Home Mathematics Non-local diffusion and pulse intervention in a faecal-oral model with moving infected fronts
Article Open Access

Non-local diffusion and pulse intervention in a faecal-oral model with moving infected fronts

  • Qi Zhou , Michael Pedersen and Zhigui Lin EMAIL logo
Published/Copyright: September 18, 2025

Abstract

How individual dispersal patterns and human intervention behaviours affect the spread of infectious diseases constitutes a central problem in epidemiological research. By incorporating these two critical factors, this article proposes an impulsive non-local faecal-oral model with free boundaries. In this framework, non-local diffusion captures the long-range movement of individuals, impulsive effects model periodic disinfection measures, and free boundaries delineate the advancing fronts of infection. We first establish the existence and uniqueness of a non-negative global classical solution to the system. Subsequently, by employing the theory of resolvent positive operators and their perturbations, we analyze the principal eigenvalue, which depends on the infected domain, impulse intensity, and non-local diffusion kernel functions. On the basis of this eigenvalue, we derive sharp threshold criteria distinguishing between disease extinction and propagation. The incorporation of non-local diffusion and impulsive interventions introduces significant mathematical complexities, which are addressed through novel analytical techniques. Finally, numerical simulations are conducted to validate the theoretical results and explore the impact of pulsed interventions. Our findings demonstrate that periodic disinfection effectively mitigates disease transmission, whereas the influence of non-local diffusion varies with the selection of kernel functions.

MSC 2010: 35R12; 35R35; 92B05

1 Introduction

Human beings have always been more seriously threatened by diseases than by wars, poverty, disasters, etc. It is estimated that there are about 1.3–4 million cases of cholera globally each year, causing about 21,000 to 143,000 deaths [2]. Typhoid disease is estimated to cause between 9.9 and 14.7 million infection cases and between 75,000 and 208,000 deaths annually [27]. Global adult mortality rates significantly increased during the COVID-19 pandemic in 2020 and 2021, reversing past decreasing trends [18]. A common feature of the diseases mentioned earlier is that they can be spread by the faecal-oral route [7,11,19]. In addition, poliomyelitis, infectious hepatitis, hand-foot-mouth diseases, and enteric viruses like norovirus, rotavirus, and astrovirus can also be transmitted by the faecal-oral route [7,11,17].

Mathematical models have been widely used to predict the trends, design the control measures, and understand the underlying mechanisms of infectious diseases [39]. The first faecal-oral model has the following form:

(1.1) u ( t ) = a 11 u + a 12 v , t > 0 , v ( t ) = a 22 v + G ( u ) , t > 0

with suitable initial conditions [8]. Model (1.1), formulated by Capasso and Paveri-Fontana, is used to understand how cholera was transmitted in the Mediterranean regions of Europe in 1973. In (1.1), u ( t ) denotes the average density of the infectious agents (bacteria, virus, etc.) at time t , while v ( t ) denotes the average density of the infected individuals at time t . The constants a 11 , a 12 , and a 22 stand for the natural mortality of the infectious agents, the growing rate of the infectious agents stemming from the infected individuals, and the death rate of the infected individuals, respectively. All these constants are positive. G ( u ) stands for the infective rate of humans, and satisfies

(G) : G C 1 ( [ 0 , ) ) , G ( 0 ) = 0 and G ( u ) > 0 for u [ 0 , ) , G ( u ) u is strongly decreasing for u > 0 and lim u + G ( u ) u < a 11 a 22 a 12 .

Taking into account the spatial movement of individuals, Capasso and Maddalena further developed the corresponding reaction–diffusion model in 1981 [7]. Subsequently, there have been a series of reaction–diffusion faecal-oral models based on the work [7]. We refer to [6] for an integro-differential model, [31] for an integro-differential model with a time delay, [36] for an model with two time delays, and [20] for a model with general growing function.

However, the infected region of infectious disease always moves gradually outwards, but the models mentioned earlier do not capture this. In 2010, Du and Lin introduced the free boundary to study the invasion of a single species [12]. Motivated by this work, Ahn et al. developed a diffusive faecal-oral model with free boundaries, where the free boundaries successfully capture the movement of the infected region [1]. The model in [1] does not take into account the movement of the infected humans. Considering the movements of the infected humans and diseases, Wang and Du [32] developed the following model:

(1.2) u t = d 1 Δ u a 11 u + a 12 v , x ( g ( t ) , h ( t ) ) , t > 0 , v t = d 2 Δ v a 22 v + G ( u ) , x ( g ( t ) , h ( t ) ) , t > 0 , u = 0 , v = 0 , x { g ( t ) , h ( t ) } , t > 0 , g ( t ) = μ 1 u x ( g ( t ) , t ) μ 2 v x ( g ( t ) , t ) , t > 0 , h ( t ) = μ 1 u x ( h ( t ) , t ) μ 2 v x ( h ( t ) , t ) , t > 0 , g ( 0 ) = h 0 , u ( x , 0 ) = u 0 ( x ) , x [ h 0 , h 0 ] , h ( 0 ) = h 0 , v ( x , 0 ) = v 0 ( x ) , x [ h 0 , h 0 ]

with suitable initial functions u 0 ( x ) and v 0 ( x ) . All constants in (1.2) are positive. The biological meanings of some of the symbols in (1.2) are listed in Table 1, and the rest have the same meanings as shown in (1.1).

Table 1

List of notations and their meanings in model (1.2)

Notation Biological meaning
u ( x , t ) Spatial density of the infectious agents at location x and time t
v ( x , t ) Spatial density of the infectious persons at location x and time t
g ( t ) Left boundary of the infected region at time t
h ( t ) Right boundary of the infected region at time t
d 1 Diffusion coefficient of the infectious agents
d 2 Diffusion coefficient of the infectious persons
μ 1 Expansion capacity of the infectious agents
μ 2 Expansion capacity of the infectious persons
h 0 Length of the right boundary of the initial infection region

Note that Fickian diffusion applies to a diffusion process that corresponds to the random walk only when the step size and time size are small compared with the spatial variable and time, respectively [42]. Therefore, the Laplace operator used in model (1.2) only describes the spatial properties of the infectious persons and the infectious agents locally. In 2003, Murray emphasised the importance and intuitively necessity of the long range effects in the biological areas [28, Chapter 17]. An extensively used non-local diffusion operator to replace the local diffusion term is given by

(1.3) ( J * u u ) ( x , t ) R J ( x y ) [ u ( y , t ) u ( x , t ) ] d y ,

where the kernel function J ( x ) for non-local diffusion satisfies

(J) : J C ( R ) L ( R ) , and + J ( x ) d x = 1 , J ( 0 ) > 0 , and J ( x ) = J ( x ) 0 for x R .

Through appropriately selecting J ( x ) , the non-local diffusion operator (1.3) may be used to describe both local and non-local diffusions. In 2019, Cao et al. introduced and studied a class of free boundary models with non-local diffusion in [5]. Based on [5], there has been extensive research on faecal-oral epidemic models with non-local diffusion and free boundaries, see, for instance, [9,13,25,33,34,41,43] and the references therein. In particular, Chang and Du considered a non-local version of the free boundary model (1.2) in [9], which has the following form:

(1.4) u t = d 1 g ( t ) h ( t ) J 1 ( x y ) u ( y , t ) d y d 1 u a 11 u + a 12 v , x ( g ( t ) , h ( t ) ) , t R + , v t = d 2 g ( t ) h ( t ) J 2 ( x y ) v ( y , t ) d y d 2 v a 22 v + G ( u ) , x ( g ( t ) , h ( t ) ) , t R + , g ( t ) = μ 1 g ( t ) h ( t ) g ( t ) J 1 ( x y ) u ( x , t ) d y d x μ 2 g ( t ) h ( t ) g ( t ) J 2 ( x y ) v ( x , t ) d y d x , t R + , h ( t ) = μ 1 g ( t ) h ( t ) h ( t ) + J 1 ( x y ) u ( x , t ) d y d x + μ 2 g ( t ) h ( t ) h ( t ) + J 2 ( x y ) v ( x , t ) d y d x , t R + , u ( g ( t ) , t ) = u ( h ( t ) , t ) = 0 , v ( g ( t ) , t ) = v ( h ( t ) , t ) = 0 , t R + , u ( x , 0 ) = u 0 ( x ) , v ( x , 0 ) = v 0 ( x ) , g ( 0 ) = h ( 0 ) = h 0 , x [ h 0 , h 0 ] ,

where the kernel functions J i ( i = 1 , 2) satisfies assumption (J). The biological meanings of the symbols in model (1.4) are the same as in the local version.

Human intervention can prevent and control the large-scale outbreaks of infectious diseases. For the diseases transmitted by the faecal-oral route, the infectious agents in the environment have a significant impact on the evolution of the diseases. The spraying of disinfectant liquids is effective to dramatically reduce the number of infectious agents rapidly. Classical reaction diffusion models do not capture this instantaneous and abrupt phenomenon. However, this phenomenon can be described well by impulse differential equation models.

The impulse differential models have the characteristics of both continuous and discrete models. In the last three decades, the theoretical research on impulsive differential equations has yielded numerous results. In the field of biomathematics, Lewis and Li proposed an impulsive reaction–diffusion equation model to study persistence and spread of species with a reproductive stage and a dispersal stage in bounded and unbounded domains [22]. Subsequently, some impulsive partial differential equation models rooted in biomathematical background were developed and analyzed. We refer to [14] for higher dimensional models extended from [22], [45] for an impulse epidemic model in a periodically evolving environment, [23] for a pest growth model with multiple pulse perturbations, [15] for a hybrid impulsive reaction-advection-diffusion model, [37] for a birth pulse population model with non-local dispersal, [40] for a birth pulse population model with shifting environments, and [38] for an impulsive hybrid population model in a heterogeneous landscape.

In this article, we develop a pulsed non-local faecal-oral model with free boundaries by introducing the impulsive intervention into model (1.4), to examine how impulsive intervention and non-local diffusion together affect the spread of the diseases. The developed model can be reduced to model (1.4) by choosing the impulse intervention function properly. The introduction of the impulsive intervention results in the appearance of a periodic solution and reduces the regularity of the solution. As a result, it may make analysis hard, and the results become complex and different and will naturally raise some new problems.

