Time Delays in Epidemic Models: Modeling and Numerical Considerations

Chapter 13
TIME DELAYS IN EPIDEMIC MODELS

Modeling and Numerical Considerations
J. Arino
Department of Mathematics,
University of Manitoba, Winnipeg, Manitoba, Canada
arionj@cc.umanitoba.ca
P. van den Driessche

Department of Mathematics and Statistics
University of Victoria BC, Canada
pvdd@math.uvic.ca
1. Introduction
Continuous time deterministic epidemic models are traditionally for-
mulated as systems of ordinary differential equations for the numbers
of individuals in various disease states, with the sojourn time in a state
being exponentially distributed. Time delays are introduced to model
constant sojourn times in a state, for example, the infective or immune
state. Models then become delay-differential and/or integral equations.
For a review of some epidemic models with delay see van den Driess-
che [228]. More generally, an arbitrarily distributed sojourn time in a
state, for example, the infective or immune state, is used by some authors
(see [69] and the references therein).
When introduced in an explicit way, time delays may change the qual-
itative behavior of a model; for example, an epidemic model with gener-
alized logistic dynamics can have periodic solutions when the time in the
infective stage is constant [112]. Qualitative differences that arise from
including time delay in an explicit way in models that include vertical
transmission are explored in [38, Chapter 4]. In population biology, a
maturation time delay is used to explain observed oscillations [192]. This
delay is included in an epidemic model by Cooke, van den Driessche and
539
O. Arino et al. (eds.), Delay Differential Equations and Applications, 539–578.
© 2006 Springer.
540 DELAY DIFFERENTIAL EQUATIONS
Zou [47]. A fixed time delay in the recruitment function for a disease
model is considered by Brauer [33].
Many disease transmission models with delay are difficult to ana-
lyze, with even the linear stability problem reducing to a hard quasi-
polynomial; see, e.g., [27, 33] and references therein. Thus a combination
of analytical and numerical techniques is often employed.
Here we proceed along the lines of van den Driessche [229], giving
details about delay in models of disease transmission by concentrating
on one particular model, namely work with two other coauthors, K. L.
Cooke and J. Velasco-Hernández [13]. In Section 2, we first motivate the
introduction of delay in epidemic models in which this delay results from
assumptions on the sojourn time in a certain epidemiological state, e.g.,
the infective state. We then (in Section 3) formulate a model including
vaccination of susceptible individuals in which the vaccine waning time is
arbitrarily distributed [13]. In Section 4, we specialize to two particular
waning functions, concentrating on the step function case that leads
to a system of delay integro-differential equations. Numerics on this
system are reported in Section 5 (with program listings in Appendix 1).
Some numerical warnings are given in Section 6, and we conclude with
an annotated listing of available delay differential equations numerical
packages (Appendix 2).
2. Origin of time delays in epidemic models

Various biological reasons lead to the introduction of time delays in
models of disease transmission. Here we concentrate on one of the pos-
sible origins: the fact that sojourn times in certain states can have a
general form, for example, can be approximately constant, as opposed
to having an exponential distribution. A formal derivation of this uses
probability theory and, in particular, survival analysis. A brief summary
of these notions is given in Section 2.1. Then, survival analysis is tied
to dynamical models of disease transmission in Section 2.2; see [225] for
a more general and more detailed presentation.
2.1 Sojourn times and survival functions

Consider a system that can be in either one of two states A and B.
Suppose that at time t = 0, the system is in state A. Suppose that
when an event E takes place, the system switches to state B, and call X
the (positive) random variable “time spent in state A before switching
to state B”. The cumulative distribution function (c.d.f.) F (t) then
characterizes the distribution of X, with F (t) = Pr{X ≤ t}, and so does
t
the probability distribution function f (t), where F (t) = 0 f (x)dx.
Time delays in epidemic models 541
Another characterization of the distribution of the random variable

X is the survival (or sojourn) function. The survival function of state
A is given by
S(t) = 1 − F (t) = Pr{X > t} (2.1)
This function gives a description of the sojourn time of a system in a
particular state. Note that S must be a nonincreasing function (since
S = 1 − F with F a c.d.f., thus a nondecreasing function), and S(0) = 1
(since X is a positive random variable).
The average sojourn time τ in state A is given by
∞
τ= tf (t)dt
0
Assuming that limt→∞ tS(t) = 0 (which is verified for most probability

distributions), ∞
τ= S(t)dt
0
Suppose that the random variable X has exponential distribution f (t) =
θe−θt for t ≥ 0, with θ > 0. Then the survival function for state A is of
the form S(t) = e−θt , for t ≥ 0, and the average sojourn time in state A
is ∞
1
τ= e−θt dt =
0 θ
If on the other hand, for some constant ω > 0,

1, 0 ≤ t ≤ ω
S(t) =
0, ω < t
which means that X has a Dirac delta distribution δω (t), then the aver-
age sojourn time is a constant, namely
ω
τ= dt = ω
0
These two distributions can be regarded as extremes.
2.2 Sojourn times in an SIS disease transmission

model
Consider a particular disease, and suppose that a population of in-
dividuals can be identified with respect to their epidemiological status:
susceptible to the disease, infected by the disease, recovered from the
disease, etc. To illustrate, consider a disease that confers no immunity.
This is modeled by a very simple SIS model. Individuals are thus either
susceptible to the disease, with the number of such individuals at time

t denoted by S(t), or infected by the disease (and are also infective in
the sense that they propagate the disease), with the number of such in-
dividuals at time t denoted by I(t). Suppose for simplicity that there is
neither birth nor death. Hence N ≡ N (t) = S(t) + I(t) is the (constant)
total population. Infection is assumed to take place following a standard
incidence pattern (see [181] for a discussion of transmission terms): the
number of new infectives resulting from random contacts between sus-
ceptible and infective individuals per unit time is given by βSI/N . Here
β is the transmission coefficient, it gives the probability of transmission
of the disease in case of a contact, times the number of such contacts
made by an infective per unit time.
Traditional epidemiological models assume that recovery from disease
occurs with a rate constant γ. However, as in [112], the assumption is
made here that, of the individuals who have become infective at time
t0 , a fraction P (t − t0 ) remain infective at time t ≥ t0 . Thus, considered
for t ≥ 0, the function P (t) is a survival function. As such, it satisfies
the properties given in Section 2.1, and in particular, P (0) = 1. The
transfer diagram for the system then has the form shown in Figure 13.1.
dS
dI
β SI/N
dN
S I
P(t)
Figure 13.1. The transfer diagram for the SIS model.
Since N is constant, it follows that S(t) = N − I(t) and the model

reduces to the following integral equation for the number of infective
individuals
t
(N − I(u))I(u)
I(t) = I0 (t) + β P (t − u)du (2.2)
0 N
Here I0 (t) is the number of individuals who were infective at time t = 0
and who still are at time t. It suffices to assume that I0 (t) is nonnegative,
nonincreasing, and such that limt→∞ I0 (t) = 0. The term P (t−u) in the
integral is the proportion of individuals who became infective at time
u and who still are at time t. Multiplying this with the contact term
β(N −I(u))I(u)/N and summing over [0, t] gives the number of infective
individuals at time t.
The two extreme cases for P (t) considered in Section 2.1 illustrate
various possibilities. First, suppose that P (t) is such that the sojourn
time in the infective state has an exponential distribution with mean

1/γ, i.e., P (t) = e−γt . Then (2.2) is
t
(N − I(u))I(u) −γ(t−u)
I(t) = I0 (t) + β e du
0 N
Taking the time derivative of I(t) yields
t
(N − I(u))I(u) −γ(t−u) (N − I(t))I(t)
I (t) = I0 (t) − γ β e du + β
0 N N
(N − I(t))I(t)
= I0 (t) + β + γ (I0 (t) − I(t))
N
In this case I0 (t) = I0 (0)e−γt , giving
(N − I(t))I(t)
I (t) = β − γI(t)
N
which is the classical logistic type ordinary differential equation (ODE)
for I in an SIS model without vital dynamics (see, e.g., [34, p. 289]).
The basic reproduction number, denoted by R0 , which is a key concept
in mathematical epidemiology, is now introduced. It is defined (see,
e.g., [12, 58]) as the expected number of secondary cases produced, in
a completely susceptible population, by the introduction of a typical
infective individual. For this ODE model, R0 = β/γ. In terms of
stability, the disease free equilibrium (DFE) with I = 0 is stable for
R0 < 1 and unstable for R0 > 1. At the threshold R0 = 1, there is a
forward bifurcation with a stable endemic equilibrium (with I > 0) for
R0 > 1. Thus the value of R0 determines whether the disease dies out
or tends to an endemic value.
The second case corresponds to P (t) being a step function:

1 if t ∈ [0, ω]
P (t) =
0 otherwise
i.e., the sojourn time in the infective state is a constant ω > 0. In this
case (2.2) becomes
t
(N − I(u))I(u)
I(t) = I0 (t) + β du
t−ω N
which when differentiated, gives for t ≥ ω

(N − I(t))I(t) (N − I(t − ω)) I(t − ω)
I (t) = I0 (t) + β −β
N N
Since I0 (t) vanishes for t > ω, this gives the delay differential equation
(DDE)
(N − I(t))I(t) (N − I(t − ω))I(t − ω)
I (t) = β −β
N N
cf. [34, Section 7.6] (where the disease transmission is modelled using
mass action). Note that every constant value of I is an equilibrium, thus
the integral form above gives a better description than the DDE. For this
case, R0 = βω again acts as a threshold. For R0 < 1, the DFE is stable;
whereas for R0 > 1, the endemic equilibrium is locally asymptotically
stable [34, Section 7.6].
More realistically, the survival function for the infective state is be-
tween an exponential and a step function (see, e.g., [12, 225]), thus the
two cases considered above can be regarded as extremes.
3. A model that includes a vaccinated state

