Differential Equations and Boundary
Differential Equations and Boundary
com
ebookmass.com
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
https://ebookmass.com/product/etextbook-978-1305965799-
differential-equations-with-boundary-value-problems/
https://ebookmass.com/product/elementary-differential-equations-
and-boundary-value-problems-12th-edition-william-e-boyce/
https://ebookmass.com/product/first-course-in-differential-
equations-11ed-differential-equations-and-boundary-value-
problems-9ed-solutions-manual-dennis-zill-roberto-martinez/
https://ebookmass.com/product/differential-equations-with-
boundary-value-problems-9th-edition-dennis-g-zill/
Boundary value problems for systems of differential,
difference and fractional equations : positive
solutions 1st Edition Johnny Henderson
https://ebookmass.com/product/boundary-value-problems-for-
systems-of-differential-difference-and-fractional-equations-
positive-solutions-1st-edition-johnny-henderson/
https://ebookmass.com/product/etextbook-978-1305965720-a-first-
course-in-differential-equations-with-modeling-applications/
https://ebookmass.com/product/transforms-and-partial-
differential-equations-t-veerarajan/
https://ebookmass.com/product/linear-algebra-and-partial-
differential-equations-t-veerarajan/
https://ebookmass.com/product/a-first-course-in-differential-
equations-with-modeling-applications-12e-metric-edition-dennis-g-
zill/
PREFACE
vii
viii Preface
see the application module for Section 5.3 (page 319). The number of terms
used from an infinite series solution of a differential equation can be varied,
and the resulting graphical change in the corresponding approximate solution
is shown immediately; see the Section 8.2 application module (page 516).
New Exposition In a number of sections, new text and graphics have been
inserted to enhance student understanding of the subject matter. For instance,
see the treatments of separable equations in Section 1.4 (page 30), linear equa-
tions in Section 1.5 (page 45), isolated critical points in Sections 6.1 (page
372) and 6.2 (page 383), and the new example in Section 9.6 (page 618)
showing a vibrating string with a momentary “flat spot.” Examples and ac-
companying graphics have been updated in Sections 2.4–2.6, 4.2, and 4.3 to
illustrate new graphing calculators.
New Content The single entirely new section for this edition is Section
5.3, which is devoted to the construction of a “gallery” of phase plane por-
traits illustrating all the possible geometric behaviors of solutions of the 2-
dimensional linear system x0 D Ax. In motivation and preparation for the
detailed study of eigenvalue-eigenvector methods in subsequent sections of
Chapter 5 (which then follow in the same order as in the previous edi-
tion), Section 5.3 shows how the particular arrangements of eigenvalues and
eigenvectors of the coefficient matrix A correspond to identifiable patterns—
“fingerprints,” so to speak—in the phase plane portrait of the system x0 D Ax.
The resulting gallery is shown in the two pages of phase plane portraits that
comprise Figure 5.3.16 (pages 315-316) at the end of the section. The new 5.3
application module (on dynamic phase plane portraits, page 319) shows how
students can use interactive computer systems to “bring to life” this gallery, by
allowing initial conditions, eigenvalues, and even eigenvectors to vary in real
time. This dynamic approach is then illustrated with several new graphics in-
serted in the remainder of Chapter 5. Finally, for a new biological application,
see the application module for Section 6.4, which now includes a substan-
tial investigation (page 423) of the nonlinear FitzHugh-Nagumo equations in
neuroscience, which were introduced to model the behavior of neurons in the
nervous system.
Computing Features
The following features highlight the computing technology that distinguishes much
of our exposition.
Modeling Features
Mathematical modeling is a goal and constant motivation for the study of differen-
tial equations. To sample the range of applications in this text, take a look at the
following questions:
What explains the commonly observed time lag between indoor and outdoor
daily temperature oscillations? (Section 1.5)
What makes the difference between doomsday and extinction in alligator pop-
ulations? (Section 2.1)
How do a unicycle and a twoaxle car react differently to road bumps? (Sec-
tions 3.7 and 5.4)
How can you predict the time of next perihelion passage of a newly observed
comet? (Section 4.3)
Why might an earthquake demolish one building and leave standing the one
next door? (Section 5.4)
What determines whether two species will live harmoniously together, or
whether competition will result in the extinction of one of them and the sur-
vival of the other? (Section 6.3)
Why and when does non-linearity lead to chaos in biological and mechanical
systems? (Section 6.5)
If a mass on a spring is periodically struck with a hammer, how does the
behavior of the mass depend on the frequency of the hammer blows? (Section
7.6)
Why are flagpoles hollow instead of solid? (Section 8.6)
What explains the difference in the sounds of a guitar, a xylophone, and drum?
(Sections 9.6, 10.2, and 10.4)
We have reshaped the usual approach and sequence of topics to accommodate new
technology and new perspectives. For instance:
treatment of the linear algebra that is needed, and then presents the eigenvalue
approach to linear systems. It includes a wide range of applications (ranging
from railway cars to earthquakes) of all the various cases of the eigenvalue
method. Section 5.5 includes a fairly extensive treatment of matrix exponen-
tials, which are exploited in Section 5.6 on nonhomogeneous linear systems.
Chapter 6 on nonlinear systems and phenomena ranges from phase plane anal-
ysis to ecological and mechanical systems to a concluding section on chaos
and bifurcation in dynamical systems. Section 6.5 presents an elementary in-
troduction to such contemporary topics as period-doubling in biological and
mechanical systems, the pitchfork diagram, and the Lorenz strange attractor
(all illustrated with vivid computer graphics).
Laplace transform methods (Chapter 7) and power series methods (Chapter 8)
follow the material on linear and nonlinear systems, but can be covered at any
earlier point (after Chapter 3) the instructor desires.
Chapters 9 and 10 treat the applications of Fourier series, separation of vari-
ables, and Sturm-Liouville theory to partial differential equations and bound-
ary value problems. After the introduction of Fourier series, the three clas-
sical equations—the wave and heat equations and Laplace’s equation—are
discussed in the last three sections of Chapter 9. The eigenvalue methods of
Chapter 10 are developed sufficiently to include some rather significant and
realistic applications.
This book includes enough material appropriately arranged for different courses
varying in length from one quarter to two semesters. The briefer version Differen-
tial Equations: Computing and Modeling (0-321-81625-0) ends with Chapter 7 on
Laplace transform methods (and thus omits the material on power series methods,
Fourier series, separation of variables and partial differential equations).
The answer section has been expanded considerably to increase its value as a learn-
ing aid. It now includes the answers to most odd-numbered problems plus a good
many even-numbered ones. The Instructor’s Solutions Manual (0-321-79701-
9) available at www.pearsonhighered.com/irc provides worked-out solutions
for most of the problems in the book, and the Student Solutions Manual (0-321-
79700-0) contains solutions for most of the odd-numbered problems. These manu-
als have been reworked extensively for this edition with improved explanations and
more details inserted in the solutions of many problems.
