1 Linear Growth and Decay Models

1.1 Population Growth Models

1.2 Effects of Immigration and Emigration on Population Size

1.3 Interest Compounded Continuously

Let the amount at time t be x(t) and let interest at rate r per unit
amount per unit time be compounded continuously then
x(t + ∆t) = x(t) + rx(t)∆t + O(∆t)
giving:
dx
= xr; x(t) = x(0)ert (1)
dt
This formula can also be derived from the formula for compound
interest:
µ ¶nt
r
x(t) = x(0) 1 + (2)
n
When interest is payable n times per unit time, by taking the limit
as n → ∞.

1.4 Decrease of Temperature

According to Newton’s law of cooling, the rate of change of tem-

perature of a body is proportional to the difference between the
temperature T of the body and the temperature Ts of the sur-
rounding medium, so that:
dT
= k(T − Ts), K < 0 (3)
dt
1
and,

T (t) − Ts = (T (0) − Ts)ekt (4)

and the excess of the temperature of the body over that of the sur-
rounding medium decays exponentially.

1.5 Diffusion

According to Fick’s law of diffusion, the time rate of movement of

a solute across a thin membrane is proportional to

• the area of the membrane, and

• to the difference in concentrations of the solute on the two

sides of the membrane.

If the area of the membrane is constant and the concentration of

solute on one side is kept fixed at a and the concentration of the
solution on the other side is initially is c0 < a, then Fick’s laws gives:
dc
= k(a − c) c(0) = c0 (5)
dt
so that:

a − c(t) = (a − c0)e−kt (6)

and we can see that c(t) → a as t → ∞, whatever be the value of

c0.

2
1.6 Change of a Price of a Commodity

Let p(t) be the price of a commodity at time t, then its rate of

change is proportional to the difference between the demand d(t)
and the supply s(t) of the commodity in the market, so that:
µ ¶
dp
= k d(t) − s(t) (7)
dt
where k > 0, since if demand is more than the supply, the price
increases.
If d(t) and s(t) are assumed linear functions of p(t), i.e., if:

d(t) = d1 + d2p(t), s(t) = s1 + s2p(t), d2 > s2 > 0 (8)

we get:
µ ¶
dp
= k (d1 − s1) + (d2 − s2)p(t) = k(a − bp(t))
dt
µ ¶
= K pe − p(t) (9)

where pe is the equilibrium price, pe = a/b and K = k/b, so:

pe − p(t) = (pe − p(0))e−Kt (10)

and p(t) → pe as t → ∞.

2 Non-Linear Growth and Decay Models

2.1 Logistic Law of Population Growth

As population increases due to overcrowding and limitations of re-

sources, the birth rate b decreases and death rate d increases with
3
the population size x. The simplest assumptions are:

b = b1 − b2x, d = d1 + d2x, b1, b2, d1, d2 > 0 (11)

so that:
µ ¶
dx
= (b1 − d1) − (b2 + d2)x x = (a − bx)x, a, b > 0 (12)
dt
Integrating, we get:
x(t) x(0)
= eat (13)
a − bx(t) a − bx(0)
Equations (12) and (13) show that:
dx
(i) x(0) < a/b =⇒ x(t) < a/b =⇒ dt > 0 =⇒ x(t) is
a monotonic increasing function of t which approaches a/b as
t → ∞.
dx
(ii) x(0) > a/b =⇒ x(t) > a/b =⇒ dt < 0 =⇒ x(t) is
a monotonic decreasing function of t which approaches a/b as
t → ∞.

Figure 1:

4
Now from (12):
d2 x
= a − 2bx (14)
dt2
d2 x
so that dt2
T 0 according as x T a/2b. Thus in case (i) the growth
curve is convex if x < a/2b and is concave if x > a/2b and it has a
point of inflexion at x = a/2b. Thus the graph of x(t) against t is as
given in Figure 1.
Moreover:

(a). If x(0) < a/2b, x(t) increases at an increasing rate till x(t)
reaches a/2b and then it increases at a decreasing rate and ap-
proaches a/b as t → ∞.

(b). If a/2b < x(0) < a/b, x(t) increases at a decreasing rate and
approaches a/b as t → ∞.

(c). If x(0) = a/b, x(t) is always equal to a/b.

(d). If x(0) > a/b, x(t) decreases at a decreasing absolute rate and
approaches a/b as t → ∞.

2.2 Spread of Technological Innovations and Infectious Diseases

Let N (t) be the number of companies which have adopted a tech-

nological innovation till time t, then the rate of change of number
of these companies depends both on the number of companies
which have adopted this innovation and those which have not

5
yet adopted it, so that if R is the total number of companies in the
region:
dN
= kN (R − N ) (15)
dt
which is the logistic law and shows that ultimately all companies will
adopt this innovation.

