CHAPTER 3

Continuous Population Models for Single Species

3.1 Introduction
The dynamics of species interactions within ecosystems have long captivated scientific in-
quiry. Single species competition models stand as pivotal tools in understanding the intricate
interplay between organisms competing for shared resources. These models delve into the
dynamics of population growth, exploring how a single species’ abundance is shaped by its
interactions with the environment and its own kind. By examining the mechanisms of com-
petition, these models shed light on fundamental ecological principles, offering insights into
the regulation and sustainability of ecosystems. This introduction sets the stage for a com-
prehensive exploration of the dynamics inherent in single species competition models.

3.2 Verhulst Logistic Growth Model

The first theory of population growth was formulated by T.R. Malthus in 1798 who concluded
that a population would grow geometrically if its birth rate was greater than the death rate.
The model is unrealistic for any long-term prediction. In reality, any particular population
can’t increase indefinitely; at least this has never been observed. The Belgian mathematician
and biologist P.F. Verhulst proposed the intrinsic growth rate,

R = R(u) = r − bu; (r > 0, b > 0)

in 1838 to overcome this deficiency in the Malthusian model. The intrinsic growth rate
decreases linearly as population size increases. Here, r stands for the growth rate independent
of environmental factors and b denotes the impact of growing population density.

0.8

0.6

R=a−bu
R(u)

0.4

0.2

−0.2
0 0.2 0.4 0.6 0.8 1
u

Figure 3.2.1: Logistic growth rate.

The model with initial population u(0) = u0 will be

u
u′ (t) = u(r − bu) = ru(1 − ), (3.2.1)
K
where K = r is called the carrying capacity of the environment which is usually determined
b
by the available sustaining resources, that is, the largest population in the environment can
sustain without loss. The population level K is sometimes called the saturation level since
for large populations there are more deaths than births.

1
 In this model, the per capita birth rate is r(1 − Ku ), that is, it is dependent on u. As the
population grows, some members of the population u infer with each other in competition
for some critical resource. This competition diminishes the growth rate. The initial growth
is approximately exponential, then as competition arises, the growth slows, and at maturity,
growth stops. The model (3.2.1) is known as the logistic equation, and its solution is called
the logistic function.
The graphs of the logistic function are called a logistic curve or S-shaped curve of growth
(logistic comes from the Greek word “logistikos” which means rational) and have proved to
be of great accuracy in predicting the growth patterns in a limited space of certain types
of bacteria or protozoa. The population satisfying (1.2.1) is said to obey a logistic law of
the population growth. Notice that if we choose b = 0 the logistic equation (3.2.1) would
reduce to (1.3.1), whose solution is the exponential law of population growth (1.3.2).
To solve the model, separating the variables, we obtain for u ̸= 0, or r − bu ̸= 0
u r
u′ (t) = u(r − bu) = ru(1 − ), K=
K b

Kdu
= rdt.
u(K − u)
Then  
1 1
+ du = rdt.
u K −u
Integrating in both sides, we get
u
ln = rt + c
K −u
that implies
u
= Cert , C = ec
K −u
Thus,
CKert
u(t) =
1 + Cert
u0
If t = 0 then u = u0 such that C = K−u0 which yields the solution

u0 Kert u0 K
u(t) = =
(K − u0 ) + u0 e rt u0 + (K − u0 )e−rt

Hence, the solution of the model

K
u(t) = . (3.2.2)
1+ ( uK0 − 1)e−rt

Taking t = t0 ̸≡ 0, then we obtain

K
u(t) = .
1+ ( uK0 − 1)e−r(t−t0 )

This equation is called the logistic law of growth.

2
Remark 3.2.1. Note that u(t) remains positive for t ≥ t0 and take the limit of (3.2.2), we
obtain that
umrx = lim u(t) = K,
t→∞

which shows that there is a limit to the growth of u as desired by the biological facts.
Also u(t) → 0 as t → −∞.

3.2.1 Behavior of the solution

From (3.2.1) and (3.2.2), we conclude that
Case 1: when u < 0 then u′ (t) > 0, and the solution u(t) decreases without bound for
increasing t and u(t) → 0 as t → −∞.
Case 2: when u < K, then u′ (t) > 0, and the solution u(t) will be increasing and u(t) → K
as t → ∞ if u0 < K.
Case 3: when u = K, then u′ (t) = 0, and the solution u(t) will be equal to K if u0 = K.
Case 4: when u > K, then u′ (t) < 0, and the solution u(t) will be decreasing and u(t) → K
as t → ∞ if u0 > K.
Thus, the population is bounded between u0 and K. In other words, the population can never
cross K, and it tends to K in the long run.
In the above observation, K is called the carrying capacity of the environment, when K → ∞
the denominator goes to unity, and the solution becomes the case for the Malthusian model.

700

P0 600

K=500

400
u(t)

300
P0
200

P0 100

0
0 2 4 6 8 10
t

Figure 3.2.2: Logistic growth model.

Point of Inflection: The solution to the logistic differential equation has a point of inflec-
tion. To find this point, set the second derivative equal to zero:

u ) and u′′ (t) = r2 u(1 − 2u )(1 − u )

We have u′ (t) = ru(1 − K K K

2
ˆ The second derivative, d u is positive for 0 < u < K
2 or u > K, and negative for
dt2
K < u < K. Therefore, u′ (t) is increasing if u(t) < K or u(t) > K, and u′ (t) is
2 2
decreasing if K
2 < u(t) < K.

3
 2
ˆ Also, d u vanishes when u = K 2 so that there is a point of inflection in the population
dt2
growth curve when half of the final population size is reached.

So, from the solution, we obtain

K u0 K
= .
2 u0 + (K − u0 )e−rt

Then  
K
1+ − 1 e−rt = 2.
u0
 
Thus, t = 1r ln K − 1 , which is the time of the point of inflection.
u0

Notice that if u0 > K 2 , then this quantity is undefined, and the graph does not have a point
of inflection. In the logistic graph, the point of inflection can be seen as the point where the
graph changes from concave up to concave down. This is where the “leveling off” starts to
occur, because the net growth rate becomes slower as the population starts to approach the
carrying capacity.

Note: The point of inflection also can be found using the 2nd derivative of the model solution
and  setting u ′′
 = 0, try to find the point of inflection time as established above is t =
1 ln K − 1 . So, we can conclude that the time period before the population reaches half
r u0
of its limiting value is a period of accelerated growth, and after this point, the rate of growth
decreases and in course of time it reaches zero, this being a period of diminishing growth.

Investigate the Concavity

The graph of u(t) verses t is increasing or decreasing depending on whether du is positive
dt
or negative. Note that du is positive then u and t are increasing or decreasing together.
dt
Consequently, if du is positive and increasing as a function of u, then it is also increasing as
dt
a function of t, and the graph of u versus t is concave upward. Similarly, if du is positive
dt
and decreasing, then the graph of u versus t is concave downward. The situation is reversed
if du is negative for u is decreasing as t is increasing and vice versa.
dt

140

120

100

80
du
dt

60

40

20

0
O K
−20
−100 0 100 200 300 400 500 600
u

Figure 3.2.3: Phase plane of the logistic equation.

