The laws of the universe are written in the language of mathematics. Algebra is sufficient to solve many static problems, but the most interesting natural phenomena involve change and are described by equations that relate changing quantities.
Because the derivative dx/dt=f′(t)
The differential equation
involves both the unknown function x(t) and its first derivative x′(t)=dx/dt
involves the unknown function y of the independent variable x and the first two derivatives y′
The study of differential equations has three principal goals:
To discover the differential equation that describes a specified physical situation.
To find—either exactly or approximately—the appropriate solution of that equation.
To interpret the solution that is found.
In algebra, we typically seek the unknown numbers that satisfy an equation such as x3+7x2−11x+41=0
—holds on some interval of real numbers. Ordinarily, we will want to find all solutions of the differential equation, if possible.
If C is a constant and
then
Thus every function y(x) of the form in Eq. (1) satisfies—and thus is a solution of—the differential equation
for all x. In particular, Eq. (1) defines an infinite family of different solutions of this differential 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).
The following three examples illustrate the process of translating scientific laws and principles into differential equations. In each of these examples the independent variable is time t, but we will see numerous examples in which some quantity other than time is the independent variable.
Rate of cooling Newton’s law of cooling may be stated in this way: The time rate of change (the rate of change with respect to time t) of the temperature T(t) of a body is proportional to the difference between T and the temperature A of the surrounding medium (Fig. 1.1.1). That is,
where k is a positive constant. Observe that if T>A
Thus the physical law is translated into a differential equation. If we are given the values of k and A, we should be able to find an explicit formula for T(t), and then—with the aid of this formula—we can predict the future temperature of the body.
Draining tank 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:
where k is a constant. If the tank is a cylinder with vertical sides and cross-sectional area A, then V=Ay
where h=k/A
Population growth 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,
where k is the constant of proportionality.
Let us discuss Example 5 further. Note first that each function of the form
is a solution of the differential equation
in (6). We verify this assertion as follows:
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=kP
Population growth Suppose that P(t)=Cekt
It follows that C=1000
Substitution of k=ln 2
that satisfies the given conditions. We can use this particular solution to predict future populations of the bacteria colony. For instance, the predicted number of bacteria in the population after one and a half hours (when t=1.5
The condition P(0)=1000
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:
The formulation of a real-world problem in mathematical terms; that is, the construction of a mathematical model.
The analysis or solution of the resulting mathematical problem.
The interpretation of the mathematical results in the context of the original real-world situation—for example, answering the question originally posed.
In the population example, the real-world problem is that of determining the population at some future time. A mathematical model consists of a list of variables (P and t) that describe the given situation, together with one or more equations relating these variables (dP/dt=kP, P(0)=P0
As an example of this process, think of first formulating the mathematical model consisting of the equations dP/dt=kP, P(0)=1000
On the other hand, it may turn out that no solution of the selected differential equation accurately fits the actual population we’re studying. For instance, for no choice of the constants C and k does the solution P(t)=Cekt
But in Example 6 we simply ignored any complicating factors that might affect our bacteria population. This made the mathematical analysis quite simple, perhaps unrealistically so. A satisfactory mathematical model is subject to two contradictory requirements: It must be sufficiently detailed to represent the real-world situation with relative accuracy, yet it must be sufficiently simple to make the mathematical 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 mathematically 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.
If C is a constant and y(x)=1/(C−x)
if x≠C
defines a solution of the differential equation
on any interval of real numbers not containing the point x=C
that satisfies the initial condition y(0)=1
Verify that the function y(x)=2x1/2−x1/2 ln x
for all x>0
First we compute the derivatives
Then substitution into Eq. (10) yields
if x is positive, so the differential equation is satisfied for all x>0
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
has no (real-valued) solution, because the sum of nonnegative numbers cannot be negative. For a variation on this theme, note that the equation
obviously has only the (real-valued) solution y(x)≡0
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
is a fourth-order equation. The most general form of an nth-order differential equation with independent variable x and unknown function or dependent variable y=y(x)
where F is a specific real-valued function of n+2
Our use of the word solution has been until now somewhat informal. To be precise, we say that the continuous function u=u(x)
for all x in I. For the sake of brevity, we may say that u=u(x)
Recall from elementary calculus that a differentiable function on an open interval is necessarily continuous there. This is why only a continuous function can qualify as a (differentiable) solution of a differential equation on an interval.
Continued Figure 1.1.5 shows the two “connected” branches of the graph y=1/(1−x)
If A and B are constants and
then two successive differentiations yield
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
on the whole real number line. Figure 1.1.6 shows the graphs of several such solutions.
Although the differential equations in (11) and (12) are exceptions to the general rule, we will see that an nth-order differential equation ordinarily has an n-parameter family of solutions—one involving n different arbitrary constants or parameters.