The approximately 45 application modules in the text contain additional prob-
lem and project material designed largely to engage students in the exploration
and application of computational technology. These investigations are expanded
considerably in the Applications Manual (0-321-79704-3) that accompanies the
text and supplements it with additional and sometimes more challenging investi-
gations. Each section in this manual has parallel subsections Using Maple, Using
Mathematica, and Using MATLAB that detail the applicable methods and tech-
niques of each system, and will afford student users an opportunity to compare the
merits and styles of different computational systems. These materials—as well as
the text of the Applications Manual itself—are freely available at the web site
www.pearsonhighered.com/mathstatsresources.
Preface xi
Acknowledgments
In preparing this revision, we profited greatly from the advice and assistance of the
following very capable and perceptive reviewers:
It is a pleasure to (once again) credit Dennis Kletzing and his extraordinary TEXpertise
for the attractive presentation of the text and the art in this book. We are grateful
to our editor, William Hoffman, for his support and inspiration of this revision; to
Salena Casha for her coordination of the editorial process and Beth Houston for her
supervision of the production of this book; and to Joe Vetere for his assistance with
technical aspects of the development of its extensive supplementary resources. Fi-
nally, we dedicate this edition to our colleague David E. Penney who passed away
on June 3, 2014.
1
2 Chapter 1 First-Order Differential Equations
are challenged to find the unknown functions y D y.x/ for which an identity such
as y 0 .x/ D 2xy.x/—that is, the differential equation
dy
D 2xy
dx
—holds on some interval of real numbers. Ordinarily, we will want to find all
solutions of the differential equation, if possible.
Example 2 If C is a constant and
2
y.x/ D C e x ; (1)
then
dy 2
2
D C 2xe x D .2x/ C e x D 2xy:
dx
Thus every function y.x/ of the form in Eq. (1) satisfies—and thus is a solution of—the
differential equation
dy
D 2xy (2)
dx
for all x . In particular, Eq. (1) defines an infinite family of different solutions of this differen-
tial equation, one for each choice of the arbitrary constant C . By the method of separation of
variables (Section 1.4) it can be shown that every solution of the differential equation in (2)
is of the form in Eq. (1).
dT
D k.T A/; (3)
dt
Temperature T
where k is a positive constant. Observe that if T > A, then d T=dt < 0, so the temperature is
a decreasing function of t and the body is cooling. But if T < A, then d T=dt > 0, so that T
is increasing.
FIGURE 1.1.1. Newton’s law of Thus the physical law is translated into a differential equation. If we are given the
cooling, Eq. (3), describes the cooling values of k and A, we should be able to find an explicit formula for T .t /, and then—with the
of a hot rock in water. aid of this formula—we can predict the future temperature of the body.
Example 4 Torricelli’s law implies that the time rate of change of the volume V of water in a draining
tank (Fig. 1.1.2) is proportional to the square root of the depth y of water in the tank:
dV p
D k y; (4)
dt
where k is a constant. If the tank is a cylinder with vertical sides and cross-sectional area A,
then V D Ay , so d V =dt D A .dy=dt /. In this case Eq. (4) takes the form
dy p
D h y; (5)
dt
where h D k=A is a constant.
1.1 Differential Equations and Mathematical Models 3
Example 5 The time rate of change of a population P .t / with constant birth and death rates is, in many
simple cases, proportional to the size of the population. That is,
dP
D kP; (6)
dt
where k is the constant of proportionality.
Let us discuss Example 5 further. Note first that each function of the form
Volume V y
P .t / D C e kt (7)
for all real numbers t . Because substitution of each function of the form given in
(7) into Eq. (6) produces an identity, all such functions are solutions of Eq. (6).
Thus, even if the value of the constant k is known, the differential equation
dP=dt D kP has infinitely many different solutions of the form P .t / D C e kt , one for
each choice of the “arbitrary” constant C . This is typical of differential equations.
It is also fortunate, because it may allow us to use additional information to select
from among all these solutions a particular one that fits the situation under study.
Example 6 Suppose that P .t / D C e kt is the population of a colony of bacteria at time t , that the pop-
ulation at time t D 0 (hours, h) was 1000, and that the population doubled after 1 h. This
additional information about P .t / yields the following equations:
1000 D P .0/ D C e 0 D C;
2000 D P .1/ D C e k :
It follows that C D 1000 and that e k D 2, so k D ln 2 0:693147. With this value of k the
differential equation in (6) is
dP
D .ln 2/P .0:693147/P:
dt
Substitution of k D ln 2 and C D 1000 in Eq. (7) yields the particular solution
C = 12 C = 6 C = 3 P .t / D 1000e .ln 2/t D 1000.e ln 2 /t D 1000 2t (because e ln 2 D 2)
8
C=1 that satisfies the given conditions. We can use this particular solution to predict future popu-
6
4 lations of the bacteria colony. For instance, the predicted number of bacteria in the population
after one and a half hours (when t D 1:5) is
2 1
C=
0 2 P .1:5/ D 1000 23=2 2828:
P
C = – 12
–2
The condition P .0/ D 1000 in Example 6 is called an initial condition because
–4 we frequently write differential equations for which t D 0 is the “starting time.”
C = –1
–6 Figure 1.1.3 shows several different graphs of the form P .t / D C e kt with k D ln 2.
–8 The graphs of all the infinitely many solutions of dP=dt D kP in fact fill the entire
–2 –1 0 1 2 3
C = –12 C = –6 C = –3 two-dimensional plane, and no two intersect. Moreover, the selection of any one
t
point P0 on the P -axis amounts to a determination of P .0/. Because exactly one
FIGURE 1.1.3. Graphs of solution passes through each such point, we see in this case that an initial condition
P .t / D C e kt with k D ln 2. P .0/ D P0 determines a unique solution agreeing with the given data.
4 Chapter 1 First-Order Differential Equations
Mathematical Models
Our brief discussion of population growth in Examples 5 and 6 illustrates the crucial
process of mathematical modeling (Fig. 1.1.4), which involves the following:
Real-world
situation
Formulation Interpretation
pare with the real-world population. Indeed, a successful population analysis may
require refining the mathematical model still further as it is repeatedly measured
against real-world experience.
But in Example 6 we simply ignored any complicating factors that might af-
fect our bacteria population. This made the mathematical analysis quite simple,
perhaps unrealistically so. A satisfactory mathematical model is subject to two con-
tradictory requirements: It must be sufficiently detailed to represent the real-world
situation with relative accuracy, yet it must be sufficiently simple to make the math-
ematical analysis practical. If the model is so detailed that it fully represents the
physical situation, then the mathematical analysis may be too difficult to carry out.