4
Proposition 3.2.1. Assume that t0 < t1 < t2 are equally spaced time values. Let the
corresponding population sizes be u0 , u1 , u2 , respectively. Then the growth rate r and the
u ),
carrying capacity K of a population that obeys the logistic law of population u′ = ru(1 − K
u(t0 ) = u0 then
2 1 1
    − −
1 1 1 1 1 u u0 u2
r= ln − / − , K= 1 .
t0 − t1 u2 u1 u1 u0 1 − 1
u21 u0 u2

Solution. We have the solution of the equation

K
u(t) = .
K − 1
e−r(t−t0 )
1+ u0

Then using the reciprocal

   
1 1 K −r(t−t0 )
= 1+ −1 e
u K u0
1 1
= [1 − e−r(t−t0 ) ] + e−r(t−t0 ) .
K u0
Also
1 1 1
1 − e−r(t1 −t0 ) + e−r(t1 −t0 ) ,

= (3.2.3)
u1 K u0
and
1 1 1
1 − e−r(t2 −t1 ) + e−r(t2 −t1 ) .

=
u2 K u1
Then  
1 1 1 1
− = − e−r(t1 −t0 ) .
u2 u1 u1 u0
Hence,    
1 1 1 1 1
r= ln − / − .
t0 − t1 u2 u1 u1 u0
From (3.2.3) we obtain
1 1 1
− e−r(t1 −t0 ) = 1 − e−r(t1 −t0 ) .

u1 u0 K
Then
1 − e−r(t1 −t0 )
K= 1 .
u1− u10 e−r(t1 −t0 )
Substituting the value of r(t1 − t0 ) then we get the value of K.

This is the appropriate formulation of doubling time for a logistic model. Since u(t) < K for
all t, the doubling time is well-defined if the initial population satisfies 0 < u0 < K
2.

5
Lemma 3.2.1. The doubling time for the logistic model is defined for 0 < u0 < K
= r,
2 2b
and by the formula  
1 2(K − u0 ) 1 2(r − bu0 )
T = ln = ln .
r K − 2u0 r r − 2bu0
In particular, the doubling time for the logistic growth model is larger than the corresponding
doubling time for the exponential growth model.
Solution. The solution of the model is
u0 K
u(t) = .
u0 + (K − u0 )e−rt
At time T , let u = 2u0
u0 K
2u0 = .
u0 + (K − u0 )e−rT
We obtain
2(K − u0 )e−rT = K − 2u0 ,
then  
1 2(K − u0 ) 1 2(r − bu0 )
T = ln = ln ,
r K − 2u0 r r − 2bu0
which is the time for doubling of the population.
Example 3.2.1. The population of a country for the years 1930-1990 is given in Table 3.2.1.
Using this data find (a) the theoretically maximum population, (b) the population in 2020,
and (c) the population of 1900.

Year Population (in Millions)

1930 76.0
1940 92.0
1950 105.7
1960 122.8
1970 131.7
1980 151.1
1990 179.3
Table 3.2.1: Population of a country.

Solution. Let t = 0, 1, and 2 correspond to the years 1930, 1960 and 1990, respectively.
Then
u0 = 76.0, u1 = 122.8, and u2 = 179.3.
(a) By using Proposition 3.2.1, we have umrx = K = 346.3, which is the maximum population
of the country.

(b) We obtain the number of population at any time is

K 346.3 346.3
u(t) = = = . (3.2.4)
1+ ( uK0 − 1)e −rt 1 + 3.557(0.5117) t 1 + 3.557e−0.6678t

where r is defined in Proposition 3.2.1.

6
(c) The year 2020 corresponds to t = 3, thus from equation (3.2.4), we get u = u3 = 234.5,
which is the population in 2020.

(d) The year 1900 corresponds, t = −1. We obtain from equation (3.2.4), then u = u−1 =
43.6, which is the population in 1900.

Remark 3.2.2. If we consider a population to be governed by

du
= f (u),
dt
where f (u) = u(r − bu), is a nonlinear function of u then the equilibrium or steady state
solutions u∗ are solutions of f (u) = 0 and are linearly stable to small perturbations if f ′ (u) <
0, and unstable if f ′ (u∗ ) > 0. There are several equilibria, and they depend on the system
f (u) models. Graphically plotting f (u) against u immediately gives the equilibrium.
We have  
′ ′ r
f (0) = r > 0 and f = −r < 0.
b
Since r and b are positive, then the critical point 0 is unstable, and critical point r is stable.
b
In particular, any solution for which u(0) > 0 must approach the value r called the carrying
b
capacity of the model. The general shape of the solution of the model 0 < u0 < r .
b

150 700

600
100
K=500
u(a−bu)

50 400
u(t)

300

0
O K 200

−50 100

0
−100 0 100 200 300 400 500 600 700 0 5 10 15
u t

Figure 3.2.4: Auxiliary graph for the logistic equation.

3.3 Stability of the Equilibrium States of Logistic Model

Graphical Analysis

200

150

100
du
dt

50

0
O K

−100 0 100 200 300 400 500 600 700 800 900
u

Figure 3.3.1: du versus u.

dt

7
Since the logistic equation is du = ru(1 − K u ). The graph is a parabola with intercepts at
dt 
(0, 0) and (K, 0) with vertex at K rK du
2 , 4 . For 0 < u < K, we see that dt > 0 and u is an
increasing function of t, that is, indicated by the rightward pointing arrow near the u-axis.
Similarly, if u > K, then du < 0 and u(t) are decreasing as indicated by the leftward pointing
dt
arrow. If u = 0 or u = K then du = 0 and u(t) does not change. The constant solution u = 0
dt
or u = K are called equilibrium solutions, and the points u = 0 or u = K on the u-axis are
called equilibrium points or critical points.
Now we sketch the graph of the solutions u versus t for t > 0 and for different initial values
u(0). It is helpful to know the relationship between the graph of du versus P and the graph
dt
of u versus t. The graph of u(t) versus t is increasing or decreasing depending on whether du
dt
is positive or negative.
Note that du > 0, then u and t are increasing, or decreasing together. Consequently, if
dt
du > 0 increases as a function of u, then it is also increasing as a function of t, and the graph
dt
of u versus t is concave upward. Similarly, if du > 0 is decreasing, the graph of u versus t is
dt
concave downward.
The situation is reversed if du < 0 for u is decreasing as t increases and vice versa.
dt
Therefore, if du < 0 and increases as a function of u, then it is decreasing as a function of t
dt
and the graph of u versus t is concave downward. Similarly, if du < 0 is decreasing, then the
dt
graph of u versus t is concave upward.
The horizontal lines are the equilibrium solution u = 0 and u = K. From Figure 3.3.2 note
that du > 0 and increasing for 0 < u < K 2,
dt

800
P0
700

600

K 500
u(t)

400

P0
300
K/2
200

P0 100

0
0 2 4 6 8 10
t

Figure 3.3.2: u versus t.

so the graph of u(t) is increasing, and concave upward there. Similarly, du > 0 and decreasing
dt
for K
2 < u < K, so the graph of u(t) is increasing and concave upward there. Thus, solutions
that start below K 2 have the S-shape shown in Figure 3.3.2. On the other hand, for u > K,
du < o and decreasing, so the graph of u(t) is decreasing and concave upward for these values
dt
of u.
Finally, the existence and uniqueness theorem guarantees that two distinct solutions never
pass through the same point. Hence while the solution approaches the equilibrium solution

8
u = K as t → ∞, thus does not attain this value at any time. Since K is the upper bound
that is approached but not exceeded by growing populations starting below this value, it is
natural to refer to K as the saturation level, as the environmental carrying capacity for the
given species.
The solution of the logistic equation (3.2.2) contains only two different equilibrium solutions
u = 0 and u = K corresponding to the initial condition u0 = 0 and u0 = K, respectively.
If u0 = 0 then u(t) = 0 for all t. If u0 > 0 then limt→∞ u(t) = K. Thus for each u0 > 0, the
solution approaches the equilibrium solution u(t) = K asymptotically as t → ∞. Thus the
constant solution u(t) = K is an asymptotically stable solution of (3.2.1). This means that
regardless of the starting population size, as long as it is positive, the population eventually
approaches the saturation level K.
On the other hand, u(t) = 0 is an unstable equilibrium solution which means the solution
remains near zero since its initial value is equal to zero. Even the solution that starts very
near zero grows as t increases, and we have seen approach K as t → ∞.

