t coefficients . The auxiliary polynomial equation is r 2 + = 0, whic h has distinct conjugate complex roots ± i \lk/m . Therefore, the general solution of this differential equation is x = c l cos
DIFFERENTIAL EQUATIONS
mt + c 2 sin
k
t
(* )
A FE W APPLICATIONS
This expression gives the displacement of the block from it s equilibrium position (which is designated x = 0) .
Example 7 : A block of mass 1 kg is attached to a spring with forc e constant k = N/m . It is pulled m from its equilibrium positio n and released from rest . Obtain an equation for its position at an y time t ; then determine how long it takes the block to complete on e cycle (one round trip) . All that is required is to adapt equation (*) to the presen t situation . First, since the block is released from rest, its intia l velocity is 0 : X(0) = 0
–c l
k sin m
k k t+c — cos NiT ±t =0 m - m m t=0
c, = 0
Since c 2 = 0, equation (*) reduces to x = c, cos \/k/m t . Now, sinc e x(0) = +m, the remaining parameter can be evaluated : x(o)
=~~I
c, cos Vm t
3
t=o
10
c
=
3 10
Finally, since k = 45 N/m and m = 1 kg, \/k/m = \/25/4 = Therefore, the equation for the position of the block as a function o f time is given by x = ,3 ,
cos 5 t
where x is measured in meters from the equilibrium position of th e block . This function is periodic, which means it repeats itself a t regular intervals . The cosine and sine functions each have a period of 27r, which means every time the argument increases by 27r, th e function returns to its previous value . (Recall that if, say, x = cos 0 , then 0 is called the argument of the cosine function .) The argumen t
176
CLIFFS QUICK REVIEW
A FE W APPLICATION S
here is ;t, and ;t will increase by 2'rr every time t increases by 4'7 . Therefore, this block will complete one cycle, that is, return to it s original position (x = 10 m), every 5'rr 2 .5 seconds . ■
The length of time required to complete one cycle (one round trip) is called the period of the motion (and denoted by T) . It can b e shown in general that for the spring-block oscillator , \
T=2'rr
/-Wi
k
Note that the period does not depend on where the block started , only on its mass and the stiffness of the spring . The maximu m distance (greatest displacement) from equilibrium is called th e amplitude of the motion . Therefore, it makes no difference whethe r the block oscillates with an amplitude of 2 cm or 10 cm ; the period will be the same in either case . This is one of the definin g characteristics of simple harmonic motion : the period is independent of the amplitude . Another important characteristic of an oscillator is the numbe r of cycles that can be completed per unit time ; this is called th e frequency of the motion [denoted traditionally by v (the Gree k letter nu) but less confusingly by the letter f] . Since the perio d specifies the length of time per cycle, the number of cycles per unit time (the frequency) is simply the reciprocal of the period : f = 1/ T. Therefore, for the spring-block simple harmonic oscillator ,
f
1
Vic
2Tr
Frequency is usually expressed in equals 1 cycle per second .
DIFFERENTIAL EQUATIONS
m hertz
(abbreviated Hz) ; 1 H z
A FE W APPLICATIONS
The quantity s/k/m (the coefficient of t in the argument of th e sine and cosine in the general solution of the differential equatio n describing simple harmonic motion) appears so often in problem s of this type that it is given its own name and symbol . It is called th e angular frequency of the motion and denoted by w (the Greek lette r omega) . Note that w = 2Trf.