Similarly, if N (t) is the number of infected persons, the rate at which

the number of infected persons increases depends on the product
of the numbers of infected and susceptible persons. As such we
again get (15), where R is the total number of persons in the system.

It may be noted in both the examples, while N (t) is essentially an

integer-valued variable, we have treated it as a continuous variable.
This can be regarded as an idealization of the situation or as an
approximation to reality.

2.3 Rate of Dissolution

Let x(t) be amount of undissolved solute in a solvent at time t and

let c0 be the maximum concentration or saturation concentration,
i.e. the maximum amount of the solute that can be dissolved in a
unit volume of the solvent. Let V be the volume of the solvent. It
is found that the rate at which the solute is dissolved is:

• proportional to the amount of undissolved solute, and

6
 • to the difference between the concentration of the solute at
time t and the maximum possible concentration.

and so we get:
µ ¶
dx x(0) − x(t)
= kx(t) − c0
dt V
µ ¶
kx(t)
= (x0 − c0V ) − x(t) (16)
V

2.4 Law of Mass Action: Chemical Reactions

Two chemical substances combine in the ratio a : b to form a third

substance Z. If z(t) is the amount of the third substance at time t,
then a proportion az(t)/(a + b) of it consists of the first substance
and a proportion bz(t)/(a + b) consists of second substance. The
rate of formation of the third substance is proportional to the
product of the amount of the two component substances which
have not yet combined together. If A and B are the initial amounts
of the two substances, then we get:
µ ¶µ ¶
dz az bz
=k A− B− (17)
dt a+b a+b
This is non-linear differential equation for a second order reaction.
Similarly, for an nth order reaction, we get the non-linear equation:
dz
= k(A1 − a1z)(A2 − a2z) . . . (An − anz) (18)
dt
where a1 + a2 + . . . + an = 1.

7
3 Compartment Models

3.1 A Simple Compartment Models

Let a vessel contain a volume V of a solution with concentration c(t)

of a substance at time t. let a solution with constant concentration
C in an overhead tank enter the vessel at a constant rate R and after
mixing thoroughly with the solution in the vessel, let the mixture
with concentration c(t) leaves the vessel at the same rate R so that
the volume of the solution in the vessel remains V .

Figure 2:

Using the principle of continuity, we get:

V (c(t + ∆t) − c(t)) = RC∆t − Rc(t)∆t + O(∆t)

giving
dc
V + Rc = RC (19)
dt

8
Integrating we get:
µ ¶ µ µ ¶¶
R R
c(t) = c(0) exp − t + C 1 − exp − t (20)
V V
As t → ∞, c(t) → C, so that ultimately the vessel has the same
concentration as the overhead tank. Since:
µ ¶
R
c(t) = C − (C − c0) exp − t (21)
V
if C > c0, the concentration in the vessel increases to C; on the other
hand if C < c0, the concentration in the vessel decreases to C.

Figure 3:

If the rate R0 at which the solution leaves the vessel is less than R,
the equation of continuity gives:
d
[(V0 + (R − R0)t)c(t)] = RC − R0c(t) (22)
dt
where V is the initial volume of the solution in the vessel. This is
also a linear differential equation of the first order and can be solved
easily.
9
3.2 Diffusion of Glucose or Medicine in Blood Stream

Let the volume of blood in the human body be V and let the initial
concentration of glucose in the blood stream be c(0). Let glucose be
introduced in the blood stream at a constant rate I. Glucose is also
removed from the blood stream due to the physiological needs of the
human body at a rate proportional to c(t), so that the continuity
principle gives:
dc
V = I − kc (23)
dt
which is similar to (19).

Now let a dose D of a medicine be given to a patient at regular

intervals of duration T each. The medicine also disappears from
the system at a rate proportional to c(t), the concentration of the
medicine in the blood stream, then the differential equation given by
the continuity principle is:
dc
V = −kc (24)
dt
Integrating, we have:
µ ¶
k
c(t) = D exp − t , 0 ≤ t < T (25)
V
µ ¶
At time T , the residue of the first dose is D exp − Vk T and now

10
another dose D is given so that we get:
µ µ ¶ ¶ µ ¶
k k
c(t) = D exp − T + D exp − (t − T )
V V
µ ¶ µ ¶
k k
= D exp − t + D exp − (t − T ) (26)
V V
T ≤ t < 2T

