A mathematical model of a vector-borne disease involving variable human population is analyzed. The varying population size includes a term for disease-related deaths. Equilibria and stability are determined for the system of ordinary differential equations. If R0≤1, the disease-“free” equilibrium is globally asymptotically stable and the disease always dies out. If R0>1, a unique “endemic” equilibrium is globally asymptotically stable in the interior of feasible region and the disease persists at the “endemic” level. Our theoretical results are sustained by numerical simulations.
1. Introduction
Vector-borne diseases such as malaria, dengue fever, plague, and West Nile fever are infectious diseases caused by the influx of viruses, bacteria, protozoa, or rickettsia which are primarily transmitted by disease-transmitting biological agents, called vectors. A vector-borne disease is transmitted by a pathogenic microorganism from an infected host to another organism results to form from an infection by blood-feeding arthropods [1].
Vector-borne diseases, in particular, mosquito-borne disease, are transmitted to humans by blood-sucker mosquitoes, which have been a big problem for the public health in the world. The literature dealing with the mathematical theory on vector-borne diseases is quite extensive. Many mathematical models concerning the emergence and reemergence of the vector-host infectious disease have been proposed and analyzed in the literature [2, 3].
By direct transmission models, we mean that the infection moves from person to person directly, with no environmental source, intermediate vector, or host. In a vector-host model, direct transmission may take place by transfusion-related transmission, transplantation-related transmission, and needle-stick-related transmission [4]. Some models have been developed to study the dynamics of a vector-borne disease that considers a direct mode of transmission in human host population [5–7].
Mathematical modeling has proven to play an important role in gaining some insights into the transmission dynamics of infectious diseases and suggest control strategies. Appropriate mathematical models can provide a qualitative assessment for the problem. Some mathematical models discussed in [8–10] provide, best understanding about the dynamics and control of infectious diseases. Immense literature on the use of mathematical models for communicable diseases is available [11, 12]. The assumption of constant population size in epidemiological models is usually valid when we study the diseases of short duration with limited effects on mortality. It may not be valid when dealing with endemic diseases such as malaria, which has a high mortality rate. Ngwa and Shu [13] assumed density-dependent death rates in both vector and human populations, so that the total populations are varying with time that includes disease-related deaths. Esteva and Vargas [14] analyzed the effect of variable host population size and disease-induced death rate. Recently Ozair et. al analyzed vector-host disease model with nonlinear incidence [15].
In this paper, based on the ideas posed in [6, 14], we develop and analyze a vector-host disease model considering a direct mode of transmission as well as a variable human population. The aim of this paper is to establish stability properties of equilibria and the threshold parameter R0 that completely determines the existence of endemic or disease-free equilibrium. If R0⩽1, the disease-free equilibrium is globally asymptotically stable. If R0>1, a unique endemic equilibrium exists and is globally asymptotically stable under parametric restrictions. However, in numerical simulations it is shown that the disease still can be “endemic” even if the conditions are violated.
The rest of the paper is organized as follows. In Section 2, we present a formulation of the extended mathematical model. The dimensionless formulation of proposed model is carried out in Section 3. Section 4 devotes existence and uniqueness of “endemic” equilibria. In Section 5, we use Lyapunov function theory to show global stability of disease-“free" equilibrium and geometric approach to prove global stability of “endemic” equilibrium. Discussions and simulations are done in Section 6.
2. Model Formulation
The human population is partitioned into subclasses of individuals who are susceptible, infectious, and recovered, with sizes denoted by Sh(t), Ih(t), and Rh(t), respectively. The vector population is subdivided into susceptible and infectious vectors, with sizes denoted by Sv(t) and Iv(t), respectively. The mosquito population does not have an immune class, since their infective period ends with their death. Thus, Nh(t)=Sh(t)+Ih(t)+Rh(t) and Nv(t)=Sv(t)+Iv(t) are, respectively, the total human and vector populations at time t. The model is given by the following system of differential equations:
(1)dSh(t)dt=b1Nh-β1ShIhNh-β2ShIvNv-μhSh,dIh(t)dt=β1ShIhNh+β2ShIvNv-μhIh-γhIh-δhIh,dRh(t)dt=γhIh-μhRh,dSv(t)dt=μvNv-β3SvIhNh-μvSv,dIv(t)dt=β3SvIhNh-μvIv.