Perturbation Analysis
u
. The equilibrium population is given by
Since the equation is u′ = ru 1 − K
u
ru(1 − ) = 0 ⇔ u = 0, u = K.
K
Thus, u = 0 and u = K are two equilibrium states of the model.
(i) Let u = 0 + u where u is very small.
Put the value of u in (1.2.1) then the equation becomes

du ru2
= ru − .
dt K
2
So, the linearized equation du = ru (neglecting ruK as so small) which is the same as
dt
the Malthusian population model for small population size.
The solution of the linear equation is u(t) = u(0)ert , that is, the population becomes
larger and larger as t → ∞. So, the state u = 0 is unstable.

(ii) Let u = K + u, then the equation (1.2.1) becomes

du ru2
= −ru − .
dt K

So, the linearized equation du = −ru. Then the solution of the linearized equation is
dt
u(t) = u(0)e−rt .
Here |u(0)| < ∂ ⇒ |u(t)| < ϵ and limt→∞ u(t) = 0. So, the equilibrium state u = K is
asymptotically stable.

Estimating the Bangladesh population by Logistic Law growth

We need three base years for a logistic model. We have chosen equally spaced years 1901,
1910, and 1921 and corresponding population is 28.92, 39.55, and 33.25, respectively. The

9
growth rate r and carrying capacity K are given by Proposition 3.2.1.
The growth rate, r = 0.0639863. The carrying capacity, K = 35.7218.
With these r and K use the following Bangladesh population by the Logistic law of population
growth
35.7218e0.0639863t
u(t) = . (3.3.1)
1.57804 × 1052 + e0.0639863t
We plot the Bangladesh population derived from the equation (3.3.1).

160

140

120

100
u(t)

80

60

40

20

0
1900 1920 1940 1960 1980 2000
t

Figure 3.3.3: Logistic curve for the Bangladesh population.

200

180

160

140

120
u(t)

100

80

60

40

20

1900 1920 1940 1960 1980 2000

t

Figure 3.3.4: Bangladesh population predicted by logistic law growth.

Consider a specific example of this type of growth.

Example 3.3.1. The population u of a city satisfies the logistic law
du u u2
= − 8,
dt 100 10
where time t is measured in years. If the population of this city was 1,00,000 in 2000,
determine the population as a function of time for t > 2000. In particular, answer the
following questions:
(a) What will be the population in 2020?
(b) In which year will be the population double of that in the year 2000?
(c) How large will the population ultimately be?
Solution. Given u(2000) = 1, 00, 000.
Separating variables, we obtain
du
= dt,
(10)−2 u[1− (10)−6 u]

10
  
1 (10)−6
and hence 100 u + 1−(10)−6 u
du = dt.
Integrating in both sides, we obtain
 
u 1
ln −6
= t + c1 ,
1 − (10) u 100

and hence
u
= cet/100 .
1 − (10)−6 u

We obtain u(t) = cet/100 .

1 + (10)−6 cet/100
Now applying the initial condition, we have

ce20.00
(10)5 = ,
1 + (10)−6 ce20.00

and we obtain
(10)5 (10)6
c= = .
e20.00 [1 − (10)5 (10)−6 ] 9e20.00
Then we get the solution in the form

(10)6
u(t) = .
1 + 9e20.00−t/100
This gives the population u as a function of time for t > 2000.

(a) The population in the year 2020 is obtained by putting thus

(10)6
u(2020) = ≈ 135, 336.
1 + 9e−0.2

(b) Solve it for doubling population, we obtain

(10)6
2(10)5 =
1 + 9e20.00−t/100 ,
from which
e20.00−t/100 = (4/9), then et/100 = (9/4)e20 ,
and hence t ≈ 2061.

(c) For large population, we find

(10)6
limt→∞ u(t) = limt→∞ = (10)6 = 1, 00, 0000.
1 + 9e20.00−t/100
Example 3.3.2. A population grows accordingly to be equation
  
du u
= 1 − exp 1 − r0 1 − ,
dt K

where r0 and K are positive constants. Determine the equilibrium state and its stability for
this population.

11
Proof. The equilibrium solution is obtained by du = 0.
dt
We have u = K 1− r10 is an equilibrium state. For stability, we use the perturbation method.

Let u = K 1 − r10 + u where u is very small, then

r0 u (r0 u)2
   
du r0 u r0 u
= 1 − exp =1− 1+ + + ...... ∼=− .
dt K K 2! K

This shows that the population decreases exponentially. Hence, the equilibrium state is
stable.

Example 3.3.3. Explain the law of infectious diseases du = Ku(R − u), and solve it for
dT
K = 0.007, R = 1000 and u(0) = 50.

Solution. By using variable separable

du
= KdT.
u(R − u)
 
1 1 1
We have R u + R − u du = KdT .
Integrate from 0 to T , we obtain

u(R − u(0))
ln = RKT.
u(0)(R − u)

We have
u(R − u(0)) = u(0)(R − u)eRKT .
Since K = 0.007, R = 1000 and u(0) = 50. Then we obtain, 19u = (1000 − u)e7T .

3.4 Von Bertalanffy model

Assume that growth is proportional to the surface area since nutrient enters through the
surface and that death is proportional to the population size. The model is also known as
the surface rule model. It has been successfully applied to describe human tumor growth.

du 2
= ru 3 − bu, (3.4.1)
dt
where r is the growth parameter and b is the growth deceleration parameter. The solution of
(3.4.1) is given by
 3
r 1 r
b
u(t) = + u03 − e− 3 (t−t0 ) .
b b
We can show the behavior of the Von Bertalanffy model in the following Figure 3.4.1.
It may be observed in Figure 3.4.1 that by taking the initial population between 0 and
( ) the population increases over time and gets stable at u = ( r )3 . Hence, u = ( r )3 is
r 3
b b b
the saturation level of the model. If we take initial volume u0 > ( r )3 the volume reduces
b
exponentially over time and is stable at u = ( r )3 . Also, if we take u0 = ( r )3 the volume
b b
remains constant.

12
 800

700

¬ u0>K3
600

u(t)
500
u0=K3

400

300 ¬ 0<u0<K3

200
0 200 400 600 800 1000
t

Figure 3.4.1: Graph of Von Bertalanffy model (3.4.1) with t0 = 0, r = 1.6 × 10−7 , b = 0.2 × 10−7 , and
K = r.
b

3.5 Richards’ model

This model was initially developed for modeling population growth. It can be used to
discuss other single-species growth also. The mathematical form of this model is given by
  u α 
du
= ru 1 − . (3.5.1)
dt K

Here α is the positive exponent. We can analytically derive the solution of the equa-
α−1
tion (3.5.1) by letting u K α = U , and by using the partial fraction  1
u  =
u 1 − ( )α
K
1+ uα−1 .
u u
K α (1 − ( )α )
K
Using these in the equation (3.5.1) and integrating both sides we obtain the solution

K
u(t) = .
 K α 1
−rα(t−t0 ) α
1 + {( ) − 1}e
u0
The graphical presentation of this function is shown in Figure 3.5.1.