If the model is too simple, the results may be so inaccurate as to be useless. Thus
there is an inevitable tradeoff between what is physically realistic and what is math-
ematically possible. The construction of a model that adequately bridges this gap
between realism and feasibility is therefore the most crucial and delicate step in
the process. Ways must be found to simplify the model mathematically without
sacrificing essential features of the real-world situation.
Mathematical models are discussed throughout this book. The remainder of
this introductory section is devoted to simple examples and to standard terminology
used in discussing differential equations and their solutions.
4x 2 y 00 C y D 0 (10)
The fact that we can write a differential equation is not enough to guarantee
that it has a solution. For example, it is clear that the differential equation
.y 0 /2 C y 2 D 1 (11)
.y 0 /2 C y 2 D 0 (12)
obviously has only the (real-valued) solution y.x/ 0. In our previous examples
any differential equation having at least one solution indeed had infinitely many.
The order of a differential equation is the order of the highest derivative that
appears in it. The differential equation of Example 8 is of second order, those in
Examples 2 through 7 are first-order equations, and
on I and
F x; u; u0 ; u00 ; : : : ; u.n/ D 0
for all x in I . For the sake of brevity, we may say that u D u.x/ satisfies the
–5
–5 0 5 differential equation in (13) on I .
x
Remark Recall from elementary calculus that a differentiable function on an open interval
FIGURE 1.1.5. The solution of is necessarily continuous there. This is why only a continuous function can qualify as a
y 0 D y 2 defined by y.x/ D 1=.1 x/. (differentiable) solution of a differential equation on an interval.
Example 7 Figure 1.1.5 shows the two “connected” branches of the graph y D 1=.1 x/. The left-hand
Continued branch is the graph of a (continuous) solution of the differential equation y 0 D y 2 that is
defined on the interval .1; 1/. The right-hand branch is the graph of a different solution of
the differential equation that is defined (and continuous) on the different interval .1; 1/. So
the single formula y.x/ D 1=.1 x/ actually defines two different solutions (with different
domains of definition) of the same differential equation y 0 D y 2 .
Example 9 If A and B are constants and
y.x/ D A cos 3x C B sin 3x; (14)
then two successive differentiations yield
y 0 .x/ D 3A sin 3x C 3B cos 3x;
y 00 .x/ D 9A cos 3x 9B sin 3x D 9y.x/
for all x . Consequently, Eq. (14) defines what it is natural to call a two-parameter family of
solutions of the second-order differential equation
y 00 C 9y D 0 (15)
on the whole real number line. Figure 1.1.6 shows the graphs of several such solutions.
1.1 Differential Equations and Mathematical Models 7
Although the differential equations in (11) and (12) are exceptions to the gen-
eral rule, we will see that an nth-order differential equation ordinarily has an n-
5 parameter family of solutions—one involving n different arbitrary constants or pa-
y3 rameters.
y1
In both Eqs. (11) and (12), the appearance of y 0 as an implicitly defined func-
y2
tion causes complications. For this reason, we will ordinarily assume that any dif-
ferential equation under study can be solved explicitly for the highest derivative that
0 appears; that is, that the equation can be written in the so-called normal form
y
y .n/ D G x; y; y 0 ; y 00 ; : : : ; y .n1/ ; (16)
–5
–3 0 3 where G is a real-valued function of n C 1 variables. In addition, we will always
x seek only real-valued solutions unless we warn the reader otherwise.
FIGURE 1.1.6. The three solutions All the differential equations we have mentioned so far are ordinary differ-
y1 .x/ D 3 cos 3x , y2 .x/ D 2 sin 3x , ential equations, meaning that the unknown function (dependent variable) depends
and y3 .x/ D 3 cos 3x C 2 sin 3x of on only a single independent variable. If the dependent variable is a function of
the differential equation y 00 C 9y D 0.
two or more independent variables, then partial derivatives are likely to be involved;
if they are, the equation is called a partial differential equation. For example, the
temperature u D u.x; t / of a long thin uniform rod at the point x at time t satisfies
(under appropriate simple conditions) the partial differential equation
@u @2 u
D k 2;
@t @x
dy
D f .x; y/: (17)
dx
We also will sample the wide range of applications of such equations. A typical
mathematical model of an applied situation will be an initial value problem, con-
sisting of a differential equation of the form in (17) together with an initial condi-
tion y.x0 / D y0 . Note that we call y.x0 / D y0 an initial condition whether or not
x0 D 0. To solve the initial value problem
dy
D f .x; y/; y.x0 / D y0 (18)
dx
y = 2/(3 – 2x) 1 2
y.x/ D D :
(1, 2)
3
2 x 3 2x
x = 3/2 Figure 1.1.7 shows the two branches of the graph y D 2=.3 2x/. The left-hand branch is
0
the graph on .1; 23 / of the solution of the given initial value problem y 0 D y 2 , y.1/ D 2.
y
The right-hand branch passes through the point .2; 2/ and is therefore the graph on . 32 ; 1/
(2, –2)
of the solution of the different initial value problem y 0 D y 2 , y.2/ D 2.
The central question of greatest immediate interest to us is this: If we are given
–5 a differential equation known to have a solution satisfying a given initial condition,
–5 0 5
x how do we actually find or compute that solution? And, once found, what can we do
FIGURE 1.1.7. The solutions of
with it? We will see that a relatively few simple techniques—separation of variables
y 0 D y 2 defined by (Section 1.4), solution of linear equations (Section 1.5), elementary substitution
y.x/ D 2=.3 2x/. methods (Section 1.6)—are enough to enable us to solve a variety of first-order
equations having impressive applications.
1.1 Problems
In Problems 1 through 12, verify by substitution that each 20. y 0 D x y ; y.x/ D C e x C x 1, y.0/ D 10
3
given function is a solution of the given differential equation. 21. y 0 C 3x 2 y D 0; y.x/ D C e x , y.0/ D 7
Throughout these problems, primes denote derivatives with re- 22. e y y 0 D 1; y.x/ D ln.x C C /, y.0/ D 0
spect to x . dy
23. x C 3y D 2x 5 ; y.x/ D 14 x 5 C C x 3 , y.2/ D 1
1. y 0 D 3x 2 ; y D x 3 C 7 dx
2. y 0 C 2y D 0; y D 3e 2x 24. xy 0 3y D x 3 ; y.x/ D x 3 .C C ln x/, y.1/ D 17
3. y 00 C 4y D 0; y1 D cos 2x , y2 D sin 2x 25. y 0 D 3x 2 .y 2 C 1/; y.x/ D tan.x 3 C C /, y.0/ D 1
4. y 00 D 9y ; y1 D e 3x , y2 D e 3x 26. y 0 C y tan x D cos x ; y.x/ D .x C C / cos x , y./ D 0
5. y 0 D y C 2e x ; y D e x e x
In Problems 27 through 31, a function y D g.x/ is described
6. y 00 C 4y 0 C 4y D 0; y1 D e 2x , y2 D xe 2x
by some geometric property of its graph. Write a differential
7. y 00 2y 0 C 2y D 0; y1 D e x cos x , y2 D e x sin x equation of the form dy=dx D f .x; y/ having the function g as
8. y 00 C y D 3 cos 2x , y1 D cos x cos 2x , y2 D sin x cos 2x its solution (or as one of its solutions).