With the introduction of the impulse intervention, it is natural to ask what kind of regularity does the solution possess? Whether the spreading-vanishing dichotomy is still true? What are the new criteria for determining spreading and vanishing? Whether the impulsive intervention and the non-local diffusion influence the movement speed of the infected area? Whether the combination of the impulsive intervention and the non-local diffusion affects or even alters the dynamical behaviour? These are the research motivations. The main contributions of this research are listed as follows:

  • Introduce pulsed intervention into a faecal-oral model with free boundaries and non-local diffusion.

  • Investigate the principal eigenvalue, which depends on the impulse intensity, the infected region, and the kernel functions.

  • Prove a vanishing-spreading dichotomy, and provide the sufficient conditions for determining vanishing and spreading.

The article is arranged as follows. Section 2 develops an impulsive faecal-oral model with free boundaries and non-local diffusion and presents some preliminaries including necessary notations, a frequently used comparison principle, a prior estimate, and the well-posedness of the solution. The related impulse problem with fixed boundaries and the corresponding eigenvalue problem with impulse are treated in Section 3. Here, the existence of the principal eigenvalue is obtained by applying the theory of resolvent positive operators with their perturbations. Section 4 proves a vanishing-spreading dichotomy and obtains the sufficient conditions for determining vanishing and spreading based on the results obtained in Section 3. In Section 5, numerical simulation is used to present the effect of the impulsive intervention and to corroborate the validity of the theoretical findings. Finally, a brief discussion and conclusion are given in Section 6.

2 Model formulation and preliminaries

In this section, we formulate a pulsed non-local faecal-oral model in a moving infected region, the global existence of its unique non-negative classical solution is then given, and finally a frequently used comparison principle is provided.

2.1 Model formulation

As stated in Section 1, the transmission of faecal-oral diseases is usually suppressed by human control measures. As an example, periodic disinfection of the infected area plays an important role in cutting off the transmission route and protecting the susceptible population. This kind of measure generally results in a dramatic decrease in the spatial density of the infectious agents within a very small time span, and there is no direct impact on the infected individuals.

Motivated by this biological phenomenon, we consider here a situation with two stages: a human intervention stage and a natural developmental stage. In a human intervention stage, the spatial density of the infectious agents decreases impulsively through a discrete-time map. We use H to describe the spatial density at the end of a human intervention stage, which is a function of the density at the beginning of the stage. The spatial density of the infected individuals remains unchanged during this stage. In a natural developmental stage, the evolution of the diseases is described by model (1.4).

Let τ be the time between two adjacent impulses. We always take k = 0 , 1 , 2 , unless otherwise specified. Denote the start and end times of the pulse by k τ and ( k τ ) + , respectively. Then, a mathematical model capturing the spatial dynamics of the diseases transmitted by faecal-oral route under impulsive intervention is given here:

(2.1) u t = d 1 g ( t ) h ( t ) J 1 ( x y ) u ( y , t ) d y d 1 u a 11 u + a 12 v , x ( g ( t ) , h ( t ) ) , t ( ( k τ ) + , ( k + 1 ) τ ] , v t = d 2 g ( t ) h ( t ) J 2 ( x y ) v ( y , t ) d y d 2 v a 22 v + G ( u ) , x ( g ( t ) , h ( t ) ) , t ( ( k τ ) + , ( k + 1 ) τ ] , g ( t ) = μ 1 g ( t ) h ( t ) g ( t ) J 1 ( x y ) u ( x , t ) d y d x μ 2 g ( t ) h ( t ) g ( t ) J 2 ( x y ) v ( x , t ) d y d x , t ( k τ , ( k + 1 ) τ ] , h ( t ) = μ 1 g ( t ) h ( t ) h ( t ) + J 1 ( x y ) u ( x , t ) d y d x + μ 2 g ( t ) h ( t ) h ( t ) + J 2 ( x y ) v ( x , t ) d y d x , t ( k τ , ( k + 1 ) τ ] , u ( g ( t ) , t ) = u ( h ( t ) , t ) = 0 , v ( g ( t ) , t ) = v ( h ( t ) , t ) = 0 , t ( k τ , ( k + 1 ) τ ] , u ( x , ( k τ ) + ) = H ( u ( x , k τ ) ) , v ( x , ( k τ ) + ) = v ( x , k τ ) , x ( g ( k τ ) , h ( k τ ) ) , u ( x , 0 ) = u 0 ( x ) , v ( x , 0 ) = v 0 ( x ) , g ( 0 ) = h ( 0 ) = h 0 , x [ h 0 , h 0 ] .

In (2.1), it is always assumed that u ( x , t ) = v ( x , t ) = 0 for x R \ [ g ( t ) , h ( t ) ] and t 0 , and u 0 ( x ) and v 0 ( x ) are assumed to satisfy

(2.2) u 0 C [ h 0 , h 0 ] , u 0 ( h 0 ) = u 0 ( h 0 ) = 0 and u 0 ( x ) > 0 in ( h 0 , h 0 ) , v 0 C [ h 0 , h 0 ] , v 0 ( h 0 ) = v 0 ( h 0 ) = 0 and v 0 ( x ) > 0 in ( h 0 , h 0 ) .

The impulsive function H ( x ) is assumed to satisfy

(H) : H C 1 ( [ 0 , ) ) , H ( 0 ) = 0 and H ( u ) > 0 for u [ 0 , ) , H ( u ) u is non-increasing and 0 < H ( u ) u 1 for u > 0 .

It should be mentioned that (H) is natural assumptions for the impulsive function. Some frequently used function forms satisfying assumption (H) are the linear function H ( u ) = c 1 u , where c 1 ( 0 , 1 ] , and the Beverton-Holt function H ( u ) = c 2 u c 3 + u , where 0 < c 2 c 3 < + . In the next content, the assumptions about the initial, kernel, impulsive, and growth functions are always satisfied without specification.

This subsection formulates a pulsed non-local epidemic model in a moving infected environment. In the next subsection, the global existence and uniqueness of the solution to model (2.1) will be investigated.

2.2 Preliminaries

We give some notations first. For any given u = ( u 1 , u 2 ) , v = ( v 1 , v 2 ) R 2 , define

u ( , , , = ) v if u i > ( < , , , = ) v i , i = 1 , 2 .

For any given T , τ , h 0 > 0 , there exists a non-negative integer N such that N τ < T ( N + 1 ) τ , and then define

G T , τ h 0 = g ( t ) g ( t ) C [ 0 , T ] , g ( 0 ) = h 0 , sup n τ < t 1 < t 2 ( n + 1 ) τ g ( t 2 ) g ( t 1 ) t 2 t 1 < 0 , sup N τ < t 1 < t 2 T g ( t 2 ) g ( t 1 ) t 2 t 1 < 0 , H T , τ h 0 = h ( t ) h ( t ) C [ 0 , T ] , h ( 0 ) = h 0 , inf n τ < t 1 < t 2 ( n + 1 ) τ h ( t 2 ) h ( t 1 ) t 2 t 1 > 0 , inf N τ < t 1 < t 2 T h ( t 2 ) h ( t 1 ) t 2 t 1 > 0 ,

where n = 0 , 1 , , N 1 . For any given ( g , h ) G T , τ h 0 × H T , τ h 0 and ( u 0 , v 0 ) satisfying (2.2), we write

Ω N τ T = Ω N τ T ( g , h ) = { ( x , t ) g ( t ) < x < h ( t ) , ( N τ ) + < t T } , Ω ¯ N τ T = Ω ¯ N τ T ( g , h ) = { ( x , t ) g ( t ) x h ( t ) , ( N τ ) + < t T } , Ω n τ ( n + 1 ) τ = Ω n τ ( n + 1 ) τ ( g , h ) = { ( x , t ) g ( t ) < x < h ( t ) , ( n τ ) + < t ( n + 1 ) τ } , Ω ¯ n τ ( n + 1 ) τ = Ω ¯ n τ ( n + 1 ) τ ( g , h ) = { ( x , t ) g ( t ) x h ( t ) , ( n τ ) + < t ( n + 1 ) τ } , X T = X T ( g , h , u 0 , v 0 ) = { ( ζ , η ) ( ζ , η ) [ C ( Ω ¯ n τ ( n + 1 ) τ ) C ( Ω ¯ N τ T ) ] 2 , ( ζ ( x , 0 ) , η ( x , 0 ) ) = ( u 0 , v 0 ) in [ h 0 , h 0 ] , ( ζ , η ) 0 in Ω n τ ( n + 1 ) τ Ω N τ T , and ζ ( x , t ) = η ( x , t ) = 0 for x { g ( t ) , h ( t ) } and t [ 0 , T ] } ,

where n = 0 , 1 , , N 1 . Define

(2.3) C 1 max u * , u 0 , a 12 a 11 v 0 , C 2 max v 0 , G ( C 1 ) a 22 ,

where u * is determined by G ( u ) u = a 11 a 22 a 12 if a 12 G ( 0 ) a 11 a 22 > 1 , and u * = 0 if a 12 G ( 0 ) a 11 a 22 1 . It is not hard to obtain that

(2.4) a 11 C 1 + a 12 C 2 0 and a 22 C 2 + G ( C 1 ) 0 .

We now prove the following estimate that has an important function in the proof of the global solution of (2.1).

Lemma 2.1

Assume that h 0 , τ > 0 , and that ( u 0 , v 0 ) satisfies (2.2). Then, for any T > 0 and ( g , h ) G T , τ h 0 × H T , τ h 0 , the following problem:

(2.5) u t = d 1 g ( t ) h ( t ) J 1 ( x y ) u ( y , t ) d y d 1 u a 11 u + a 12 v , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , v t = d 2 g ( t ) h ( t ) J 2 ( x y ) v ( y , t ) d y d 2 v a 22 v + G ( u ) , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , u ( x , ( n τ ) + ) = H ( u ( x , n τ ) ) , v ( x , ( n τ ) + ) = v ( x , n τ ) , x [ g ( n τ ) , h ( n τ ) ] , u ( g ( t ) , t ) = u ( h ( t ) , t ) = 0 , v ( g ( t ) , t ) = v ( h ( t ) , t ) = 0 , t ( 0 , T ] , u ( x , 0 ) = u 0 ( x ) , v ( x , 0 ) = v 0 ( x ) , x [ h 0 , h 0 ]

admits a unique solution ( u , v ) X T = X T ( g , h , u 0 , v 0 ) . Moreover, ( u , v ) satisfies

0 < u ( x , t ) C 1 , 0 < v ( x , t ) C 2 for ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T .