600
¬ u0>K
® K=500 u0=K

400

¬ u0=300
u(t)

300

200
¬ u0=50
100

0
0 0.5 1 1.5 2 2.5 3
t

Figure 3.5.1: Richards’ model (3.5.1) for various initial population with r = 5 × 10−6 , b = 0.1 × 10−7 ,
α = 1.001, and K = r .
b

It can be observed from Figure 3.5.1 that this model behaves just like the logistic model.
However, in this model, the system can be made stable more quickly by increasing the value
of α suitably.

13
3.6 Generalised logistic model
The generalised form of the logistic equation can be written as
   β γ
du α u
= ru 1 − , (3.6.1)
dt K
where α, β, and γ are non-negative exponents and r is the growth rate parameter. From this
generalized form, it is possible to derive all the models described above by choosing the values
of α, β, and γ suitably.
For α = 1 and γ = 0, one gets the exponential growth equation
du
= ru.
dt
The graph of its solution behaves exactly like the exponential growth curve that increases in
an unbounded manner, as t → ∞ for r > 0.
For α = 1, β = 1 and γ = 1, equation (3.6.1) reduces to the logistic equation
  u 
du
= ru 1 − .
dt K
The logistic curve can also be generated by setting all the exponents equal to 1. The graph
is stable at u = K
2 for different initial volumes.
If we take α = 2 1
3 , β = 3 , and γ = 1 we get from (3.6.1),
  1 
du 2 u 3
= ru 1 −
3 ,
dt K
which is the equation of the Von Bertalanffy model. We can observe that Von Bertalanffy’s
curve gets stable at a slower rate than the logistic curve.
Lastly, we can consider α = 1, and γ = 1. Then we get Richard’s equation
  β 
du u
= ru 1 − .
dt K
The graph of its solution becomes stable at a faster speed than the normal logistic equation
with the smallest increment of β.
One can now analyze the behavior of the generalized logistic model by considering different
values of the parameters and observing the nature of the graphs of the solution.
We notice that the model behaves differently with different parameter values as presented
in Figure 3.6.1. By increasing the value of α slightly one can observe that the size of the
population reaches the carrying capacity K quickly. Similarly, an increase in the value of β
causes the size to increase faster. But for γ the observation is to the contrary. The higher
the exponent, the slower the growth rate becomes.
2
Moreover, differentiating (3.6.1) with respect to t and considering d u = 0, one can obtain
dt2
the inflection population
βγ −1/β
 
uinf = K 1 + .
α

14
 110

a=2
100

90 ¬ b=0.9

u(t)
80 ¬ b=0.15

¬ g=1.5
70

60 ¬ g=2
a=0.1

50
0 2 4 6 8 10
t

Figure 3.6.1: Generalised logistic model (3.6.1) with r = 3, K = 100, u0 = 50, and t = 0 to t = 10.

It may further be observed that for all three exponents α, β, and γ, uinf → K as α, β → ∞
and γ → 0; that is,
lim uinf = lim uinf = lim uinf = K.
α→∞ β→∞ γ→0

Also,
lim uinf = lim uinf = 0, and lim uinf = Ke−γ/α .
α→0 γ→∞ β→0

The observation presented above can be schematically shown by using the following dia-
gram

Generalised
Logistic Equation

α = 1, γ = 0 α = 1, α = 2/3, α = 1,
Exponential β = 1, γ = 1 β = 1/3, γ = 1 γ = 1
Growth Logistic Growth Von Bertalanffy Richard

Figure 3.6.2: Schematic diagram of generalised logistic model (3.6.1).

3.7 Smith’s Model

The Malthusian model
dx
= bx − dx = rx,
dt
where b, d, and r are positive constants. In general, b is a monotonic decreasing function of x
and d is a monotonic increasing function of x so that r is a monotonic decreasing function of
x. Hence, we obtain
dx
= r(x)x.
dt
For the logistic growth model, we have
dx
= x(r − cx)
dt
1 dx
⇒ = r − cx,
x dt

15
where x1 dx should be a linear function of x.
dt
In 1963, Smith show that x1 dx should be a concave function of x. He argued that the specific
dt
growth rate function r(x) must be proportional to the rate of food supply which is used by
the population, so that  
T −F F
r(x) = r =r 1− ,
T T
where F is the rate at which a population size x uses food, and T is the corresponding rate
when the population has a limiting or saturation size. Now F depends on x and dx . Hence,
dt
the simplest assumption is
dx
F = c1 x + c2 .
dt
If the saturation population is K then

T = c1 K.

Thus, we obtain

c1 x + c2 dx
   
dx dt x c rx(K − x) c1
= rx 1 − = rx 1 − − = , where c = ,
dt c1 K K K dx K + rcx c2
dt
which is the Smith’s model for a single species population growth.
By using variable separable, we obtain

(K + rcx)dx = xr(K − x)dt.

We have  
1 1 + rc
+ dx = r dt.
x K −x
Integrate in both sides, we obtain
x
= ert .
A(K − x)1+rc

If t → ∞ then
x
→ ∞.
A(K − x)1+rc
We obtain A(K − x)1+rc → 0 then x → K.
For equilibrium points, we have dx = 0 then x = 0, or x = K are equilibrium points.
dt
Using the perturbation method, one can see x grows exponentially for any small initial value
for x = 0. Hence, x = 0 is unstable. Similarly, x = K is stable.

3.8 Assumption regarding the model u′ = uf (u) of function f (u)

The density-dependent model is u′ = uf (u). In the Malthusian model, it is assumed that
f (u) = r (constant). But this makes the model unrealistic. Because population size increases
to infinity as t → ∞.

16
In practice, it is observed that the growth rate is controlled by population size, that is, the
df
more the population increases, the more the growth decreases. Indeed, ≤ 0.
du
But the function f (u) depends on u, and cannot be the final form for the growth function.
Because growth function not only depends on population size, it also depends on the environ-
mental situation. Since the environment is not constant, it varies concerning time t. Thus in
general, the form of the growth function is modified as

f (u) → f (u, t, r) where r is a parameter.