Damped oscillations . The spring-block oscillator discussed above i s an idealized example of a frictionless system . In real life, however , frictional (or dissipative) forces must be taken into account, particu larly if you want to model the behavior of the system over a lon g period of time . Unless the block slides back and forth on a frictionless table in a room evacuated of air, there will be resistanc e to the block's motion due to the air (just as there is for a falling sk y diver) . This resistance would be rather small, however, so you ma y want to picture the spring-block apparatus submerged in a larg e container of clear oil . The viscosity of the oil will have a profoun d effect upon the block's oscillations . The air (or oil) provides a damping force, which is proportional to the velocity of the object . (Again, recall the sky diver falling with a parachute . At th e relatively low speeds attained with an open parachute, the force du e to air resistance was given as Ky, which is proportional to th e velocity . ) With a restoring force given by –kx and a damping force give n by –Ky (the minus sign means that the damping force opposes th e velocity), Newton's Second Law (Fnet = ma) becomes –kx – Ky = ma, or, since v = .X and a = I , –kx – K.z=ml
This second-order linear differential equation with constant coefficients can be expressed in the more standard for m mx+KX+kx= 0
178
CLIFFS QUICK REVIEW
A FEW APPLICATIONS
The auxiliary polynomial equation is mr2 are r=
+ Kr +
k = 0, whose root s
–K ± JK2 – 4mk 2m
The system will exhibit periodic motion only if these roots ar e distinct conjugate complex numbers, because only then will th e general solution of the differential equation involve the periodi c functions sine and cosine . In order for this to be the case, th e discriminant K2 – 4mk must be negative ; that is, the damping constant K must be small ; specifically, it must be less than 2Jmk . When this happens, the motion is said to be underdamped, becaus e the damping is not so great as to prevent the system from oscillating ; it just causes the amplitude of the oscillations to gradually die out . [If the damping constant K is too great, then the discriminant i s nonnegative, the roots of the auxiliary polynomial equation are rea l (and negative), and the general solution of the differential equatio n involves only decaying exponentials . This implies there would be n o sustained oscillations . ] In the underdamped case (K < 2 Jmk), the roots of the auxiliary polynomial equation can be written a s
/4mk_K2
K 2m
r= – ±
2m
=–
K 1k K– ± 2m i m 4m2
and consequently, the general solution of the defining differentia l equation is X=e-(ki?m)t
Cocos
DIFFERENTIAL EQUATIONS
1k K2 -,t+c, sin m 4m-
k K2 ., t -m 4m -
179
A FE W APPLICATION S
Example 8 : (Compare to Example 7 .) A block of mass 1 kg i s attached to a spring with force constant k = N/m . It is pulled m from its equilibrium position and released from rest . If this spring-block apparatus is submerged in a viscous fluid mediu m which exerts a damping force of -4v (where v is the instantaneou s velocity of the block), sketch the curve that describes the position o f the block as a function of time . The net force on the block is — 4x — 4v = — x — 4.x, so Newton's Second Law become s
x+4.z+4x= 0 because m = 1 . Since the roots of the auxiliary polynomial equation , r2 = O, ar e r_—
-4±N/422 —4 . 1•--~ -4±3i -2±4 i 2 . 1 — 2 _— -
the general solution of the differential equation i s x=e - ''(c,
cos4t+c2sin4t)
Because the block is released from rest, v(0) = di(0) = 0 : [e- 2 '(-4c,
sin 4t + 4c2 cos 4t) — 2e-''(c, cos t +
c2
sin 4 t )],-,) = 0
This implies 4c2 — 2c l = 0. And, since x(0) = m , [e- 2'(c,
cos 4t +
c2
sin 4t)], =0 =
c,
c2 = (2c,) = 1(-h) = and the equation =Therfo, that gives th e position of the block as a function of time i s x = Tbe -2'(3 cos 4t + 4 sin 4t )
where x is measured in meters from the equilibrium position of th e block .
180
CLIFFS QUICK REVIEW
A FE W
APPLICATION S
This expression for the position function can be rewritten usin g the trigonometric identity cos(a — /3) = cos a cos /3+ sin a sin [3, a s follows: x = -Me -2'(3
cos 4t + 4 sin 4t)
;e-''[(cos _ ;e - ''[(cos x = ;e
cos
40(0 + (sin 3t)(O] 4t)(cos
(4t —
4) +
(sin 4t)(sin ~) ]
cf) )
