Escolar Documentos
Profissional Documentos
Cultura Documentos
Abstract
In a recent paper a new model for the Aedes aegypti mosquito dispersal dynamics was proposed
and its Lie point symmetries were investigated. According to the carried group classification,
the maximal symmetry Lie algebra of the nonlinear cases is reached whenever the advection
term vanishes. In this work we analyze the family of systems obtained when the wind effects
on the proposed model are neglected. Wide new classes of solutions to the systems under
consideration are obtained.
Keywords: Aedes aegypti mosquitoes, symmetries, exact solutions
2008 MSC: 92D25, 76M60, 58J70, 35A30, 70G65
1. Introduction
Aedes aegypti mosquitoes can be considered the primary vector for dengue and urban yellow
fever. It is possible to find them around the world, not only in the tropical regions but also
beyond them, reaching temperate climates.
Because of its importance as a vector of deadly diseases, the significance of its distribution
in urban areas and the possibility of breeding in laboratory facilities, Aedes aegypti is one of
the best-focused and studied mosquitoes [1].
To the best of our knowledge, the first partial differential equation model for the Aedes
aegypti was introduced in [2, 3] by writing a semilinear system of two partial differential equa-
tions. In recent works, see [4, 5], taking into account the modeling for Proteus mirabilis bacterial
colonies (see [7]) we introduced the following quasilinear system for the dispersal dynamics of
the Aedes aegypti,
p q γ γ
ut = (u ux )x − 2νu ux + v +
− µ1 u,
k k
(1)
vt = ku + (k − µ2 − γ)v.
2
2. The symmetries for ν = 0
System (1) was studied from the point of view of Lie symmetry theory in [4] and there it
was completely classified with respect to the constitutive parameters p, q, ν.
For ν = 0, apart from the case p = 0, the largest Lie symmetry algebra is reached when
p = −4/3. In this case the system (1) reads
−4 γ γ
ut = (u 3 ux )x + v + ( − µ1 )u,
k k
(2)
vt = ku + (k − µ2 − γ)v,
whose Lie algebra, taking the results obtained in [4] into account, is spanned by
X3 = x2 ∂x − 3xu∂u − 3xv∂v .
The symmetries of (4) generates a 3-dimensional Lie algebra spanned by the generators (3).
X = (c1 + 2xc4 + c3 x2 )∂x + c2 ∂t − (3xuc3 + 3uc4 )∂u − (3xvc3 + 3vc4 )∂v (5)
of the basis of the Lie symmetry algebra for the system (2).
Then in order to determine the invariant solutions, we must solve the characteristic system
dt dx du
= = − , (6)
c2 c3 x2 + 2xc4 + c1 3xuc3 + 3uc4
dt dx dv
= 2
=− , (7)
c2 c3 x + 2xc4 + c1 3xvc3 + 3vc4
obtained from the invariant surface conditions.
One can assume, without lost of generality, that c3 = 1. Then, the solutions of the last
equation of (6) and (7) are, respectively,
A(φ) B(φ)
u= , v= 2 , (8)
(x2 + 2c4 x + c1 )3/2 (x + 2c4 x + c1 )3/2
3
while
c2
t+ , if c24 − c1 = 0,
x + c4
Z
!
dx
c2 x + c4
φ = t − c2 = t− p arctan p , if c24 − c1 < 0, (9)
x2 + 2c4 x + c1 c1 − c4 2
c1 − c42
p
2
c x + c − c − c
2 4 1
4
t − ln , if c24 − c1 > 0.
p p
2 2
2 c4 − c1 x + c4 + c4 − c1
In order to use the transformations (8) and (9) to construct exact solutions of (2), we avoid
from (5) time translations substituting c2 = 0 into (9). Therefore, after setting c2 = 0 in (9),
substituting it into (8) and, in the following, into (2), it is obtained the following ordinary
differential equation system for A and B:
1 γ γ
A0 = −3A− 3 (c1 − c24 ) + B + − µ1 A,
k k
(10)
B 0 = kA + (k − µ2 − γ)B.
If we consider the case c1 = c24 , by isolating B in the first equation of (10) and by substituting
into the second one, the following linear second order ODE is obtained
A00 (t) + bA0 (t) + cA(t) = 0, (11)
where
γ γ
b = µ1 + µ2 + γ − k − , c = −γ + − µ1 (k − µ2 − γ). (12)
k k
The cases c2 6= 0 and the cases c1 6= c24 of the system (10) will not be considered in this paper.
The solution of (11) is
√ ! √ !
bt ∆t bt ∆t
A(t) = a1 e− 2 cosh + a2 e− 2 sinh , (13)
2 2
Going back to the original variables, taking (9) into account, we are able to write u(x, t)
and v(x, t).