Derivation of logistic model

From density dependent growth, we have f (u) = the growth function.
df
Then < 0, and the model is u′ = uf (u).
du
Now, the simple linear function

f (u) = r − bu where r, b > 0,

df
satisfies the above condition, since = −b < 0.
du
Therefore, considering f (u) = r − bu as the growth function depending a population density,
the model becomes
u
u′ = u(r − bu) = ru(1 − ),
K
where
r = intrinsic growth rate per individual.
= birth rate per individual-death rate per individual.
b = intra-specific competition factor.
K = r = carrying capacity of the environment.
b
Verhulst, a Dutch mathematician, proposed this equation, which is known as the logistic
model.
The solution of the logistic model with u(0) = u0 is
u0
u(t) = u0 .
+ 1 − uK0 e−rt


K

We obtain that
umrx = lim u(t) = K.
t→∞

Observe that
u(t) ↑ K as t → ∞ if u0 < K since u′ (t) > 0, and
u(t) ↓ K as t → ∞ if u0 > K since u′ (t) < 0,
which implies that the population remains limited. So, the logistic equation is more realistic
than the Malthusian equation. There are so many examples that the population dynamics
follow a logistic equation. But it cannot express any disaster’s effect on the population. In
nature, populations have to face so many disasters, and these disasters make the population
size, periodic fluctuate whereas, in the logistic model, the population attains a constant value
as t → ∞.
u

If we consider f (u) = r(u − M ) 1 − where r, M, K > 0, and M < K. (3.8.1)
K

17
Solution of the equation
Consider u(t) ̸= M , or K is the solution of
du r
= (u − M )(K − u).
dt K
We obtain from (3.8.1) by variable separable
du r
= dt,
(u − M )(K − u) K
then  
K 1 1
+ du = r dt.
K −M u−M K −u
Integrating from 0 to t, we get
(u − M )(K − u0 ) K −M
ln = rbt where u(0) = u0 > 0, and b = > 0.
(K − u)(u0 − M ) K

u = K − u0 e−rbt = ce−rbt where c = K − u0 . Then, 2u − (M + K) = 1 − ce−rbt .

Also, uK−−M u0 − M u0 − M K −M 1 + ce−rbt
 
−rbt
Hence, the solution of the model is u(t) = 21 Kb 1 − ce−rbt + (M + K) .
1 + ce
Also,
umrx = lim u(t) = K.
t→∞

For constant solution, we have du = 0.

dt
Then u = M , or K since K r > 0. Therefore, the equilibrium solutions are u(t) = M and
u(t) = K for all t.
Observe that
u(t) ↑ K as t → ∞ since u′ (t) > 0 for M < u < K,
u(t) ↓ K as t → ∞ since u′ (t) < 0 for u > K, and
u(t) ↓ as t → ∞ since u′ (t) > 0 for u < K.
For the third case, there is no limiting value. Since for some value of t
K − u0 −rt
1 + ce−rbt = 0 ⇒ 1 − e = 0.
u0 − M
− M . Hence, t = 1 ln K − u0 .
So, e−rt = uK0 − u0 r u0 − M
We have from (3.8.1), we obtain

u′ (t) > 0 for M < u < K,

u′ (t) < 0 for u > K, and
u′ (t) < 0 for u < M. Also,
du′ r  r 
= (u − M )(−1) + (K − u) = (K + M ) − 2u .
du K K
′ ′
Now, u′ ↑ for u < 21 (M + K) since du > 0 and u′ ↓ for u > 12 (M + K) since du < 0 and
du du
u(t) = 12 (M + K), these have a critical value.

18
 2 ′
Since d u2 = −2 K
r < 0 at u = 1 (M + K).
2
du
Thus, the population size u(t) = 21 (M + K), the net growth rate attains its maximum value.

100

80

60

du
dt
40

20

0
M 1/2(M+K) K

−20
−200 0 200 400 600 800
u(t)

Figure 3.8.1: du versus u.

dt

This model is most interesting because the extinction of a small population is an unsuitable
environment. Another reason is this growth becomes negative. From the above calculation,
it is shown that at the population size u(t) = 21 (M + K), the net growth rate attains its
maximum value.
The equilibrium solutions of the models are u = 0, M , and K.

20

15

10

5
f (u)

0
O M K
−5

−10

−15

−20
−2 0 2 4 6 8 10
u

Figure 3.8.2: f (u) versus u.

It is clear that f (u) > 0 for M < u < K and u is increasing. The reverse is true for u < M
and u > K. Hence the equilibrium solution u = 0 and K are stable, and the equilibrium
solution u = M is unstable.

Example 3.8.1. A population is governed by the logistic law

u
(i)u′ = r(1 − )u, (r > 0, K > 0),
K
with the initial population u(0) = u0 being less than K. Suppose that an epidemic strikes as
soon as u reaches a certain value u1 < K and at this stage, the dynamics are governed by the
logistic law
u

(ii)u′ = r1 1 − u, (0 < r1 < r, 0 < K1 < u),
K1
shows that the population then starts decreasing. Suppose that when the population falls to
u2 with K1 < u2 < u1 the epidemic ceases, and the population begins to grow the following
equation (i) until the incidence of a fresh epidemic that the population undergoes periodic

19
fluctuations between u2 and u1 . Calculate the period of these fluctuations. Comment on the
appropriateness of the model.
Solution. The graph of the population is drawn below

100

80
K

60 P1

u(t)
40

P2
20 K1
P0
0
t1 t2 t3
−20
−2 −1 0 1 2 3 4 5 6 7
t

Figure 3.8.3: Described population over time.

The solution of (i) can be written as

K
u1 (t) = ,
K − u0 −rt
1+ e
u0
and solution of (ii) is
K1
u2 (t) = .
K1 − u0 −r1 t
1+ e
u0
For curve AB????, we have u(t1 ) = u1 , u(t2 ) = u2 then
K1
u2 =
K1 − u1 −r1 (t2 −t1 )
1+ e
u1
K1
= ,
K1 − u1 −r1 T1
1+ e
u1
where T = t1 − t2 therefore, we obtain
K1
u1
e−r1 T1 = −1 .
u2 K1 − u1
Hence,
1 u2 (K1 − u1 )
T1 = ln .
r1 u1 (K1 − u2 )
Similarly,
1 u1 (K − u2 )
T2 = ln .
r u2 (K − u1 )
So, T = T1 + T2 where T1 , and T2 are as in above.
Since u(t) as soon as attain the value u1 then it decreases that is u1 is the maximum value.
Similarly, u2 is the minimum value. Therefore, the population undergoes periodic fluctuation
between u2 and u1 .

20
Remark 3.8.1. In nature, the environmental situations are not constant. A population has
to face so many disasters. So, the population model described by one logistic model can
not express this disaster’s effect. But the population is described by two different logistic
equations that express a population periodically. This periodic fluctuation can the effect of
disasters more appropriately. So, this model is more realistic. Then the model is described
by one logistic equation.

3.9 Non-autonomous Logistic Models

Consider the logistic growth model
u′ = αu − f (t)u2 , u(0) = u0 , (3.9.1)
where a constant specific growth rate α and f is a function of t.
The equation (3.9.1) is Bernoulli’s equation. Putting v(t) = 1 , then
u(t)
v ′ (t) + αv(t) = f (t), (3.9.2)
which is a linear differential equation and its integrating factor is eαt .
Multiply (3.9.2) by integrating factor, then we obtain
Z t
1 αt 1
e = + eαs f (s)ds.
u u0 0
Thus, the solution of equation (3.9.1) is
u0 eαt
u(t) = Z t .
αs
1 + u0 e f (s)ds
0

3.10 Logistic equation in a time-varying environment

Although the logistic curve (3.2.2) has been applied, with remarkable success, to fit growth
curves of various populations, many of the hypotheses that are employed in formulating the
Verhulst model (3.2.1) seem somewhat unrealistic. For instance, the growth parameters r
and b in (3.2.1) are assumed to be constant for all time. But the ability of a population to
grow depends ultimately on the nature of the surrounding environment. Any change in the
environment may have a considerable effect on the growth process. So, it is more realistic
to consider time-dependent growth parameters leading to non-autonomous equations. An
ecological interesting case occurs when the parameters are periodic functions. It has been
observed that many populations fluctuate periodically, and this is often attributed to periodic
fluctuations in the growth parameters.