We now use the ideas of the previous section in a different setting.
Consider a disease for which there exists a vaccine. Suppose that, al-
though there exists a vaccine, we can assume that developing the disease
confers no immunity. For example, at a given time, there are several
strains of influenza circulating in a given population. Vaccination usu-
ally focuses on particular strains, which are expected to be the dominant
ones in a particular year. Vaccination gives partial protection from other
strains as does contracting the disease. However, this protection is only
partial, and some individuals can contract the disease several times.
Thus if considered as one single disease, influenza can fit the above de-
scription. The assumptions also apply to models in which individuals
can be in two groups depending on their transmission coefficients with
respect to a given disease. They can move between these groups as
education campaigns or policies influence their behavior.
Our model, which is similar to that in [13], has the transfer diagram
shown in Figure 13.2. The number of individuals in the susceptible,
infective and vaccinated states are given by S(t), I(t), V (t), respectively.
As noted above, V (t) may alternatively correspond to an educated state,
but we refer to it as vaccinated. Individuals move from one state to the
other as their status with respect to the disease evolves. New individuals
are born into the susceptible state with a birth rate constant d > 0, and
all individuals, whatever their status, are subject to death with the same
natural death rate constant d. It is assumed that the disease does not
cause death, thus the total population N = S(t)+I(t)+V (t) is constant,
allowing for the simplification that the number of individuals in the S
state is given by S(t) = N − I(t) − V (t). Susceptible individuals are
vaccinated with rate constant φ, and enter the V state. Note that the
model in [13] further assumes that a fraction of newborns are vaccinated.
dS
dI
βSI/N
dN
S I
γI
φS
σβVI/N
P(t)
V
dV
Figure 13.2. The transfer diagram for the SIV model.
As in Section 2.2, disease transmission is assumed to be of standard in-

cidence type, thus susceptibles enter the infective state at a rate βSI/N ,
where β > 0 is the transmission coefficient. In addition, it is assumed
that successfully vaccinated individuals may only be partially protected
from infection (i.e., the vaccine is leaky). Vaccinated individuals can
contract the disease, but vaccination reduces transmission by a factor
σ ∈ [0, 1). Thus the number of new infectives produced by random con-
tacts between I infectives and V vaccinated individuals per unit time
is σβSI/N , and vaccinated individuals enter the infective state at this
rate.
Many vaccines wane with time, and so vaccinated individuals return
to the susceptible state. In [130], this waning is assumed to be expo-
nential but here we assume a more general waning function P (t). We
suppose that, at a given time t, there is a fraction P (t) of the vaccinated
individuals who are still under protection of the vaccine t units after
being vaccinated. Since the waning function P (t) is a survival function
∞ to be nonnegative and nonincreasing with P (0) = 1, and
it is assumed
moreover 0 P (u)du is positive and finite. Finally, it is assumed that
the infective individuals can be cured, so that members of the I state
return to the susceptible state, with rate constant γ ≥ 0 (the recovery
rate).
Since the total population remains constant, it is more convenient to
use proportions (rather than number of individuals) in each state. Here-
after, we use I(t) and V (t) to denote the proportion of infective and
vaccinated individuals, respectively, with S(t) = 1 − I(t) − V (t), the
proportion of susceptibles. Let the initial susceptible and infective pro-

portions be S(0) > 0, I(0) > 0 and let V0 (t) be the proportion of individ-
uals who are initially in the vaccinated state and for whom the vaccine
is still effective at time t. With the above assumptions, the following
integro-differential system describes the model depicted in Figure 13.2.
dI(t)
= β(1 − I(t) − (1 − σ)V (t))I(t) − (d + γ)I(t) (3.1a)
dt
t t
V (t) = V0 (t) + φS(u)P (t − u)e−d(t−u) e−σβ u I(x)dx du(3.1b)
0
The integral in (3.1b) sums the proportion of those who were vaccinated
at time u and remain in the V state at time t. Specifically, φS(u) is
the proportion of vaccinated susceptibles, P (t − u) is the fraction of the
proportion vaccinated still protected by the vaccine t−u time units after
going in (i.e., not returned to S), e−d(t−u) is the fraction of
the propor-
−σβ ut I(x)dx
tion vaccinated not dead due to natural causes, and e is the
fraction of the proportion vaccinated not infective (i.e., not progressed
to the I state). An expression for V0 (t) can be obtained by formulating
the model with vaccination state-age (see, e.g., [13, 112]) as
∞
− 0t (σβI(x)+d)dx P (t + u)
V0 (t) = e v(0, u) du (3.2)
0 P (u)
where v(0, u) ≥ 0 is the density at t = 0 of the
∞proportion of individu-
als in vaccination state-age u; thus V0 (0) = 0 v(0, u)du. The above
integral converges, and thus V0 (t) is nonnegative, nonincreasing and
limt→∞ V0 (t) = 0.
Define the subset D of the nonnegative orthant by
D = {(S, I, V ); S ≥ 0, I ≥ 0, V ≥ 0, S + I + V = 1}
It is easy to show (see [13]) that the set D is positively invariant under
the flow of (3.1) with I(0) > 0, S(0) > 0.
Differentiating (3.1b) gives
d d
V (t) = V0 (t) + φS(t) − (d + σβI(t))(V (t) − V0 (t)) + Q(t) (3.3)
dt dt
where
t t
Q(t) = φS(u)dt (P (t − u))e−d(t−u) e−σβ u I(x)dx
du
0
With the assumed initial conditions in D, the system defined by (3.1a)

and (3.1b) is equivalent to the system defined by (3.1a) and (3.3). This
latter system is of standard form, therefore results of Hale and Verd1uyn

Lunel [105, p. 43] ensure the local existence, uniqueness and continua-
tion of solutions of model (3.1).
Equation (3.1a) has I = 0 as an equilibrium and using I = 0 in
equation (3.1b) as t → ∞ gives the disease free equilibrium (DFE) as
IDF E = 0,
1 φP̃
SDF E = , VDF E =
1 + φP̃ 1 + φP̃
Here t
P̃ = lim P (v)e−dv dv
t→∞ 0
which is the average length of time that an individual remains vaccinated
(before losing vaccination protection or dying).
The basic reproduction number with vaccination is defined in terms
of P̃ as
σφP̃ + 1
Rvac = R0 (3.4)
φP̃ + 1
β
in which R0 = d+γ is the basic reproduction number with natural death
but no vaccination. The number Rvac is the important quantity in this
model that includes vaccination; Rvac is equal to the product of the mean
infective period 1/(d + γ) and the sum of the contact rate constant in
each of the susceptible and vaccinated states multiplied respectively by
the proportion in that state at the DFE, namely βSDF E + σβVDF E .
Note that Rvac ≤ R0 , and in the case of no vaccination, that is φ = 0,
Rvac = R0 .
If R0 < 1, then the only equilibrium of (3.1a) is IDF E = 0, thus the
DFE is the only equilibrium of system (3.1) when R0 < 1. In this case,
(3.1a) gives
dI
< (d + γ) ((S + σV ) − 1) I
dt
which implies that dI/dt < 0, and so I(t) → 0 = IDF E as t → ∞, for all
initial conditions I(0) > 0. Thus the disease dies out if R0 < 1.
The importance of Rvac can be seen from the following linear stability
result.
Theorem 1 For model ( 3.1) with a general waning function, if Rvac <
1, then the DFE is locally asymptotically stable (l.a.s.); if Rvac > 1, then
it is unstable.
Proof. Linearize (3.1a) and (3.1b) about the DFE, taking t → ∞.
Then the eigenvalues z of the linearized system at the DFE are given by
z = β(SDF E + σVDF E ) − (d + γ) = (d + γ)(Rvac − 1) (3.5a)
and the roots of

∞
1 = −φ P (v)e−(d+z)v dv (3.5b)
0
Let z = x + iy be a root of equation (3.5b). Then by the proof of

Lemma 2 in [230], if x ≥ 0, then y = 0. But since φ ≥ 0, equation
(3.5b) has no nonnegative real root, thus all of its roots have negative
real parts. Hence, from (3.5a), the DFE is l.a.s if Rvac < 1, and unstable
if Rvac > 1.
4. Reduction of the system by using specific

P (t) functions
Here we show two examples of models resulting from the choice of
specific vaccine waning functions P (t) as the two extreme cases in Sec-
tion 2.2. The first example (Section 4.1) is obtained when the distri-
bution of waning times is exponential, and leads to the ODE system
studied in [130]. As discussed in [130], for some parameter values, there
is a backward bifurcation, a rather uncommon phenomenon in epidemio-
logical models. This backward bifurcation is also present when the sys-
tem consists of delay integro-differential equations, such as is the case in
Section 4.2 when the waning function is assumed to be a step function
corresponding to a constant sojourn time in the vaccinated state.
4.1 Case reducing to an ODE system

Assuming that the vaccine waning
rate is a constant θ > 0, i.e., P (t) =
− 0t σβI(x)dx
e−θt ,
then V0 (t) = V0 (0)e−(d+θ)t e from (3.2), equations (3.1a)
and (3.3) give the ODE system
dI
= β(1 − I − (1 − σ)V )I − (d + γ)I (4.1a)
dt
dV
= φ(1 − I − V ) − σβIV − (d + θ)V (4.1b)
dt
which is the model studied in [130]. The DFE with IDF E = 0,
θ+d φ
SDF E = , VDF E =
d+θ+φ d+θ+φ
always exists, and from (3.4) the basic reproduction number is
d + θ + σφ
Rvac = R0
d+θ+φ
Assume that R0 > 1, then endemic equilibria (I > 0) can be obtained

analytically from a quadratic equation, and for σ > 0 (i.e., a leaky
vaccine) it is possible to have a backward bifurcation leading to two
endemic equilibria for some parameter values. This occurs for a range
of Rvac , namely Rc < Rvac < 1 where Rc is the value of Rvac at the
saddle node bifurcation point where the two endemic equilibria coincide;
see [130] for details.
4.2 Case reducing to a delay integro-differential

system
Assume that the vaccine waning period is constant and equal to ω > 0,
that is the function P (t) takes the form of a step function on a finite
interval:
1 if t ∈ [0, ω]
P (t) =
0 otherwise
Since S = 1 − I − V , and V0 (t) = 0 for t > ω, the integral equation
(3.1b) becomes, for t > ω
t t
V (t) = φ(1 − I(u) − V (u))e−d(t−u) e−σβ u I(x)dx du (4.2)
t−ω
Differentiating this last expression (see equation (3.3)), the model can
be written as the two dimensional integro-differential equation system
for t > ω
dI(t)
= β(1 − I(t)−(1−σ)V (t))I(t)−(d + γ)I(t) (4.3a)
dt
dV (t) t
= φ(1−I(t)−V (t))−φ(1−I(t−ω)−V (t−ω))e−dω e−σβ t−ω I(x)dx
dt
−σβIV − dV (4.3b)
Hereafter, we shift time by ω so that these equations hold for t > 0. By

introducing a third state variable
t
X(t) = I(x)dx (4.4)
t−ω
which gives dX(t)

dt = I(t) − I(t − ω), the system can be regarded as a
three dimensional DDE system.
For a constant waning period, the basic reproduction number from
(3.4) is
d + σφ(1 − e−dω )
Rvac = R0 (4.5)
d + φ(1 − e−dω )
The DFE is IDF E = 0,
d φ(1 − e−dω )
SDF E = , VDF E = (4.6)
d + φ(1 − e−dω ) d + φ(1 − e−dω )
Note that the delay ω enters into these equilibrium values. If R0 < 1,
then the system tends to the DFE and the disease dies out (see Sec-
tion 3). For R0 > 1, from nullclines, there exists one (or more) (EEP)
iff there exists I ∗ ∈ (0, 1] such that
∗
1 − 1/R0 − I ∗ φ(1 − I ∗ )(1 − e−dω−σβωI )
= (4.7)
1−σ φ(1 − e−dω−σβωI ∗ ) + d + σβI ∗
5. Numerical considerations
We give some insights into numerical aspects by considering the de-
lay integro-differential model (4.3). First, in Section 5.1 we set up the
algorithm that is used to study the occurence of forward and backward
bifurcations at Rvac = 1. We use this algorithm in Section 5.2, and in-
vestigate the dynamical behavior of system (4.3) by running numerical
integrations.
5.1 Visualising and locating the bifurcation

An EEP exists iff there exists an I ∗ ∈ (0, 1] such that (4.7) holds. So
we study the zeros of
1 − 1/R0 − I φ(1 − I)(1 − e−dω−σβωI )