In model (1), b1 is the recruitment rate of humans into the population which is assumed to be susceptible. Susceptible hosts get infected via two routes of transmission, through a contact with an infected individual and through being bitten by an infectious vector. We denote the infection rate of susceptible individuals which results from effective contact with infectious individuals by β1 and β2 is the infection rate of susceptible humans resulting due to the biting of infected vectors. The incidence of new infections via direct and indirect route of transmission is given by the standard incidence form β1(ShIh/Nh) and β2(ShIv/Nh), respectively. The term μh is the natural mortality rate of humans. We assume that infectious individuals acquire permanent immunity by the rate γh. The infectious humans suffer from disease-induced death at a rate δh. The recruitment and natural death rate of vector population is assumed to be μv. The susceptible vectors become infectious as a result of biting effect of infectious humans at a rate β3, so that the incidence of newly infected vectors is again given by standard incidence form β3(SvIh/Nh). The total human population is governed by the following equation:
(2)dNhdt=b1Nh-μhNh-δhIh.
3. Dimensionless Formulation
Denote sh=Sh/Nh, ih=Ih/Nh, rh=Rh/Nh, sv=Sv/Nv, and iv=Iv/Nv. It is easy to verify that sh, ih, rh, sv, and iv satisfy the following system (see the Appendix details for):
(3)dsh(t)dt=b1(1-sh)-β1shih-β2shiv+δhshih,dih(t)dt=β1shih+β2shiv-(b1+γh+δh)ih+δhih2,drh(t)dt=γhih-b1rh+δhihrh,dsv(t)dt=μv(1-sv)-β3svih,div(t)dt=β3svih-μviv,
where solutions are restricted to sh+ih+rh=1 and sv+iv=1. Before analyzing the unnormalized model (1) and (2), we consider the normalized model (3) by scaling, and so we can study the following reduced system that describes the dynamics of the proportion of individuals in each class
(4)dsh(t)dt=b1(1-sh)-β1shih-β2shiv+δhshih,dih(t)dt=β1shih+β2shiv-(b1+γh+δh)ih+δhih2,div(t)dt=β3(1-iv)ih-μviv,
determining rh from
(5)drh(t)dt=γhih-b1rh+δhihrh,
or from rh=1-sh-ih and sv from sv=1-iv, respectively. The correlation between normalized and unnormalized models is explained in the Appendix. Throughout this work, we study the reduced system (4) in the closed, positively invariant set Γ={(sh,ih,iv)∈R+3,0≤sh+ih≤1,0≤iv≤1}, where R+3 denotes the nonnegative cone of R3 with its lower dimensional faces.
4. Existence of Equilibria
We seek the conditions for the existence and stability of the disease-“free” equilibrium (DFE) E0(sh0,0,0) and the “endemic” proportion equilibrium E*(sh*,ih*,iv*). Obviously, E0(1,0,0)∈Γ is the DFE of (4), which exists for all positive parameters. The Jacobian matrix of (4) at an arbitrary point E(sh,ih,iv) takes the following form:(6)J(E)=(-b1-β1ih-β2iv+δhih-(β1-δh)sh-β2shβ1ih+β2ivβ1sh-(b1+γh+δh)+2δhihβ2sh0β3(1-iv)-β3ih-μv).To analyze the stability of DFE, we calculate the characteristic equation of J(E) at E=E0 as follows:
(7)(λ+b1)(λ2+λ(μv+b1+γh+δh-β1)+μv(b1+γh+δh)(1-R0)λ2),
where
(8)R0=β1b1+γh+δh+β2β3μv(b1+γh+δh).
By Routh Hurwitz criteria [16], all roots of (7) have negative real parts if and only if R0<1. So, E0 is locally asymptotically stable for R0<1. If R0>1, the characteristic equation (7) has positive eigenvalue, and E0 is thus unstable. We established the following theorem.
Theorem 1.
The disease-free equilibrium is locally asymptotically stable whenever R0<1 and unstable for R0>1.