1
9. y 0 C 2xy 2 D 0; y D 27. The slope of the graph of g at the point .x; y/ is the sum
1 C x2
1 of x and y .
10. x 2 y 00 C xy 0 y D ln x ; y1 D x ln x , y2 D ln x
x 28. The line tangent to the graph of g at the point .x; y/ inter-
1 ln x sects the x -axis at the point .x=2; 0/.
11. x 2 y 00 C 5xy 0 C 4y D 0; y1 D 2 , y2 D 2
x x 29. Every straight line normal to the graph of g passes through
12. x 2 y 00 xy 0 C 2y D 0; y1 D x cos.ln x/, y2 D x sin.ln x/ the point .0; 1/. Can you guess what the graph of such a
function g might look like?
In Problems 13 through 16, substitute y D e rx into the given
30. The graph of g is normal to every curve of the form
differential equation to determine all values of the constant r
y D x 2 C k (k is a constant) where they meet.
for which y D e rx is a solution of the equation.
31. The line tangent to the graph of g at .x; y/ passes through
13. 3y 0 D 2y 14. 4y 00 D y the point .y; x/.
15. y 00 C y 0 2y D 0 16. 3y 00 C 3y 0 4y D 0
In Problems 32 through 36, write—in the manner of Eqs. (3)
In Problems 17 through 26, first verify that y.x/ satisfies the through (6) of this section—a differential equation that is a
given differential equation. Then determine a value of the con- mathematical model of the situation described.
stant C so that y.x/ satisfies the given initial condition. Use a
32. The time rate of change of a population P is proportional
computer or graphing calculator (if desired) to sketch several
to the square root of P .
typical solutions of the given differential equation, and high-
light the one that satisfies the given initial condition. 33. The time rate of change of the velocity v of a coasting
motorboat is proportional to the square of v .
17. y 0 C y D 0; y.x/ D C e x , y.0/ D 2 34. The acceleration dv=dt of a Lamborghini is proportional
18. y 0 D 2y ; y.x/ D C e 2x , y.0/ D 3 to the difference between 250 km/h and the velocity of the
19. y 0 D y C 1; y.x/ D C e x 1, y.0/ D 5 car.
1.1 Differential Equations and Mathematical Models 9
35. In a city having a fixed population of P persons, the time rodents, and their number is increasing at the rate of
rate of change of the number N of those persons who have dP=dt D 1 rodent per month when there are P D 10 ro-
heard a certain rumor is proportional to the number of dents. Based on the result of Problem 43, how long will it
those who have not yet heard the rumor. take for this population to grow to a hundred rodents? To
36. In a city with a fixed population of P persons, the time rate a thousand? What’s happening here?
of change of the number N of those persons infected with 46. Suppose the velocity v of a motorboat coasting in water
a certain contagious disease is proportional to the product satisfies the differential equation dv=dt D kv 2 . The ini-
of the number who have the disease and the number who tial speed of the motorboat is v.0/ D 10 meters per sec-
do not. ond (m/s), and v is decreasing at the rate of 1 m/s2 when
v D 5 m/s. Based on the result of Problem 43, long does
In Problems 37 through 42, determine by inspection at least it take for the velocity of the boat to decrease to 1 m/s? To
one solution of the given differential equation. That is, use 1
10 m/s? When does the boat come to a stop?
your knowledge of derivatives to make an intelligent guess. 47. In Example 7 we saw that y.x/ D 1=.C x/ defines a
Then test your hypothesis. one-parameter family of solutions of the differential equa-
37. y 00 D 0 38. y 0 D y tion dy=dx D y 2 . (a) Determine a value of C so that
39. xy 0 C y D 3x 2 40. .y 0 /2 C y 2 D 1 y.10/ D 10. (b) Is there a value of C such that y.0/ D 0?
41. y 0 C y D e x 42. y 00 C y D 0 Can you nevertheless find by inspection a solution of
dy=dx D y 2 such that y.0/ D 0? (c) Figure 1.1.8 shows
Problems 43 through 46 concern the differential equation typical graphs of solutions of the form y.x/ D 1=.C x/.
Does it appear that these solution curves fill the entire xy -
dx plane? Can you conclude that, given any point .a; b/ in
D kx 2 ;
dt the plane, the differential equation dy=dx D y 2 has ex-
where k is a constant. actly one solution y.x/ satisfying the condition y.a/ D b ?
48. (a) Show that y.x/ D C x 4 defines a one-parameter fam-
43. (a) If k is a constant, show that a general (one-parameter) ily of differentiable solutions of the differential equation
solution of the differential equation is given by x.t / D xy 0 D 4y (Fig. 1.1.9). (b) Show that
1=.C k t /, where C is an arbitrary constant.
(
(b) Determine by inspection a solution of the initial value x 4 if x < 0,
problem x 0 D kx 2 , x.0/ D 0. y.x/ D
x 4 if x = 0
44. (a) Assume that k is positive, and then sketch graphs of
solutions of x 0 D kx 2 with several typical positive defines a differentiable solution of xy 0 D 4y for all x , but is
values of x.0/. not of the form y.x/ D C x 4 . (c) Given any two real num-
(b) How would these solutions differ if the constant k bers a and b , explain why—in contrast to the situation in
were negative? part (c) of Problem 47—there exist infinitely many differ-
45. Suppose a population P of rodents satisfies the differen- entiable solutions of xy 0 D 4y that all satisfy the condition
tial equation dP=dt D kP 2 . Initially, there are P .0/ D 2 y.a/ D b .
C = –2 C = –1 C = 0 C = 1 C = 2 C = 3
3
100
2 80
60
1 C=4 40
20
0
y
0
y
C = –4
–1 –20
– 40
–2 –60
–80
–3
–3 –2 –1 0 1 2 3 –100
C = –3 C = –2 C = –1 C = 0 C = 1 C = 2 –5 – 4 –3 –2 –1 0 1 2 3 4 5
x x
FIGURE 1.1.8. Graphs of solutions of the FIGURE 1.1.9. The graph y D C x 4 for
equation dy=dx D y 2 . various values of C .