H(I) = −
1−σ φ(1 − e−dω−σβωI ) + d + σβI
Rvac −1
Note that H(0) = (1−σ)R 0
and H(1) < 0.
Let A = {β, d, γ, φ, ω, σ} be the set of parameters of the model. When
needed, we denote H(I, a) and Rvac (a), with a ∈ A a parameter, to
indicate that the bifurcation is considered as a function of this parameter
a; e.g., Rvac (β) indicates that β is the bifurcation parameter that varies.
For a totally effective vaccine (σ = 0), a unique I ∗ ∈ (0, 1] such that
H(I ∗ ) = 0 can be found explicitly for Rvac > 1, and the bifurcation
is forward with Rvac behaving as a (local) threshold [13]. For a leaky
vaccine, σ ∈ (0, 1), the zeros of H(I) for I ∈ (0, 1] cannot be found
analytically. We proceed to obtain numerical estimates by using the
following algorithm.
Choose a parameter a ∈ A.
Fix the value of all other elements of A.
Choose amin , amax and ∆a for a.
for ak = amin to amax do

Compute I ∗ such that H(I ∗ , ak ) = 0, using MatLab’s fzero func-
tion.
Compute Rvac (a) for this value ak .
ak = ak + ∆a.
end for
Results of the use of this procedure give zero, one or two values of
I ∗ . Thus two bifurcation scenarios are possible, as summarized in Fig-
ure 13.3. Example bifurcation diagrams are plotted in Figures 13.4(a)
and 13.5.
In order to be able to characterize the nature of the bifurcation, we
then need to define Rc as in Section 4.1. To obtain a numerical estimate
of Rc , we use the same procedure as for the visualization of the bifurca-
tion: we find the value I ∗ such that H(I ∗ , a) = 0 and dH(I ∗ , a)/dI = 0,
for a given parameter a ∈ A with all other elements of A fixed.
Forward bifurcation
0 EEP 1 EEP
R vac
R c =1
Backward bifurcation
0 EEP 2 EEP 1 EEP
R vac
Rc 1
Figure 13.3. Possible bifurcation scenarios.
Suppose that R0 > 1 (otherwise there is no EEP, as was remarked

in Section 3). When Rvac < Rc , there is no EEP as H(0) < 0 and
numerical simulations indicate that H < 0 on (0, 1); when Rvac > 1,
H(0) > 0 so there is an odd number of EEP (numerical simulations
indicate this number is 1). When Rc = 1, we are then in the case of a
forward bifurcation, as illustrated in the first part of Figure 13.3 and in
Figure 13.4(c). The backward bifurcation arises when Rc < 1. In this
case, when Rc < Rvac < 1, H(0) < 0 so there is an even number of zeros
of H in (0, 1]. Numerical simulations indicate that the number of EEP
is 2. The system then undergoes the transitions shown in the second
part of Figure 13.3.
5.2 Numerical bifurcation analysis and

integration
We use the following parameter values. We suppose a 3 weeks disease
duration (γ = 1/21), taking the time unit as one day. The average
lifetime is assumed to be 75 years (d = 1/(75 × 365)), and the average
number of adequate contacts per day is estimated as β = 0.4. The

vaccine is assumed to be 10% leaky (σ = 0.1), and susceptibles are
vaccinated at the rate φ = 0.1. Finally, we assume that the vaccine
stops being effective after 5 years, i.e., ω = 1825.
These parameters give R0 = 8.3936 and Rvac (β) = 0.8819 from (4.5),
which is in the range of the backward bifurcation since (using the above
method) Rc (β) is estimated as 0.78. The bifurcation diagram is depicted
in Figure 13.4(a). Note that in the vicinity of Rc , it is very difficult for
MatLab’s fzero function to find solutions (since it detects sign changes
and Rvac = Rc corresponds to tangency); hence the non-closed curve.
Numerical simulations of the DDE model indicate that there are no
additional bifurcations; solutions either go to the DFE or to the (larger)
EEP, as depicted in Figure 13.4(b), which shows some solutions for I(t)
with the above parameter values. These same parameter values, except
that σ = 0.3, give Rvac (β) = 2.55, and there is a forward bifurcation
(see Figure 13.4(c)) with solutions going to the endemic equilibrium as
depicted in Figure 13.4(d).
To obtain Figures 13.4(b), 13.4(d), system (4.3) is integrated numeri-
cally. These numerical simulations are run using dde23 [205], an example
code (as well as an example code with XPPAUT) being given in Appen-
dix 1. Initial data is I(t) = c, for t ∈ [−ω, 0], with c varying from 0 to 1
by steps of 0.02.
Figure 13.5 shows the bifurcation for these parameter values as a func-
tion of ω. The situation is clearly different from that of Figure 13.4(a),
since in Figure 13.5 every value of ω gives at least one endemic equi-
librium. Let ωm be the value of ω determined by solving Rvac (ω) = 1
with Rvac given by (4.5). If all other parameters are fixed as given
at the beginning of this section, and for small waning time, 0 < ω <
ωm = 457.032, giving Rvac (ω) > 1, the only stable equilibrium is a large
endemic one. This is of course a highly undesirable state in terms of
epidemic control. Then increasing ω (i.e., increasing the waning time)
past ωm allows the DFE to become locally stable, and it is found nu-
merically that solutions starting with I(0) below the unstable endemic
equilibrium tend to the DFE. Increasing ω beyond 1000 days seems inef-
fective in terms of disease control, since there is no increase in the initial
value of infectives that tend to the DFE (see Figure 13.5).
6. A few words of warning

Even more so than with ordinary differential equations, great care
has to be taken when running numerical integrations of delay differential
equations. In [47], Cooke, van den Driessche and Zou study the dynamics
1
1
0.9
0.9
0.8
0.8
0.7
0.7
I(t)
0.6
0.6
I* 0.5
0.5
0.4
0.4
0.3 0.3
0.2 0.2
0.1 0.1
R (β)
c
0 0
0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 1.1 0 100 200 300 400 500 600 700 800
R (β) t
vac
(a) (b)
1
1
0.9
0.9
0.8
0.8
0.7
0.7
I*
0.6 I(t)
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1 0.1
0 0
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 50 100 150
Rvac(β) t
(c) (d)
Figure 13.4. Bifurcation diagram and some solutions of (4.3). (a) and (b): Backward
bifurcation case, parameters as in the text. (c) and (d): Forward bifurcation case,
parameters as in the text except that σ = 0.3.
of the following equation for an adult population N (t) with maturation

delay:
N (t) = be−aN (t−T ) N (t − T )e−d1 T − dN (t) (6.1)
Here d > 0 is the death rate constant, b > d and a > 0 are parameters
in the birth function, T is a developmental or maturation time and d1
is the death rate constant for each life stage prior to the adult stage.
In particular, they prove [47, Corollary 3.4] that Hopf bifurcation may
occur for (6.1) even for d1 = 0. For fixed values of the parameters, as
T increases the equilibrium may switch from being stable to unstable,
giving rise to periodic solutions. For d1 > 0, it is possible for stability
of the equilibrium to be regained as T increases further. They then
proceed to illustrate the stability switches by numerical simulations of
0.9
0.8
0.7
I*
0.6
0.5
0.4
0.3
0.2
0.1
0
0 200 400 ω 600 800 1000 1200 1400 1600 1800 2000
m ω
Figure 13.5. Value of I ∗ as a function of ω by solving H(I, ω) = 0, parameters as in

text.
(6.1) using XPPAUT. For d1 > 0, equation (6.1) has a delay dependent
parameter. The introduction of delay dependent parameters can lead to
dramatic differences in dynamics, see [27].
Using the demography of (6.1), the authors of [47] formulate the fol-
lowing SIS model with maturation delay [47, (4.2)]
I (t) = β(N (t) − I(t)) NI(t)

(t) − (d + + γ)I(t)
−a(t−T
(6.2)
N (t) = be ) N (t − T )e−d1 T − dN (t) − I(t)
where S(t) = N (t) − I(t), ≥ 0 is the disease induced death rate con-
stant, γ ≥ 0 is the recovery rate constant, and standard incidence βSI/N
is assumed. They perform numerical simulations of (6.2), and, in par-
ticular, obtain periodic solutions for parameter values a = d = d1 = 1,
b = 80, γ = 0.5, T = 0.2, = 10 and β = 20.
But... When documenting their delay differential equation numerical
integrator dde23 [205], Shampine and Thompson tested their algorithm
on a large number of delay systems, among which were equation (6.1)
and system (6.2). With parameters as in the paragraph above, they
obtain a figure similar to Figure 13.6, which shows damped oscillations
to an endemic steady state.
So, what is wrong? For delay differential equations, XPP (the nu-
merical integrator part of XPPAUT) uses a fixed step-size numerical
integrator, whereas dde23 uses a variable step-size. With the particular
values of the parameters chosen for β = 20, the fixed step-size is too
4
I
N
3.5
2.5
1.5
0.5
0 5 10 15 20 25
t
Figure 13.6. Plot of the solution of (6.2), with parameters as in the text, using dde23.
large (its default value is 0.05). In a case in which variables I and N

undergo a very quick initial drop, this is overlooked by the first integra-
tion step of XPP, and the solver ends caught in the solution curve of a
nearby periodic solution. Setting the step size in XPP to 0.005, as in
the Erratum of [47], is sufficient to obtain a correct solution as shown in
Figure 13.6.
Acknowledgements
This research was supported in part by MITACS and NSERC. The
authors acknowledge helpful discussions on probabilistic aspects of dis-
ease transmission with J. Hsieh, and thank K.L. Cooke and J. Velasco-
Hernández for collaboration on the vaccination model.
Appendix
1. Program listings
The following gives examples of code used with MatLab and XPPAUT to run
numerical integrations of system (4.3). In both cases, constant initial data has been
used, though both do allow for initial data of functional or of numerical type. In the
case of constant initial data, both programs have the same behavior: they extend the
given initial point to the interval [−ω, 0]. Note that we make use of the third “fake”
state variable X(t) introduced in (4.4), and of its time derivative, in order to take
care of the integral term in (4.3b).
1.1 MatLab code

The following is called vaccddeRHS.m. It defines the vector field of (4.3). This is
done in a very similar manner to the definition of the vector field that would be used
in a MatLab program with an ordinary differential equation solver. The one important
difference is in the variable Z. dde23 can handle many discrete delays. The variable
Z, which is passed as an argument to the function, contains the state of the system
at the different delays. Here, we have only one delay. But suppose we had two delays
ω1 and ω2 . Then each column of Z would contain the state variables corresponding
to one of the delays:
Iω1 Iω2
Z=
Vω1 Vω2
function v = vaccddeRHS ( t , y , Z , params )