From the equation it is clear that the first term gives the residual of
the first dose and the second term gives the residual of the second
dose. Proceeding in the same way, we get the residual after n doses:
µ ¶ µ ¶
k k
c(t) = D exp − t + D exp − (t − T )
V V
µ ¶ µ ¶
k k
+ D exp − (t − 2T ) + . . . + D exp − (t − (n − 1)T )
V V
µ ¶· µ ¶ µ ¶
k k k
= D exp − t 1 + exp T + exp 2T
V V V
µ ¶¸
k
+ . . . + exp (n − 1)T
V
µ ¶
µ ¶ exp Vk nT − 1
k
= D exp − t µ ¶ , (n − 1)T ≤ t < nT (27)
V
exp Vk T − 1
µ ¶
1 − exp − Vk nT
c(nT − 0) = D µ ¶ (28)
exp Vk T − 1

11
and
µ ¶ µ ¶
k
exp VT − exp − Vk nT
c(nT + 0) = D µ ¶ (29)
exp Vk T − 1
µ ¶¶ µ
Thus the concentration never exceeds D/ 1 − exp − Vk T . The
graph of c(t) is shown in Figure.

Figure 4:

• Thus in each interval, concentration decreases

• In any interval, the concentration is maximum at the beginning

of this interval and thus maximum concentration at the begin-
ning of an interval goes on increasing as the number of intervals
increases
k
• The maximum concentration is always below than D/(1−e− V T )
12
 • The minimum value in an interval occurs at the end of each
k
interval. This also increases but it lies below D/(e V T − 1)

The concentration curve is piecewise continuous and has points of

discontinuity at T, 2T, 3T, . . ..

By injecting glucose or pencillin in blood and fitting curve (25) to

the data, we can estimate the value of k and V . In particular this
gives a method for finding the volume of blood in the human body.

3.3 The Case of a Succession of Compartments

Let a solution with concentration c(t) of a solute pass successively

into n tanks in which the initial concentration of the solutions are
c1(0), c2(0), . . . , cn(0). The rate of inflow in each tank is the same as
the rate of outflow from the tank. We have to find out the concen-
trations c1(t), c2(t), . . . , cn(t) at time t. We get the equations:
dc1
V = Rc − Rc1
dt
dc2
V = Rc1 − Rc2 (30)
dt
... ... ...
dcn
V = Rcn−1 − Rcn
dt
By solving the first equation, we get c1(t). Substituting the value of
c1(t) and proceeding in the same way, we can find c2(t), . . . , cn(t).

13
4 Modeling through ODEs of first order

4.1 Simple Harmonic Motion

A particle moves in a straight line in such a manner that its acceler-

ation is always proportional to its distance from the origin and
is always directed towards origin, so that:
dv
v = −µx (31)
dx
Integrating:

v 2 = µ(a2 − x2) (32)

where the particle is initially at rest at x = a. Equation (31) gives:

dx √ p 2
= − µ a − x2 (33)
dt
we take negative sign since velocity increases as x decreases.

Figure 5:

Thus in simple harmonic motion, both displacement and velocity are

√
periodic functions with period 2π/ µ.

The particle starts from A with zero velocity and moves towards 0
√
with increasing velocity and reaches 0 at time π/2 µ with veloc-
√
ity µa. It continues to move in the same direction, but now with
14
decreasing velocity till it reaches A’ (OA’=a), where its velocity is
again zero. It then begins moving towards 0 with increasing velocity
√
and reaches 0 with velocity µa and again comes to rest at A after
a total time period 2π/µ. the periodic motion then repeats itself.

As one example of SHM, consider a particle of mass m attached

to one end of a perfectly elastic string, the other end of which is
attached to a fixed point 0 as given in Figure. The particle moves
under gravity in vacuum. Let l0 be the natural length of the string

Figure 6:

and let a be its extension when the particle is in equilibrium so that

by Hook’s law:
a
mg = T0 = λ (34)
l0
where λ is the coefficient of elasticity. Now let the string be further
stretched a distance c and then the mass be left free. The equation
of motion which states that:
15
mass×acceleration in any direction=force on the particle in that di-
rection, gives:
dv a+x λx
mv = mg − T = mg − λ =− (35)
dx l0 l0
or
dv λx gx
v = =− (36)
dx ml0 a
q
a
which gives a simple harmonic motion with time period 2π g.

4.2 Motion under gravity in a resisting medium

A particle falls under gravity in a medium in which the resistance

is proportional to the velocity. The equation of motion is:
dv
m = mg − mkv
dt
or
dv g
= kdt; V = (37)
V −v k
Integrating:

V − v = V e−kt (38)

If the particle starts from rest with zero velocity. Equation (38) gives:

v = V (1 − e−kt) (39)

so that the velocity goes on increasing and approaches the limiting

velocity g/k as t → ∞. Replacing v by dx/dt, we get:
dx
= V (1 − e−kt) (40)
dt
16
Integrating using x = 0 when t = 0, we get:
V e−kt V
x=Vt+ − (41)
k k
4.3 Motion of a rocket

As a first idealization, we neglect both gravity and air resistance. A

rocket moves forward because of large supersonic velocity with which
gases produces by burning of the fuel inside the rocket come out of
the converging-diverging nozzle of the rocket.