10 Chapter 1 First-Order Differential Equations
–1
–2 This is a general solution of Eq. (1), meaning that it involves an arbitrary constant
C = –3
–3 C , and for every choice of C it is a solution of the differential equation in (1). If
–4 G.x/ is a particular antiderivative of f —that is, if G 0 .x/ f .x/—then
– 4 –3 –2 –1 0 1 2 3 4
x
y.x/ D G.x/ C C: (3)
FIGURE 1.2.1. Graphs of
y D 14 x 2 C C for various values of C . The graphs of any two such solutions y1 .x/ D G.x/ C C1 and y2 .x/ D G.x/ C
C2 on the same interval I are “parallel” in the sense illustrated by Figs. 1.2.1 and
1.2.2. There we see that the constant C is geometrically the vertical distance be-
6 tween the two curves y.x/ D G.x/ and y.x/ D G.x/ C C .
C=4 To satisfy an initial condition y.x0 / D y0 , we need only substitute x D x0 and
4
C=2 y D y0 into Eq. (3) to obtain y0 D G.x0 / C C , so that C D y0 G.x0 /. With this
2
choice of C , we obtain the particular solution of Eq. (1) satisfying the initial value
C=0 problem
0 dy
y
C = –2 D f .x/; y.x0 / D y0 :
dx
–2
We will see that this is the typical pattern for solutions of first-order differential
–4 equations. Ordinarily, we will first find a general solution involving an arbitrary
C = –4
constant C . We can then attempt to obtain, by appropriate choice of C , a particular
–6
–6 –4 –2 0 2 4 6 solution satisfying a given initial condition y.x0 / D y0 .
x
Remark As the term is used in the previous paragraph, a general solution of a first-order
FIGURE 1.2.2. Graphs of differential equation is simply a one-parameter family of solutions. A natural question is
y D sin x C C for various values of C . whether a given general solution contains every particular solution of the differential equa-
tion. When this is known to be true, we call it the general solution of the differential equation.
For example, because any two antiderivatives of the same function f .x/ can differ only by a
constant, it follows that every solution of Eq. (1) is of the form in (2). Thus Eq. (2) serves to
define the general solution of (1).
Figure 1.2.3 shows the graph y D x 2 C 3x C C for various values of C . The particular solution
we seek corresponds to the curve that passes through the point .1; 2/, thereby satisfying the
initial condition
y.1/ D .1/2 C 3 .1/ C C D 2:
It follows that C D 2, so the desired particular solution is
y.x/ D x 2 C 3x 2:
1.2 Integrals as General and Particular Solutions 11
–4
C = –2
–6
C = –4
in which the function g on the right-hand side involves neither the dependent vari-
–8 able y nor its derivative dy=dx . We simply integrate once to obtain
C = –6
–10 Z Z
–6 –4 –2 0 2 4 dy 00
x D y .x/ dx D g.x/ dx D G.x/ C C1 ;
dx
FIGURE 1.2.3. Solution curves for
the differential equation in Example 1. where G is an antiderivative of g and C1 is an arbitrary constant. Then another
integration yields
Z Z Z
y.x/ D y 0 .x/ dx D ŒG.x/ C C1 dx D G.x/ dx C C1 x C C2 ;
dx
v.t / D f 0 .t /I that is, vD : (6)
dt
dv d 2x
aD D : (7)
dt dt 2
R Equation (6) is sometimes applied either in the indefinite integral form x.t / D
v.t/ dt or in the definite integral form
Z t
x.t / D x.t0 / C v.s/ ds;
t0
where m is the mass of the particle. If the force F is known, then the equation
x 00 .t / D F .t /=m can be integrated twice to find the position function x.t / in terms
of two constants of integration. These two arbitrary constants are frequently deter-
mined by the initial position x0 D x.0/ and the initial velocity v0 D v.0/ of the
particle.
Constant acceleration. For instance, suppose that the force F , and therefore the
acceleration a D F=m, are constant. Then we begin with the equation
dv
Da (a is a constant) (9)
dt
and integrate both sides to obtain
Z
v.t / D a dt D at C C1 :
x.t / D 21 at 2 C v0 t C x0 : (11)
Thus, with Eq. (10) we can find the velocity, and with Eq. (11) the position, of
the particle at any time t in terms of its constant acceleration a, its initial velocity
v0 , and its initial position x0 .
Example 2 A lunar lander is falling freely toward the surface of the moon at a speed of 450 meters per
second (m=s). Its retrorockets, when fired, provide a constant deceleration of 2.5 meters per
second per second (m=s2 ) (the gravitational acceleration produced by the moon is assumed
to be included in the given deceleration). At what height above the lunar surface should the
retrorockets be activated to ensure a “soft touchdown” (v D 0 at impact)?
Solution We denote by x.t / the height of the lunar lander above the surface, as indicated in Fig. 1.2.4.
We let t D 0 denote the time at which the retrorockets should be fired. Then v0 D 450
(m=s, negative because the height x.t / is decreasing), and a D C2:5, because an upward
thrust increases the velocity v (although it decreases the speed jvj). Then Eqs. (10) and (11)
become
v.t / D 2:5t 450 (12)
a υ and
x.t / D 1:25t 2 450t C x0 ; (13)
Lunar surface
where x0 is the height of the lander above the lunar surface at the time t D 0 when the
retrorockets should be activated.
FIGURE 1.2.4. The lunar lander of From Eq. (12) we see that v D 0 (soft touchdown) occurs when t D 450=2:5 D 180 s
Example 2. (that is, 3 minutes); then substitution of t D 180, x D 0 into Eq. (13) yields
x0 D 0 .1:25/.180/2 C 450.180/ D 40;500
meters—that is, x0 D 40.5 km 25 16 miles. Thus the retrorockets should be activated when
the lunar lander is 40.5 kilometers above the surface of the moon, and it will touch down
softly on the lunar surface after 3 minutes of decelerating descent.
1.2 Integrals as General and Particular Solutions 13
Physical Units
Numerical work requires units for the measurement of physical quantities such as
distance and time. We sometimes use ad hoc units—such as distance in miles or
kilometers and time in hours—in special situations (such as in a problem involving
an auto trip). However, the foot-pound-second (fps) and meter-kilogram-second
(mks) unit systems are used more generally in scientific and engineering problems.
In fact, fps units are commonly used only in the United States (and a few other
countries), while mks units constitute the standard international system of scientific
units.
The last line of this table gives values for the gravitational acceleration g at
the surface of the earth. Although these approximate values will suffice for most
examples and problems, more precise values are 9:7805 m=s2 and 32:088 ft=s2 (at
sea level at the equator).
Both systems are compatible with Newton’s second law F D ma. Thus 1 N is
(by definition) the force required to impart an acceleration of 1 m=s2 to a mass of 1
kg. Similarly, 1 slug is (by definition) the mass that experiences an acceleration of
1 ft=s2 under a force of 1 lb. (We will use mks units in all problems requiring mass
units and thus will rarely need slugs to measure mass.)