Proof

Because the initial function ( u 0 , v 0 ) satisfies (2.2) and the impulse function satisfies assumption (H), it is not hard to derive that

u ( x , 0 + ) C [ h 0 , h 0 ] , u ( ± h 0 , 0 + ) = 0 and u ( x , 0 + ) > 0 in ( h 0 , h 0 ) , v ( x , 0 + ) C [ h 0 , h 0 ] , v ( ± h 0 , 0 + ) = 0 and v ( x , 0 + ) > 0 in ( h 0 , h 0 ) ,

which means that ( u ( x , 0 + ) , v ( x , 0 + ) ) satisfies (2.2) and can be viewed as an initial function. Then, using similar techniques as in [33, Lemma 5.1] yields that problem (2.5) admits a unique solution ( u , v ) X τ and

0 < v ( x , t ) max { v ( x , 0 + ) , G ( M 1 ) a 22 } for ( x , t ) Ω 0 τ , 0 < u ( x , t ) max u * , u ( x , 0 + ) , a 12 a 11 v ( x , 0 + ) M 1 for ( x , t ) Ω 0 τ .

The hypothesis of 0 < H ( u ) u 1 in (H) shows that

max u * , u ( x , 0 + ) , a 12 a 11 v ( x , 0 + ) C 1 and max { v ( x , 0 + ) , G ( M 1 ) a 22 } C 2 ,

which implies that the conclusion holds for ( t , x ) Ω 0 τ .

Since ( u ( x , τ + ) , v ( x , τ + ) ) satisfies

u ( x , τ + ) C [ g ( τ ) , h ( τ ) ] , u ( g ( τ ) , τ + ) = u ( h ( τ ) , τ + ) = 0 and u ( x , τ + ) > 0 in ( g ( τ ) , h ( τ ) ) , v ( x , τ + ) C [ g ( τ ) , h ( τ ) ] , v ( g ( τ ) , τ + ) = v ( h ( τ ) , τ + ) = 0 and v ( x , τ + ) > 0 in ( g ( τ ) , h ( τ ) ) ,

it can be viewed as a new initial function for t ( τ + , 2 τ ] . By the same procedure, it follows that problem (2.5) admits a unique solution ( u , v ) for t ( τ + , 2 τ ] and

0 < v ( x , t ) max { v ( x , τ ) , G ( M 2 ) a 22 } for ( x , t ) Ω τ 2 τ , 0 < u ( x , t ) max u * , u ( x , τ ) , a 12 a 11 v ( x , τ ) M 2 for ( x , t ) Ω τ 2 τ .

Since ( u ( x , τ ) , v ( x , τ ) ) ( C 1 , C 2 ) in ( g ( τ ) , h ( τ ) ) , we have that

0 < v ( x , t ) max { C 2 , G ( M 3 ) a 22 } for ( x , t ) Ω τ 2 τ , 0 < u ( x , t ) max { C 1 , a 12 G ( C 1 ) ( a 11 a 22 ) } M 3 for ( x , t ) Ω τ 2 τ .

With the help of (2.4), it follows that a 12 G ( C 1 ) ( a 11 a 22 ) C 1 and G ( C 1 ) a 22 C 2 . Therefore, the conclusion also holds for ( x , t ) Ω τ 2 τ . Step by step, problem (2.5) provides a unique solution ( u , v ) X T and

0 ( u ( x , t ) , v ( x , t ) ) ( C 1 , C 2 ) for ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T .

This ends the proof.□

On the basis of the preceding lemma, we can now establish the following theorem.

Theorem 2.1

For any given h 0 , τ > 0 , the solution of model (2.1), denoted by ( u , v , g , h ) , exists and is unique for all t ( 0 , ) . Moreover, for any T > 0 , we have that

( u , v , g , h ) X T × G T , τ h 0 × H T , τ h 0 .

Proof

Because the initial function ( u 0 , v 0 ) satisfies (2.2) and the impulse function satisfies assumption (H), it is not hard to derive that

u ( x , 0 + ) C [ h 0 , h 0 ] , u ( ± h 0 , 0 + ) = 0 and u ( x , 0 + ) > 0 in ( h 0 , h 0 ) , v ( x , 0 + ) C [ h 0 , h 0 ] , v ( ± h 0 , 0 + ) = 0 and v ( x , 0 + ) > 0 in ( h 0 , h 0 ) ,

which means that ( u ( x , 0 + ) , v ( x , 0 + ) ) satisfies (2.2) and can be viewed as an initial function. Then, using similar techniques as in [33, Theorem 1.1], we find that the conclusion holds for t ( 0 + , τ ] .

Since ( u ( x , τ + ) , v ( x , τ + ) ) satisfies

(2.6) u ( x , τ + ) C [ g ( τ ) , h ( τ ) ] , u ( g ( τ ) , τ + ) = u ( h ( τ ) , τ + ) = 0 and u ( x , τ + ) > 0 in ( g ( τ ) , h ( τ ) ) , v ( x , τ + ) C [ g ( τ ) , h ( τ ) ] , v ( g ( τ ) , τ + ) = v ( h ( τ ) , τ + ) = 0 and v ( x , τ + ) > 0 in ( g ( τ ) , h ( τ ) ) ,

it can be viewed as a new initial function for t ( τ + , 2 τ ] . Then, the conclusion also holds for t ( τ + , 2 τ ] with the help of Lemma 2.1. By the same procedures, we can find a maximum time T max such that model (2.1) has a unique solution ( u , v , g , h ) for t ( 0 + , T max ) and ( u , v , g , h ) X T × G T , τ h 0 × H T , τ h 0 for T ( 0 + , T max ) . Next, we use a proof by contradiction to show that T max = . Assume that T max < . This assumption implies that there exists a non-negative integer k 0 such that T max ( ( k 0 τ ) + , ( k 0 + 1 ) τ ] . We consider the following two cases.

(1) When T max ( ( k 0 τ ) + , ( k 0 + 1 ) τ ) , ( u ( x , ( k 0 τ ) + ) , v ( x , ( k 0 τ ) + ) ) is viewed as a new initial vector-valued function. In virtue of Lemma 2.1 and the similar proof method as in [33, Theorem 1.1], one can obtain that the solution of model (2.1), denoted by ( u , v , g , h ) , exists and is unique for all t ( 0 , ( k 0 + 1 ) τ ] and ( u , v , g , h ) X T × G T , τ h 0 × H T , τ h 0 for T ( 0 , ( k 0 + 1 ) τ ] . This is a contradiction to the definition of T max .

(2) When T max = ( k 0 + 1 ) τ , we have that (2.6) with τ + replaced by ( ( k 0 + 1 ) τ ) + still holds. Taking ( u ( x , ( ( k 0 + 1 ) τ ) + ) , v ( x , ( ( k 0 + 1 ) τ ) + ) ) to a new initial function and repeating the procedure in (1) yield that the solution of model (2.1), denoted by ( u , v , g , h ) , exists and is unique for all t ( 0 , ( k 0 + 1 ) τ ] and ( u , v , g , h ) X T × G T , τ h 0 × H T , τ h 0 for T ( 0 , ( k 0 + 2 ) τ ] . This is also a contradiction to the definition of T max . The proof is completed.□

Finally, we present the following comparison principle, which is used frequently in later analysis.

Lemma 2.2

Assume that ( g , h ) G T , τ h 0 × H T , τ h 0 for some h 0 , τ , T > 0 , and that c 12 , c 21 0 , u , v , u t , v t C ( Ω ¯ n τ ( n + 1 ) τ ) C ( Ω ¯ N τ T ) , and c i j L ( Ω ¯ n τ ( n + 1 ) τ Ω ¯ N τ T ) for i , j = 1 , 2 , and n = 0 , 1 , , N 1 . If

u t d 1 g ( t ) h ( t ) J 1 ( x y ) u ( y , t ) d y d 1 u + c 11 u + c 12 v , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , v t d 2 g ( t ) h ( t ) J 2 ( x y ) v ( y , t ) d y d 2 v + c 21 u + c 22 v , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , u ( x , ( n τ ) + ) H ( u ( x , n τ ) ) , v ( x , ( n τ ) + ) v ( x , n τ ) , x [ g ( n τ ) , h ( n τ ) ] , u ( g ( t ) , t ) , u ( h ( t ) , t ) 0 , v ( g ( t ) , t ) , v ( h ( t ) , t ) 0 , t ( 0 , T ] , u ( x , 0 ) 0 , v ( x , 0 ) 0 , x [ h 0 , h 0 ] .

Then, ( u , v ) 0 in Ω ¯ n τ ( n + 1 ) τ Ω ¯ N τ T . Moreover, if u ( x , 0 ) ( v ( x , 0 ) ) 0 in [ h 0 , h 0 ] , then u ( x , t ) ( v ( x , t ) ) > 0 in Ω n τ ( n + 1 ) τ Ω N τ T .

Proof

When ( x , t ) Ω 0 τ , it follows from assumption (H) that ( u ( x , 0 + ) , v ( x , 0 + ) ) 0 in [ h 0 , h 0 ] . Then treat ( u ( x , 0 + ) , v ( x , 0 + ) ) as an initial function and the conclusion holds by using a similar proof method as in [33, Lemma 2.1]. For ( x , t ) Ω τ 2 τ , we take ( u ( x , τ + ) , v ( x , τ + ) ) as a new initial function, which satisfies ( u ( x , τ + ) , v ( x , τ + ) ) 0 in [ g ( τ ) , h ( τ ) ] , and u ( x , τ + ) ( v ( x , τ + ) ) 0 in [ g ( τ ) , h ( τ ) ] as u ( x , 0 ) ( v ( x , 0 ) ) 0 in [ h 0 , h 0 ] . Therefore, the conclusion holds for ( x , t ) Ω τ 2 τ by using a similar proof method as in [33, Lemma 2.1] again. By the use of mathematical induction, it follows that ( u , v ) 0 in Ω ¯ n τ ( n + 1 ) τ Ω ¯ N τ T , and u ( x , t ) ( v ( x , t ) ) > 0 in Ω n τ ( n + 1 ) τ Ω N τ T when u ( x , 0 ) ( v ( x , 0 ) ) 0 in [ h 0 , h 0 ] . This ends the proof.□

3 The associated problem with fixed boundaries

In this section, we investigate the fixed boundary problem corresponding to model (2.1), that is, model (2.1) with

g ( t ) r < 0 and h ( t ) s > 0 ,

which is represented by the following equations:

(3.1) u t = d 1 r s J 1 ( x y ) u ( y , t ) d y d 1 u a 11 u + a 12 v , x [ r , s ] , t ( ( k τ ) + , ( k + 1 ) τ ] , v t = d 2 r s J 2 ( x y ) v ( y , t ) d y d 2 v a 22 v + G ( u ) , x [ r , s ] , t ( ( k τ ) + , ( k + 1 ) τ ] , u ( x , ( k τ ) + ) = H ( u ( x , k τ ) ) , v ( x , ( k τ ) + ) = v ( x , k τ ) , x [ r , s ] , u ( x , 0 ) = u 0 ( x ) , v ( x , 0 ) = v 0 ( x ) , x [ r , s ] .