Figure 7:

Let m(t) bet he mass of the rocket at time t and let it moves forward
with velocity v(t) so that the momentum at time t is m(t)v(t).

In the interval of time (t, t + ∆t), the mass of rocket becomes:

dm
m(t + ∆t) = m(t) + ∆t + O(∆t)
dt
Since the rocket is losing mass, dm
dt is negative and the mass of gases
− dm
dt ∆t moves with velocity u relative to the rocket, i.e. with a

17
velocity u(u+∆t)−u relative to the earth so the the total momentum
of the rocket and the gases at time t + ∆t is:
dm
m(t + ∆t)v(t + ∆t) − ∆t(v(t + ∆t) − u) (42)
dt
Since we are neglecting air resistance and gravity, there is no external
force on the rocket and as such the momentum is conserved, giving
the equation:
µ ¶µ ¶
dm dv
m(t)v(t) = m(t) + ∆t v(t) + ∆t
dt dt
dm
− ∆t(v − u) + O(∆t)2 (43)
dt
Dividing by ∆t and proceeding to the limit ∆t → 0, we get:
dv dm
m(t) = −u (44)
dt dt
or
dm 1
= − dv (45)
m u
or
m(t) v(t)
ln =− (46)
m(0) u
assuming that the rocket starts with zero velocity.

As the fuel burns, the mass of the rocket decreases. Initially the
mass of the rocket was m = mP + mF + mS , where mP is the mass
of pay-load, mF is the mass of the fuel and mS is the mass of the

18
structure.

When the fuel is completely burnt out, mF becomes zero and if vB

is the velocity of the rocket at this stage, when the fuel is all burnt,
then (45) gives:
µ ¶
mP + mF + mS mF
vB = u ln = u ln 1 + (47)
mP + mS mP + mS
This is the maximum velocity that the rocket can attain and it de-
mF
pends on the velocity u of efflux of gases and the ratio mP +mS . The
mF
larger the values of u and mP +mS , the larger will be the maximum
velocity attained.

For the best modern fuels and structural materials, the maximum ve-
locity this gives is 7 km/sec. In practice, it would be much less since
we have neglected air resistance and gravity, both of which tend to
reduce the velocity. However a rocket is to place a satellite in orbit,
we require a velocity of more than 7 km/sec.

The problem can be overcome by using the concept of multi-stage

rockets.

The fuel may be carried in a number of containers and when the fuel
of a container is burnt up, the container is thrown away, so that the
rocket has not to carry any dead weight.

19
Thus in a three-stage rocket, let mF1 , mF2 , mF3 be the masses of the
fuels and mS1 , mS2 , mS3 be the corresponding masses of containers,
then the velocity at the end of the first stage is:
mP + mF1 + mS1 + mF2 + mS2 + mF3 + mS3
v1 = u ln (48)
mP + mF2 + mS2 + mF3 + mS3
At the end of the second stage, the velocity is:
mP + mF2 + mF3 + mS3
v2 = v1 + u ln (49)
mP + mF3 + mS3
and at the end of the third stage, the velocity:
mP + mF3
v3 = v2 + u ln (50)
mP
In this way, a much larger velocity is obtained that can be obtained
by a single stage rocket.

5 Simple Geometric Problems through ODEs

5.1 Some Geometric Problems

(i) Find the curves for which the tangent at a point is always per-
pendicular to the line joining the point to the origin.

the slop of the tangent is dy/dx and the slope of the line joining the
point (x, y) to the origin is y/x. Since these lines are orthogonal:
dy x
=− (51)
dx y
20
Integrating:
x2 + y 2 = a2 (52)
which represents a family of concentric circles.

(ii) Find a curve which passes through the point (1, 1) in the xy−plane
having at each of its points the slop −y/x.

The function giving the desired curve must be a solution of the dif-
ferential equation:
y
y0 = − (53)
x
Integrating we get:
c
y= (54)
x
But the curve passes through (1, 1), so c = 1, hence the final answer
is y = x1 .

(iii) Find the curves for which the projection of the normal on the
x-axis is of constant length.

This condition gives:

dy
y =k (55)
dx
Integrating:
y 2 = 2kx + A (56)
21
which represents a family of parabolas, all with the same axis and
same length of latus rectum.

(iv) Find curves for which tangent makes a constant angle with the
radius vector.

This condition gives:

dθ
r = tan α (57)
dr
Integrating:

r = Aeθ cot α (58)

which represents a family of equiangular spirals.

22