Inches and centimeters (as well as miles and kilometers) also are commonly
used in describing distances. For conversions between fps and mks units it helps to
remember that
For instance,
cm
1 ft D 12 in. 2:54 D 30.48 cm;
in.
and it follows that
cm
1 mi D 5280 ft 30:48 D 160934.4 cm 1.609 km:
ft
Thus a posted U.S. speed limit of 50 mi=h means that—in international terms—the
legal speed limit is about 50 1:609 80:45 km=h.
W D mg (14)
14 Chapter 1 First-Order Differential Equations
for the weight W of the mass m at the surface of the earth (where g 32 ft=s2 9:8
m=s2 ). For instance, a mass of m D 20 kg has a weight of W D (20 kg)(9.8 m=s2 ) D
196 N. Similarly, a mass m weighing 100 pounds has mks weight
so its mass is
W 444.8 N
mD D 45.4 kg:
g 9.8 m=s2
To discuss vertical motion it is natural to choose the y -axis as the coordinate
system for position, frequently with y D 0 corresponding to “ground level.” If we
choose the upward direction as the positive direction, then the effect of gravity on a
vertically moving body is to decrease its height and also to decrease its velocity v D
dy=dt . Consequently, if we ignore air resistance, then the acceleration a D dv=dt of
the body is given by
dv
D g: (15)
dt
This acceleration equation provides a starting point in many problems involving
vertical motion. Successive integrations (as in Eqs. (10) and (11)) yield the velocity
and height formulas
y.t / D 12 gt 2 C v0 t C y0 : (17)
Here, y0 denotes the initial (t D 0) height of the body and v0 its initial velocity.
Example 3 (a) Suppose that a ball is thrown straight upward from the ground (y0 D 0) with initial
velocity v0 D 96 (ft=s, so we use g D 32 ft=s2 in fps units). Then it reaches its maximum
height when its velocity (Eq. (16)) is zero,
v.t / D 32t C 96 D 0;
and thus when t D 3 s. Hence the maximum height that the ball attains is
y.3/ D 12 32 32 C 96 3 C 0 D 144 (ft)
FIGURE 1.2.5. A swimmer’s You can use Eq. (18) to verify that the water does flow the fastest at the center,
problem (Example 4). where vR D v0 , and that vR D 0 at each riverbank.
1.2 Integrals as General and Particular Solutions 15
Suppose that a swimmer starts at the point .a; 0/ on the west bank and swims
due east (relative to the water) with constant speed vS . As indicated in Fig. 1.2.5, his
velocity vector (relative to the riverbed) has horizontal component vS and vertical
component vR . Hence the swimmer’s direction angle ˛ is given by
vR
tan ˛ D :
vS
Because tan ˛ D dy=dx , substitution using (18) gives the differential equation
dy v0 x2
D 1 2 (19)
dx vS a
y.x/ D 3x 4x 3 C 1:
Then 3
1 1 1
y 2 D3 2 4 2 C 1 D 2;
so the swimmer drifts 2 miles downstream while he swims 1 mile across the river.
1.2 Problems
In Problems 1 through 10, find a function y D f .x/ satisfy- In Problems 11 through 18, find the position function x.t / of a
ing the given differential equation and the prescribed initial moving particle with the given acceleration a.t /, initial posi-
condition. tion x0 D x.0/, and initial velocity v0 D v.0/.
dy 11. a.t / D 50, v0 D 10, x0 D 20
1. D 2x C 1; y.0/ D 3
dx
12. a.t / D 20, v0 D 15, x0 D 5
dy
2. D .x 2/2 ; y.2/ D 1 13. a.t / D 3t , v0 D 5, x0 D 0
dx
dy p 14. a.t / D 2t C 1, v0 D 7, x0 D 4
3. D x ; y.4/ D 0
dx 15. a.t / D 4.t C 3/2 , v0 D 1, x0 D 1
dy 1
4. D 2 ; y.1/ D 5 1
dx x 16. a.t / D p , v0 D 1, x0 D 1
t C4
dy 1
5. D p ; y.2/ D 1 1
dx xC2 17. a.t / D , v0 D 0, x0 D 0
.t C 1/3
dy p
6. D x x 2 C 9; y.4/ D 0 18. a.t / D 50 sin 5t , v0 D 10, x0 D 8
dx
dy 10 dy In Problems 19 through 22, a particle starts at the origin and
7. D 2 ; y.0/ D 0 8. D cos 2x ; y.0/ D 1
dx x C1 dx travels along the x -axis with the velocity function v.t / whose
dy 1 dy graph is shown in Figs. 1.2.6 through 1.2.9. Sketch the graph
9. D p ; y.0/ D 0 10. D xe x ; y.0/ D 1 of the resulting position function x.t / for 0 5 t 5 10.
dx 1 x2 dx
16 Chapter 1 First-Order Differential Equations
19. 10 23. What is the maximum height attained by the arrow of part
(b) of Example 3?
8 24. A ball is dropped from the top of a building 400 ft high.
How long does it take to reach the ground? With what
6 speed does the ball strike the ground?
(5, 5)
25. The brakes of a car are applied when it is moving at 100
v
36. Suppose a woman has enough “spring” in her legs to jump 41. A bomb is dropped from a helicopter hovering at an alti-
(on earth) from the ground to a height of 2.25 feet. If tude of 800 feet above the ground. From the ground di-
she jumps straight upward with the same initial velocity rectly beneath the helicopter, a projectile is fired straight
on the moon—where the surface gravitational acceleration upward toward the bomb, exactly 2 seconds after the bomb
is (approximately) 5.3 ft/s2 —how high above the surface is released. With what initial velocity should the projectile
will she rise? be fired in order to hit the bomb at an altitude of exactly
37. At noon a car starts from rest at point A and proceeds at 400 feet?
constant acceleration along a straight road toward point 42. A spacecraft is in free fall toward the surface of the moon
B . If the car reaches B at 12:50 P.M. with a velocity of at a speed of 1000 mph (mi/h). Its retrorockets, when
60 mi=h, what is the distance from A to B ? fired, provide a constant deceleration of 20,000 mi/h2 . At
38. At noon a car starts from rest at point A and proceeds with what height above the lunar surface should the astronauts
constant acceleration along a straight road toward point C , fire the retrorockets to insure a soft touchdown? (As in
35 miles away. If the constantly accelerated car arrives at Example 2, ignore the moon’s gravitational field.)
C with a velocity of 60 mi=h, at what time does it arrive 43. Arthur Clarke’s The Wind from the Sun (1963) describes
at C ? Diana, a spacecraft propelled by the solar wind. Its alu-
minized sail provides it with a constant acceleration of
39. If a D 0:5 mi and v0 D 9 mi=h as in Example 4, what must
0:001g D 0:0098 m/s2 . Suppose this spacecraft starts from
the swimmer’s speed vS be in order that he drifts only 1
rest at time t D 0 and simultaneously fires a projectile
mile downstream as he crosses the river?