It turns out that the methods required here are quite different from previous works where the impulsive function H ( u ) in (3.1) is replaced by u . A good understanding of the dynamical behaviours of (3.1) is remarkably significant in determining the dynamical behaviours of model (2.1). Therefore, we next present some results for (3.1).

3.1 The corresponding eigenvalue problem

There are some works focusing on the existence of principal eigenvalues of the time-periodic cooperative problems with long-range diffusion, see, for instance, [3,16,24]. Motivated by these works, the existence of the principal eigenvalue of the linearised system corresponding to (3.1) will be established next. For the specific definitions of some terms mentioned in this subsection, we refer to [29, Section 3.1] for positive perturbation, [30, Definition 4.5] for essentially compact perturbation, [16, Section 2.1] for positive operator, strongly positive operator, resolvent positive operator, spectral bound, real spectral bound, spectral radius, and principal eigenvalue.

It is a well-known fact that the dynamical behaviours of model (3.1) are controlled by the principal eigenvalue of its linearised eigenvalue problem at ( 0 , 0 ) , which has the following form:

(3.2) ζ t = d 1 r s J 1 ( x y ) ζ ( y , t ) d y d 1 ζ a 11 ζ + a 12 η + λ ζ , x [ r , s ] , t ( 0 + , τ ] , η t = d 2 r s J 2 ( x y ) η ( y , t ) d y d 2 η a 22 η + G ( 0 ) ζ + λ η , x [ r , s ] , t ( 0 + , τ ] , ζ ( x , 0 + ) = H ( 0 ) ζ ( x , 0 ) , η ( x , 0 + ) = η ( x , 0 ) , x [ r , s ] , ζ ( x , 0 ) = ζ ( x , τ ) , η ( x , 0 ) = η ( x , τ ) , x [ r , s ] .

To overcome the difficulties presented by the impulse intervention, we consider the following eigenvalue problem:

(3.3) ϕ t τ + d 1 r s J 1 ( x y ) ϕ ( y , t ) d y d 1 ϕ + a 22 ϕ ln H ( 0 ) τ ϕ + a 12 φ = μ ϕ , x [ r , s ] , t ( 0 , 1 ] , φ t τ + d 2 r s J 2 ( x y ) φ ( y , t ) d y d 2 φ + a 11 φ ln H ( 0 ) τ φ + G ( 0 ) ϕ = μ φ , x [ r , s ] , t ( 0 , 1 ] , ϕ ( x , 0 ) = H ( 0 ) ϕ ( x , 1 ) , φ ( x , 0 ) = φ ( x , 1 ) , x [ r , s ] ,

where μ = a 11 + a 22 ln H ( 0 ) τ λ . With the help of [45, Section 3], problems (3.3) and (3.2) are equivalent.

For any given α ( 0 , ) , let

X α = { ( ϕ , φ ) C ( [ r , s ] × [ 0 , α ] , R 2 ) : ϕ ( x , 0 ) = H ( 0 ) ϕ ( x , α ) , φ ( x , 0 ) = φ ( x , α ) , x [ r , s ] }

with norm ( ϕ , φ ) X α = sup [ r , s ] × [ 0 , α ] ϕ 2 + φ 2 , and

X α + = { ( ϕ , φ ) X : ( ϕ , φ ) 0 , ( x , t ) [ r , s ] × [ 0 , α ] } , X α + + = { ( ϕ , φ ) X : ( ϕ , φ ) 0 , ( x , t ) [ r , s ] × [ 0 , α ] } .

Let X = C ( [ r , s ] , R 2 ) with norm ( ϕ , φ ) X = sup [ r , s ] ϕ 2 + φ 2 , and

X + = { ( ϕ , φ ) X : ( ϕ , φ ) 0 , x [ r , s ] } , X + + = { ( ϕ , φ ) X : ( ϕ , φ ) 0 , x [ r , s ] } .

Now, define two operators A , : X 1 X 1 by

A [ ϕ , φ ] ( x , t ) = d 1 r s J 1 ( x y ) ϕ ( y , t ) d y d 2 r s J 1 ( x y ) φ ( y , t ) d y

and

[ ϕ , φ ] ( x , t ) = ϕ t τ d 1 ϕ + a 22 ϕ ln H ( 0 ) τ ϕ + a 12 φ φ t τ d 2 φ + a 11 φ ln H ( 0 ) τ φ + G ( 0 ) ϕ

for ( ϕ , φ ) X 1 . Then, problem (3.3) can be rewritten as follows:

( A + ) [ ϕ , φ ] C [ ϕ , φ ] = μ ( ϕ , φ ) .

Let { G μ ( t , t 0 ) 0 t 0 t 1 } and { μ ( t , t 0 ) 0 t 0 t 1 } , respectively, be the evolution family on X determined by

ϕ t τ = d 1 ϕ + a 22 ϕ ln H ( 0 ) τ ϕ + a 12 φ μ ϕ , x [ r , s ] , t ( 0 , 1 ] , φ t τ = d 2 φ + a 11 φ ln H ( 0 ) τ φ + G ( 0 ) ϕ μ φ , x [ r , s ] , t ( 0 , 1 ] ,

and

ϕ t τ = d 1 r s J 1 ( x y ) ϕ ( t , y ) d y d 1 ϕ + a 22 ϕ ln H ( 0 ) τ ϕ + a 12 φ μ ϕ , x [ r , s ] , t ( 0 , 1 ] , φ t τ = d 2 r s J 2 ( x y ) φ ( t , y ) d y d 2 φ + a 11 φ ln H ( 0 ) τ ϕ + G ( 0 ) ϕ μ φ , x [ r , s ] , t ( 0 , 1 ] .

With the help of [16, Lemma 2.8] (or [3, Lemma 2.1]), we have the following lemma.

Lemma 3.1

The eigenvalue problem [ ϕ , φ ] = η ( ϕ , φ ) has a principal eigenvalue, denoted by η 1 , which has a unique corresponding eigenfunction in X 1 + + , denoted by ( ϕ 1 , φ 1 ) , with ( ϕ 1 , φ 1 ) X 1 = 1 . Furthermore, we have that

η 1 = a 11 + a 22 ( d 1 + d 2 + 2 ln H ( 0 ) τ ) + ( d 1 + d 2 + 2 ln H ( 0 ) τ a 11 a 22 ) 2 + 4 a 12 G ( 0 ) 2 ,

and for any ( x , t ) [ r , s ] × [ 0 , 1 ] ,

ϕ 1 ( x , t ) ϕ 1 ( t ) and φ 1 ( x , t ) φ 1 ( t ) .

Based on the aforementioned preparations, some properties of the operator can now be given.

Proposition 3.1

The resolvent operator ( α ) 1 exists for every α C with Re α > η 1 . Furthermore, is a resolvent positive operator and the spectral bound s ( ) of the operator satisfies s ( ) = η 1 .

Proof

To begin with, the fundamental theorem of ordinary differential equations gives that for every α C with Re α > η 1 , ( α ) 1 exists in X 1 . This implies that s ( ) η 1 . Then, it follows from [24, Lemma B.3] that the evolution family { G 0 ( t , 0 ) t ( 0 , 1 ] } generated by is positive on X 1 for any t ( 0 , 1 ] . With the help of [29, Theorem 3.12], it can be immediately obtained that is a resolvent positive operator.

To prove that s ( ) = η 1 , we only need to show that s ( ) η 1 . Arguing indirectly, assume that η 1 ρ ( ) , where ρ ( ) denotes the resolvent set of the operator . Then ( η 1 ) u = v has a unique solution u X 1 with u ( x , t ) u ( t ) for any v X 1 with v ( x , t ) v ( t ) . In virtue of the Fredholm alternative [4, Theorem 6.6], it follows that ( η 1 ) u = 0 has no non-trivial solution u ( t ) = ( u 1 ( t ) , u 2 ( t ) ) with u 1 ( 0 ) = H ( 0 ) u 1 ( 1 ) and u 2 ( 0 ) = u 2 ( 1 ) . Recall that ( η 1 , ϕ 1 , φ 1 ) is the principal eigenpair of the eigenvalue problem [ ϕ , φ ] = η ( ϕ , φ ) . Then, u ( t ) = ( ϕ 1 ( t ) , φ 1 ( t ) ) satisfy ( η 1 ) u = 0 , which is a contradiction. Thus, η 1 C ρ ( ) and η 1 s ( ) . This proof is completed.□

With the help of the evolution family { μ ( t , t 0 ) 0 t 0 t 1 } , we define an operator T μ by

T μ ϕ = e μ τ 0 ( 1,0 ) ( H ( 0 ) ϕ 1 , ϕ 2 ) for ϕ = ( ϕ 1 , ϕ 2 ) X .

Proposition 3.2

There exists μ 0 ( , + ) such that the spectral radius r ( T μ 0 ) of the operator T μ 0 satisfies r ( T μ 0 ) = 1 . Moreover, the operator C is resolvent positive and s ( C ) = μ 0 .

Proof

We consider the following equation:

(3.4) v = ( μ C ) 1 u , μ ρ ( C ) , u = ( u 1 , u 2 ) X 1 .

With the help of the variation of constant formula, it can be immediately obtained that

(3.5) v = e μ τ t 0 ( t , 0 ) v ( x , 0 ) + τ 0 t e μ τ ( t s ) 0 ( t , s ) u ( x , s ) d s .