Then the Verhulst logistic model in a time-varying environment can be reduced to

u′ (t) = (r(t) − b(t)u(t))u(t), u(0) = u0 , (3.10.1)
where r(t) > 0 the natural growth rate and b(t) ≥ 0 the inhibition coefficient and assume that
they are piecewise continuous functions on ℜ+ = [0, ∞). The equation (3.9.1) is Bernoulli’s
equation. Putting v(t) = 1 , then
u(t)
v ′ (t) + r(t)v(t) = b(t),

21
 Rt
r(u)du
which is linear differential equation and its integrating factor is e 0 . Multiplying by
integrating factor Z t Z t
 r(u)du  ′ r(u)du
v(t)e 0 = b(t)e 0 . (3.10.2)

Integrating on both sides of the equation (3.10.2). Hence, we obtain the unique solution of
the model Z t
r(s)ds
u0 e 0
u(t) = Z s .
Z t r(u)du
1 + u0 e 0 b(s)ds
0

3.10.1 Behavior of the solution

We have seen that the solution u(t) stay positive for all t > 0. Since r(t) and b(t) are
Z t
non-negative piecewise continuous function on [0, ∞), the functions t → r(s)ds and t →
Z s 0
Z t r(u)du
e 0 b(s)ds are non-decreasing on [0, ∞).
0
Z s
Z ∞ Z ∞ r(u)du
Hence, r(s)ds and e 0 b(s)ds exists on [0, ∞). It follows that
0 0
Z ∞
r(s)ds
Proposition 3.10.1. If e 0 < ∞, then the solution of the equation (3.10.1) con-
verges, that is
lim u(t) = u∞ (u0 ) < ∞.
t→∞
Z ∞
Proposition 3.10.2. If r(s)ds < ∞ and the carrying capacity K(t) = K0 > 0 then
0

(i) u0 > K0 ⇒ K0 < u∞ (u0 ) < ZK∞

0 ,
− r(s)ds
1−e 0

(ii) u0 = K0 ⇒ u∞ (u0 ) = K0 , and

(iii) u0 < K0 ⇒ u∞ (u0 ) < K0 .

22
Solution. We have the solution for the non-autonomous logistic model
Z t
r(s)ds
u0 e 0
u(t) = Z s
Z t r(u)du
1 + u0 e 0 b(s)ds
0
Z t
r(s)ds
u0 e 0
= Z 1 . (3.10.3)
Z t r(u)du
1 + u0 K0−1 e 0 r(u)du
0
Comparing (3.10.3) with the Verhulst model, then environment carrying capacity K(t) =
r(t)
> 0 and K(t) = K0 .
b(t) Z u
Z u Z t r(s)ds Z t
Let v(u) = r(s)ds, then v ′ (u) = r(u), and e 0 r(u)du = ev(u) .v ′ (u)du =
Z 0t 0 0

 v(u) t r(u)du
e 0
= e 0 − 1.
From (3.10.3), we obtain
u0
u(t) = Z t Z t .
− r(s)ds  − r(s)ds 
e 0 + u0 K0−1 1 − e 0

Hence, u∞ (u0 ) = u0 .
Z ∞
 − r(s)ds
u0 K0−1 + 1 − u0 K0−1 e

0

If u0 > K0 then K0 = u0−1 < u∞ (u0 ) <  u0 Z

∞ = ZK∞
0 .
u0 K0 − r(s)ds  − r(s)ds
u0 K0−1 1 − e 0 1−e 0



If u0 = K0 , then u∞ (u0 ) = u0 = K .
0
u0 K0−1
If u0 < K0 , then u∞ (u0 ) < u0−1 = K0 .
u0 K0
Z ∞
Remark 3.10.1. If r(s)ds is sufficiently small, then the bound M = ZK∞
0
0 − r(s)ds
1−e 0
can be arbitrarily large. If the inhibition coefficient b(t) > 0 on ℜ+ and r(t) = ◦(b(t)) as
t → ∞ then limt→∞ K(t) = K(∞) = 0 where K(t) = r(t)/b(t) is the environmental carrying
capacity. The situation is called a deteriorating environment.
Z t
Proposition 3.10.3. If r(s)ds = ∞ (that is, deteriorating environment), then each solu-
0
tion u(t) satisfies
lim u(t) = 0.
t→∞

23
 r(t) r(t)
Solution. Here, limt→∞ = 0 so ∃t1 > 0 such that for all t > t1 , , that is, b(t) > r(t).
b(t) b(t)
For all t > t1 ,
Z t R Z t1 R s Z t Rs
s
r(u)du r(u)du r(u)du
e0 b(s)ds = e 0 b(s)ds + e 0b(s)ds,
0 0 t1
Z t Rs Z t Rs Z t  
r(u)du r(u)du d R s r(u)du Rt R t1
and e 0 b(s)ds > e 0 r(s)ds = e 0 ds = e 0 r(u)du − e 0 r(s)ds .
t1 t1 t1 dt
Rt
Z t
r(u)du
The fact that limt→∞ e 0 = ∞. Since r(u)du = ∞.
0
So we have, Z ∞ Rs
r(u)du
e 0 b(s)ds = ∞.
0
We have (3.10.3) is ∞/∞ form, then by using L’Hospital rule, we obtain
Rt
r(s)ds
u0 e 0 r(t)
lim u(t) = lim Rs
r(u)du
u0 e 0 t→∞ b(t) t→∞

r(t)
= lim = 0.
t→∞ b(t)

Z t
Remark 3.10.2. A relatively large growth rate, that is, r(s)ds = ∞ coupled with a
0
deteriorating environment causes a population to track back its environment to extinction.

Disadvantage of the Model: The disadvantage of this model is that it does not take into
account the age distribution of the population. Birth and death rates are independent of the
age of the individuals in this model. However, it is obvious that, in many situations, age
distribution does influence population size and behavior through the birth and death rates.

3.11 Population Models with Age Distribution

Let x(t, r) be the population density at time t in the age range r to r + ds. Let v(r) and µ(r)
be the birth and death rates, respectively which are functions of age a. A small increment
of time dt the number of the population of age r that dies is µ(r)x(t, r)dt. The birth rate
is x(t, 0); there can be no births of age r > 0. The conservation law for the population says
that population Models with Age Distribution

De x(t, r) = −µ(r)x(t, r)dt, (3.11.1)

where De x is the directional derivative of x and is defined by the linear partial differential
equations

∂x ∂x
+ , De x(t, r) = (3.11.2)
∂t ∂r
which holds for t > 0, and r > 0. The equation (3.11.1) is a first-order partial differential
equation that requires condition on x(t, r) in t and r. The initial condition

x(0, r) = ϕ(r),

24
the population at time t = 0 has a given age distribution ϕ(r).
The boundary condition on r comes from the birth rate and is
Z ∞
x(t, 0) = ν(r)x(t, r)dr, (3.11.3)
0
where birth rate ν(r) of course will tend to zero for large r, and birth rate ν(r) only appears in
the integral equation (3.11.3) not in the differential equation (3.11.1). The equation (3.11.1)
is called the Von-Foerster equation. The birth and death rates ν(r) and µ(r) affect the growth
of the population after a long time.
The solution of (3.11.1) is
Z r
− µ(s)ds
x(t, r) = x(0, r0 )e r0 , r > t,

where x(0, r0 ) = x(0, r − t) = ϕ(r − t), then

Z r
− µ(s)ds
x(t, r) = ϕ(r − t)e r−t , r > t.