In order to find the “endemic” equilibrium of (4), we set the right hand side of (4) equal to zero and get
(9)iv*=β3ih*μv+β3ih*,sh*=b1(μv+β3ih*)(μv+β3ih*)(b1+(β1-δh)ih*)+β2β3ih*,
where ih* is a positive solution of the equation
(10)f(ih*)=A3ih*3+A2ih*2+A1ih*+A0=0,
where
(11)A3=β3δh(β1-δh),A2=β2β3δh+(μvδh-β3(b1+γh+δh))(β1-δh)+b1β3δh,A1=b1(μvδh+β1β3)-(b1+γh+δh)(β2β3+b1β3+μv(β1-δh)),A0=(b1+γh+δh)b1μv(R0-1).
From right hand side of (5), we have γhih*=(b1-δhih*)rh*>0 and second equation of (9) β1μv+β2β3-μvδh>β3δhih*, which means that
(12)0<ih*<min{1,b1δh,(β1μv+β2β3μvδh-1)μvβ3}.
If (β1μv+β2β3)/μvδh≤1, there is no positive ih*, and therefore the only equilibrium point in Γ is E0. Note that this is a special case of R0<1.
Assume that R0>1.
If β1>δh, then A3>0, we have f(-∞)<0, f(∞)>0 and f(0)=A0>0. Further,
f(1)<0(ifb1/δh≥1) and f(b1/δh)<0. Thus, there exists unique ih* such that f(ih*)=0 (see Figure 1).
If β1=δh, then A3=0 and f(ih*)=A2ih*2+A1ih*+A0, where A2=β2β3δh+b1β3δh>0. We observe that
f(-∞)>0, f(∞)>0 and f(0)=A0>0. Moreover, f(1)<0(ifb1/δh≥1) and f(b1/δh)<0. Therefore, there exists unique ih* such that f(ih*)=0 (see Figure 2).
If β1<δh, then A3<0, we have f(-∞)>0, f(∞)<0 and still f(0)=A0>0, f(1)<0(ifb1/δh≥1), f(b1/δh)<0. In this case, we can say that there is only one root or three roots in the interval (0,1)ifb1/δh≥1 or (0,b1/δh)ifb1/δh<1.
We know that f(ih*)=0 has three real roots if and only if
(13)q24+p327≤0,
where
(14)p=A1A3-(A2)23(A3)2,q=A0A3-A1A23(A3)2+2(A2)327(A3)3,
or
(15)R^1=18A0A1A2A3-4A0(A2)3-4(A1)3A3+(A1)2(A2)227(A0)2(A3)2≥1.
β1>δh.
β1=δh.
If R^1<1, there is unique ih* such that f(ih*)=0 in the feasible interval.
If R^1>1, there are three different real roots for f(ih*)=0 say ih1*,ih2*,ih3*(ih1*<ih2*<ih3*). Note that, differentiating with respect to ih*, we obtain
(16)f′(ih*)=3A3ih*2+2A2ih*+A1.
The three different real roots for f(ih*)=0 are in the feasible interval if and only if the following inequalities are satisfied:
(17)0<-A23A3<1,f′(0)=A1<0,f′(1)=3A3+2A2+A1<0(ifb1δh≥1),f′(b1δh)=3A3(b1δh)2+2A2(b1δh)+A1<0(ifb1δh<1).
If R^1=1, there are three real roots for f(ih*)=0, in which at least two are identical. Similarly, if inequalities (17) are satisfied, then there are three real roots for f(ih*)=0 in the feasible interval, say ih1*,ih2*,ih3*(ih1*=ih2*).
Assume that R0=1.
If β1=δh, then A3=0 and (10) reduces to ih*(A2ih*+A1)=0, which implies that ih*=0 or ih*=-A1/A2, which is positive but it lies outside the interval (0,1)ifb1/δh≥1 or (0,b1/δh)ifb1/δh<1.
If β1>δh, then A3>0, we have ih*(A3ih*2+A2ih*+A1)=0, which implies that ih*=0 or ih* is the solution of the equation
(18)g(ih*)=A3ih*2+A2ih*+A1=0,
where g(-∞)>0, g(∞)>0, and g(0)=A1<0. Moreover, g(1)<0(ifb1/δh≥1) and g(b1/δh)<0ifb1/δh<1. Therefore, there exists no ih* such that g(ih*)=0 in the interval (0,1)ifb1/δh≥1 or (0,b1/δh)ifb1/δh<1.