Let v = ( v 1 , v 2 ) . Noting that v 1 ( x , 0 ) = H ( 0 ) v 1 ( x , 1 ) and v 2 ( x , 0 ) = v 2 ( x , 1 ) for any x [ r , s ] , it follows from (3.5) that

(3.6) ( T μ ) v ( x , 0 ) = τ 0 1 e μ τ ( 1 s ) 0 ( 1 , s ) u ( x , s ) d s diag ( H ( 0 ) , 1 ) .

In virtue of (3.4)–(3.6), it follows that if 1 ρ ( T μ ) , then

(3.7) ( μ C ) 1 u = τ 0 t e μ τ ( t s ) 0 ( t , s ) u ( x , s ) d s + τ e μ τ t 0 ( t , 0 ) ( T μ ) 1 0 1 e μ τ ( 1 s ) 0 ( 1 , s ) u ( x , s ) d s diag ( H ( 0 ) , 1 ) .

Therefore, 1 ρ ( T μ ) if and only if μ ρ ( C ) .

A simple calculation yields that the matrix

A a 22 a 12 G ( 0 ) a 11

has a real eigenvalue

μ 1 ( A ) = a 11 + a 22 + ( a 11 a 22 ) 2 + 4 a 12 G ( 0 ) 2 ,

which has a positive eigenvector

x = ( a 12 , μ 1 ( A ) a 22 ) .

In addition, assumption (J) yields that ( μ 1 ln H ( 0 ) τ , x ) is an eigenpair of the operator

D [ ϕ , φ ] ( x ) d 1 r s J 1 ( x y ) [ ϕ ( y ) ϕ ( x ) ] d y + a 22 ϕ ln H ( 0 ) τ ϕ + a 12 φ d 2 r s J 1 ( x y ) [ φ ( y ) φ ( x ) ] d y + a 11 φ ln H ( 0 ) τ φ + G ( 0 ) ϕ .

In virtue of [29, Lemma 5.8], one can obtain that

e σ ( τ D ) t = σ ( 0 ( t , 0 ) ) { 0 } for all t ( 0 , 1 ] ,

where σ ( ) denotes the spectrum of operator . Recalling that 0 ( t , 0 ) is a positive operator, it follows from [26, Proposition 4.1.1] that r ( 0 ( t , 0 ) ) σ ( 0 ( t , 0 ) ) . Then, we have that

r ( 0 ( t , 0 ) ) = e r ( τ D ) t e τ [ μ 1 ln H ( 0 ) τ ] t for all t ( 0 , 1 ] .

Therefore, we have that

r ( T μ 1 ) H ( 0 ) r ( e μ 1 τ 0 ( 1,0 ) ) 1 .

On the other hand, the definition of the operator T μ implies that lim μ + r ( T μ ) = 0 . Since r ( T μ ) is strongly decreasing about μ , a unique μ 0 can be found such that r ( T μ 0 ) = 1 and when μ > μ 0 , r ( T μ ) < r ( T μ 0 ) = 1 . This means that ρ ( C ) contains a ray ( μ 0 , + ) . In addition, it follows from (3.7) that ( μ C ) 1 is a positive operator. Hence, C is a resolvent positive operator. Recalling that T μ 0 is a positive operator, it follows from [26, Proposition 4.1.1] that r ( T μ 0 ) = 1 σ ( T μ 0 ) . This implies that μ 0 σ ( C ) and then μ 0 = s R ( C ) . Finally, [30, Theorem 3.5] gives that s ( C ) = s R ( C ) = μ 0 . This proof is complected.□

Now we present the main theorem of this subsection.

Theorem 3.1

The eigenvalue problem (3.2) has a unique eigenvalue λ 1 , and the corresponding eigenfunction satisfies ( ζ 1 , η 1 ) X τ + + . Moreover, we have that

λ 1 = a 11 + a 22 ln H ( 0 ) τ s ( C ) .

Proof

In virtue of Proposition 3.1, one can obtain that is a resolvent positive operator. In addition, A is a positive linear operator. Therefore, C = A + is a positive perturbation of . By Proposition 3.2, we now have that the operator C is resolvent positive. This implies that case (i) in [29, Theorem 3.6] is impossible.

By standard calculation, the eigenvalue problem

ϕ t τ = a 22 ϕ ln H ( 0 ) τ ϕ + a 12 φ , x [ r , s ] , t ( 0 , 1 ] , φ t τ = a 11 φ ln H ( 0 ) τ φ + G ( 0 ) ϕ , x [ r , s ] , t ( 0 , 1 ] , ϕ ( x , 0 ) = H ( 0 ) ϕ ( x , 1 ) , φ ( x , 0 ) = φ ( x , 1 ) , x [ r , s ]

has a principal eigenpair ( η 2 , ( ϕ 2 , φ 2 ) ) . Then C [ ϕ 2 , φ 2 ] = η 2 ( ϕ 2 , φ 2 ) and s ( C ) η 2 > s ( ) . This implies that case (ii) in [29, Theorem 3.6] is also impossible. Hence, case (iii) in [29, Theorem 3.6] will happen. Define

μ = A ( μ ) 1 , μ > s ( ) .

Case (iii) in [29, Theorem 3.6] tells us that there exist s ( ) < μ 1 < μ 2 such that r ( μ 2 ) < 1 r ( μ 1 ) .

Let { u n } X 1 i X 1 be a bounded sequence. Define v n = ( α ) 1 u n , where α C and Re α > η 1 . According to the boundedness of + t τ , both v n and t v n are bounded sequences in X 1 i X 1 . Then, assumption (J) yields that A v n is equicontinuous and uniformly bounded. It follows from the Arzelà-Ascoli theorem that A v n is precompact in X . Therefore, A ( α ) 1 is a compact operator in X . Noticing that ( α ) 1 is a linear bounded operator in X , ( α ) 1 A ( α ) 1 is compact for any α > s ( ) . This implies that A is an essentially compact perturbator of . Now, applying [30, Theorem 4.7] yields that s ( C ) is a principal eigenvalue of C with an eigenfunction ( ζ 0 , η 0 ) in X 1 + + . By using a similar proof method as in [3, Theorem 2.3], it can be obtained that s ( C ) is an algebraically simple principal eigenvalue.

Next, we prove uniqueness. Let ( η , ( ϕ , φ ) ) be the principal eigenpairs of the operator C , where ( ζ , η ) X 1 + + . A standard calculation gives that

(3.8) 0 ( t , 0 ) ( ζ ( x , 0 ) , η ( x , 0 ) ) = e τ η t ( ζ ( x , t ) , η ( x , t ) ) .

The strong positivity of the eigenfunction ( ζ , η ) yields that for any given ( u 1 ( x ) , u 2 ( x ) ) X + ,

( u 1 ( x ) , u 2 ( x ) ) C 0 ( ζ ( x , 0 ) , η ( x , 0 ) ) ( u 1 ( x ) , u 2 ( x ) ) X min [ min [ r , s ] ζ ( x , 0 ) , min [ r , s ] η ( x , 0 ) ] ( ζ ( x , 0 ) , η ( x , 0 ) ) , x [ r , s ] .

Since the operator 0 ( t , 0 ) is positive, we have that

(3.9) 0 ( t , 0 ) ( u 1 ( x ) , u 2 ( x ) ) C 0 0 ( t , 0 ) ( ζ ( x , 0 ) , η ( x , 0 ) ) , t ( 0 , 1 ] .

Now (3.8) together with (3.9) yields that

0 ( t , 0 ) ( u 1 ( x ) , u 2 ( x ) ) C 0 e τ η t ( ζ ( x , t ) , η ( x , t ) ) , t ( 0 , 1 ] .

This implies that

(3.10) τ s ( C ) ω ( 0 ( t , 0 ) ) τ η ,

where ω ( 0 ( t , 0 ) ) denotes the growth bound of the operator 0 ( t , 0 ) . Combining (3.10) and the definition of s ( C ) gives that η = s ( C ) . Finally, from the relationship between problems (3.2) and (3.3), we see that the desired conclusion holds. This ends the proof.□

Remark 3.1

By the proof of Theorem 3.1, it follows that the unique eigenvalue λ 1 of the eigenvalue problem (3.2) is an algebraically simple principal eigenvalue, and its corresponding vector-valued eigenfunction ( ζ 1 , η 1 ) is uniquely determined in the sense that the norm is one.

The aforementioned theorem exhibits that the eigenvalue problem (3.2) corresponding to (3.1) has a unique principal eigenvalue λ 1 . The next two lemmas will provide a method for estimating the principal eigenvalue.

Lemma 3.2

If there exists a non-negative vector-valued function ( ζ , η ) with ζ , η 0 and ( ζ t , η t ) C ( [ r , s ] × [ 0 , τ ] , R 2 ) , and a number λ ¯ satisfy

(3.11) ζ t d 1 r s J 1 ( x y ) ζ ( y , t ) d y d 1 ζ a 11 ζ + a 12 η + λ ¯ ζ , x [ r , s ] , t ( 0 + , τ ] , η t d 2 r s J 2 ( x y ) η ( y , t ) d y d 2 η a 22 η + G ( 0 ) ζ + λ ¯ η , x [ r , s ] , t ( 0 + , τ ] , ζ ( x , 0 + ) H ( 0 ) ζ ( x , 0 ) , η ( x , 0 + ) η ( x , 0 ) , x [ r , s ] , ζ ( x , 0 ) = ζ ( x , τ ) , η ( x , 0 ) = η ( x , τ ) , x [ r , s ] ,

then λ 1 λ ¯ holds. Moreover, λ 1 = λ ¯ iff the equalities of (3.11) all hold.

Proof

We first prove the first assertion. Let

ϕ ( x , t ) = ζ 1 ( x , τ t ) and φ ( x , t ) = η 1 ( x , τ t ) ,

where ( ζ 1 , η 1 ) is the eigenvector pair corresponding to the unique principal eigenvalue λ 1 of the eigenvalue problem (3.2). Then, ϕ 1 and φ 1 satisfy the following equation:

(3.12) ϕ t = d 1 r s J 1 ( x y ) ϕ ( y , t ) d y d 1 ϕ a 11 ϕ + a 12 φ + λ 1 ϕ , x [ r , s ] , t [ 0 , τ ) , φ t = d 2 r s J 2 ( x y ) φ ( y , t ) d y d 2 φ a 22 φ + G ( 0 ) ϕ + λ 1 φ , x [ r , s ] , t [ 0 , τ ) , ϕ ( x , τ ) = H ( 0 ) ϕ ( x , τ ) , φ ( x , τ ) = φ ( x , τ ) , x [ r , s ] , ϕ ( x , 0 ) = ϕ ( x , τ ) , φ ( x , 0 ) = φ ( x , τ ) , x [ r , s ] .

