MAT 127: Calculus C, Spring 2022

Course Summary I

Extremely Important: what it means for a function to solve a differential equation or initial-
value problem and how to check this; implicitly defined solutions; general solution vs. specific
solution.

Very Important: descriptive analysis of first-order differential equations, including

y ′ = rate of change of y = slope of graph of y at (x, y)

and sketches as in Figure 1 on page 7; direction fields and solution curves; separable equations;
general structure of solutions (the number of C’s); linear approximations of functions; Euler’s for-
mula.

Important: finding solutions to separable first-order equations and linear homogeneous second-
order equations with constant coefficients; finding solutions to initial-value problems involving such
equations; Euler’s method and its geometric interpretation; autonomous first-order equations; ap-
plication problems (leading to separable equations).

A: Terminology

A.1 A first-order differential equation is an equation that involves a function, its derivative, and
the independent variable,
R(x, y, y ′ ) = 0, y = y(x), (A1)
and cannot be simplified, through algebraic means, to a relation R̃(x, y) = 0. In (A1), R is a
function of three variables. Most first-order differential equations arising in applications can be
put into the form
y ′ = f (x, y), y = y(x), (A2)
where f is a function of two variables. An initial-value problem, for a first-order differential
equation, is a set of conditions:

R(x, y, y ′ ) = 0 or y ′ = f (x, y), y = y(x), y(x0 ) = y0 . (A3)

The last condition in (A3) is the initial-value requirement for (A3).

A.2 A solution of (A1), or of (A2), is a function y = y(x) that satisfies (A1), or (A2). In order
to check if a given function y = y(x) is a solution of (A1), or of (A2), one must compute y ′ (x)
and plug in y and y ′ into (A1), or (A2), to see if the equality holds. A solution of the initial-value
problem (A3) is a function y = y(x) that satisfies the differential equation and the initial-value re-
quirement in (A3). In order to check if a given function y = y(x) is a solution of (A3), one must check
whether y = y(x) is a solution of the differential equation and that y(x0 ) = y0 ; the latter is usually
easier to do and so should be done first. Typically, but not always, (A3) will have a unique solution.

A.3 A solution curve for the first-order differential equation (A1), or for (A2), is the graph, in
the xy-plane, of a solution y = y(x) of (A1), or of (A2). Typically, but not always, solution curves
for the same first-order differential equation do not intersect, because typically (A3) has a unique
solution. A solution curve for the initial-value problem (A3) is the graph of a solution y = y(x)
of (A3). Such a graph must pass through the point (x0 , y0 ).
Caution: While the solution curves for the simplest first-order differential equations, i.e. (B1) be-
low, differ by vertical shifts, this is not the case for other first-order differential equations.

A.4 The direction field for (A2) is usually thought of as a diagram, in the xy-plane, consisting of
short line segments of slope y ′ = f (x, y) through a number of points (x, y). Since the derivative of
a function y = y(x) is the slope of the tangent line to the graph of y, a solution curve for (A2) is
everywhere tangent to the direction field. In particular, if the direction field is drawn at sufficiently
many points, one can pretty much see solution curves.

B: Finding Solutions of Some First-Order Differential Equations

B.1 The simplest first-order differential equations to solve are those of the form

y ′ = f (x), y = y(x). (B1)

They are solved by taking the indefinite integral of both sides:

R
y ′ = f (x), y = y(x) =⇒ y = f (x)dx

The solution curves of (B1) differ by vertical shifts. An initial-value problem for (B1) is solved by
Rx
y ′ = f (x), y(x0 ) = y0 =⇒ y(x) = y0 + x0 f (u)du

B.2 Separable first-order differential equations are the equations of the form

y ′ = f (x) · g(y), y = y(x). (B2)

dy
Equation (B2) is solved by writing y ′ = dx , moving all expressions involving y to LHS and all
expressions involving x to RHS, and integrating both sides:

dy dy R dy R
dx = f (x) · g(y), y = y(x) =⇒ g(y) = f (x)dx =⇒ g(y) = f (x)dx

Once the two integrals are computed, one obtains a relation between y and x of the form

G(y) = F (x) + C ⇐⇒ G(y) − F (x) = C. (B3)

These relations define solutions y = y(x) of (B2) implicitly. In some cases, it is possible to solve
(B3) for y = y(x). An initial-value problem for (B2) is solved by

dy dy Ry dz
Rx
dx = f (x) · g(y), y(x0 ) = y0 =⇒ g(y) = f (x)dx =⇒ y0 g(z) = x0 f (u)du

Alternatively, one can first find the general solution and then find the constant C by plugging in the
initial conditions. It is the easiest to find C as soon as it appears, i.e. plug in the initial conditions

2
into (B3) and then solve for y = y(x), instead of first solving (B3) for y = y(x) and then solving for C.