In both Eqs. (11) and (12), the appearance of y′
where G is a real-valued function of n+1
All the differential equations we have mentioned so far are ordinary differential equations, meaning that the unknown function (dependent variable) depends on only a single independent variable. If the dependent variable is a function of 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=u(x,t)
where k is a constant (called the thermal diffusivity of the rod). In Chapters 1 through 8 we will be concerned only with ordinary differential equations and will refer to them simply as differential equations.
In this chapter we concentrate on first-order differential equations of the form
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, consisting of a differential equation of the form in (17) together with an initial condition y(x0)=y0
means to find a differentiable function y=y(x)
Given the solution y(x)=1/(C−x)
We need only find a value of C so that the solution y(x)=1/(C−x)
so 2C−2=1
Figure 1.1.7 shows the two branches of the graph y=2/(3−2x)
The central question of greatest immediate interest to us is this: If we are given a differential equation known to have a solution satisfying a given initial condition, how do we actually find or compute that solution? And, once found, what can we do with it? We will see that a relatively few simple techniques—separation of variables (Section 1.4), solution of linear equations (Section 1.5), elementary substitution methods (Section 1.6)—are enough to enable us to solve a variety of first-order equations having impressive applications.
In Problems 1 through 12, verify by substitution that each given function is a solution of the given differential equation. Throughout these problems, primes denote derivatives with respect to x.
y′=3x2
y′+2y=0
y″+4y=0
y″=9y
y′=y+2e−x
y″+4y′+4y=0
y″−2y′+2y=0
y″+y=3 cos 2x, y1=cos x−cos 2x, y2=sin x−cos 2x
y′+2xy2=0
x2y″+xy′−y=ln x
x2y″+5xy′+4y=0
x2y″−xy′+2y=0
In Problems 13 through 16, substitute y=erx
3y′=2y
4y″=y
y″+y′−2y=0
3y″+3y′−4y=0
In Problems 17 through 26, first verify that y(x) satisfies the given differential equation. Then determine a value of the constant C so that y(x) satisfies the given initial condition. Use a computer or graphing calculator (if desired) to sketch several typical solutions of the given differential equation, and highlight the one that satisfies the given initial condition.
y′+y=0
y′=2y
y′=y+1
y′=x−y
y′+3x2y=0
eyy′=1
xdydx+3y=2x5
xy′−3y=x3
y′=3x2(y2+1)
y′+ytanx=cos x
In Problems 27 through 31, a function y=g(x)
The slope of the graph of g at the point (x, y) is the sum of x and y.
The line tangent to the graph of g at the point (x, y) intersects the x-axis at the point (x/2,0)
Every straight line normal to the graph of g passes through the point (0, 1). Can you guess what the graph of such a function g might look like?
The graph of g is normal to every curve of the form y=x2+k
The line tangent to the graph of g at (x, y) passes through the point (−y,x)
In Problems 32 through 36, write—in the manner of Eqs. (3) through (6) of this section—a differential equation that is a mathematical model of the situation described.
The time rate of change of a population P is proportional to the square root of P.
The time rate of change of the velocity v of a coasting motorboat is proportional to the square of v.
The acceleration dv/dt
In a city having a fixed population of P persons, the time rate of change of the number N of those persons who have heard a certain rumor is proportional to the number of those who have not yet heard the rumor.
In a city with a fixed population of P persons, the time rate of change of the number N of those persons infected with a certain contagious disease is proportional to the product of the number who have the disease and the number who do not.
In Problems 37 through 42, determine by inspection at least one solution of the given differential equation. That is, use your knowledge of derivatives to make an intelligent guess. Then test your hypothesis.
y″=0
y′=y
xy′+y=3x2
(y′)2+y2=1
y′+y=ex
y″+y=0
Problems 43 through 46 concern the differential equation
where k is a constant.
If k is a constant, show that a general (one-parameter) solution of the differential equation is given by x(t)=1/(C−kt)
Determine by inspection a solution of the initial value problem x′=kx2, x(0)=0
Assume that k is positive, and then sketch graphs of solutions of x′=kx2
How would these solutions differ if the constant k were negative?
Suppose a population P of rodents satisfies the differential equation dP/dt=kP2
Suppose the velocity v of a motorboat coasting in water satisfies the differential equation dv/dt=kv2
In Example 7 we saw that y(x)=1/(C−x)
Determine a value of C so that y(10)=10
Is there a value of C such that y(0)=0
Figure 1.1.8 shows typical graphs of solutions of the form y(x)=1/(C−x)
Show that y(x)=Cx4
Show that
defines a differentiable solution of xy′=4y
Given any two real numbers a and b, explain why—in contrast to the situation in part (c) of Problem 47—there exist infinitely many differentiable solutions of xy′=4y
3.15.202.214