In summary, regarding the existence and the number of the “endemic” equilibria, we have the following.
Theorem 2.
Suppose that β1≥δh. There is always a disease-“free” equilibrium for system (4); if R0>1, then there is a unique “endemic” equilibrium E*(sh*,ih*,iv*) with coordinates satisfying (9) and (10) besides the disease-“free” equilibrium.
5. Global Dynamics5.1. Global Stability of the Disease-“Free” Equilibrium
In this subsection, we analyze the global behavior of the equilibria for system (4). The following theorem provides the global property of the disease-free equilibrium E0 of the system.
Theorem 3.
If R0≤1, then the infection-free equilibrium E0 is globally asymptotically stable in the interior of Γ.
Proof.
To establish the global stability of the disease-free equilibrium, we construct the following Lyapunov function:
(19)L(t)=μvih(t)+β2iv(t).
Calculating the time derivative of L along (4), we obtain
(20)L′(t)=μvih′(t)+β2iv′(t)=μv[δhih2β1shih+β2shiv-(b1+γh+δh)ih+δhih2]+β2(β3svih-μviv)=μv[ih2δhih2β1(1-ih)ih+β2(1-ih)iv-(b1+γh+δh)ih+δhih2]+β2[β3(1-iv)ih-μviv]=μv[β1ih-β1ih2+β2iv-β2ihiv-(b1+γh+δh)ih+δhih2]+β2(β3ih-β3ivih-μviv)=μvβ1ih-μvβ1ih2+μvβ2iv-μvβ2ihiv-μv(b1+γh+δh)ih+μvδhih2+β2β3ih-β2β3ivih-β2μviv=-μv(b1+γh+δh)(1-R0)ih-μv(β1-δh)ih2-μvβ2ihiv-β2β3ivih.
Thus, L′(t) is negative if R0≤1 and L′=0 if and only if ih=0. Consequently, the largest compact invariant set in {(Sh,Ih,Iv)∈Γ,L′=0}, when R0≤1, is the singelton {E0}. Hence, LaSalle’s invariance principle [16] implies that “E0" is globally asymptotically stable in Γ. This completes the proof.
5.2. Global Stability of “Endemic” Equilibrium
Here, we use the geometrical approach of Li and Muldowney to investigate the global stability of the endemic equilibrium E* in the feasible region Γ. We have omitted the detailed introduction of this approach, and we refer the interested readers to see [17]. We summarize this approach below.
Consider a C1 map f:x↦f(x) from an open set D⊂Rn to Rn such that each solution x(t,x0) to the differential equation
(21)x′=f(x).
is uniquely determined by the initial value x(0,x0). We have the following assumptions:
(H1)D is simply connected;
(H2) there exists a compact absorbing set K⊂D;
(H3) (21) has unique equilibrium x- in D.
Let P:x↦P(x) be a nonsingular (n2)×(n2) matrix-valued function which is C1 in D and a vector norm |·| on RN, where N=(n2).
Let μ be the Lozinskiĭ measure with respect to the |·|. Define a quantity q-2 as
(22)q-2=limsupt→∞supx0∈K1t∫0tμ(B(x(s,x0)))ds,
where B=PfP-1+PJ[2]P-1, the matrix Pf is obtained by replacing each entry p of P by its derivative in the direction of f, (pij)f, and J[2] is the second additive compound matrix of the Jacobian matrix J of (21). The following result has been established in Li and Muldowney [17].
Theorem 4.
Suppose that (H1), (H2), and (H3) hold, the unique endemic equilibrium E* is globally stable in Γ if q-2<0.
Obviously Γ is simply connected and E* is a unique endemic equilibrium for R0>1 in Γ. To apply the result of the above theorem for global stability of endemic equilibrium E*, we first prove the uniform persistence of (4) when the threshold parameter R0>1, by applying the acyclicity Theorem (see [18]).
Definition 5 (see [19]).
The system (4) is uniformly persistent, that is, there exists c>0 (independent of initial conditions), such that liminft→∞sh≥c,liminft→∞ih≥c,liminft→∞iv≥c.