Caution: (i) This separation-of-variables method involves division by g = g(y) and may miss some
of the constant solutions of (B2). Such solutions are necessarily of the form y = y ∗ , where y ∗ is a
real number such that g(y ∗ ) = 0.
(ii) If you are solving an initial-value problem and it is possible to solve for y = y(x) explicitly,
make sure you take the correct branch, if there is more than one, of the appropriate level curve of
H = F −G, e.g. the positive or negative square root, and not both. The correct branch is the one
satisfying the initial condition y(x0 ) = y0 .

C: Finding Solutions of Some Second-Order Differential Equations

The general solution of a second-order linear homogeneous equation with constant coefficients

y ′′ + by ′ + cy = 0, b, c = const, y = y(x), (C1)

is determined by the two roots, r1 and r2 , of the associated polynomial

r2 + br + c = 0. (C2)

The general solution can be of two or three different forms, depending on whether one is looking
for complex or real solutions:

y ′′ + by ′ + cy = 0, y = y(x) =⇒ y(x) = C1 er1 x + C2 xer1 x if r1 = r2 ⇐⇒ b2 = 4c

y ′′ + by ′ + cy = 0, y = y(x) =⇒ y(x) = C1 er1 x + C2 er2 x if r1 6= r2 ⇐⇒ b2 6= 4c

If the coefficients b and c are real, the roots r1 and r2 of (C2) are either real or complex conjugates
of each other. In the latter case, Euler’s formula,

eiθ = cos θ + i sin θ

can be used to extract the general real solution from the general complex solution:
√
y ′′ +by ′ +cy = 0 =⇒ y(x) = C1 epx cos qx + C2 epx sin qx, p = − 12 b, q = 21 4c−b2 , if b2 < 4c

The numbers p and q are related to the roots r1 and r2 by r1 , r2 = p±iq.

3
 D: Euler’s Method
D.1 Euler’s method is used to estimate the value y(b) of the solution y = y(x) to a first-order
initial-value problem
y ′ = f (x, y), y(a) = y0 , (D1)
at b, for b > a. This is especially useful when the explicit solution y = y(x) of (D1) cannot be found.
Euler’s method can also be used if f is known only at a discreet grid of points (xi , yj ), as the case
may well be in an experimental setting.

D.2 Euler’s method is a fixed-step method. This means that we break up the interval [a, b] into n
segments [xi , xi+1 ] of equal length h = (b−a)/n, i.e.
x0 = a, x1 = x0 +h = a+h, . . . xn−1 = xn−2 +h = a+(n−1)h, xn = xn−1 +h = a+nh = b.
We then give an estimate yi for the value of the function y at xi . More precisely, we give an
estimate y1 for y(x1 ), where y = y(x) is the solution to the initial-value problem
y ′ = f (x, y), y(x0 ) = y0 .
We then use the same procedure to give an estimate y2 for ỹ1 (x2 ), where ỹ1 = ỹ1 (x) is the solution
to the initial-value problem
y ′ = f (x, y), y(x1 ) = y1 .
Since y1 is an estimate for y(x1 ), y2 will also be an estimate for y(x2 ). At the i-th step of this
construction, we give an estimate yi+1 for ỹi (xi+1 ), where ỹi = ỹi (x) is the solution to the initial-
value problem
y ′ = f (x, y), y(xi ) = yi .
After n steps, we end up with an estimate yn for ỹn−1 (xn ), where ỹn−1 = ỹn−1 (x) is the solution to
the initial-value problem
y ′ = f (x, y), y(xn ) = yn−1 .
This number yn will also be an estimate for y(b).