b e t a=params ( 1 ) ;
d=params ( 2 ) ;
g=params ( 3 ) ; %MatLab h a t e s gamma ’ s o t h e r than gamma f u n c t i o n s .
p h i=params ( 4 ) ;
omega=params ( 5 ) ;
sigma=params ( 6 ) ;
ylag = Z ( : , 1 ) ;
v = zeros ( 3 , 1 ) ;
v ( 1 ) = b e t a ∗(1−y(1) −(1 − sigma ) ∗ y ( 2 ) ) ∗ y (1) −( d+g ) ∗ y ( 1 ) ;
v ( 2 ) = p h i ∗(1−y(1) − y ( 2 ) ) . . .
−p h i ∗(1− y l a g (1) − y l a g ( 2 ) ) ∗ exp(−d∗omega ) ∗ exp(−sigma ∗ b e t a ∗y ( 3 ) ) . . .
−sigma ∗ b e t a ∗y ( 1 ) ∗ y (2) −d∗y ( 2 ) ;
v (3) = y (1) − ylag ( 1 ) ;
This function is then used by the main calling routine, which follows. This partic-
ular procedure will run a certain number of integrations of system (4.3). The initial
condition for V (0) (lines 16 and 17) is obtained from (4.2) by setting t = 0, that of
X(0) follows from (4.4).
path ( path , ’ /home/ j a r i n o / programs / matlab / dde23 / d d e a l l ’ )
beta =0.4;
d =3.65297E−05;
g =0.047619048;
phi =0.1;
omega =1825;
sigma = 0 . 1 ;
ylim ( [ 0 , 1 ] ) ;
hold on ;
f o r I 0 = 0 : 0 . 0 2 : 1 , %Loop on i n i t i a l c o n d i t i o n s
% The d e l a y must be added t o t h e parameter v e c t o r s i n c e i t
% i s used i n t h e v e c t o r f i e l d .
params =[ beta , d , g , phi , omega , sigma ] ;
% I n i t i a l c o n d i t i o n s : I 0 i s g i v e n , V0 and X0 a r e computed .
V0=( p h i ∗(1− I 0 )∗(1 −exp(−omega ∗ ( d+b e t a ∗ sigma ∗ I 0 ) ) ) ) . . .
/ ( d+b e t a ∗ sigma ∗ I 0+p h i ∗(1−exp(−omega ∗ ( d+b e t a ∗ sigma ∗ I 0 ) ) ) ) ;
X0=I 0 ∗omega ; % I n i t i a l c o n d i t i o n on X i s e a s y t o compute .
IC=[ I0 , V0 , X0 ] ; % Extended t o [−omega , 0 ] i f o n l y g i v e n a t 0 .
t s p a n = [ 0 , 8 0 0 ] ; % S e t i n t e g r a t i o n time r ange .
% Call the numerical routine .
s o l = dde23 ( ’ vaccddeRHS ’ , omega , IC , tspan , [ ] , params ) ;
plot ( s o l . t , s o l . y ( 1 , : ) ) ; %P l o t I ( t ) v e r s u s t
end ;
xlabel ( ’ t ’ ) ; ylabel ( ’ I ’ , ’ R o t a t i o n ’ , 0 ) ;
1.2 XPPAUT code

The following code allows the integration of system (4.3) with XPPAUT. Note that
the initial conditions for V and X have to be computed explicitly from equations (4.2)
and (4.4), since XPP does not allow inclusion of unevaluated formula in the code.
# Constants
p b e t a = 0 . 4 , d =3.65E−05 , g = 0 . 0 4 7 6 2 , p h i = 0 . 1 , omega =1825 , sigma = 0 . 1 ;
# The system
d I / dt = b e t a ∗(1− I−V+sigma ∗V) ∗ I − ( d+g ) ∗ I
dV/ dt = p h i ∗(1− I−V)− p h i ∗(1− d e l a y ( I , omega)− d e l a y (V, omega ) ) \
∗ exp(−d∗omega ) ∗ exp(−sigma ∗ b e t a ∗X) −( sigma ∗ b e t a ∗ I ∗V)−d∗V
dX/ dt = I−d e l a y ( I , omega )
# I n i t i a l conditions
I (0)=0.1
V( 0 ) = 0 . 8 6 5 0 5 9 5 3 3 4
X( 0 ) = 1 8 2 . 5
# s e t maxdelay
@ d e l a y =2000
@ ylow=0
@ b e l l =0
@ bound=500
@ XP=I ,YP=V
@ XHI=1 ,YHI=1
# done
d
2. Delay differential equations packages

Several packages and even software are available for the numerical integration
and/or the study of bifurcations in delay differential equations. Here is a short list,
elaborated from the list given by Koen Engelborghs1 .
2.1 Numerical integration

The following are numerical solvers for DDE’s.
Archi (C.A.H. Paul) (Fortran 77) simulates a large class of functional differential
equations. In particular, Archi can be used to estimate unknown scalar parameters
in delay and neutral differential equations.
dde23 (L. Shampine, S. Thompson) (MatLab) is a MatLab package that integrates
delay differential equations. It is integrated in the latest versions of MatLab (starting
with Release 13).
DDVERK (H. Hiroshi, W. Enright) (Fortran 77) simulates retarded and neutral
differential equations with several fixed discrete delays.
DifEqu (G. Makay) (DOS, Windows) simulates differential equations with discrete
possibly varying delays.
∗ http://www.cs.kuleuven.ac.be/ koen/delay
DKLAG6 (S. Thompson) (Fortran 77, Fortran 90, C) simulates retarded differential
equations with several fixed discrete delays.
Dynamics Solver (J. M. Aguirregabiria) simulates differential equations with dis-
crete possibly varying delays.
RETARD (E. Hairer, G. Wanner) simulates retarded differential equations with
several fixed discrete delays.
RADAR5 (N. Guglielmi, E. Hairer) (Fortran 90) simulates retarded differential-
algebraic equations, including neutral problems with vanishing or small delays.
XPPAUT (G.B. Ermentrout) (Unix, Windows) simulates differential equations with
several fixed discrete delays. XPPAUT is a standalone software.
2.2 Bifurcation analysis

The following software packages provide some means to carry out numerical bifur-
cation analysis of delay differential equations.
BIFDD (B.D. Hassard) (Fortran 77) normal form analysis of Hopf bifurcations of
differential equations with several fixed discrete delays.
DDE-BIFTOOL (K. Engelborghs) (MatLab) allows computation and stability
analysis of steady state solutions, their fold and Hopf bifurcations and periodic solu-
tions of differential equations with several fixed discrete delays.
XPPAUT (G.B. Ermentrout) (Unix, Windows) allows limited stability analysis of
steady state solutions of differential equations with several fixed discrete delays.
References
[1] M. Adimy and K. Ezzinbi, Spectral decomposition for some func-

tional differential equations, Submitted.
[2] E. Ait Dads, Contribution à l’existence de solutions pseudo
presque périodiques d’une classe d’équations fonctionnelles, Doc-
torat d’Etat Unversité Cadi Ayyad, Marrakech (1994).
[3] E. Ait Dads, K. Ezzinbi, O. Arino, Pseudo almost periodic solu-
tions for some differential equations in a Banach space, Nonlinear
Analysis, Theory, Methods & Applications 28 (1997), 1141–1155.
[4] E. Ait Dads and K. Ezzinbi, Boundedness and almost periodicity
for some state-dependent delay differential equations, Electron. J.
Differential Equations 2002 (2002), No. 67, 1–13.
[5] W.C. Allee, Animal Aggregations: A Study in General Sociology,
Chicago University Press, Chicago (1931).
[6] H. Amann, Linear and Quasilinear Parabolic Problems,
Birkhäuser (1995).
[7] L. Amerio, and G. Prouse, Almost Periodic Functions and Func-
tional Analysis, Van Nostrand (1971).
[8] B. Amir and L. Maniar, Composition of pseudo almost periodic
functions and semilinear Cauchy problems with non dense domain,
Ann. Math. Pascal 6 (1999), 1–11.
[9] B. Amir and L. Maniar, Asymptotic behavior of hyperbolic
Cauchy problems for Hille-Yosida operators with an application
to retarded differential equations, Questiones Mathematicae 23
(2000), 343–357.
[10] B. Amir and L. Maniar, Existence and asymptotic behavior of
solutions of semilinear Cauchy problems with non dense domain
via extrapolation spaces, Rend. Circ. Mat. Palermo XLIX (2000),
481–496.
559
[11] Y. Ammar, Eine dreidimensionale invariante Mannigfaltigkeit für

autonome Differentialgleichungen mit Verzögerung, Doctoral dis-
sertation, München (1993).
[12] R.M. Anderson and R.M. May, Infectious Diseases of Humans,
Oxford University Press (1991).
[13] J. Arino, K.L. Cooke, P. van den Driessche, and J. Velasco-
Hernández, An epidemiology model that includes vaccine efficacy
and waning, Discr. Contin. Dyn. Systems B 4 (2003), 479–495.
[14] O. Arino and M. Kimmel, Stability analysis of models of cell pro-
duction systems, Math. Modelling 7 (1986), 1269–1300.
[15] O. Arino, E. Sánchez and A. Fathallah, State-dependent delay dif-
ferential equations in population dynamics: modeling and analysis,
in “Topics in Functional Differential and Difference Equations”,
ed. T. Faria and P. Freitas, Fields Institute Commun. 29 (2001),
Amer. Math. Soc., Providence, pp. 19–36.
[16] O. Arino and E. Sanchez, A saddle point theorem for functional
state-dependent delay equations, preprint (2002).
[17] A. Bátkai, L. Maniar and A. Rhandi, Regularity properties of per-
turbed Hille-Yosida operators and retarded differential equations,
Semigroup Forum 64 (2002), 55–70.
[18] M. Bardi, An equation of growth of a single species with realistic
dependence on crowding and seasonal factors, J. Math. Biol. 17
(1983), 33–43.
[19] M. Bardi and A. Schiaffino, Asymptotic behavior of positive solu-
tions of periodic delay logistic equations, J. Math. Biol. 14 (1982),
95–100.
[20] M. Bartha, Convergence of solutions for an equation with state-
dependent delay, J. Math. Anal. Appl. 254 (2001), 410–432.
[21] M. Bartha, Periodic solutions for differential equations with state-
dependent delay and positive feedback, Nonlinear Analysis - TMA
53 (2003), 839–857.
[22] C.J.K. Batty, and R. Chill, Bounded convolutions and solutions of
inhomogeneous Cauchy problems, Forum Math. 11 (1999), 253–
277.
[23] J.R. Beddington and R.M. May, Time delays are not necessarily
destabilizing, Math. Biosci. 27 (1975), 109–117.
[24] J. Bélair, Population modles with state-dependent delays, in Math-
ematical Population Dynamics, ed. O. Arino, D. E. Axelrod and
M. Kimmel, Marcel Dekker, New York (1991), pp. 165–176.
REFERENCES 561
[25] J. Bélair and S. A. Campbell, Stability and bifurcations of equi-