Let X be a locally compact metric space with metric d and let Γ be a closed nonempty subset of X with boundary Γ and interior Γ∘. Clearly, Γ∘ is a closed subset of Γ. Let Φt be a dynamical system defined on Γ. A set B in X is said to be invariant if Φ(B,t) = B. Define M∂:={x∈Γ:Φ(t,x)∈Γ,forallt≥0}.
Lemma 6 (see [18]).
Assume that
Φt has a global attractor;
there exists M={M1,…,Mk} of pair-wise disjoint, compact and isolated invariant set on ∂Γ such that
⋃x∈M∂ω(x)⊆⋃j=1kMj;
no subsets of M form a cycle on ∂Γ;
each Mj is also isolated in Γ;
Ws(Mj)⋂Γ∘=ϕ for each 1≤j≤k, where Ws(Mj) is stable manifold of Mj. Then Φt is uniformly persistent with respect to Γ.
Proof.
We have Γ={(sh,ih,iv)∈R+3,0≤sh+ih≤1,0≤iv≤1}, Γ∘={(sh,ih,iv)∈R+3sh,ih>0}, ∂Γ=Γ/Γ∘. Obviously M∂=∂Γ. Since Γ is bounded and positively invariant, so there exists a compact set M in which all solutions of system (4) initiated in Γ ultimately enter and remain forever. On sh-axis we have sh′=b1(1-sh) which means sh→1 as t→∞. Thus, E0 is the only omega limit point on ∂Γ, that is, ω(x)=E0 for all x∈M∂. Furthermore, M=E0 is a covering of Ω=⋃x∈M∂ω(x), because all solutions initiated on the sh-axis converge to E0. Also E0 is isolated and acyclic. This verifies that hypothesis (1) and (2) hold. When R0>1, the disease-“free” equilibrium (DFE) E0 is unstable from Theorem 1 and also Ws(M)=∂Γ. Hypothesis (3) and (4) hold. Therefore, there always exists a global attractor due to ultimate boundedness of solutions.
The boundedness of Γ and the above lemma imply that (4) has a compact absorbing set K⊂Γ [19]. Now we shall prove that the quantity q-2<0. We choose a suitable vector norm |·| in R3 and a 3×3 matrix-valued function
(23)P(x)=(1000ihiv000ihiv).
Obviously, P is C1 and nonsingular in the interior of Ω. Linearizing system (4) about an endemic equilibrium E* gives the following Jacobian matrix:(24)J(E*)=(-b1sh-(β1-δh)sh-β2shβ1ih+β2ivβ1sh-(b1+γh+δh)+2δhihβ2sh0β3(1-iv)-β3ih-μv).The second additive compound matrix of J(E*) is given by
(25)J[2]=(M11β2shβ2shβ3(1-iv)M22-(β1-δh)sh0β1ih+β2ivM33),
where
(26)M11=-b1sh+β1sh-(b1+γh+δh)+2δhih,M22=-b1sh-β3ih-μv,M33=β1sh-(b1+γh+δh)+2δhih-β3ih-μv.
The matrix B=PfP-1+PJ[2]P-1 can be written in block form as
(27)B=(B11B12B21B22),
with
(28)B11=-b1sh+β1sh-(b1+γh+δh)+2δhih,B12=(β2shivih,β2shivih),B21=((ihiv)β3(1-iv)0),B22=(Q11Q12Q21Q22),
where
(29)Q11=-b1sh-β3ih-μv,Q12=-(β1-δh)sh,Q21=β1ih+β2iv,Q22=β1sh-(b1+γh+δh)+2δhih-β3ih-μv,ivih(ihiv)f=ih′ih-iv′iv.
Consider the norm in R3 as |(u,v,w)|=max(|u|,|v|+|w|) where (u,v,w) denotes the vector in R3. The Lozinskiĭ, measure with respect to this norm is defined as μ(B)≤sup(g1,g2), where
(30)g1=μ1(B11)+|B12|,g2=μ1(B22)+|B21|.
From system (4), we can write
(31)ih′ih=β1sh+β2shivih-(b1+γh+δh)+δhih,iv′iv=β3(1-iv)ihiv-μv.