The phase angle, 0, is defined here by the equations cos (P = and sin 4 = 5, or, more briefly, as the first-quadrant angle whose tangen t is 4 (it's the larger acute angle in a 3-4-5 right triangle) . The presence of the decaying exponential factor e -2 ' in the equation for x(t) means that as time passes (that is, as t increases), the amplitud e of the oscillations gradually dies out . See Figure 15 . envelope : x = Ae_R 2m) t position of block x
•,
time, t
■ Figurc 15
DIFFERENTIAL EQUATIONS
■
181
A FE W APPLICATION S
The angular frequency of this periodic motion is the coefficien t of t in the cosine, w' = rad/s, which implies a period o f 1 27 2-rr T ' =—_ =err 4.2 second s f' w' _ Compare this to Example 7, which described the same spring, block , and initial conditions but with no damping . The position functio n there was x = cos ;t, it had constant amplitude, an angula r frequency of w = rad/s, and a period of just 5 Tr 2 .5 seconds . Therefore, not only does (under)damping cause the amplitude t o gradually die out, but it also increases the period of the motion . But this seems reasonable : Damping reduces the speed of the block, so it takes longer to complete a round trip (hence the increase in th e period) . This will always happen in the case of underdamping, sinc e w' = \/k/m – K 2 /4m 2 will always be lower than w = \/k/m . ■
Electric circuits and resonance. When an electric circuit containin g an ac voltage source, an inductor, a capacitor, and a resistor i n series is analyzed mathematically, the equation that results is a second-order linear differential equation with constant coefficients . The voltage v(t) produced by the ac source will be expressed by th e equation z' = V sin wt, where V is the maximum voltage generated . An inductor is a circuit element that opposes changes in current , causing a voltage drop of L (di /dt), where i is the instantaneou s current and L is a proportionality constant known as the inductance . A capacitor stores charge, and when each plate carries a magnitud e of charge q, the voltage drop across the capacitor is q/C, where C i s a constant called the capacitance . Finally, a resistor opposes the flo w of current, creating a voltage drop equal to iR, where the constant R is the resistance . Kirchhoff's Loop Rule states that the algebraic su m of the voltage differences as one goes around any closed loop in a circuit is equal to zero . Therefore, if the voltage source, inductor , capacitor, and resistor are all in series, the n
182
CLIFFS QUICK REVIEW
A FE W
APPLICATIONS
di
q
t C –iR= 0
Vsinwt–Ld –
which can be rewritten a s L
di
dt
+Ri+
1
C q=
Vsin cut
Now, if an expression for 40—the current in the circuit as a function of time is desired, then the equation to be solved must b e written in terms of i . To this end, differentiate the equation directl y above, and use the definition i = dq/dt : L
d2i
di
1
dt , + Rdt + Ci =
wVcos cat
(* )
This differential equation governs the behavior of an LRC serie s circuit with a source of sinusoidally varying voltage . The first step in solving this equation is to obtain the genera l solution of the corresponding homogeneous equatio n d2i di 1 Ldt'-+Rdt+Ci=0
(** )
But notice that this differential equation has exactly the sam e mathematical form as the equation for the damped oscillator , m
d2x
dx
dt2 +Kdt
+kx=0
(*** )
By comparing the two equations, it is easy to see that the current (i ) is analogous to the position (x), the inductance (L) is analogous t o the mass (m), the resistance (R) is analogous to the dampin g constant (K), and the reciprocal capacitance (1/C) is analogous t o the spring constant (k) . Since the general solution of (* * *) wa s found to b e
DIFFERENTIAL EQUATIONS
183
-(x/ -"' f' c, cos
K K , t + c-) sin mkx=e — ,t m 4m
V4
ifK<2 m k the general solution of (* *) must be, by analogy ,
r~,_ e -(R/ 21-)r
Cl
1 _ R-t + c, sin cos LC 4L'
V
1 R-' _ t LC 4L' if R < 2 ~Ll c
But the solution does not end here . The original differentia l equation (*) for the LRC circuit was nonhomogeneous, so a particular solution must still be obtained . The family of th e nonhomogeneous right-hand term, c'V cos a, is {sin wt, cos a l , so a particular solution will have the for m i =A sin
a + B cos at
where A and B are the undetermined coefficients . Given thi s expression for i, it is easy to calculat e wA cos
a — (vB si n