By multiplying the first equations of (3.11) and (3.12) by ϕ and ζ , respectively, it follows that

(3.13) ζ t ϕ d 1 r s J 1 ( x y ) ζ ( y , t ) d y ϕ d 1 ζ ϕ a 11 ζ ϕ + a 12 η ϕ + λ ¯ ζ ϕ , ϕ t ζ = d 1 r s J 1 ( x y ) ϕ ( y , t ) d y ζ + d 1 ϕ ζ + a 11 ϕ ζ a 12 φ ζ λ 1 ϕ ζ .

Then, integrating the two sides of (3.13) on [ r , s ] × ( 0 + , τ ) and summing the outcomes yield that

(3.14) r s 0 + τ ( ζ t ϕ + ϕ t ζ ) d t d x ( λ ¯ λ 1 ) r s 0 + τ ζ ϕ d t d x + a 12 r s 0 + τ ( η ϕ φ ζ ) d t d x + d 1 r s 0 + τ r s J 1 ( x y ) ζ ( y , t ) d y ϕ r s J 1 ( x y ) ϕ ( y , t ) d y ζ d t d x .

Recalling that the assumption J 1 ( x ) = J 1 ( x ) 0 for x R in (J), one can obtain that

(3.15) r s 0 + τ r s J 1 ( x y ) ζ ( y , t ) d y ϕ r s J 1 ( x y ) ϕ ( y , t ) d y ζ d t d x = 0 .

In addition, the impulse and period conditions of (3.11) and (3.12) yield that

(3.16) r s 0 + τ ( ζ t ϕ + ϕ t ζ ) d t d x = r s [ ζ ( x , τ ) ϕ ( x , τ ) ζ ( x , 0 + ) ϕ ( x , 0 + ) ] d x H ( 0 ) r s [ ζ ( x , τ ) ϕ ( x , 0 ) ζ ( x , τ ) ϕ ( x , 0 ) ] d x = 0 .

Inserting (3.15) and (3.16) into (3.14) yields that

(3.17) λ ¯ λ 1 a 12 r s 0 + τ ζ ϕ d t d x + r s 0 + τ ( η ϕ φ ζ ) d t d x 0 .

A similar process results in

(3.18) λ ¯ λ 1 G ( 0 ) r s 0 + τ η φ d t d x + r s 0 + τ ( ζ φ ϕ η ) d t d x 0 .

Adding (3.17) and (3.18) gives that

(3.19) ( λ ¯ λ 1 ) r s 0 + τ ζ ϕ a 12 + η φ G ( 0 ) d t d x 0 .

Noting that ( ζ 1 , η 1 ) X τ + + and ζ , η ( ) 0 , it follows from (3.19) that λ ¯ λ 1 .

The proof of the remaining conclusion is shown next. When the equalities of (3.11) all hold, the conclusion λ ¯ = λ 1 can be obtained since the eigenvalue of the eigenvalue problem (3.2) is unique. Finally, when λ ¯ = λ 1 , we prove that the equalities of (3.11) must all hold. Arguing indirectly, assume that there exists an inequality holding in (3.11). Then, the same method as for the first assertion yields that λ ¯ > λ 1 , which is a contradiction to λ ¯ = λ 1 . This proof is completed.□

Lemma 3.3

If there exists a non-negative vector-valued function ( ζ , η ) with ζ , η 0 and ( ζ t , η t ) C ( [ 0 , τ ] × [ r , s ] , R 2 ) , and a number λ ̲ satisfies

(3.20) ζ t d 1 r s J 1 ( x y ) ζ ( y , t ) d y d 1 ζ a 11 ζ + a 12 η + λ ̲ ζ , x [ r , s ] , t ( 0 + , τ ] , η t d 2 r s J 2 ( x y ) η ( y , t ) d y d 2 η a 22 η + G ( 0 ) ζ + λ ̲ η , x [ r , s ] , t ( 0 + , τ ] , ζ ( x , 0 + ) H ( 0 ) ζ ( x , 0 ) , η ( x , 0 + ) η ( x , 0 ) , x [ r , s ] , ζ ( x , 0 ) = ζ ( x , τ ) , η ( x , 0 ) = η ( x , τ ) , x [ r , s ] ,

then λ 1 λ ̲ holds. Moreover, λ 1 = λ ̲ iff the equalities of (3.11) all hold.

Proof

This proof is similar to Lemma 3.2, and we omit it here.□

Subsequently, we will investigate the monotonicity and continuity of the eigenvalue λ 1 of problem (3.2) with respect to the intensity of the impulse function H and the length of the interval [ r , s ] .

Lemma 3.4

Assume that [ r , s ] = [ l , l ] . Let H ( 0 ) = z . Denote the eigenvalue of (3.2) by λ 1 ( z ) and λ 1 ( l ) to emphasise its dependency on z and l. Then, we have the following conclusions:

  1. λ 1 ( z ) is strongly decreasing and continuous with respect to z ( 0 , 1 ] .

  2. λ 1 ( l ) is strongly decreasing and continuous with respect to l ( 0 , ) .

Proof

(1) First, we show that λ 1 ( z ) is monotone with respect to z . For any given z 1 , z 2 ( 0 , 1 ] and z 1 < z 2 , let ( λ 1 ( z 2 ) , ζ 2 , η 2 ) be the principal eigenpair of (3.2). According to problem (3.2), we can obtain that

ζ 2 t = d 1 r s J 1 ( x y ) ζ 2 ( y , t ) d y d 1 ζ 2 a 11 ζ 2 + a 12 η 2 + λ ( z 2 ) ζ 2 , x [ r , s ] , t ( 0 + , τ ] , η 2 t = d 2 r s J 2 ( x y ) η 2 ( y , t ) d y d 2 η 2 a 22 η 2 + G ( 0 ) ζ 2 + λ ( z 2 ) η 2 , x [ r , s ] , t ( 0 + , τ ] , ζ 2 ( x , 0 + ) > z 1 ζ 2 ( x , 0 ) , η 2 ( x , 0 + ) = η 2 ( x , 0 ) , x [ r , s ] , ζ 2 ( x , τ ) = ζ 2 ( x , τ ) , η 2 ( x , 0 ) = η 2 ( x , τ ) , x [ r , s ] .

Then, Lemma 3.3 yields that λ ( z 2 ) < λ ( z 1 ) .

Subsequently, we show that λ 1 ( z ) is continuous with respect to z . To do this, we will prove that for any given z 1 , z 2 ( 0 , 1 ] , if z 1 z 2 , then

(3.21) λ 1 ( z 1 ) λ 1 ( z 2 ) ln z 2 ln z 1 τ ,

while if z 1 z 2 , then

(3.22) λ 1 ( z 1 ) λ 1 ( z 2 ) ln z 2 ln z 1 τ .

Let ( λ 1 ( z 1 ) , ζ 1 , η 1 ) be the principal eigenpair of (3.2). Now, we construct the following functions:

ϕ ( x , t ) = z 1 z 2 ζ 1 ( x , t ) , x [ r , s ] , t = 0 , ζ 1 ( x , t ) , x [ r , s ] , t = 0 + , e ln ( z 1 z 2 ) t τ ζ 1 ( x , t ) , x [ r , s ] , t ( 0 + , τ ] ,

and

ψ ( x , t ) = z 1 z 2 η 1 ( x , t ) , x [ r , s ] , t = 0 , η 1 ( x , t ) , x [ r , s ] , t = 0 + , e ln ( z 1 z 2 ) t τ η 1 ( x , t ) , x [ r , s ] , t ( 0 + , τ ] .

Then, the functions ϕ and ψ satisfy

(3.23) ϕ t = d 1 r s J 1 ( x y ) ϕ ( y , t ) d y d 1 ϕ a 11 ϕ + a 12 ψ + λ ( z 1 ) + ln z 1 ln z 2 τ ϕ , ( x , t ) Ω 0 , ψ t = d 2 r s J 2 ( x y ) ψ ( y , t ) d y d 2 ψ a 22 ψ + G ( 0 ) ϕ + λ ( z 1 ) + ln z 1 ln z 2 τ ψ , ( x , t ) Ω 0 , ϕ ( x , 0 + ) = z 2 ϕ ( x , 0 ) , ψ ( x , 0 + ) = z 2 z 1 ψ ( x , 0 ) , x [ r , s ] , ϕ ( x , 0 ) = ϕ ( x , τ ) , ψ ( x , 0 ) = ψ ( x , τ ) , x [ r , s ] ,

where Ω 0 = [ r , s ] × ( 0 + , τ ] . When z 1 z 2 , applying Lemma 3.3 to (3.23) yields that

λ 1 ( z 2 ) λ 1 ( z 1 ) + ln z 1 ln z 2 τ ,

which implies that (3.21) holds. When z 1 z 2 , applying Lemma 3.2 to (3.23) yields that

λ 1 ( z 2 ) λ 1 ( z 1 ) + ln z 1 ln z 2 τ ,

which implies that (3.22) also holds. We have now proved the desired continuity.

(2) This proof is similar to [33, Lemma 3.8], and we omit it here.□

Consider the following eigenvalue problem:

(3.24) ϕ t = a 12 φ a 11 ϕ + μ ϕ , t ( 0 + , τ ] , φ t = G ( 0 ) ϕ a 22 φ + μ φ , t ( 0 + , τ ] , ϕ ( 0 ) = ϕ ( τ ) , φ ( 0 ) = φ ( τ ) , ϕ ( 0 + ) = H ( 0 ) ϕ ( 0 ) , φ ( 0 + ) = φ ( 0 ) .

By the standard method (see, e.g., the proof of [44, Lemma 3.3]), it follows that problem (3.24) has a unique principal eigenvalue μ 1 , which has a strongly positive eigenfunction pair ( ϕ 1 , φ 1 ) . At the end of this subsection, we will establish a relationship between the principal eigenvalues of (3.2) and (3.24).