Since B11 is a scalar, its Lozinskiĭ measure with respect to any vector norm in R1 will be equal to B11. Thus
(32)B11=-b1sh+β1sh-(b1+γh+δh)+2δhih,|B12|=β2shivih,
and g1 will become
(33)g1=-b1sh+β1sh-(b1+γh+δh)+2δhih+β2shivih=ih′ih-b1sh+δhih≤ih′ih-b1+δhih.
Also |B21|=(ih/iv)β3(1-iv), |B12| and |B21| are the operator norms of B12 and B21 which are mapping from R2, to R and from R to R2 respectively, and R2 is endowed with the l1 norm. μ1(B22) is the Lozinskiĭ measure of 2×2 matrix B22 with respect to l1 norm in R2. (34)μ(B22)=Sup{ivih(ihiv)f-b1sh-β3ih-μv+β1ih+β2iv,ivih(ihiv)f+(β1-δh)sh+β1sh-(b1+γh+δh)+2δh-β3ih-μvihivih}≤ih′ih-iv′iv-b1+δhih-β3ih-μv,
if β1≤γh/2. Hence
(35)g2≤ih′ih-iv′iv-b1+δhih-β3ih-μv+(ihiv)β3(1-iv)=ih′ih-b1+δhih-β3ih.
Thus,
(36)μ(B)=Sup{g1,g2}≤ih′ih-b1+δh≤ih′ih-β-1,
where β-1=min(γh/2,b1/2). Since (4) is uniformly persistent when R0>1, so for T>0 such that t>T implies ih(t)≥c, iv(t)≥c and (1/t)logih(t)<(β-1/2) for all (sh(0),ih(0),iv(0))∈K. Thus,
(37)1t∫0tμ(B)dt<logih(t)t-β-1<-β-12,
for all (sh(0),ih(0),iv(0))∈K, which further implies that q-2<0. Therefore, all the conditions of Theorem 4 are satisfied. Hence, unique endemic equilibrium E* is globally stable in Γ.
6. Discussions and Simulations
This paper deals with a vector-host disease model which allows a direct mode of transmission and varying human population. It concerns diseases with long duration and substantial mortality rate (e.g., malaria). Our main results are concerned with the global dynamics of transformed proportionate system. We have constructed Lyapunov function to show the global stability of disease-“free” equilibrium and the geometric approach is used to prove the global stability of “endemic” equilibrium. The epidemiological correlations between the two systems (normalized and unnormalized) have also been discussed. The dynamical behavior of the proportionate model is determined by the basic reproduction number of the disease. The model has a globally asymptotically stable disease-“free” equilibrium whenever R0≤1 (Figures 3 and 4). When R0>1, the disease persists at an “endemic” level (Figures 5 and 6) if β1<min(b1/2,γh/2). Figures 7, 8, 9, and 10 describe numerically “endemic” level of infectious individuals and infectious vectors under the condition β1<min(b1/2,γh/2). We here question that what are the dynamics of the proportionate system (4) even if the condition β1<min(b1/2,γh/2) is not satisfied? We see numerically that if δh,γh/2<β1<b1/2 or γh/2<β1=δh<b1/2 then infectious individuals and infectious vectors will also approach to endemic level for different initial conditions (Figures 11, 12, 13, and 14). It is also numerically shown that the same is true for the case δh,b1/2<β1<γh/2 or b1/2<β1=δh<γh/2 (Figures 15, 16, 17, and 18). This implies that the condition β1<min(b1/2,γh/2) is weak for the global stability of unique “endemic” equilibrium.