For r < t, Z r
− µ(s)ds
x(t, r) = x(t0 , 0)e 0 ,
where x(t0 , 0) = x(t − r, 0) then
Z r
− µ(s)ds
x(t, r) = x(t − r, 0)e 0 , r < t.

Thus,
Z r Z r
Z t − µ(s)ds Z ∞ − µ(s)ds
x(t, 0) = ν(r)x(t − r, 0)e 0 dr + ν(r)ϕ(r − t)e r−t dr. (3.11.4)
0 t

Sharpe-Lotka proposed the linear, continuous, and age-dependent model in 1911; it is difficult
to solve but it can be done by iteration. We are interested in the long time behavior of the
population and in particular whether or not it will increase or decline.
Let the solution of (3.11.2) in the form

x(t, r) = eγt r(r). (3.11.5)

We obtain from (3.11.2), and (3.11.5) that

dr
= −[µ(r) + γ]r,
dr
and so,

Z r 
r(r) = r(0) exp − γr − µ(s)ds .
0

25
The boundary condition gives
Z ∞ Z r
γt γt
 
e r(0) = ν(r)e r(0) exp − γr − µ(s)ds dr,
0 0

and hence, Z ∞ 
Z r 
1= ν(r) exp − γr − µ(s)ds dr = ϕ(γ), is a monotonic decreasing function
0 0
of γ which decreases from ∞ to 0 as r increases from −∞ to ∞, and therefore there can one
and only one real value of γ for which ϕ(γ) = 1.
Thus, the critical threshold S for population growth is
Z ∞ Z r
 
S = ϕ(0) = ν(r) exp − µ(s)ds dr,
0 0

where S > 1 implies growth, and S < 1 implies decay.

3.12 Some Applications of Logistic Equation

3.12.1 Logistic Equation in Dynamics of Spruce Budworms
The dynamics of spruce budworm for a single species of prey can be described by using the
logistic equation
du  u βu2
= ru 1 − − 2 , (3.12.1)
dt K α + u2
where r is the intrinsic growth rate of the spruce budworm. K is the carrying capacity, which
depends on the availability of the foliage of the balsam firs. The rate of predation

βu2
h(u) = (3.12.2)
α2 + u2
is from avian predators, which associate reward with prey, a learned behavior. They focus
on their best prey sources, allowing low-density prey to escape notice. The functional form
of this type of predation follows Holling’s Type III S-shaped response as defined by h(u). β
represents the saturation level of the predator, e.g., a constant population can eat so much
prey; this saturation level works like an upper limit on the predation. α determines the
densities of spruce budworms that cause avian predation.
Now if taking the limiting value of the predation function,

βu2 2βu
lim h(u) = lim = lim = β.
u→∞ u→∞ α2 + u2 u→∞ 2u

26
 Figure 3.12.1: Rate of Predation.

It is clear from Figure 3.12.1 that the function increases monotonically until it reaches its
asymptote at the value β. This indicates that no matter how much the budworm population
increases the birds can eat a limited quantity. Using basic calculus the first and second deriva-
tive is calculated to find the point of inflection. This point determines when the budworm
population increases rapidly.
Differentiating equation (3.12.2) with respect to u

(α2 + u2 ) × 2βu − βu2 × (2u)

h′ (u) =
(α2 + u2 )2
2α2 βu
= 2 . (3.12.3)
(α + u2 )2

It is obvious that h′ (u) > 0 since all the constants are positive. Now, again differentiating
with respect to u

(α2 + u2 )2 × 2α2 β − 2α2 βu × 2(α2 + u2 ) × 2u

h′′ (u) =
(α2 + u2 )4
2α2 β(α2 − 3u2 )
= . (3.12.4)
(α2 + u2 )3

Substituting equation (3.12.4) to zero, yields

2α2 β(α2 − 3u2 )

=0
(α2 + u2 )3
α
⇒ u = ±√ . (3.12.5)
3
The population can not be negative, so the negative is omitted. Thus, predation dramat-
ically climbs upwards to √α from below. After crossing the inflection point, the function
3
gradually approaches its asymptote at h = β.

3.12.2 Chemostat Modeling

Numerous significant sectors, including the pharmaceutical and food industries, use the
growth of bacteria and other microorganisms. The process is maintained in a bioreactor

27
which is known as the chemostat. The reactor of the chemostat is continuously stirred, and
the environment is adjusted to culture the desired organism. The bacteria and microorgan-
isms produced in the chamber are then collected at an optimal outflow rate to have maximum
production.
The chemostat modeling can be described by the following model using the logistic equa-
tion
du u
= ru(1 − ) − bu, (3.12.6)
dt K
where r is the growth function of the bacteria and b is the bacterial outflow rate. The con-
centration of the nutrients for the bacteria is assumed to be constant in the whole container.
The model has two equilibrium points at u1 = 0 and u2 = (1 − rb )K. For the r > b
condition, the equilibrium point u1 is unstable, and u2 is stable. On the other hand, for
r < b, u1 is the stable, and u2 is the unstable equilibrium point.
The solution for r > b and for initial population u0 < (1 − rb )K the solution of the
equation (3.12.6) is u(t) = r r−b , where C is a constant. The solution converges
1
+ exp−(r−b)t
K C
to (1 − rb )K as t → ∞ and hence u = (1 − rb )K is the stable equilibrium.
The model is also used to derive that, the optimal outflow rate of a chemostat is s = 2r ,
i.e, the bacterial outflow rate is half of the bacterial growth.

3.13 Problems
1. Suppose that a certain population obeys the Verhulst model with intrinsic growth rate r
and carrying capacity K. Find the complete solution of the model. Discuss the behavior
of the population as t → ∞. Obtain a formula for the time t when the population size
u(t)βK, given that the initial population u(0) = αK with 0 < α < β < 1.

β(1 − α)
Ans. t = 1r ln .
α(β − 1)
2. A population in a varying environment is governed by the logistic equation where the
natural growth rate r(t) is a piecewise continuous non-negative function such that
Z ∞
r(s)ds < ∞ and the carrying capacity K a positive constant. Show that the
0
population size u(t) is bounded and obtain the bounds.
3. Discuss the non-autonomous logistic model U ′ (t) = [r(t) − b(t)U (t)]U (t). Find its
solution and then indicate the behavior of the solution.
4. Describe the Verhulst or logistic model for the dynamics of a single species population.
Obtain a formula for the population size. Comment on the plausibility of the model.
How is this model an improvement over the Malthusian model? Determine the steady
states and discuss the stability of the model.
u )u for the growth of a single species population
5. Consider the logistic model u′ = r(1 − K
where a and K are positive constants. Explain the biological concepts behind the
model. Show that according to this model, the population tends to a constant saturation
level/carrying capacity irrespective of the initial states. Discuss the stability of the
equilibrium states using graphical and perturbation analysis.

28
 6. If the carrying capacity of the logistic model is K and 0 < u0 < K 2 . Determine the
double time of the population u0 in this model. Hence, if there are 100 rabbits in the
colony which multiply according to the rate constant K = 3 × 10−6 /sec, how long does
it take for the rabbit population to double to 200?
  
7. A population grows according to the equation dU = 1 − exp − r 1 − K U where r
dt
and K are positive constants

(i) Determine the equilibrium population size.

(ii) Decide whether the equilibrium is stable, unstable, or neutral.