4
4. Reductions and exact solutions of the system (4)
4.1. Reductions
We now consider the system (4) with the following general infinitesimal operator
X = pc1 X1 + c2 X2 + Xp (15)
where c1 and c2 are arbitrary real constants.
By applying the invariant surface conditions we get that the invariant solutions, with respect
to the transformations generated by the operator X, must be of the following form:
2 2
u(x, t) = A(σ)(x + c1 ) p , v(x, t) = B(σ)(x + c1 ) p ,
where
et
σ= c2
(c1 + x) p
is the so-called similarity variable and with A(σ) and B(σ) satisfying the following general
reduced system of ODEs
0 0 2 γ γ
A σ = (2A − A σ) + B + − µ 1 A
k k
c2 A00 σ 2
p 4 2 0σ 2 (16)
+ A 2
− A + A (1 − c2 ) 1 + + ,
p p p p p
0
B σ = kA + (k − µ2 − γ)B.
As in the previous case, by neglecting traslations in time, the reduced system (16) can be
simplified. The similarity variable becomes σ = t, while the solutions are separable variable
solutions of the form:
2 2
u(x, t) = A(t)(x + c1 ) p , v(x, t) = B(t)(x + c1 ) p , (17)
where A(t) and B(t) satisfy the following first order system of ODEs
A0 (t) = 4 + 2p A(t)p+1 + γ B(t) + γ − µ1 A(t),
p2 k k
(18)
B 0 (t) = kA(t) + (k − µ − γ)B(t).
2
It is easy to ascertain that by a simple calculations it is possible to reduce the search for
solutions of (18) to the search or solutions of the following second order ODE:
00 (p + 1)(4 + 2p) p γ
A (t) − A (t) + − µ1 + k − µ2 − γ A0 (t) +
p2 k
4 + 2p p
γ
+ (k − µ2 − γ) 2 A(t) − γ + (k − µ2 − γ) − µ1 A(t) = 0, (19)
p k
5
that is a family of Lienard equations. After having put
(p + 1)(4 + 2p) 4 + 2p
a= , d = (k − µ2 − γ) , (20)
p2 p2
we are able to write (19) as
where b and c are given by (12). Some classes of solutions of this family will be investigated in
the following.
This equation is identical to equation (11) whose solutions have been shown in the previous
section. In particular, if the parameters are related by (see [4])
α2 α2 α2
γ= , µ1 = , k = ρ, µ2 = ρ − ,
4 4ρ 4
where α and ρ are positive constants, the solutions obtained in Section 3 are specialized as it
follows
1 αt 2ρ αt
u(x, t) = sinh , v(x, t) = cosh
(x + c1 ) 2 α(x + c1 ) 2
or
1 αt 2ρ αt
u(x, t) = cosh , v(x, t) = sinh .
(x + c1 ) 2 α(x + c1 ) 2
where
d 2(k − µ2 − γ)
= .
−γ + γk − µ1 (k − µ2 − γ)
c
After substituting (22) into the second equation of the system (18) we can easily obtain the
function B(t).
6
4.4. Solutions to Eq. (21) with p 6= −1; −2
In this subsection, once considered the previous special cases, the analysis of Eq. (21) is
concerned with the study of a Lienard equation family. After having studied some special cases,
where we have been able to get solutions in a closed form, in this subsection we show some
special solutions concerned with this type of equations.
We observe that for values of the constitutive parameters of the same order of magnitude
of those given by (25) it is possible to assume that
Let w be a function such that w(A) = A0 (t). Then equation (21) becomes
ww0 − (a Ap − b) w + (d Ap + c) A = 0. (27)
7
It is a simple matter to ascertain that the equation (27) admits the solutions w = mA, m ∈
R, provided that the following conditions are satisfied
ma − d = 0, m2 + bm + c = 0.
In the same way after substituting A1,2 in the second equation of the system (18) it is a simple
matter to solve it and get the corresponding B1,2 (t).
equation (27) becomes an Abel equation of the second kind in the canonical form
cz + dz p+1
ww0 = w + , (29)
b − az p
8
where w = w(z).
We observe that equation (29) is always defined because we are in the case p 6= −1; −2
which implies a 6= 0.
In the case b = 0, that is when
γ
k − µ2 − γ = − µ1 ,
k
the equation (29) specializes as
with g = − ad and h = − ac .
Once equation (29) is solved, we can determine the function A and, consequently, sys-
tem (18) is solved. The solutions of (30) are, usually, determined in a parametric form and,
not rarely, they depend on special functions, such as Elliptic Weierstrass function and Bessel
function. A large number of equations of the type (30) are discussed in [13].
k − µ2 − γ < 0.