Lemma 3.5

Assume that [ r , s ] = [ l , l ] . Let λ 1 ( l ) and μ 1 denote the principal eigenvalues of (3.2) and (3.24), respectively. Then, we have that

lim l + λ 1 ( l ) = μ 1 .

Proof

For any given l R , we choose ( ζ , η ) = ( ϕ 1 , φ 1 ) , where ( ϕ 1 , φ 1 ) is the positive eigenfunction pair corresponding to μ 1 . From assumption (J), it follows that

ζ t = d 1 + J 1 ( x y ) [ ζ ( t ) ζ ( t ) ] d y a 11 ζ + a 12 η + μ 1 ζ d 1 l l J 1 ( x y ) ζ ( t ) d y d 1 ζ a 11 ζ + a 12 η + μ 1 ζ ,

and

η t = d 2 + J 2 ( x y ) [ η ( t ) η ( t ) ] d y a 22 η + G ( 0 ) ζ + μ 1 η d 2 l l J 2 ( x y ) η ( t ) d y d 2 η a 22 η + G ( 0 ) ζ + μ 1 η .

Lemma 3.3 gives that λ 1 ( l ) μ 1 . Then, Lemma 3.4(2) yields that lim l + λ 1 ( l ) μ 1 .

Subsequently, we show that μ 1 = inf l R λ 1 ( l ) . To do this, we first define

θ l ( x ) max { 0 , l x } , x R .

It follows from the proof of [33, Lemma 3.9(2)] that for any sufficiently small ε > 0 , there exists a sufficiently large l * > 0 such that for any l l * , we have that

l l J i ( x y ) θ l ( y ) d y θ l ( x ) > ε max { d 1 , d 2 } θ l ( x ) for x [ l , l ] , i = 1 , 2 .

For any l l * , we choose ( ζ , η ) = ( ϕ 1 θ l , φ 1 θ l ) . Then, a standard calculation yields that

ζ t = d 1 ζ d 1 ζ a 11 ζ + a 12 η + μ 1 ζ < d 1 l l J 1 ( x y ) ζ ( t ) d y d 1 ζ a 11 ζ + a 12 η + ( μ 1 + ε ) ζ ,

and

η t = d 2 η d 2 η a 22 η + G ( 0 ) ζ + μ 1 η < d 2 l l J 2 ( x y ) η ( t ) d y d 2 η a 22 η + G ( 0 ) ζ + ( μ 1 + ε ) η .

With the help of Lemma 3.2, it follows that λ 1 ( l ) < μ 1 + ε . This implies that μ 1 = inf l R λ 1 ( l ) . This ends the proof.□

3.2 The asymptotic profile in a fixed domain

In this subsection, we still use some notations from Section 2.2 with the free boundaries g ( t ) and h ( t ) replaced by the fixed boundaries r and s . By a similar discussion as in Lemma 2.1, for any given τ > 0 , [ r , s ] R , and ( u 0 ( x ) , v 0 ( x ) ) satisfying (2.2) with [ h 0 , h 0 ] replaced by [ r , s ] , problem (3.1) has a unique non-negative solution defined for all t > 0 . In this subsection, we always denote this solution by ( u ( x , t ) , v ( x , t ) ) or ( u , v ) . Moreover, for any T > 0 , one can obtain that

( u , v ) [ C ( Ω ¯ n τ ( n + 1 ) τ ) C ( Ω ¯ N τ T ) ] 2

and

0 < u ( x , t ) C 1 , 0 < v ( x , t ) C 2 for ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T ,

where C 1 and C 2 are defined in (2.3). The corresponding steady-state problem for (3.1) can be written as follows:

(3.25) U t = d 1 r s J 1 ( x y ) U ( y , t ) d y d 1 U a 11 U + a 12 V , x [ r , s ] , t ( 0 + , τ ] , V t = d 2 r s J 2 ( x y ) V ( y , t ) d y d 2 V a 22 V + G ( U ) , x [ r , s ] , t ( 0 + , τ ] , U ( x , 0 + ) = H ( U ( x , 0 ) ) , V ( x , 0 + ) = V ( x , 0 ) , x [ r , s ] , U ( x , 0 ) = U ( x , τ ) , V ( x , 0 ) = V ( x , τ ) , x [ r , s ] .

Before discussing the asymptotic profile of model (3.1), we first give the definitions of the ordered upper and lower solutions of problems (3.1) and (3.25) and then prove that the comparison principles of problems (3.1) and (3.25) are also valid. They form the basis of this subsection.

Definition 3.1

For any given T > 0 , non-negative vector-valued functions ( u ¯ , v ¯ ) and ( u ̲ , v ̲ ) with

( u ¯ t , v ¯ t ) , ( u ̲ t , v ̲ t ) [ C ( Ω ¯ n τ ( n + 1 ) τ ) C ( Ω ¯ N τ T ) ] 2

are called the ordered upper and lower solutions of (3.1), respectively, if ( u ¯ , v ¯ ) and ( u ̲ , v ̲ ) satisfy

u ¯ t d 1 r s J 1 ( x y ) u ¯ ( y , t ) d y d 1 u ¯ a 11 u ¯ + a 12 v ¯ , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , v ¯ t d 2 r s J 2 ( x y ) v ¯ ( y , t ) d y d 2 v ¯ a 22 v ¯ + G ( u ¯ ) , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , u ¯ ( x , ( n τ ) + ) H ( u ¯ ( x , n τ ) ) , v ¯ ( x , ( n τ ) + ) v ¯ ( x , n τ ) , x [ r , s ] , u ¯ ( x , 0 ) u 0 ( x ) , v ¯ ( x , 0 ) v 0 ( x ) , x [ r , s ] ,

and

u ̲ t d 1 r s J 1 ( x y ) u ̲ ( y , t ) d y d 1 u ̲ a 11 u ̲ + a 12 v ̲ , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , v ̲ t d 2 r s J 2 ( x y ) v ̲ ( y , t ) d y d 2 v ̲ a 22 v ̲ + G ( u ̲ ) , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , u ̲ ( x , ( n τ ) + ) H ( u ̲ ( x , n τ ) ) , v ̲ ( x , ( n τ ) + ) v ̲ ( x , n τ ) , x [ r , s ] , u ̲ ( x , 0 ) u 0 ( x ) , v ̲ ( x , 0 ) v 0 ( x ) , x [ r , s ] ,

respectively.

Remark 3.2

Denote the ordered upper and lower solutions of problem (3.25) by the pairs ( U ¯ , V ¯ ) and ( U ̲ , V ̲ ) , respectively. Similarly, an upper (lower) solution of problem (3.25) can also be defined by substituting the initial condition ( u ¯ ( x , 0 ) , v ¯ ( x , 0 ) ) ( u 0 ( x ) , v 0 ( x ) ) ( ( u ̲ ( x , 0 ) , v ̲ ( x , 0 ) ) ( u 0 ( x ) , v 0 ( x ) ) ) with the periodic condition ( U ¯ ( x , 0 ) , V ¯ ( x , 0 ) ) ( U ¯ ( x , τ ) , V ¯ ( x , τ ) ) ( ( U ̲ ( x , 0 ) , V ̲ ( x , 0 ) ) ( U ̲ ( x , τ ) , V ̲ ( x , τ ) ) ) .

Lemma 3.6

Assume that r, s, τ , T > 0 and c 12 , c 21 0 , that u , v , u t , v t C ( Ω ¯ n τ ( n + 1 ) τ ) C ( Ω ¯ N τ T ) , and that c i j L ( Ω ¯ n τ ( n + 1 ) τ Ω ¯ N τ T ) for i , j = 1 , 2, and n = 0 , 1 , , N 1 . If

u t d 1 r s J 1 ( x y ) u ( y , t ) d y d 1 u + c 11 u + c 12 v , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , v t d 2 r s J 2 ( x y ) v ( y , t ) d y d 2 v + c 21 u + c 22 v , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , u ( x , ( n τ ) + ) 0 , v ( x , ( n τ ) + ) 0 , x [ r , s ] , u ( x , 0 ) 0 , v ( x , 0 ) 0 , x [ r , s ] ,

then, ( u , v ) 0 in Ω ¯ n τ ( n + 1 ) τ Ω ¯ N τ T . Moreover, if u ( x , 0 ) 0 ( v ( x , 0 ) 0 ) in [ h 0 , h 0 ] , then u ( x , t ) > 0 ( v ( x , t ) > 0 ) in Ω n τ ( n + 1 ) τ Ω N τ T .

Proof

By the similarity to the proof of Lemma 2.2, the desired conclusion can be obtained. Since it is actually much simpler, we omit the details here.□

Remark 3.3

Note that the fixed boundary problem (3.1) and the corresponding steady-state problem (3.25) have a very similar structure. Therefore, when the initial conditions u ( x , 0 ) 0 and v ( x , 0 ) 0 are replaced by the periodic conditions U ( x , 0 ) U ( x , τ ) and V ( x , 0 ) V ( x , τ ) , the conclusions of Lemma 3.6 also hold for problem (3.25).

Now, we present the major results of the subsection.

Theorem 3.2

Let λ 1 be the unique algebraically simple principal eigenvalue of (3.2). Then the conclusions below hold:

  1. If λ 1 0 , then (0, 0) is the unique solution to problem (3.25). Moreover, the solution of problem (3.1) satisfies

    lim t + ( u ( x , t ) , v ( x , t ) ) = ( 0 , 0 ) uniformly for x [ r , s ] .

  2. If λ 1 < 0 , then the problem (3.25) has a unique solution ( U , V ) . Moreover, the solution of problem (3.1) satisfies

    lim m + ( u ( x , t + m τ ) , v ( x , t + m τ ) ) = ( U ( x , t ) , V ( x , t ) ) uniformly for ( x , t ) [ r , s ] × [ 0 , + ) .

Proof

(1) We first prove the first assertion. Clearly, 0 is a non-negative solution of problem (3.25). Next, we prove that it is the only one. Arguing indirectly, suppose that ( U , V ) is a strongly positive solution pair of (3.25). By assumptions (G) and (H), it follows that

(3.26) ( G ( U ) , H ( U ) ) ( G ( 0 ) U , H ( 0 ) U ) .