librium in a multiple-delayed differential equation, SIAM J. Appl.
Math. 54 (1994), 1402–1424.
[26] J. Bélair, M.C. Mackey and J.M. Mahaffy, Age-structured and
two delay models for erythropoiesis, Math. Biosci. 128 (1995),
317–346.
[27] E. Beretta and Y. Kuang, Geometric stability switch criteria in
delay differential systems with delay dependent parameters, SIAM
J. Math. Anal. 33 (2002), 1144–1165.
[28] A. Beuter, J. Bélair and C. Labrie, Feedback and delay in neuro-
logical diseases: a modeling study using dynamical systems, Bull.
Math. Biol. 55 (1993), 525–541.
[29] R.H. Bing, The Geometric Topology of 3-Manifolds, A.M.S. Col-
loquium Publ., Vol. 40, Providence (1983).
[30] S.P. Blythe, R.M. Nisbet and W.S.C. Gurney, Instability and com-
plex dynamic behaviour in population models with long time de-
lays, Theor. Pop. Biol. 22 (1982), 147-176.
[31] L.L. Bonilla and A. Liñán, Relaxation oscillations, pulses, and
traveling waves in the diffusive Volterra delay-differential equation,
SIAM J. Appl. Math. 44 (1984), 369-391.
[32] R.D. Braddock and P. van den Driessche, On a two lag differential
delay equation, J. Austral. Math. Soc. Ser. B 24 (1983), 292-317.
[33] F. Brauer. Time lag in disease models with recruitment. Mathe-
matics and Computer Modeling 31 (2000), 11–15.
[34] F. Brauer and C. Castillo-Chávez, Mathematical Models in Pop-
ulation Biology and Epidemiology, Springer (2001).
[35] N.F. Britton, Reaction-Diffusion Equations and their Applications
to Biology, Academic Press, London (1986).
[36] N.F. Britton, Spatial structures and periodic traveling waves in
an integro-differential reaction-diffusion population model, SIAM
J. Appl. Math. 50 (1990), 1663-1688.
[37] S. Busenberg and K.L. Cooke, Periodic solutions of a periodic non-
linear delay differential equation, SIAM J. Appl. Math. 35 (1978),
704-721.
[38] S. Busenberg and K.L. Cooke. Vertically Transmitted Diseases.
Springer-Verlag (1993).
[39] S. Busenberg and W. Huang, Stability and Hopf bifurcation for a
population delay model with diffusion effects, J. Differential Equa-
tions 124 (1996), 80-107.
[40] Y. Cao and T.C. Gard, Ultimate bounds and global asymptotic
stability for differential delay equations, Rocky Mountain J. Math.
25 (1995), 119-131.
[41] M.-P. Chen, J.S. Yu, X.Z. Qian and Z.C. Wang, On the stability of
a delay differential population model, Nonlinear Analysis - TMA,
25 (1995), 187–195.
[42] Y. Chen, Global attractivity of a population model with state-
dependent delay, in “Dynamical Systems and Their Applications
in Biology”, ed. S. Ruan, G. Wolkowicz and J. Wu, Fields Institute
Communications 36 (2003), Amer. Math. Soc., Providence, pp.
113–118.
[43] S.N. Chow and H.O. Walther, Characteristic multipliers and sta-
bility of symmetric periodic solutions of ẋ(t) = g(x(t − 1)). Trans-
actions of the A.M.S. 307 (1988), 127–142.
[44] Ph. Clément, O. Diekmann, M. Gyllenberg, H.J.A.M. Heijmans
and H.R. Thieme, Perturbation theory for dual semigroups. I. The
sun-reflexive case, Math. Ann. 277 (1987), 709–725.
[45] D.S. Cohen and S. Rosenblat, A delay logistic equation with
growth rate, SIAM J. Appl. Math. 42 (1982), 608-624.
[46] K.L. Cooke, Stability analysis for a vector disease model, Rocky
Mountain J. Math. 9 (1979), 31-42.
[47] K.L. Cooke, P. van den Driessche, and X. Zou. Interaction of
maturation delay and nonlinear birth in population and epidemic
models. J. Math. Biol. 39 (1999) 332–352. Erratum: J. Math.
Biol., 45 (2002), 470.
[48] K.L. Cooke and W. Huang, On the problem of linearization for
state-dependent delay differential equations. Proceedings of the
A.M.S. 124 (1996), 1417-1426.
[49] K.L. Cooke and J.A. Yorke, Some equations modelling growth
processes and gonorrhea epidemics, Math. Biosci. 16 (1973), 75-
101.
[50] C. Corduneanu, Almost-Periodic Functions, 2nd edition, Chelsea ,
New York (1989).
[51] C. Corduneanu, Integral Equations and Stability of Feedback Sys-
tems, Academic Press New York and London (1973).
[52] J.M. Cushing (1977a), Integrodifferential Equations and Delay
Models in Population Dynamics, Lecture Notes in Biomathematics
20 (1977), Springer-Verlag, Heidelberg.
REFERENCES 563
[53] J.M. Cushing (1977b), Time delays in single growth models, J.

Math. Biol. 4 (1977), 257-264.
[54] F.A. Davidson and S.A. Gourley, The effects of temporal delays
in a model for a food-limited diffusion population, J. Math. Anal.
Appl. 261 (2001), 633-648
[55] W. Desch, G. Gühring and I. Györi, Stability of nonau-
tonomous delay equations with a positive fundamental solution,
Tübingerberichte zur Funktionalanalysis 9 (2000).
[56] O. Diekmann, S.A. van Gils, S.M. Verduyn Lunel and H.O. Wal-
ter, Delay Equations Functional, complex and nonlinear Analysis,
Springer Verlag (1991).
[57] O. Diekmann, S.A. van Gils, S.M. Verduyn Lunel and H.O.
Walther, Delay Equations: Functional-, Complex- and Nonlinear
Analysis. Springer, New York (1995).
[58] O. Diekmann and J.A.P. Heesterbeek, Mathematical Epidemiol-
ogy of Infectious Diseases: Model Building, Analysis and Interpre-
tation, Wiley (2000).
[59] P. Dormayer, An attractivity region for characteristic multipliers
of special symmetric periodic solutions of ẋ(t) = α f (x(t−1)) near
critical amplitudes, J. Math. Anal. Appl. 169 (1992), 70–91.
[60] P. Dormayer, Floquet multipliers and secondary bifurcation of pe-
riodic solutions of functional differential equations, Habilitation
thesis, Gießen, 1996.
[61] R.D. Driver, Existence theory for a delay-differential system, Con-
tributions to Differential Equations 1 (1963), 317–336.
[62] Eichmann, M., A local Hopf bifurcation theorem for differen-
tial equations with state-dependent delays, Doctoral dissertation,
Justus-Liebig-Universität, Gießen (2006).
[63] K.J. Engel and R. Nagel, One-Parameter Semigroups for Lin-
ear Evolution Equations, Graduate Texts in Mathematics 194
Springer-Verlag (2000).
[64] K. Ezzinbi, Contribution à l’étude des équations différentielles à
retard en dimension finie et infinie : existence et aspects qualitatifs,
application à des problèmes de dynamique de populations, Thesis,
Marrakesh (1997).
[65] M. D. Fargue, Reducibilité des systèmes héréditaires à des
systèmes dynamiques, C. R. Acad. Sci. Paris. Sér. B 277 (1973),
471-473.
[66] T. Faria and Huang, Stability of periodic solutions arising from

Hopf bifurcation for a reaction-diffusion equation with time delay,
in “Differential Equations ad Dynamical Systems”, ed. A. Galves,
J. K. Hale and C. Rocha, Fields Institute Communications 31
(2002), Amer. Math. Soc., Providence, pp. 125-141.
[67] W. Feng and X. Lu, Asymptotic periodicity in diffusive logis-
tic equations with discrete delays, Nonlinear Analysis - TMA 26
(1996), 171-178.
[68] W. Feng and X. Lu, Global periodicity in a class of reaction-
diffusion systems with time delays, Discrete Contin. Dynam. Sys-
tems - Ser. B 3 (2003), 69-78.
[69] Z. Feng and H.R. Thieme, Endemic models with arbitrarily dis-
tributed periods of infection I: Fundamental properties of the
model, SIAM J. Appl. Math. 61 (2000), 803–833.
[70] P. Fife, Mathematical Aspects of Reacting and Diffusing Systems,
Lecture Notes in Biomathematics 28, Springer-Verlag, Berlin
(1979).
[71] R.A. Fisher (1937), The wave of advance of advantageous genes,
Ann. Eugenics 7, 353-369.
[72] A.C. Flower and M.C. Mackey, Relaxation oscillations in a class
of delay differential equations, SIAM J. Appl. Math. 63 (2002),
299-323.
[73] H.I. Freedman and K. Gopalsamy, Global stability in time-delayed
single-species dynamics, Bull. Math. Biol. 48 (1986), 485-492.
[74] H.I. Freedman and J. Wu, Periodic solutions of single-species mod-
els with periodic delay, SIAM J. Math. Anal. 23 (1992), 689-701.
[75] H.I. Freedman and H. Xia, Periodic solutions of single species mod-
els with delay, in “Differential Equations, Dynamical Systems, and
Control Science”, ed. K. D. Elworthy, W. N. Everitt and E. B. Lee,
Marcel Dekker, New York, pp. 55-74 (1994).
[76] H.I. Freedman and X. Zhao, Global asymptotics in some quasi-
monotone reaction-diffusion systems with delays, J. Differential
Equations 137 (1997), 340-362.
[77] R.E. Gaines and J. L. Mawhin, Coincidence Degree and Nonlinear
Differential Equations, Springer-Verlay, Berlin (1977).
[78] K. Gopalsamy, Global stability in the delay-logistic equation with
discrete delays, Houston J. Math. 16 (1990), 347-356.
[79] K. Gopalsamy, Stability and Oscillations in Delay Differential
Equations of Population Dynamics, Mathematics and its Appli-
cations 74, Kluwer Academic Pub., Dordrecht (1992).
REFERENCES 565
[80] K. Gopalsamy and B.D. Aggarwala, The logistic equation with a

diffusionally coupled delay, Bull. Math. Biol. 43 (1981), 125-140.
[81] K. Gopalsamy, M.R.S. Kulenović and G. Ladas, Time lags in a
“food-limited” population model, Appl. Anal. 31 (1988), 225-237.
[82] K. Gopalsamy, M.R.S. Kulenović and G. Ladas (1990a), Enrivon-
mental periodicity and time delays in a “food-limited” population
model, J. Math. Anal. Appl. 147 (1990), 545-555.
[83] K. Gopalsamy, M.R.S. Kulenović and G. Ladas (1990b), Oscilla-
tions and global attractivity in models of hematopoiesis, J. Dy-
namics Differential Equations 2 (1990), 117-132.
[84] K. Gopalsamy and G. Ladas, On the oscillation and asymptotic
behavior of Ṅ (t) = N (t)[a + bN (t − τ ) − cN 2 (t − τ )], Quart. Appl.
Math. 48 (1990), 433-440.
[85] S.A. Gourley and N.F. Britton, On a modified Volterra population
equation with diffusion, Nonlinear Analysis - TMA 21 (1993), 389-
395.
[86] S.A. Gourley and M.A.J. Chaplain, Travelling fronts in a food-
limited population model with time delay, Proc. Royal Soc. Edin-
burgh 132A (2002), 75-89.
[87] S.A. Gourley and S. Ruan, Dynamics of the diffusive Nicholson’s
blowflies equation with distributed delay, Proc. Royal Soc. Edin-
burgh 130A (2000), 1275-1291.
[88] S.A. Gourley and J.W.-H. So, Dynamics of a food-limited popu-
lation model incorporating nonlocal delays on a finite domain, J.
Math. Biol. 44 (2002), 49-78.
[89] J.R. Greaf, C. Qian and P.W. Spikes, Oscillation and global attrac-
tivity in a periodic delay equation, Canad. Math. Bull. 38 (1996),
275-283.
[90] D. Green and H.W. Stech, Diffusion and hereditary effects in a
class of population models, in “Differential Equations and Appli-
cations in Ecology, Epidemics, and Population Problems,” ed. S.
Busenberg and K. Cooke, Academic Press, New York, pp. 19-28
(1981).
[91] E.A. Grove, G. Ladas and C. Qian, Global attractivity in a “food-
limited” population model, Dynamic. Systems Appl. 2 (1993), 243-
249.
[92] G. Gühring and F. Räbiger, Asymptotic properties of mild solu-
tions of evolution equations with applications to retarded differ-
ential equations, Abstr. Apl. Anal. 4 (1999), 169–194.
[93] G. Gühring, F. Räbiger and W. Ruess, Linearized stability for