The biological condition for the existence of the mosquito population, see in [14] and [6], is
γ − µ1 (µ2 + γ) > 0 which is easily obtained once it is assumed
γ
k< , k > µ2 + γ.
µ1
These assumptions imply that A0 (t) > 0 and B 0 (t) > 0, for any p > −2. Such conditions
occur when there is an infestation of mosquitos.
5. Conclusions
In this paper we recall a model concerned with Aedes aegypti introduced in [4] and look
for similarity solutions, in absence of wind effects, by using the symmetries found in [4]. The
classification performed there shows that the system considered admits a four dimensional
algebra for p = − 34 and a three-dimensional algebra for all p 6= − 34 ; 0. After having found
the most general reduction for both systems we look for separable variable solutions. Several
classes of exact solutions have been found. In particular for p 6= − 34 ; 0 the solution of reduced
system is brought to the search for solutions of a class of Lienard equations, which was widely
discussed.
Finally some remarks on the behavior of invariant solutions of the system (4) and its bi-
ological compatibility are done. In particular we observed that, as expected, the boundness
of solutions, as well as the biological condition (see [14, 6]) for the existence of the mosquito
population γ − µ1 (µ2 + γ) > 0, are strictly related to the features of the per capita death rate
γ
k
− µ1 in adult phase and to the features of the per capita death rate k − µ2 in aquatic phase.
We wish to stress that the solutions found here, as well as those obtained in [4], are, until
now the only exact solutions concerned with the considered mathematical model for Aedes
aegypti dispersal dynamics. These special invariant solutions we derived here, in general, do
not satisfy arbitrary initial or boundary conditions prescribed for a given problem. However
they might be useful for a benchmark test for larger numerical schemes devised to solve our
system in a realistic case.
10
Moreover, as subject of further investigations, it will be useful to study and look for additionl
solutions of the Lienard equations (19), Abel equations (24) and Abel equations of the second
kind (29). We plan to continue in the aforesaid researches together with the search, concerned
with model (1), for additional exact solutions and additional news about the structure of its
constitutive functions.
Acknowledgements
The authors would like to thank FAPESP for financial support (grants 2011/20072-0 and
2011/19089-6). Mariano Torrisi would also like to thank CMCC-UFABC for its warm hospi-
tality and GNFM (Gruppo Nazionale per Fisica-Matematica) for its support. He also thanks
the support from University of Catania through PRA and from MIUR through PRIN: Modelli
cinetici e macroscopici per il transporto di particelle in semiconductori: aspetti modellistici,
analitici e computazionali.
We would like to thank the referees for their useful comments and suggestions which have
considerably improved the paper.
[4] I. L. Freire and M. Torrisi, Symmetry methods in mathematical modeling Aedes ae-
gypti dispersal dynamics, Nonlin. Anal. RWA, 14, 1300-1307, (2013), DOI: 10.1016/
j.nonrwa.2012.09.018.
[5] I. L. Freire and M. Torrisi, On a model for Aedes aegypti, Numerical Analysis and
Applied Mathematics ICNAAM 2012: International Conference of Numerical Analysis
and Applied Mathematics, AIP Conf. Proc., vol. 1479, 1373-1376; doi:http://dx.doi.org/
10.1063/1.4756412.
[6] N. A. Maidana, H. M. Yang, Describing the geographic spread of dengue disease by trav-
eling waves, Mathematical Biosciences, vol. 215, 64–77, (2008).
[7] M. Torrisi and R. Tracinà, Exact solutions of a reaction-difusion system for Proteus
Mirabilis bacterial colonies, Nonlin. Anal.: Real World Appl., vol. 12, 1865–1874, (2011).
11
[9] P. J. Olver, Applications of Lie groups to differential equations, Springer, New York,
(1986).
[10] G. W. Bluman and S. Kumei, Symmetries and Differential Equations, Applied Mathemat-
ical Sciences 81, Springer, New York, (1989).
[12] N. H. Ibragimov, CRC Handbook of Lie group analysis of differential equations, vol. 1,
CRC Press, (1994).
[13] A. D. Polyanin and V. F. Zaitsev, Handbook of Exact Solutions for Ordinary Differential
Equations, 2nd Edition , Chapman & Hall/CRC, Boca Raton, 2003.
[14] N.A. Maidana, H.M. Yang, A Spatial Model to Describe the Dengue Propagation, TEMA
Tend. Mat. Apl. Comput., 8, No. 1 (2007), 83-92.
12