D.3 In Euler’s method, we take

yi+1 = yi + si h, where si = f (xi , yi ).
Since ỹi (xi ) = yi and ỹi′ (xi ) = f (xi , yi ) = si , yi+1 estimates ỹi (xi+1 ) by a linear approximation in h.
This means that instead of moving from xi to xi+1 along the graph of ỹi (which we do not know),
we move along the tangent line to this graph at (xi , yi ); we know what the tangent line is, since it
is determined by the condition that it passes through (xi , yi ) and its slope is si . This computation
can be carried out using a five-column table, as done in the solutions to PS2. The columns are
label i, xi , yi , si = f (xi , yi ), and yi+1 = yi + si h, with f (xi , yi ) replaced by the expression from
(D1) and h by the actual step size (a small number). The rows in the first columns are then filled
out by the integers from 0 to n − 1. The corresponding entries in the second columns are filled
with the numbers x0 , x1 , . . . , xn−1 ; these numbers begin with the given x0 = a, increase by h from
row to row, and end with xn−1 = b−h. The third entry in the i = 0 row is the given initial value y0 .
From here one computes to the right as indicated by the column labels and copies the last entry in
each to the third column of the following row. The entry in the bottom right corner is our estimate
yn for y(b).

4
 E: Applications

E.1 Find an equation for the curve with slope f (x, y) at (x, y) and passing through a point (x0 , y0 ).
Since the slope of the graph of y = y(x) at (x, y(x)) is y ′ (x), this reduces to solving the initial-value
problem
y ′ = f (x, y), y = y(x), y(x0 ) = y0 ;
the last condition means that the graph of y passes through (x0 , y0 ). You should simplify your
final answer as much as possible, without necessarily finding an explicit expression for y = y(x)
since you are asked for an equation of a√ curve, not of a function. For example, x2 + y 2 = 1 is a
nicer expression for a curve than y = ± 1 − x2 .

E.2 Mixing problems. In these problems the independent variable is usually t (time), measured in
some units, and y = y(t) is the amount of some substance in a thoroughly mixed solution/mixture
contained in some “reservoir” (water tank, room, etc.). The mixture might be leaving the reservoir
at some fixed rate, while another mixture (or several other mixtures) might be entering the reservoir.
The main difficulty in such problems is to set up a differential equation for y(t). It will be of the form

y ′ (t) = yin
′ ′
(t) − yout (t), y = y(t).
′ (t), is the concentration of the substance in the mixture
The outgoing flow rate of the substance, yout
times the outgoing flow rate of the mixture; while the outgoing flow rate of the mixture is typically
fixed, the concentration typically changes with time. The incoming flow rate of the substance,
′ (t), is the concentration of the substance in the incoming mixture times the incoming flow rate
yin
of the mixture (this might have to be summed up over several incoming mixtures); this number is
typically constant. So you should end up with an initial-value problem of the form

y ′ = a − by, y = y(t), y(0) = y0 , (E1)

where y0 is the initial amount of the substance in the reservoir and a, b > 0; both are constants if
the volume in the reservoir is kept constant (otherwise, only a is constant). If a, b are constants,
the above equation is separable and so can be solved. If you do this correctly, the concentration
of the substance in the reservoir should approach the (weighted) concentration of the substance in
the incoming mixture(s). Make sure your final answer has the correct physical units.

E.3 Exponential growth/decay equation. In this case, the independent variable is usually t (time),
measured in some units, and the main dependent variable y(t) is described

y(t) = y(0)ert

which satisfies the differential equation

y ′ = ry, y = y(t). (E2)

The first equation is the equation you should start with and then try to figure out what the num-
bers y(0) and r are. These might be given in the statement of the problem or you may have to
find them knowing that y(t) is given by the above formula for some y(0) and r. The function y(t)
will typically be the size of a population, amount of a radioactive substance, difference with the
ambient temperature, or bank/loan balance at time t measured in certain units (which you may
need to specify). In all of these cases y(t) > 0 for all t; in the middle two cases r < 0, while in the

5
other two cases r > 0. Make sure your final answer has the correct physical units.