semilinear non-autonomous equations with applications to re-
tarded differential equations, to appear in Diff. Integ. Equat.
[94] G. Gühring, F. Räbiger and R. Schnaubelt, A characteristic equa-
tion for non-autonomous partial functional differential equations,
J. Differential Equations, to appear.
[95] W.S.C. Gurney, S.P. Blythe and R.M. Nisbet, Nicholson’s blowflies
revisited, Nature 287 (1980), 17-21.
[96] I. Györi and S. I. Trofimchuk, Global attractivity in x (t) =
−δx(t) + pf (x(t − τ )), Dynamic. Systems Appl. 8 (1999), 197-210.
[97] I. Györi and S. I. Trofimchuk, On existence of rapidly oscillatory
solutions in the Nicholson blowflies equation, Nonlinear Analysis
- TMA 48 (2002), 1033-1042.
[98] K.P. Hadeler, On the stability of the stationary state of a pop-
ulation growth equation with time-lag, J. Math. Biol. 3 (1976),
197-201.
[99] K.P. Hadeler and J. Tomiuk, Periodic solutions of difference-
differential equations, Arch. Rational Mech. Anal. 65 (1977), 82-
95.
[100] J.K. Hale Ordinary Differential Equations, Wiley-Inter Science,
A Division of John Wiley & Sons, New York, London, Sydney,
Toronto (1969).
[101] J.K. Hale, Theory of Functional Differential Equations, Springer
Verlag (1977).
[102] J. K. Hale and W. Huang, Global geometry of the stable regions for
two delay differential equations, J. Math. Anal. Appl. 178 (1993),
344-362.
[103] J.K. Hale and X.B. Lin, Symbolic dynamics and nonlinear semi-
flows. Annali di Matematica Pura ed Applicata 144 (1986), 229–
259.
[104] J.K. Hale and S. Verduyn Lunel, Introduction to Functional Dif-
ferential Equations, Springer Verlag (1991).
[105] J.K. Hale and S. Verduyn Lunel, Introduction to Functional Dif-
ferential Equations. Springer, New York (1993).
[106] T.G. Hallam and J.T. DeJuna, Effects of toxicants on popula-
tions: A qualitative approach III. Environmental and food chain
pathways, J. Theor. Biol. 109 (1984), 411-429.
REFERENCES 567
[107] F. Hartung, T. Krisztin, H.O. Walther and J. Wu, Functional dif-

ferential equations with state-dependent delay: Theory and applica-
tions, To appear in Handbook of Differential Equations - Ordinary
Differential Equations, A. Canada, P. Drabek. and A. Fonda eds.,
Elsevier Science B. V., North Holland.
[108] B.D. Hassard, N.D. Kazarinoff and Y.H. Wan (1981), Theory and
Applications of Hopf Bifurcation, London Mathematical Society
Lecture Note Series 41, Cambridge University Press, Cambridge.
[109] E. Hille and R.S. Phillips, Functional Analysis and Semigroups,
Amer. Math. Soc. Providence (1975).
[110] Y. Hino, S. Murakami and T. Naito, Functional Differential Equa-
tions with Infinite Delay, Lecture Notes in Mathematics 1473,
Springer Verlag (1991).
[111] W. Huang, Global dynamics for a reaction-diffusion equation with
time delay, J. Differential Equations 143 (1998), 293-326.
[112] H.W. Hethcote and P. van den Driessche. Two SIS epidemiologic
models with delays. J. Math. Biol., 40 (2000), 3–26.
[113] G.E. Hutchinson, Circular cause systems in ecology, Ann. N. Y.
Acad. Sci. 50 (1948), 221–246.
[114] G.E. Hutchinson, An Introduction to Population Ecology, Yale
University Press, New Haven (1978).
[115] A.F. Ivanov, B. Lani-Wayda and H.O. Walther, Unstable hyper-
bolic periodic solutions of differential delay equations. In Recent
Trends in Differential Equations, R.P. Agarwal ed., 301–316, WS-
SIAA vol. 1, World Scientific, Singapore (1992).
[116] A.F. Ivanov and J. Losson, Stable rapidly oscillating solutions in
delay differential equations with negative feedback. Differential and
Integral Equations 12 (1999), 811–832.
[117] G.S. Jones (1962a), On the nonlinear differential-difference equa-
tion f (t) = −αf (x − 1)[1 + f (x)], J. Math. Anal. Appl. 4 (1962),
440–469.
[118] G.S. Jones (1962b), The existence of periodic solutions of f (t) =
−αf (x − 1)[1 + f (x)], J. Math. Anal. Appl. 5 (1962), 435–450.
[119] S. Kakutani and L. Markus, On the non-linear difference-
differential equation y (t) = [A − By(t − τ )]y(t), in “Contributions
to the Theory of Nonlinear Oscillations”, Vol. 4, ed. S. Lefschetz,
Princeton University Press, New Jersey, 1-18 (1958).
[120] J.L. Kaplan and J.A. Yorke, Ordinary differential equations which
yield periodic solutions of differential delay equations. J. Math.
Anal. Appl. 48 (1974), 317-324.
[121] J.L. Kaplan and J.A. Yorke, On the stability of a periodic solution
of a differential delay equation, SIAM J. Math. Anal. 6 (1975),
268-282.
[122] F. Kappel, Linear Autonomous Functional Differential Equations,
Summer School on Delay Differential Equations Marrakesh June
7 − 16 (1995).
[123] G. Karakostas, The effect of seasonal variations to the delay pop-
ulation equation, Nonlinear Analysis -TMA 6 (1982), 1143-1154.
[124] G. Karakostas, Ch.G. Philos and Y.G. Sficas, Stable steady state
of some population models, J. Dynamics Differential Equations 4
(1992), 161-190.
[125] J. Kirk, J.S. Orr and J. Forrest, The role of chalone in the control
of the bone marrow stem cell population, Math. Biosci. 6 (1970),
129-143.
[126] R.L. Kitching, Time, resources and population dynamics in in-
sects, Austral. J. Ecol. 2 (1977), 31-42.
[127] A.N. Kolmogorov, I.G. Petrovskii and N.S. Piskunov, Étude de
l’équation de le diffusion avec croissance de la quantité de mateère
et son application à un problème biologique, Moscow Univ. Bull.
1 (1937), 1-25.
[128] V.A. Kostitzin, Sur les équations intégrodifférentielles de la théorie
de l’action toxique du milieu, C. R. Acad. Sci. 208 (1939), 1545-
1547.
[129] J. Kreulich, Eberlein weak almost periodicity and differential equa-
tions in Banach spaces, Ph.D. Thesis, Essen (1992).
[130] C. Kribs-Zaleta and J. Velasco-Hernández. A simple vaccination
model with multiple endemic states. Math. Biosci. 164 (2000),
183–201.
[131] H.P. Krishnan, An analysis of singularly perturbed delay-
differential equations and equations with state-dependent delays,
Ph.D. thesis, Brown University, Providence (R.I.) (1998).
[132] T. Krisztin, A local unstable manifold for differential equations
with state-dependent delay, Discrete and Continuous Dynamical
Systems 9 (2003), 993–1028.
[133] T. Krisztin, Invariance and noninvariance of center manifolds of
time-t maps with respect to the semiflow, SIAM J. Math. Analysis
36 (2004), 717–739.
REFERENCES 569
[134] T. Krisztin, C 1 -smoothness of center manifolds for differential

equations with state-dependent delay, Preprint, to appear in Non-
linear Dynamics and Evolution Equations, Fields Institute Com-
munications.
[135] T. Krisztin and O. Arino, The 2-dimensional attractor of a differ-
ential equation with state-dependent delay, J. Dynamics Differen-
tial Equations 13 (2001), 453–522.
[136] T. Krisztin and H.O. Walther, Unique periodic orbits for delayed
positive feedback and the global attractor, J. Dynamics and Dif-
ferential Equations 13 (2001), 1–57.
[137] T. Krisztin, H.O. Walther and J. Wu, Shape, Smoothness, and In-
variant Stratification of an Attracting Set for Delayed Monotone
Positive Feedback. Fields Institute Monograph series vol. 11,
A.M.S., Providence (1999).
[138] T. Krisztin, H.O. Walther and J. Wu, The structure of an attract-
ing set defined by delayed and monotone positive feedback, CWI
Quarterly 12 (1999), 315–327.
[139] T. Krisztin and J. Wu, Monotone semiflows generated by neutral
equations with different delays in neutral and retarded parts, Acta
Mathematicae Universitatis Comenianae 63 (1994), 207–220.
[140] Y. Kuang, Global attractivity in delay differential equations re-
lated to models of physiology and population biology, Japan. J.
Indust. Appl. Math. 9 (1992), 205–238.
[141] Y. Kuang, Delay Differential Equations with Applications in Pop-
ulation Dynamics, Academic Press, New York (1993).
[142] Y. Kuang and H.L. Smith (1992a), Slowly oscillating periodic so-
lutions of autonomous state-dependent delay equations, Nonlinear
Analysis - TMA 19 (1992), 855–872.
[143] Y. Kuang and H.L. Smith (1992b), Periodic solutions of differ-
ential delay equations with threshold-type delays, Contemporay
Mathematics 129 (1992), 153–175.
[144] M.R.S. Kulenović, G. Ladas and Y.G. Sficas, Global attractivity
in Nicholson’s blowflies, Appl. Anal. 43 (1992), 109-124.
[145] G. Ladas and C. Qian, Oscillation and global stability in a delay
logistic equation, Dynamics and Stability of Systems 9 (1994), 153-
162.
[146] V. Lakshmikantham and S. Leela, Differential and integral inequal-
ities, theory and applications, Volume 1, Academic Press, (1969).
[147] B.S. Lalli and B.G. Zhang, On a periodic delay population model,
Quart. Appl. Math. 52 (1994), 35-42.
[148] K.A. Landman, Bifurcation and stability theory of periodic so-
lutions for integrodifferential equations, Stud. Appl. Math. 62
(1980), 217-248.
[149] B. Lani-Wayda, Hyperbolic Sets, Shadowing and Persistence for
Noninvertible Mappings in Banach Spaces, Pitman Research Notes
in Math., Vol. 334, Longman, Essex (1995).
[150] B. Lani-Wayda, Erratic solutions of simple delay equations, Trans-
actions of the A.M.S. 351 (1999), 901–945.
[151] B. Lani-Wayda, Wandering solutions of equations with sine-like
feedback. Memoirs of the A.M.S. Vol. 151, No. 718, 2001.
[152] B. Lani-Wayda and R. Srzednicki, The Lefschetz fixed point the-
orem and symbolic dynamics in delay equations, Ergodic Theory
and Dynamical Systems 22 (2002), 1215–1232.
[153] B. Lani-Wayda and H.O. Walther, Chaotic motion generated by
delayed negative feedback. Part I: A transversality criterion, Dif-
ferential and Integral Equations 8 (1995), 1407-1452.
[154] B. Lani-Wayda and H.O. Walther, Chaotic motion generated by
delayed negative feedback. Part II: Construction of nonlinearities,
Mathematische Nachrichten 180 (1996), 141-211.
[155] Y. Latushkin, T. Randolph and R. Schnaubelt, Exponential di-
chotomy and mild solutions of nonautonomous equations in Ba-
nach spaces, preprint.
[156] Y. Latushkin and R. Schnaubelt, The spectral mapping theorem
for evolution semigroups on Lp associated with strongly continu-
ous cocycles, preprint.
[157] S.M. Lenhart and C.C. Travis, Global stability of a biological
model with time delay, Proc. Amer. Math. Soc. 96 (1986), 75-78.
[158] X. Li, S. Ruan and J. Wei, Stability and bifurcation in delay-
differential equations with two delays, J. Math. Anal. Appl. 236
(1999), 254-280.
[159] Y. Li, Existence and global attractivity of a positive periodic so-
lution of a class of delay differential equation, Sci. China Ser. A
41 (1998), 273–284.
[160] Y. Li and Y. Kuang, Periodic solutions in periodic state-dependent
delay equations and population models, Proc. Amer. Math. Soc.
130 (2001), 1345–1353.
REFERENCES 571
[161] J. Lin and P.B. Kahn, Phase and amplitude instability in delay-
diffusion population models, J. Math. Biol. 13 (1982), 383–393.
[162] E. Liz, C. Martı́nez and S. Trofimchuk, Attractivity properties of
infinite delay Mackey-Glass type equations, Differential Integral
Equations 15 (2002), 875–896.
[163] E. Liz, M. Pinto, G. Robledo, S. Trofimchuk and V. Tkachenko,
Wright type delay differential equations with negative Schwarzian,
Discrete Contin. Dynam. Systems 9 (2003), 309–321.
[164] M. Louihi, M.L. Hbid and O. Arino, Semigroup properties and the
Crandall-Liggett approximation for a class of differential equations
with state-dependent delays, J. Differential Equations 181 (2002),
1–30.
[165] S. Luckhaus, Global boundedness for a delay differential equation,
Trans. Amer. Math. Soc. 294 (1986), 767–774.
[166] N. MacDonald, Time Lags in Biological Models, Lecture Notes in
Biomathematics 27 (1978), Springer-Verlag, Heidelberg.
[167] M.C. Mackey and L. Glass, Oscillation and chaos in physiological
control systems, Science 197 (1977), 287–289.
[168] M.C. Mackey and J.G. Milton, Dynamics disease, Ann. N. Y.
Acad. Sci. 504 (1988), 16-32.
[169] M.C. Mackey and J.G. Milton, Feedback delays and the origins of
blood cell dynamics, Comm. Theor. Biol. 1 (1990), 299–327.
[170] J.M. Mahaffy, P.J. Zak and K.M. Joiner, A geometric analysis
of the stability regions for a linear differential equation with two
delays, Internat. J. Bifur. Chaos 5 (1995), 779–796.
[171] J. Mallet-Paret, Morse decompositions for differential delay equa-
tions, J. Differential Equations 72 (1988), 270–315.
[172] J. Mallet-Paret and R. Nussbaum, Boundary layer phenomena
for differential-delay equations with state-dependent time lags, I,
Arch. Rational Mech. Anal. 120 (1992), 99-146.
[173] J. Mallet-Paret and R. Nussbaum, Boundary layer phenomena for
differential-delay equations with state-dependent time lags, II, J.
reine angew. Math. 477 (1996), 129-197.
[174] J. Mallet-Paret and R. Nussbaum, Boundary layer phenomena for
differential-delay equations with state-dependent time lags, III, J.
Differential Equations 189 (2003), 640-692.
[175] J. Mallet-Paret, R. Nussbaum and P. Paraskevopoulos, Periodic
solutions for functional differential equations with multiple state-
dependent time lags, Topol. Methods Nonlinear Anal. 3 (1994),
101-162.
[176] J. Mallet-Paret and G. Sell, Systems of differential delay equations:

Floquet multipliers and discrete Lyapunov functions, J. Differen-
tial Equations 125 (1996), 385–440.
[177] J. Mallet-Paret and H.O. Walther, Rapidly oscillating soutions
are rare in scalar systems governed by monotone negative feedback
with a time lag. Preprint, 1994.
[178] L. Maniar and A. Rhandi, Extrapolation and inhomogeneous re-
tarded differential equations in infinite dimensional Banach space
via extrapolation spaces, Rend. Circ. Mat. Palermo 47 (1998),
331-346.
[179] J. E. Marsden and M. McCracken, The Hopf Bifurcation and Its
Applications, Applied Mathematical Sciences 19, Springer-Verlag,
New York (1976).
[180] R.M. May, Stability and Complexity in Model Ecosystems, 2nd ed.,
Princeton University Press, Princeton (1974).
[181] H. McCallum, N. Barlow, and J. Hone, How should pathogen
transmission be modelled? Trends Ecol. Evol., 16 (2001), 295–
300.
[182] R.K. Miller, On Volterra’s population equation, SIAM J. Appl.
Math. 14 (1966), 446-452.
[183] R.K. Miller, Nonlinear Volterra Integral Equations, Benjamin
Press, Menlo Park, California (1971).
[184] Y. Morita, Destabilization of periodic solutions arising in delay-
diffusion systems in several space dimensions, Japan J. Appl.
Math. 1 (1984), 39-65.
[185] J.D. Murray, Mathematical Biology, Biomathematics 19, Springer-
Verlag, Berlin (1989).
[186] R. Nagel and E. Sinestrari, Inhomogeneous Volterra integrodiffer-
ential equations for Hille-Yosida operators, in Functional Analysis
(ed. K.D. Bierstedt, A. Pietsch, W.M. Ruess and D. Vogt), Lecture
Notes Pure Appl. Math. 150, Marcel Dekker (1994), 51-70.
[187] J. van Neerven, The adjoint of a Semigroup of Linear Operators,
Lecture Notes in Mathematics 1529, Springer-Verlag (1992).
[188] A.J. Nicholson, The balance of animal population, J. Animal Ecol.
2 (1933), 132-178.
[189] A.J. Nicholson, An outline of the dynamics of animal populations,
Austral. J. Zoo. 2 (1954), 9-65.
REFERENCES 573
[190] G. Nickel and A. Rhandi, On the essential spectral radius of semi-

groups generated by perturbations of Hille-Yosida operators, J.
Diff. Integ. Equat., to appear.
[191] R.M. Nisbet and W.S.C. Gurney, Population dynamics in a peri-
odically varying environment, J. Theor. Biol. 56 (1976), 459-475.
[192] R.M. Nisbet and W.S.C. Gurney, Modelling Fluctuating Popula-
tions, John Wiley and Sons (1982).
[193] R. Nussbaum, Periodic solutions of some nonlinear autonomous
functional differential equations, Ann. Mat. Pura Appl. 101
(1974), 263-306.
[194] C.V. Pao, Systems of parabolic equations with continuous and
discrete delays, J. Math. Anal. Appl. 205 (1997), 157-185.
[195] A. Pazy, Semigroups of linear operators and applications to partial
differential equations, Springer Verlag (1983).
[196] E.R. Pianka, Evolutionary Ecology, Harper and Row, New York
(1974).
[197] C. Qian, Global attractivity in nonlinear delay differential equa-
tions, J. Math. Anal. Appl. 197 (1996), 529-547.
[198] R. Redlinger, On Volterra’s population equation with diffusion,
SIAM J. Math. Anal. 16 (1985), 135-142.
[199] A. Rhandi, Extrapolation methods to solve non-autonomous re-
tarded differential equations, Studia Math. 126 (1997), 219-233.
[200] A. Rhandi and R. Schnaubelt, Asymptotic behaviour of a
non-autonomous population equation with diffusion in L1 ,
Disc. Cont. Dyn. Syst. 5 (1999), 663-683.
[201] A. Rhandi, Positivity and Stability for a population equation with
diffusion on L1 , Positivity 2 (1998), 101-113.
[202] G. Rosen, Time delays produced by essential nonlinearity in pop-
ulation growth models, Bull. Math. Biol. 28 (1987), 253-256.
[203] S. Ruan and D. Xiao, Stability of steady states and existence of
traveling waves in a vector disease model, Proceedings of the royal
Society of Edinburgh: Sect. A - Mathematics 134 (2004), 991-1011.
[204] S.H. Saker and S. Agarwal, Oscillation and global attractivity in
a periodic nicholson’s blowflies model, Math. Comput. Modelling
35 (2002), 719-731.
[205] L.F. Shampine and S. Thompson, Solving DDEs in MATLAB,
Appl. Numer. Math. 37 (2001), 441–458.
[206] H.C. Simpson, Stability of periodic solutions of integrodifferential

equations, SIAM J. Appl. Math. 38 (1980), 341-363.
[207] A. Schiaffino, On a diffusion Volterra equation, Nonlinear Analysis
- TMA 3 (1979), 595-600.
[208] A. Schiaffino and A. Tesei, Monotone methods and attractivity
for Volterra integro-partial differential equations, Proc. Royal Soc.
Edinburgh 89A (1981), 135-142.
[209] D. Schley and S.A. Gourley, Linear stability criteria for population
models with periodic perturbed delays, J. Math. Biol. 40 (2000),
500-524.
[210] A. Schoenfliess, Die Entwicklung der Lehre von den Punktman-
nigfaltigkeiten. Bericht erstattet der Deutschen Mathematiker-
Vereinigung, Teil II, Jahres-Berichte der DMV, Ergänzungsband
II (1908).
[211] A.L. Skubachevsky and H.O. Walther, On Floquet multipliers for
slowly oscillating periodic solutions of nonlinear functional differ-
ential equations, Trudy Moskov. Mat. Obshch. 64 (2002), 3–54;
English translation in: Transactions of the Moscow Mathematical
Society (2003), 1–44.
[212] H.L. Smith, Monotone Dynamical Systems, An Introduction to
the Theory of Competitive and Cooperative Systems, Mathemat-
ical Surveys and Monographs 41, Amer. Math. Soc., Providence
(1995).
[213] J.W.-H. So, J. Wu and Y. Yang, Numerical steady state and Hopf
bifurcation analysis on the diffusive Nicholson’s blowflies equation,
Appl. Math. Comput. 111 (2000), 33-51.
[214] J.W.-H. So and Y. Yang, Dirichlet problem for the diffusive Nichol-
son’s blowflies equation, J. Differential Equations 150 (1998), 317-
348.
[215] J.W.-H. So and J.S. Yu, Global attractivity and uniform persis-
tence in Nicholson’s blowfies, Differential Equations Dynamical
Systems 2 (1994), 11-18.
[216] J.W.-H. So and J.S. Yu, On the uniform stability for a “food-
limited” population model with time delay, Proc. Royal Soc. Ed-
inburgh 125A (1995), 991-1002.
[217] J.W.-H. So and X. Zou, Traveling waves for the diffusive Nichol-
son’s blowflies equation, Appl. Math. Comput. 122 (2001), 385-
392.
REFERENCES 575
[218] H. Stech, The effect of time lags on the stability of the equilibrium
state of a population growth equation, J. Math. Biol. 5 (1978),
115-120.
[219] H. Steinlein and H.O. Walther, Hyperbolic sets and shadowing for
noninvertible maps. In Advanced Topics in the Theory of Dynami-
cal Systems, Fusco, G., Iannelli, M., and L. Salvadori eds., 219-234,
Academic Press, New York (1989).
[220] H. Steinlein and H.O. Walther, Hyperbolic sets, transversal homo-
clinic trajectories, and symbolic dynamics for C 1 -maps in Banach
spaces, J. Dynamical Systems and Differential Equations 2 (1990),
325–365.
[221] J. Sugie, On the stability for a population growth equation with
time delay, Proc. Royal Soc. Edinburgh 120A (1992), 179-184.
[222] C.E. Taylor and R.R. Sokal, Oscillations in housefly population
sizes due to time lags, Ecology 57 (1976), 1060-1067.
[223] A. Tesei, Stability properties for partial Volterra integrodifferential
equations, Ann. Mat. Pura Appl. 126 (1980), 103-115.
[224] H.R. Thieme, Semiflows generated by Lipschitz perturbations of
non-densely defined linear operators, Diff. Integ. Equat. 3 (1990),
1035-1066.
[225] H.R. Thieme, The transition through stages with arbitrary length
distributions, and applications in epidemics, In C. Castillo-Chavez,
with S. Blower, P. van den Driessche, D. Kirschner, and A.-
A. Yakubu, editors, Mathematical Approaches for Emerging and
Reemerging Infectious Diseases: Models, Methods, and Theory,
volume 126 of IMA Volumes in Mathematics and its Applications,
pages 45–84 (2002).
[226] C.C. Travis and G.F. Webb, Existence and stability, for partial
functional differential equations, Tran. Amer. Math. Soc. 200
(1974), 395–418.
[227] C.C. Travis and G.F. Webb, Existence, stability, and compacteness
in the α-norm for partial functional differential equations, Tran.
Amer. Math. Soc. 240 (1978), 129–143.
[228] P. van den Driessche, Some epidemiological models with delays, In
Differential Equations and Applications to Biology and to Industry
(Claremont, CA, 1994), pages 507–520. World Sci. Publishing,
River Edge, NJ (1996).
[229] P. van den Driessche, Time delay in epidemic models, in
C. Castillo-Chavez, with S. Blower, P. van den Driessche,
D. Kirschner, and A.-A. Yakubu, editors, Mathematical Ap-