Using the transformation sh=Sh/Nh, ih=Ih/Nh, rh=Rh/Nh, sv=Sv/Nv, and iv=Iv/Nv for scaling, their differentials: dsh(t)/dt=(1/Nh)(dSh/dt)-(Sh/Nh2)(dNh/dt), dih(t)/dt=(1/Nh)(dIh/dt)-(Ih/Nh2)(dNh/dt), drh(t)/dt=(1/Nh)(dRh/dt)-(Rh/Nh2)(dNh/dt), dsv(t)/dt=(1/Nv)(dSv/dt)-(Sv/Nv2)(dNv/dt), and div(t)/dt=(1/Nv)(div/dt)-(iv/Nv2)(dNv/dt), and the system (1) and (2), we obtain the dimensionless form (3). If δh=0 and b1=μh, then Nh(t)′=0 and so Nh(t) remains fixed at its initial value Nh0. In this case, the system (1) becomes the model with constant population whose dynamics are the same as the proportionate system (3). Hence, the solutions with initial conditions Sh0+Ih0+Rh0=Nh0 tend to (Nh0,0,0) if R0≤1 and to Nh0(sh*,ih*,rh*) if R0>1. In the rest of this section, we suppose that δh>0. From system (1) and (2), the trivial equilibrium E=(0,0,0,0,0) can be easily obtained. Assume that E*=(Nh*,Sh*,Ih*,Rh*,Iv*) is the endemic equilibrium of system (1) and (2), where Nh*=Sh*+Ih*+Rh*. This equilibrium exists if and only if the following equations are satisfied
(A.1)Sh*Nh*=Q(β3αh+μvδh)β1(β3αh+μvδh)+β2β3δh,Ih*Nh*=αhδh,Rh*Nh*=γhαhμhδh,Iv*Nh*=β3αhNv(β3αh+μv)Nh*,
where αh=b1-μh and Q=μh+γh+δh. We introduce the parameters
(A.2)R1={b1μh,ifR0≤1b1μh+δhih*,ifR0>1,R2={β1μh+γh+δh+β2β3μv(μh+γh+δh),ifR0≤1β1sh*μh+γh+δh+β2β3sh*(1-iv*)μv(μh+γh+δh),ifR0>1.
From (2) we have for t→∞(A.3)dNhdt=Nh(b1-μh-δhih)⟶{Nh(b1-μh),ifR0≤1Nh(b1-μh-δhih*),ifR0>1.
By the definition of R1, we have following threshold result.
Theorem A.1.
The total population Nh(t) for the system (1) decreases to zero if R1<1 and increases to ∞ if R1>1 as t→∞. The asymptotic rate of decrease is μh(R1-1) if R0≤1, and the asymptotic rate of increase is (μh+δhih*)(R1-1) when R0>1.
Theorem A.2.
Suppose R1>1, for t→∞, (Sh(t),Ih(t),Rh(t)) tend to (∞,0,0) if R2<1 and tend to (∞,∞,∞) if R2>1.
Proof.
Since iv′→0 as t→∞, so in the limiting case the proportion of infectious mosquitoes is related to the proportion of infectious humans as
(A.4)iv=β3(1-iv)ihμv,
thus, the equation for Ih(t) has limiting form
(A.5)dIh(t)dt=(μh+γh+δh)(R2-1)Ih,
which shows that Ih(t) decreases exponentially if R2<1 and increases exponentially if R2>1.
The solution Rh(t) is given by
(A.6)Rh(t)=Rh0e-μht+γhe-μht∫0tIh(s)eμhsds.
From the exponential nature of Ih(t), it follows that Ih(t) declines exponentially if R2<1 and grows exponentially if R2>1.
Suppose R1=1, then b1=μh corresponding to R0<1 and the differential equation for Nh(t) will have the form
(A.7)dNhdt=-δhIh,
which means that Nh(t) is bounded for all t>0, the equilibria (Nh*,0,0,0) have one eigenvalue zero, and the other eigenvalues have negative real parts. Therefore, each orbit approaches an equilibrium point.
If R0>1, the disease becomes “endemic.” From the global stability of E* and the equation(A.8)dNhdt=δh[(b1-μhδh-ih*)-(ih-ih*)]Nh,
we observe that (Nh,Sh,Ih,Rh,Iv) approaches to (0,0,0,0,0) or (∞,∞,∞,∞,∞) if R1<1 or R1>1. From the global stability of ih*, we have Nh(t) converges to some Nh* as t approaches to ∞. Since sh=Sh/Nh,ih=Ih/Nh,rh=Rh/Nh, so we have Sh*=sh*Nh*,Ih*=ih*Nh*, Rh*=rh*Nh*. All the above discussion is summarized in Table 1.
Asymptotic behavior with threshold criteria.