E.4 Logistic growth equation. In this case, the independent variable is usually t (time), measured
in some units, and the main dependent variable y(t) is normally the size of a population at time t
growing with limited resources. You should begin any such problem with the formula

K Ky(0)
y(t) = y(0)−K −rt
==
1− y(0) + (K −y(0))e−rt
y(0) e

which solves the differential equation

 
′ y
y = ry 1 − , y = y(t). (E3)
K
If you forget the first formula, you can recover it by solving the differential equation. Since the
equation is separable, this is doable, though requires some effort. In this setting, K > 0 represents
the carrying capacity of the available resources, measured in whatever units the population y(t) is
measured; y(t) approaches K as t −→ ∞. Furthermore, r > 0. You may be given the numbers
y(0), r, and K in the statement of the problem or you may have to find (some of) them based on
the information provided. Make sure your final answer has the correct physical units.

F: Autonomous First-Order ODEs

F.1 An autonomous first-order ODE is an ODE of the form

y ′ = g(y), y = y(t). (F1)

Such equations often model natural phenomena, expressing the fact these are governed by the
same principles whether they start today or tomorrow. The equations (E1), (E2), and (E3) are
autonomous (assuming a, b are constant in (E1)). Equation (F1) is separable and we can solve it
implicitly for y = y(t) as G(y) = t+C. However, a lot of descriptive information about (F1) can be
obtained without solving it.

F.2 Since RHS of (F1) does not involve t, the direction field of (F1) does not change under hori-
zontal shifts. Thus, a horizontal shift of a solution curve is again a solution curve. Furthermore, if
y ∗ is a real number such that g(y ∗ ) = 0, then the constant function y(t) = y ∗ is a solution of (F1).
Such a number y ∗ is an equilibrium point for (F1) and y(t) = y ∗ is an equilibrium solution of (F1).
The corresponding solution curve is the horizontal line y = y ∗ in (t, y)-plane. The horizontal graphs
of the equilibrium solutions of (F1) partition the (t, y)-plane into horizontal bands y1∗ < y < y2∗ . In
each band, the function g(y) does not change sign. Thus, in each single band, all solution curves
of (F1) either descend and approach the line y = y1∗ or ascend and approach the line y = y2∗ as t
approaches ∞.

F.3 Here is an example. The equilibrium solutions of

y ′ = (y + 3)2 (y + 1)(y − 3), y = y(t), (F2)

are y = −3, y = −1, and y = 3. The graphs of these solutions are the horizontal lines y = −3,
y = −1, and y = 3, shown in the third plot in Figure 1. These lines partition the ty-plane into

6
 g(y) y y
3 y=3

y t
−3 −1 3
−1 y = −1

−3 y = −3

Figure 1: Plots for the differential equation y ′ = g(y) = (y + 3)2 (y + 1)(y − 3)

horizontal bands y1∗ < y < y2∗ . Since solution curves of the differential equation (F2) do not intersect,
no solution curve can cross the graphs of the equilibrium solutions. For example, if y = y(t) is a
solution of (F2) such that y(t0 ) ∈ (−1, −3) for some t0 , then y(t) ∈ (−1, −3) for all t. In each band,
the function g(y) does not change sign. Thus, in each single band, all solution curves of (F2) either
descend or ascend. Furthermore, each solution curve must approach either an equilibrium solution
curve or ±∞ as t −→ ±∞. The best way to tell whether the solution curves in the given band
descend or ascend is by sketching the graph of the function g = g(y), as done for the differential
equation (F2) in the first plot in Figure 1. Note that the y-intercepts of this graph correspond
to the equilibrium solutions of the differential equation. The phase line in the middle of Figure 1
shows the equilibrium points for the differential equation (F2), or the y-intercepts of the graph of g.
It also indicates, using arrows, whether the solution curves in each band cut out by the horizontal
equilibrium-solution lines ascend and descend. The arrow corresponding to a segment of the phase
line points up (down) if g(y) is positive (negative) on the this segment.

G: Euler’s Formula and its Implications

eiθ = cos θ + i sin θ

Here are some consequences:

e−iθ = cos θ − i sin θ
eiθ + e−iθ eiθ − e−iθ
cos θ = and sin θ =
2 2i

The double-angle formulas follow from Euler’s formula:

cos 2θ = cos2 θ − sin2 θ = 2 cos2 θ − 1 = 1 − 2 sin2 θ and sin 2θ = 2 cos θ · sin θ,
as do the more general formulas:
cos(α±β) = cos α · cos β ∓ sin α · sin β and sin(α±β) = sin α · cos β ± cos α · sin β.

Please derive all these from Euler’s formula.