proaches for Emerging and Reemerging Infectious Diseases: An
Introduction, volume 125 of IMA Volumes in Mathematics and its
Applications, pages 119–128 (2002).
[230] P. van den Driessche and J. Watmough, A simple SIS epidemic
model with a backward bifurcation, J. Math. Biol. 40 (2000),
525–540.
[231] Vanderbauwhede, A., and S.A. van Gils, Center manifolds and
contractions on a scale of Banach spaces, J. Functional Analysis
71 (1987), 209–224.
[232] F.F. Verhulst, Notice sur la loi que la population suit dans son
acroissement, Corr. Math. Phys. 10 (1838), 113-121.
[233] V. Volterra, Remarques sur la note de M. Régnier et Mlle. Lambin
(Étude d’un cas d’antagonisme microbien), C. R. Acad. Sci. 199
(1934), 1684-1686.
[234] H.O. Walther, Existence of a non-constant periodic solution of a
nonlinear autonomous functional differential equation representing
the growth of a single species population, J. Math. Biol. 1 (1975),
227-240.
[235] H.O. Walther, Stable periodic motion of a system with state de-
pendent delay, Differential Integral Equations 15 (2002), 923-944.
[236] H.O. Walther, Über Ejektivität und periodische Lösungen bei Funk-
tionaldifferentialgleichungen mit verteilter Verzögerung, Habilita-
tion thesis, München (1977).
[237] H.O. Walther, On instability, ω-limit sets and periodic solutions of
nonlinear autonomous differential delay equations, in Functional
Differential Equations and Approximation of Fixed Points, H.O.
Peitgen and H.O. Walther eds., Lecture Notes in Math. 730, 489–
503, Springer, Heidelberg (1979).
[238] H.O. Walther, Bifurcation from periodic solutions in functional
differential equations, Mathematische Zeitschrift 182 (1983), 269–
289.
[239] H.O. Walther, A differential delay equation with a planar attrac-
tor, Accepted for the Proceedings of the Int. Conf. on Differential
Equations Marrakech 1991.
[240] H.O. Walther, The two-dimensional attractor of x (t) = −µx(t) +
f (x(t − 1)), Memoirs of the A.M.S. Vol. 113, No. 544, 1995.
[241] H.O. Walther, Stable sets of periodic solutions of a delay differen-
tial equation. To appear in Proceedings of the Internat. Conf. on
Differential Equations Marrakesh 1995.
REFERENCES 577
[242] H.O. Walther, The singularities of an attractor of a delay differen-

tial equation, Functional Differential Equations 5 (1998), 513–548.
[243] H.O. Walther, Contracting return maps for some delay differen-
tial equations, In Topics in Functional Differential and Difference
Equations, T. Faria and P. Freitas eds., Fields Institute Commu-
nications series 29, A.M.S., Providence (2001), 349–360.
[244] H.O. Walther, Contracting return maps for monotone delayed
feedback, Discrete and Continuous Dynamical Systems 7 (2001),
259–274.
[245] H.O. Walther, Stable periodic motion of a system with state-
dependent delay, Differential and Integral Equations 15 (2002),
923–944.
[246] H.O. Walther, Stable periodic motion of a delayed spring, Topo-
logical Methods in Nonlinear Analysis 21 (2003), 1–28.
[247] H.O. Walther, The solution manifold and C 1 -smoothness of so-
lution operators for differential equations with state dependent
delay, J. Differential Eqs. 195 (2003), 46–65.
[248] H.O. Walther, Stable periodic motion of a system using echo for
position control, J. Dynamics and Differential Eqs. 15 (2003),
143–223.
[249] H.O. Walther, Smoothness properties of semiflows for differential
equations with state dependent delay, Russian, in Proceedings of
the International Conference on Differential and Functional Dif-
ferential Equations, Moscow, 2002 1, pp. 40–55, Moscow State
Aviation Institute (MAI), Moscow (2003). English version: Jour-
nal of the Mathematical Sciences 124 (2004), 5193–5207.
[250] H.O. Walther and M. Yebdri, Smoothness of the attractor of al-
most all solutions of a delay differential equation, DISSERTA-
TIONES MATHEMATICAE CCCLXVIII (1997).
[251] M. Wazewska-Czyzewska and A. Lasota, Mathematical problems
of the dynamics of the red-blood cells system, Ann. Polish Math.
Soc. III Appl. Math. 6 (1976), 23-40.
[252] E. M. Wright, A non-linear difference-differential equation, J.
reine angew. Math. 194 (1955), 66–87.
[253] J. Wu, Theory and Applications of Partial Functional Differential
Equations, Applied Mathematical Sciences 119, Springer-Verlag,
New York (1996).
[254] Wu, J., Stable phase-locked periodic solutions in a delay differential
system, J. Differential Eqs. 194 (2003), 237–286.
[255] X. Xie, Uniqueness and stability of slowly oscillating periodic so-

lutions of delay differential equations with bounded nonlinearity,
J. Dynamics and Differential Equations 3 (1991), 515–540.
[256] J. Yan and Q. Feng, Global attractivity and oscillation in a nonlin-
ear delay equation, Nonlinear Analysis - TMA 43 (2001), 101–108.
[257] Y. Yamada, Asymptotic behavior of solutions for semilinear
Volterra diffusion equations, Nonlinear Analysis - TMA 21 (1993),
227–239.
[258] K. Yoshida, The Hopf bifurcation and its stability for semilinear
differential equations with time delay arising in ecology, Hiroshima
Math. J. 12 (1982), 321–348.
[259] J.S. Yu (1996), Global attractivity of the zero solution of a class
of functional-differential equations and its applications, Sci. China
Ser. A 39, 225-237.
[260] S.D. Zaidman, Abstract differential equations, Research notes in
Mathematics (1994).
[261] C. Zhang, Pseudo almost periodic functions and their applications,
Ph.D thesis, University of Western Ontario (1992).
[262] C. Zhang, Pseudo almost-periodic solutions of some differential
equations, J. Math. Anal. Appl. 181 (1994), 62–76.
[263] B. G. Zhang and K. Gopalsamy, Global attractivity and oscilla-
tions in a periodic delay-logistic equation, J. Math. Anal. Appl.
150 (1990), 274-283.
[264] Lin Zhensheng, Theory of linear systems, Ann. of Diff. Equ. 6
(1990).

Time Delays in Epidemic Models: Modeling and Numerical Considerations

Uploaded by

Copyright:

Available Formats

Time Delays in Epidemic Models: Modeling and Numerical Considerations

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Time Delays in Epidemic Models: Modeling and Numerical Considerations

Uploaded by

Copyright:

Available Formats

Chapter 13

TIME DELAYS IN EPIDEMIC MODELS

P. van den Driessche

2. Origin of time delays in epidemic models

2.1 Sojourn times and survival functions

Another characterization of the distribution of the random variable

Assuming that limt→∞ tS(t) = 0 (which is veriﬁed for most probability

These two distributions can be regarded as extremes.

2.2 Sojourn times in an SIS disease transmission

susceptible to the disease, with the number of such individuals at time

Figure 13.1. The transfer diagram for the SIS model.

Since N is constant, it follows that S(t) = N − I(t) and the model

time in the infective state has an exponential distribution with mean

which when diﬀerentiated, gives for t ≥ ω

3. A model that includes a vaccinated state

Figure 13.2. The transfer diagram for the SIV model.

As in Section 2.2, disease transmission is assumed to be of standard in-

proportion of susceptibles. Let the initial susceptible and infective pro-

With the assumed initial conditions in D, the system deﬁned by (3.1a)

latter system is of standard form, therefore results of Hale and Verd1uyn

and the roots of

Let z = x + iy be a root of equation (3.5b). Then by the proof of

4. Reduction of the system by using speciﬁc

4.1 Case reducing to an ODE system

always exists, and from (3.4) the basic reproduction number is

Assume that R0 > 1, then endemic equilibria (I > 0) can be obtained

4.2 Case reducing to a delay integro-diﬀerential

Hereafter, we shift time by ω so that these equations hold for t > 0. By

which gives dX(t)

The DFE is IDF E = 0,

5.1 Visualising and locating the bifurcation

1 − 1/R0 − I φ(1 − I)(1 − e−dω−σβωI )

for ak = amin to amax do

Figure 13.3. Possible bifurcation scenarios.

Suppose that R0 > 1 (otherwise there is no EEP, as was remarked

5.2 Numerical bifurcation analysis and

number of adequate contacts per day is estimated as β = 0.4. The

6. A few words of warning

of the following equation for an adult population N (t) with maturation

Figure 13.5. Value of I ∗ as a function of ω by solving H(I, ω) = 0, parameters as in

I (t) = β(N (t) − I(t)) NI(t)

large (its default value is 0.05). In a case in which variables I and N

1.1 MatLab code

function v = vaccddeRHS ( t , y , Z , params )

1.2 XPPAUT code

2. Delay diﬀerential equations packages

2.1 Numerical integration

2.2 Bifurcation analysis

[1] M. Adimy and K. Ezzinbi, Spectral decomposition for some func-

[11] Y. Ammar, Eine dreidimensionale invariante Mannigfaltigkeit für

[25] J. Bélair and S. A. Campbell, Stability and bifurcations of equi-

[53] J.M. Cushing (1977b), Time delays in single growth models, J.

[66] T. Faria and Huang, Stability of periodic solutions arising from

[80] K. Gopalsamy and B.D. Aggarwala, The logistic equation with a

[93] G. Gühring, F. Räbiger and W. Ruess, Linearized stability for

[107] F. Hartung, T. Krisztin, H.O. Walther and J. Wu, Functional dif-

[134] T. Krisztin, C 1 -smoothness of center manifolds for diﬀerential

[176] J. Mallet-Paret and G. Sell, Systems of diﬀerential delay equations:

[190] G. Nickel and A. Rhandi, On the essential spectral radius of semi-