8. The growth of a single species population is modeled by the equation u′ = ug(u). Make
plausible assumptions regarding the function g(u). Discuss the case g(u) = r Ku −
0


1 1− K u
where 0 < K < K. Investigate the stability of the equilibrium states of this
0
later model. Find also the complete solution analytically.

9. Consider the autonomous differential equation

X ′ = k(α − X)(β − X)

where k > 0 and β > α > 0.

(i) From the phase portrait of the differential equation, can you predict the behavior
of X as t → ∞?
(ii) Consider the case α = β. What is the behavior of X as t → ∞ if X(0) < α? From
the phase portrait of the differential equation, can you predict the behavior of X
as t → ∞ if X(0) > α?
(iii) Verify the solution of the differential equation when k = 1 and α = β is X(t) =
α− t+ 1 . Find solutions satisfying X(0) = α and X(0) = 2α.
c 2

Ans. (i) X = β − 1 [e(α−β)(kt+c) − 1].

α−β
10. Comment on the density dependent model u′ = ug(u). Describing the dynamics of
a single species population. Make plausible assumption regarding the function g(u).
u
u and interpret the parameters r and
Derive the logistic growth model u′ = r 1 − K
K. Comment on the appropriateness of the model.

11. A single species population grows according to the law

u
u′ = r(u − M )(1 − )
K
where r, M, and K are positive constants with M < K.

(i) Find the complete solution analytically and discuss the limiting behavior of the
population.
(ii) Find the population size when the net population growth rate is maximum.
(iii) Investigate the stability of the equilibrium populations.

29
 (iv) Comment on the appropriateness of the model.

12. A population is governed by the logistic law

u

(i) u′ = r 1 − u, (r > 0, K > 0)
K
the initial population u(0) = u0 being less than K. Suppose that an epidemic strikes
as soon as u reaches a certain value u1 < K and that at this stage the dynamics are
governed by the logistic law
u

(ii) u′ = r1 1 − u, (0 < r1 < r, 0 < K1 < u1 ).
K
Show that the population then starts decreasing. Suppose that when the population
falls to u2 with K1 < u2 < u1 the epidemic ceases and the population begins to grow
again till the onset of a fresh epidemic. Show that the population undergoes periodic
fluctuations between u2 and u1 . Calculate the period of this fluctuation.

13. (a) Formulate a model to describe the dynamics of a single species population subject
to the conditions:

(i) If few members are present, the population becomes extinct, as it cannot propagate
itself successfully.
(ii) The population cannot become unbounded due to resource constraints.

(b) Locate the equilibrium states for the model (in part (a)) and discuss their stability.

14. The population of a city would satisfy the logistic law

du 1 1
= u− u2 .
dt 25 (25)106

(a) Modify this equation to take into account the fact that 6,000 people move from the
city, and 4,000 people per year are murdered.
(b) Assume that the population of a city was 8, 000,000 in 1970. Find the population
for all future time. What happens as t → ∞?

15. The population of a city would satisfy the logistic law

du u u2
= − 8.
dt 100 10
If the population of this city was 1,00,000 in 1980, what will be the population in 2020?

16. Suppose that the growth of a single-species population in a time-varying environment

is modeled by
u′ (t) = (r(t) − b(t)u(t))u(t),
where r(t) and b(t) are positive continuous functions on the half-line [0, ∞).

(a) Show that if r(t) is integrable on [0, ∞), then the population converges to a limiting
size N∞ which depends on the initial population.

30
 (b) Discuss a relatively large growth rate coupled with a deteriorating environment
leads the population to extinction.

17. Consider du = ru2 − bu, (r > 0, b > 0).

dt
(a) How does the growth rate depend on the population?
(b) Sketch the solution in the phase plane.
(c) Obtain the exact solution.
(d) Find the behavior of the model.

18. Consider du = ru + bu2 , (r > 0, b > 0).

dt
(a) How does the growth rate depend on the population?
(b) What would you expect to happen to the population?
(c) Using the phase plane, show that the population tends towards infinity.
(d) Considering the exact solution, show that the population reaches infinity in a finite
time.

19. The model of population growth is described by

du
= u(r − b ln u),
dt
where r and b are positive constants.

(a) Find the critical points and discuss their stability.

r r
(b) Show that if 0 < u(0) < e b , then u(t) tends to e b when t → ∞.
(c) Solve the differential equation.
(d) Plot the solution when r = 1 and b = 4.

20. Suppose that the logistic equation is modified to

du
= −u(r − bu),
dt
(a) Find the critical points and discuss their stability.
(b) Show that if 0 ≤ u(0) < r , then u(t) tends to zero when t → ∞.
b
r
(c) Show that if u(0) > , then u(t) becomes infinite at some positive time.
b
21. Suppose that the logistic equation is further modified to
du
= −u(r − bu)(r − cu),
dt
where r, b, and c are positive constants with b > c.

(a) Find the critical points and discuss their stability.

(b) Show that if 0 ≤ u(t0 ) < r , then u(t) tends to zero when t → ∞.
b

31
 (c) Show that if u(t0 ) > r , then u(t) tends to cb when t → ∞.
b
22. Suppose that the logistic equation is further modified to
du
= −u(r − bu)2 ,
dt
where r, b, and c are positive constants.

(a) Find the critical points and discuss their stability.

(b) Solve the differential equation.
(c) Find the time t for which u(t) = 0.9r , if u(t0 ) = 0.1r .
b b
23. Suppose that a population grows according to the logistic law
du
= 10(1 − 0.1u)u
dt
(a) What is the carrying capacity?
(b) What is the growth rate?
(c) For what value of u is the population growing most rapidly?

24. Suppose that the growth of the population is given by the logistic equation
1000
u= .
1 + 999e−0.9t
(a) What is the population at time t = 0?
(b) What is the carrying capacity?
(c) What is the growth rate?
(d) When does the population each 75% of the carrying capacity?
(e) Find the initial value problem whose solution is u(t).

25. Describe Smith’s model for a single species population growth. Describe the behavior of
the population as t → ∞. Find the points of equilibrium and investigate their stability.

26. Assuming the population growth model with migration.

27. Explain the logistic model in a time-varying environment.

28. Bangladesh’s population was estimated to be 800 million in 1980 and 1200 million in
2000. Estimate the population in the year 2020 by using Malthusian (or logistic) model.

29. Assuming that the population is growing at a rate proportional to the present number.
It has an initial population of 1,00,000 that increase by 15% in 10 years. What will be
the population in 30 years?

30. Suppose that a population grows according to the logistic law

du
= 10(1 − 0.1u)u.
dt

32
 (a) What is the carrying capacity?
(b) What is the growth rate?
(c) For what value does the population grow most rapidly?

31. (a) Suppose that a population has a growth rate of r per member per unit of time with
log 2
r < 0. Show that the time required for the population to double its initial size is r
log 2
and the time required for the population to decrease to half its initial size is − r .
(b) Show that for a population that satisfies the logistic model the maximum rate of
growth of population size is rK K
4 , attained when population size is 2 .
32. (a) Suppose a population satisfies a differential equation having the form of the logistic
equation but with
an intrinsic growth rate that depends on t:
Kx0 R
x′ = r(t)x 1 − K
x
, x(0) = x0 . Show that the solution is x(t) = t r(s)ds .
x0 +(K−x0 )e 0

(b) Use a logistic model with an assumed carrying capacity of 100 × 109 an observed
population of 5 × 109 in 1986, and an observed rate of growth of 2 percent per year
when the population size is 5 × 109 to predict the population of the earth in the year
2014.

33