R0
R1
R2
Nh
(sh,ih,rh,iv)→
(Sh,Ih,Rh)→
≤1
=1, δh=0
≤1
Nh=Nh0
(1,0,0,0)
(Nh0,0,0)
>1
=1, δh=0
=1
Nh=Nh0
(sh*,ih*,rh*,iv*)
Nh0(sh*,ih*,rh*)
≤1
<1
<1
Nh→0
(1,0,0,0)
(0,0,0)
>1
<1
<1
Nh→0
(sh*,ih*,rh*,iv*)
(0,0,0)
≤1
>1
<1
Nh→∞
(1,0,0,0)
(∞,0,0)
≤1
>1
>1
Nh→∞
(1,0,0,0)
(∞,∞,∞)
<1
=1
<1
Nh→Nh*
(1,0,0,0)
(Nh*,0,0)
>1
>1
>1
Nh→∞
(sh*,ih*,rh*,iv*)
(∞,∞,∞)
>1
=1
=1
Nh→Nh*
(sh*,ih*,rh*,iv*)
(sh*,ih*,rh*)
Acknowledgment
This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (MEST) (2012-000599).
YangH.WeiH.LiX.Global stability of an epidemic model for vector-borne disease201023227929210.1007/s11424-010-8436-7MR2653590ZBL1197.93059FengZ.Velasco-HernándezJ. X.Competitive exclusion in a vector-host model for the dengue fever199735552354410.1007/s002850050064MR1479326ZBL0878.92025QiuZ.Dynamical behavior of a vector-host epidemic model with demographic structure200856123118312910.1016/j.camwa.2008.09.002MR2474567ZBL1165.34382WiwanitkitV.Unusual mode of transmission of dengue2009305154WeiH.-M.LiX.-Z.MartchevaM.An epidemic model of a vector-borne disease with direct transmission and time delay2008342289590810.1016/j.jmaa.2007.12.058MR2445247ZBL1146.34059CaiL.LiX.Analysis of a simple vector-host epidemic model with direct transmission201020101267961310.1155/2010/679613MR2644227ZBL1190.92029LashariA. A.ZamanG.Global dynamics of vector-borne diseases with horizontal transmission in host population201161474575410.1016/j.camwa.2010.12.018MR2770478ZBL1217.34064de la SenM.AgarwalR. P.IbeasA.Alonso-QuesadaS.On the existence of equilibrium points, boundedness, oscillating behavior and positivity of a SVEIRS epidemic model under constant and impulsive vaccination2011201132748608MR2780672ZBL1219.34066de la SenM.Alonso-QuesadaS.Vaccination strategies based on feedback control techniques for a general SEIR-epidemic model201121873888390410.1016/j.amc.2011.09.036MR2851486ZBL1238.92030ZhangT.LiuJ.TengZ.Dynamic behavior for a nonautonomous SIRS epidemic model with distributed delays2009214262463110.1016/j.amc.2009.04.029MR2541699ZBL1168.92327AronJ. L.MayR. M.1982London, UKChapman & HallHethcoteH. W.The mathematics of infectious diseases200042459965310.1137/S0036144500371907MR1814049ZBL0993.92033NgwaG. A.ShuW. S.A mathematical model for endemic malaria with variable human and mosquito populations2000327-874776310.1016/S0895-7177(00)00169-2MR1792096ZBL0998.92035EstevaL.VargasC.A model for dengue disease with variable human population199938322024010.1007/s002850050147MR1684881ZBL0981.92016OzairM.LashariA. A.JungI. H.OkosunK. O.Stability analysis and optimal control of a vector-borne disease with nonlinear incidence201220122110.1155/2012/595487LaSalleJ. P.1976Philadelphia, Pa, USASIAMv+76Regional Conference Series in Applied MathematicsMR0481301LiM. Y.MuldowneyJ. S.A geometric approach to global-stability problems19962741070108310.1137/S0036141094266449MR1393426ZBL0873.34041ZhaoX.-Q.200316New York, NY, USASpringerxiv+276CMS Books in MathematicsMR1980821ButlerG.FreedmanH. I.WaltmanP.Uniformly persistent systems198696342543010.2307/2046588MR822433ZBL0603.34043