(straight ahead in the same direction) that travels at one-
40. Suppose that a D 0:5 mi, v0 D 9 mi=h, and vS D 3 mi=h tenth of the speed c D 3 108 m/s of light. How long will
as in Example 4, but that the velocity of the river is given it take the spacecraft to catch up with the projectile, and
by the fourth-degree function how far will it have traveled by then?
! 44. A driver involved in an accident claims he was going only
x4 25 mph. When police tested his car, they found that when
vR D v0 1 4
a its brakes were applied at 25 mph, the car skidded only
45 feet before coming to a stop. But the driver’s skid
rather than the quadratic function in Eq. (18). Now find marks at the accident scene measured 210 feet. Assum-
how far downstream the swimmer drifts as he crosses the ing the same (constant) deceleration, determine the speed
river. he was actually traveling just prior to the accident.
dy
D f .x; y/ (1)
dx
where the right-hand function f .x; y/ involves both the independent variable x and
the dependent variable y . We mightRthink of integrating both sides in (1) with re-
spect to x , and hence write y.x/ D f .x; y.x// dx C C . However, this approach
does not lead to a solution of the differential equation, because the indicated integral
involves the unknown function y.x/ itself, and therefore cannot be evaluated explic-
itly. Actually, there exists no straightforward procedure by which a general differen-
tial equation can be solved explicitly. Indeed, the solutions of such a simple-looking
differential equation as y 0 D x 2 C y 2 cannot be expressed in terms of the ordinary
elementary functions studied in calculus textbooks. Nevertheless, the graphical and
numerical methods of this and later sections can be used to construct approximate
solutions of differential equations that suffice for many practical purposes.
point .x; y.x// through which it passes—that is, y 0 .x/ D f .x; y.x//. Thus a so-
lution curve of the differential equation y 0 D f .x; y/—the graph of a solution of
the equation—is simply a curve in the xy-plane whose tangent line at each point
.x; y/ has slope m D f .x; y/. For instance, Fig. 1.3.1 shows a solution curve of
the differential equation y 0 D x y together with its tangent lines at three typical
points.
(x1, y1)
x
(x2, y2)
(x3, y3)
y
–1 –1 –1
–2 –2 –2
–3 –3 –3
–4 –4 –4
–4 –3 –2 –1 0 1 2 3 4 – 4 –3 –2 –1 0 1 2 3 4 – 4 –3 –2 –1 0 1 2 3 4
x x x
FIGURE 1.3.2(a) Slope field and FIGURE 1.3.2(b) Slope field and FIGURE 1.3.2(c) Slope field and
solution curves for y 0 D 2y . solution curves for y 0 D .0:5/y . solution curves for y 0 D y .
x ny 4 3 2 1 0 1 2 3 4
4 4 0 1 2 3 4 5 6 7 8
3 3 1 0 1 2 3 4 5 6 7
2 2 2 1 0 1 2 3 4 5 6
1 1 3 2 1 0 1 2 3 4 5
0
1 2 3 4
y
0 4 3 2 1 0
–1
1 5 4 3 2 1 0 1 2 3
–2
2 6 5 4 3 2 1 0 1 2
–3
3 7 6 5 4 3 2 1 0 1
–4
– 4 –3 –2 –1 0 1 2 3 4
x 4 8 7 6 5 4 3 2 1 0
FIGURE 1.3.2(d) Slope field FIGURE 1.3.3. Values of the slope y 0 D x y for 4 x; y 4.
and solution curves for y 0 D 3y .
was easily and quickly constructed. (Of course, a more complicated function f .x; y/ on the
right-hand side of the differential equation would necessitate more complicated calculations.)
Figure 1.3.4 shows the corresponding slope field, and Fig. 1.3.5 shows an approximate so-
lution curve sketched through the point .4; 4/ so as to follow this slope field as closely as
possible. At each point it appears to proceed in the direction indicated by the nearby line
segments of the slope field.
Although a spreadsheet program (for instance) readily constructs a table of
slopes as in Fig. 1.3.3, it can be quite tedious to plot by hand a sufficient number
5 5 (– 4, 4)
4
3
2
1
0
y
0
y
–1
–2
–3
–4
–5 –5
–5 0 5 –5 0 5
x x
FIGURE 1.3.4. Slope field for y 0 D x y FIGURE 1.3.5. The solution curve
corresponding to the table of slopes in Fig. 1.3.3. through .4; 4/.
20 Chapter 1 First-Order Differential Equations
of slope segments as in Fig. 1.3.4. However, most computer algebra systems in-
clude commands for quick and ready construction of slope fields with as many line
4 segments as desired; such commands are illustrated in the application material for
3 this section. The more line segments are constructed, the more accurately solution
2 curves can be visualized and sketched. Figure 1.3.6 shows a “finer” slope field for
1 the differential equation y 0 D x y of Example 2, together with typical solution
0 curves treading through this slope field.
y
–1 If you look closely at Fig. 1.3.6, you may spot a solution curve that appears
–2 to be a straight line! Indeed, you can verify that the linear function y D x 1 is
–3 a solution of the equation y 0 D x y , and it appears likely that the other solution
–4 curves approach this straight line as an asymptote as x ! C1. This inference
–4 –3 –2 –1 0 1 2 3 4 illustrates the fact that a slope field can suggest tangible information about solutions
x that is not at all evident from the differential equation itself. Can you, by tracing the
FIGURE 1.3.6. Slope field and appropriate solution curve in this figure, infer that y.3/ 2 for the solution y.x/ of
typical solution curves for y 0 D x y . the initial value problem y 0 D x y , y.4/ D 4?
dt
100
The result is shown in Fig. 1.3.7, together with a number of solution curves correspond-
0 ing to different values of the initial velocity v.0/ with which you might throw the baseball
0 5 10 15 20 25
t downward. Note that all these solution curves appear to approach the horizontal line v D 200
as an asymptote. This implies that—however you throw it—the baseball should approach the
FIGURE 1.3.7. Slope field and
typical solution curves for limiting velocity v D 200 ft/s instead of accelerating indefinitely (as it would in the absence
v 0 D 32 0:16v . of any air resistance). The handy fact that 60 mi/h D 88 ft/s yields
ft 60 mi/h mi
v D 200 136.36 :
s 88 ft/s h
Perhaps a catcher accustomed to 100 mi/h fastballs would have some chance of fielding this
speeding ball.