—w '-Asinak—
(0 2 Bcosa t
Substituting these last three expressions into the given nonhomogeneous differential equation (*) yield s
I
—Ico ZL — 1
1
C~A — ujRB
l
J
sinwt
I 11 l + (wR~4 — I ce zL — C B cos at = ceV cos
1 J
Therefore, in order for this to be an identity, A and the simultaneous equations
B
a
must satisfy
CLIFFS QUICK REVIEW
I — w-L—C, A—(wR)B= 0 (wR)A—
1
W2L--
B=w V
The solution of this system i s w'R
A=
-WV
V B=
and
1 a) L — C, + (wR )'
w2L —
1 1W
2L —
C,
1 C
+ ((vR) 2
These expressions can be simplified by invoking the followin g standard definitions : ■ wL is called the inductive reactance and denoted X1. l
■, is called the capacitive reactance and denoted X( , ■ XI. — X( , is simply called the reactance and denoted X ■ ~Xj, - X( )' +
R2 =
+
R '- is called the impedance and
denoted Z Therefore, 1 '-
w2L—C
~
1
+(wR)_= wwL —wC + ((()R ) =wz
COL
= w2 [(XI,
1 WC +R-' XC,)2 + R2 ]
= w2 [X2 + R' ] =
DIFFERENTIAL EQUATIONS
.2 -z 2
M
and the expressions for the coefficientsA and B given above can b e written as A=
w-R V 1 C, + (wR ) w-L —
1 — wV w'L — C B= 1 w-L — C, + ((,)R)2
w'RV
RV
w'Z
Z-'
1 — wV • w wL -- wC
,
w-Z-
- w2 V(X1, — X( , ) VX — w'Z- - Z-' These simplifications yield the following particular solution of th e given nonhomogeneous differential equation : RV
VX
V/R
X
i = Z, sinat— Z, cos(t = Z~Zsinut—Zcosu t Combining this with the general solution of the correspondin g homogeneous equation gives the complete solution of the nonhomo geneous equation : i= ih + i o r i=
e-(R1'r_)t
c, cos
1 R' , t + c-) sin LC — 4L
1 R2 LC — ,t 4L V ~R
X
1
+Z Zsinuat —Zcoscet ~
Despite its rather formidable appearance, it lends itself easily t o analysis . The first term [the one with the exponential-decay facto r e -(Rl'1 )t] goes to zero as t increases, while the second term remain s indefinitely . For these reasons, the first term is known as th e transient current, and the second is called the steady-state current :
M
CLIFFS QUICK REVIEW
A FE W APPLICATIONS
i=
e-(R~ '--1.)' CI
1
cos
1
R'
LC, — 4 , t + c-) sin
—
R'
LC 4L
L
t
transient curren t (goes to 0)
vR
+ Z Z sin
ak
x
cos wt
steady-state curren t
Example 9: Consider the underdamped LRC series circuit discussed above . Once the transient current becomes so small that i t may be neglected, under what conditions will the amplitude of th e oscillating steady-state current be maximized? In particular, assuming that the inductance L, capacitance C, resistance R, and voltag e amplitude V are fixed, how should the angular frequency w of the voltage source be adjusted to maximize the steady-state current i n the circuit ? The steady-state curent is given by the equatio n V /R
1 =
X
Z I Z—Slll a — Z cos wt
1
By analogy with the phase-angle calculation in Example 8, thi s equation is rewritten as follows : V /R X 1 i =Z~Zsincot —Zcoscet l V R Z
X -+R'-
sin
V
wt —
X ~_ X '+R-)
= Z (cos ¢sin a — sin 0 cos V = Z—sin (cot —
DIFFERENTIAL EQUATIONS
~)
ut )
cos cot
(where cos 0 = R/ X-" + R 2 and sin 0 = X/ X--+R2) . Therefore , the amplitude of the steady-state current is VIZ, and, since V i s fixed, the way to maximize VIZ is to minimize Z . Becaus e Z= J -2+R2, Z will be minimized if X = U . And because w i s necessarily positive, X
OaX, .—Xc=OawL=
1
<-*
&J-
=
1 LC
WC
p w =
1
JC
This value of to is called the resonant angular frequency . When th e underdamped circuit is "tuned" to this value, the steady-stat e current is maximized, and the circuit is said to be in resonance . Thi s is the principle behind tuning a radio, the process of obtaining th e strongest response to a particular transmission . In this case, th e frequency (and therefore angular frequency) of the transmission i s fixed (an FM station may be broadcasting at a frequency of, say, 95 . 5 MHz, which actually means that it's broadcasting in a narrow band around 95 .5 MHz), and the value of the capacitance C or inductance L can be varied by turning a dial or pushing a button . According to the calculation above, resonance is achieved whe n
Therefore, in terms of a (relatively) fixed w and a variabl e capacitance, resonance will occur whe n C=
1
1
w'L
(2-rrf )2 L
(where f is the frequency of the broadcast) . Or in terms of a variabl e inductance, the circuitry will resonate to a particular station when L is adjusted to the value 1 L w-C
1 ( 21Tf) "C
CUFFS QUICK REVIEW