Comment If the ball’s initial velocity is v.0/ D 200, then Eq. (4) gives v 0 .0/ D 32
.0:16/.200/ D 0, so the ball experiences no initial acceleration. Its velocity therefore remains
unchanged, and hence v.t / 200 is a constant “equilibrium solution” of the differential
equation. If the initial velocity is greater than 200, then the initial acceleration given by
Eq. (4) is negative, so the ball slows down as it falls. But if the initial velocity is less than
1.3 Slope Fields and Solution Curves 21
200, then the initial acceleration given by (4) is positive, so the ball speeds up as it falls. It
therefore seems quite reasonable that, because of air resistance, the baseball will approach a
limiting velocity of 200 ft/s—whatever initial velocity it starts with. You might like to verify
that—in the absence of air resistance—this ball would hit the ground at over 300 mi/h.
In Section 2.1 we will discuss in detail the logistic differential equation
dP
D kP .M P / (5)
dt
dP
D 0:0004P .150 P / D 0:06P 0:0004P 2 : (6)
dt
300 The positive term 0:06P on the right in (6) corresponds to natural growth at a 6% annual rate
250 (with time t measured in years). The negative term 0:0004P 2 represents the inhibition of
200 growth due to limited resources in the environment.
Figure 1.3.8 shows a slope field for Eq. (6), together with a number of solution curves
P
150
100 corresponding to possible different values of the initial population P .0/. Note that all these
50
solution curves appear to approach the horizontal line P D 150 as an asymptote. This implies
0
that—whatever the initial population—the population P .t / approaches the limiting popula-
0 25 50 75 100 tion P D 150 as t ! 1.
t
FIGURE 1.3.8. Slope field and Comment If the initial population is P .0/ D 150, then Eq. (6) gives
typical solution curves for
P 0 D 0:06P 0:0004P 2 . P 0 .0/ D 0:0004.150/.150 150/ D 0;
1
y0 D ; y.0/ D 0 (7)
x
R
has no solution, because no solution y.x/ D .1=x/ dx D ln jxj C C of the differential equation
is defined at x D 0. We see this graphically in Fig. 1.3.9, which shows a direction field and
some typical solution curves for the equation y 0 D 1=x . It is apparent that the indicated
direction field “forces” all solution curves near the y -axis to plunge downward so that none
can pass through the point .0; 0/.
22 Chapter 1 First-Order Differential Equations
2 1
(0, 0) y1(x) = x2
0
y
y
y2(x) = 0
0
(0, 0)
–2
–1 0 1 0 1
x x
FIGURE 1.3.9. Direction field and typical FIGURE 1.3.10. Direction field and two
solution curves for the equation y 0 D 1=x . different solutionpcurves for the initial value
problem y 0 D 2 y , y.0/ D 0.
(b) [Failure of uniqueness] On the other hand, you can readily verify that the initial value
problem
p
y 0 D 2 y; y.0/ D 0 (8)
has the two different solutions y1 .x/ D x 2 and y2 .x/ 0 (see Problem 27). Figure 1.3.10
shows a direction field and these two different solution curves for the initial value problem in
(8). We see that the curve y1 .x/ D x 2 threads its way through the indicated direction field,
p
whereas the differential equation y 0 D 2 y specifies slope y 0 D 0 along the x -axis y2 .x/ D 0.
Example 5 illustrates the fact that, before we can speak of “the” solution of
an initial value problem, we need to know that it has one and only one solution.
Questions of existence and uniqueness of solutions also bear on the process of
mathematical modeling. Suppose that we are studying a physical system whose be-
havior is completely determined by certain initial conditions, but that our proposed
mathematical model involves a differential equation not having a unique solution
satisfying those conditions. This raises an immediate question as to whether the
mathematical model adequately represents the physical system.
The theorem stated below implies that the initial value problem y 0 D f .x; y/,
y.a/ D b has one and only one solution defined near the point x D a on the x -axis,
provided that both the function f and its partial derivative @f =@y are continuous
near the point .a; b/ in the xy-plane. Methods of proving existence and uniqueness
y theorems are discussed in the Appendix.
R
THEOREM 1 Existence and Uniqueness of Solutions
(a, b)
b Suppose that both the function f .x; y/ and its partial derivative Dy f .x; y/ are
y = y(x)
continuous on some rectangle R in the xy -plane that contains the point .a; b/
in its interior. Then, for some open interval I containing the point a, the initial
I value problem
a x dy
D f .x; y/; y.a/ D b (9)
FIGURE 1.3.11. The rectangle R dx
and x -interval I of Theorem 1, and the has one and only one solution that is defined on the interval I . (As illustrated in
solution curve y D y.x/ through the
point .a; b/. Fig. 1.3.11, the solution interval I may not be as “wide” as the original rectangle
R of continuity; see Remark 3 below.)
— Ei, Ivan, sinä olet itse sanonut itsellesi useita kertoja, että sinä
olet murhaaja.
— Sinä olet sanonut tätä itsellesi monta kertaa, kun olet jäänyt
yksiksesi näitten kauheitten kahden kuukauden aikana, — jatkoi
Aljoša hiljaa ja selvästi kuten ennenkin. Mutta hän puhui aivan kuin
ulkopuolelta itseään, aivan kuin ei puhuisi omasta tahdostaan, vaan
tottelisi jotakin vastustamatonta käskyä.
Mutta yhtäkkiä hän aivan kuin hillitsi itsensä. Hän seisoi ja näytti
miettivän jotakin. Omituinen nauru vääristi hänen huuliaan.
6.
Smerdjakov huokasi.
— Jos sinä siis itse sanot, ettei voinut arvata, niin kuinka sitten
minä saatoin arvata ja jäädä? Mitä sinä sotket? — lausui Ivan
Fjodorovitš ajatuksissaan.
— Siitä juuri olisitte voinut hoksata, että jos kerran minä koetan
taivuttaa teitä menemään Tšermašnjaan Moskovan asemasta, niin
siis haluan teidän olevan täällä mahdollisimman lähellä, sillä
Moskova on kaukana, mutta kun Dmitri Fjodorovitš tietävät teidän
olevan lähiseuduilla, niin he ovat vähemmän rohkea. Ja minuakin
olisitte voinut, jos jotakin olisi tapahtunut, nopeasti tulla suojelemaan,
sillä minä itse viittasin teille sen lisäksi Grigori Vasiljitšin sairauteen ja
siihen, että pelkäsin kaatuvatautia. Ja kun minä selitin teille ne
koputukset, joiden avulla saattoi päästä sisälle vainajan luo, ja että
Dmitri Fjodorovitš oli saanut ne minun kauttani tietää, niin ajattelin
teidän itsennekin jo arvaavan, että he aikovat tehdä jotakin, ja silloin
ette olisi lähtenyt edes Tšermašnjaan, vaan olisitte jäänyt kokonaan
tänne.
— Mitenkä moitteeksi!