Substituting (3.26) into (3.25) yields that

(3.27) U t = d 1 r s J 1 ( x y ) U ( y , t ) d y d 1 U a 11 U + a 12 V , x [ r , s ] , t ( 0 + , τ ] , V t < d 2 r s J 2 ( x y ) V ( y , t ) d y d 2 V a 22 V + G ( 0 ) U , x [ r , s ] , t ( 0 + , τ ] , U ( x , 0 + ) < H ( 0 ) U ( x , 0 ) , V ( x , 0 + ) = V ( x , 0 ) , x [ r , s ] , U ( x , 0 ) = U ( x , τ ) , V ( x , 0 ) = V ( x , τ ) , x [ r , s ] .

In virtue of Lemma 3.2, we have from (3.27) that λ 1 < 0 , which is a contradiction to λ 1 0 .

The second assertion is proved next. Its proof is divided into the two situations below.

Case 1: λ 1 > 0 .

It follows from Remark 3.1 that the unique principal eigenvalue λ 1 has a unique eigenfunction pair ( ζ 1 , η 1 ) . With the help of this eigenfunction pair, let

u ¯ ( x , t ) = K e λ 1 t ζ 1 ( x , t ) and v ¯ ( x , t ) = K e λ 1 t η 1 ( x , t ) ,

where K is a sufficiently large positive constant such that K ( ζ 1 ( x , 0 ) , η 1 ( x , 0 ) ) ( u 0 ( x ) , v 0 ( x ) ) . By a standard calculation, it can be obtained that the function pair ( u ¯ , v ¯ ) is an upper solution of (3.1). Now applying Lemma 3.6 to ( u ¯ u , v ¯ v ) , it follows that

( u , v ) ( u ¯ , v ¯ ) for all ( x , t ) [ r , s ] × R + .

This combined with the condition λ 1 > 0 yields that

lim t + ( u ( x , t ) , v ( x , t ) ) = ( 0 , 0 ) uniformly for x [ r , s ] .

Case 2: λ 1 = 0 .

By a standard calculation, the function pair ( C 1 , C 2 ) is an upper solution of (3.1), where C 1 and C 2 are defined in (2.3). By selecting ( C 1 , C 2 ) as an initial iteration, we can obtain the iteration sequence { ( u ¯ ( j ) , v ¯ ( j ) ) } j = 1 + from the following law:

(3.28) u ¯ ( j ) t d 1 r s J 1 ( x y ) u ¯ ( j ) d y + m 1 u ¯ ( j ) = a 12 v ¯ ( i 1 ) , ( x , t ) Ω k , v ¯ ( j ) t d 2 r s J 1 ( x y ) v ¯ ( j ) d y + m 2 v ¯ ( j ) = G ( u ¯ ( i 1 ) ) , ( x , t ) Ω k , u ¯ ( j ) ( x , 0 ) = u ¯ ( j 1 ) ( x , τ ) , u ¯ ( j ) ( x , ( k τ ) + ) = H ( u ¯ ( j 1 ) ( x , ( k + 1 ) τ ) ) , x [ r , s ] , v ¯ ( j ) ( x , 0 ) = v ¯ ( j 1 ) ( x , τ ) , v ¯ ( j ) ( x , ( k τ ) + ) = v ¯ ( j 1 ) ( x , ( k + 1 ) τ ) , x [ r , s ] ,

where Ω k = [ r , s ] × ( ( k τ ) + , ( k + 1 ) τ ] , m 1 = d 1 + a 11 , and m 2 = d 2 + a 22 . Next, we prove that the obtained sequence { ( u ¯ ( j ) , v ¯ ( j ) ) } j = 0 + is monotonically decreasing. Let

ϕ ( j ) = u ¯ ( j ) u ¯ ( j + 1 ) and ψ ( j ) = v ¯ ( j ) v ¯ ( j + 1 ) .

Then, the initial conditions satisfy

ϕ ( 0 ) ( x , 0 ) = u ¯ ( 0 ) ( x , 0 ) u ¯ ( 1 ) ( x , 0 ) = u ¯ ( 0 ) ( x , 0 ) u ¯ ( 0 ) ( x , τ ) = 0 , x [ r , s ] , ψ ( 0 ) ( x , 0 ) = v ¯ ( 0 ) ( x , 0 ) v ¯ ( 1 ) ( x , 0 ) = v ¯ ( 0 ) ( x , 0 ) v ¯ ( 0 ) ( x , τ ) = 0 , x [ r , s ] .

The impulse conditions satisfy

ϕ ( 0 ) ( x , 0 + ) = u ¯ ( 0 ) ( x , 0 + ) u ¯ ( 1 ) ( x , 0 + ) = u ¯ ( 0 ) ( x , 0 + ) H ( u ¯ ( 0 ) ( x , τ ) ) 0 , x [ r , s ] , ψ ( 0 ) ( x , 0 + ) = v ¯ ( 0 ) ( x , 0 + ) v ¯ ( 1 ) ( x , 0 + ) = v ¯ ( 0 ) ( x , 0 + ) v ¯ ( 0 ) ( x , τ ) = 0 , x [ r , s ] .

The equations satisfy

ϕ t ( 0 ) d 1 r s J 1 ( x y ) ϕ ( 0 ) ( y , t ) d y d 1 ϕ ( 0 ) a 11 ϕ ( 0 ) , ( x , t ) Ω 0 , ψ t ( 0 ) d 2 r s J 1 ( x y ) ψ ( 0 ) ( y , t ) d y d 2 ψ ( 0 ) a 22 ϕ ( 0 ) , ( x , t ) Ω 0 .

Based on these, Lemma 3.6 yields that ( ϕ ( 0 ) , ψ ( 0 ) ) 0 for x [ r , s ] and t [ 0 , τ ] . Taking ( ϕ ( 0 ) ( x , τ + ) , ψ ( 0 ) ( x , τ + ) ) as a fresh start function for t ( τ + , 2 τ ] , a similar discussion yields that ( ϕ ( 0 ) , ψ ( 0 ) ) 0 for ( x , t ) [ r , s ] × ( τ , 2 τ ] . From the method of induction, ( ϕ ( 0 ) , ψ ( 0 ) ) 0 for ( x , t ) Ω k . Using the method of induction again, the sequence ( ϕ ( j ) , ψ ( j ) ) 0 for ( x , t ) Ω k , which implies that { ( u ¯ ( j ) , v ¯ ( j ) ) } j = 0 + is monotonically decreasing. Noting that ( u ¯ ( 0 ) , v ¯ ( 0 ) ) is an upper solution of (3.1), applying Lemma 3.6 to ( u ¯ ( 0 ) u , v ¯ ( 0 ) v ) yields that

(3.29) ( u , v ) ( u ¯ ( 0 ) , v ¯ ( 0 ) ) for all ( x , t ) Ω k .

Then, it follows from (3.29) that

(3.30) ( u ( x , τ ) , v ( x , τ ) ) ( u ¯ ( 0 ) ( x , τ ) , v ¯ ( 0 ) ( x , τ ) ) for all x [ r , s ] .

This together with ( u ¯ ( 0 ) ( x , τ ) , v ¯ ( 0 ) ( x , τ ) ) = ( u ¯ ( 1 ) ( x , 0 ) , v ¯ ( 1 ) ( x , 0 ) ) yields that

(3.31) ( u ( x , τ ) , v ( x , τ ) ) ( u ¯ ( 1 ) ( x , 0 ) , v ¯ ( 1 ) ( x , 0 ) ) for all x [ r , s ] .

Recalling that H is a monotonically increasing function with respect to u , it follows from (3.30), the third equation of (3.1), and the last two equations of (3.27) that

(3.32) ( u ( x , τ + ) , v ( x , τ + ) ) ( H ( u ( x , τ ) ) , v ( x , τ ) ) ( H ( u ¯ ( 0 ) ( x , τ ) ) , v ¯ ( 0 ) ( x , τ ) ) ( u ¯ ( 1 ) ( x , 0 + ) , v ¯ ( 1 ) ( x , 0 + ) )

for all x [ r , s ] . In addition, (3.29) and the first two equations of (3.1) and (3.27) yield that

(3.33) ζ t d 1 r s J 1 ( x y ) ζ ( y , t ) d y d 1 ζ a 11 ζ , ( x , t ) Ω 0 , η t d 2 r s J 1 ( x y ) η ( y , t ) d y d 2 η a 22 η , ( x , t ) Ω 0 ,

where ( ζ ( x , t ) , η ( x , t ) ) = ( u ¯ ( 1 ) ( x , t ) u ( x , t + τ ) , v ¯ ( 1 ) ( x , t ) v ( x , t + τ ) ) . Based on (3.31)–(3.33), applying Lemma 3.6 to ( u ¯ ( 1 ) ( x , t ) u ( x , t + τ ) , v ¯ ( 1 ) ( x , t ) v ( x , t + τ ) ) yields that

u ( x , t + τ ) u ¯ ( 1 ) ( x , t ) and v ( x , t + τ ) v ¯ ( 1 ) ( x , t )

for all ( x , t ) [ r , s ] × [ 0 , τ ] . By the method of induction, it follows that

u ( x , t + τ ) u ¯ ( 1 ) ( x , t ) and v ( x , t + τ ) v ¯ ( 1 ) ( x , t )

for all ( x , t ) [ r , s ] × [ 0 , + ) . By using mathematical induction again, we arrive at

(3.34) u ( x , t + m τ ) u ¯ ( m ) ( x , t ) and v ( x , t + m τ ) v ¯ ( m ) ( x , t )

for all ( x , t ) [ r , s ] × [ 0 , τ ] , where m is an arbitrary natural number. Recalling that ( 0 , 0 ) is the unique solution to problem (3.25) when λ 1 = 0 , it follows from (3.34) and the non-negativity of the solution to problem (3.1) that

lim m + ( u ( x , t + m τ ) , v ( x , t + m τ ) ) = ( 0 , 0 ) for all ( x , t ) [ r , s ] × [ 0 , + ) ,

which implies that

lim t + ( u ( x , t ) , v ( x , t ) ) = ( 0 , 0 ) uniformly for x [ r , s ] .

(2) We first prove the first assertion. To begin with, we construct the following functions:

( U ¯ ( x , t ) , V ¯ ( x , t ) ) = ( C 1 , C 2 ) for all ( x , t ) [ r , s ] × [ 0 , τ ] ,

where C 1 and C 2 are defined in (2.3). For any λ 1 < 0 , there exist θ > 0 such that λ 1 + θ < 0 . On this basis, we define

U ̲ ( x , t ) = ε ζ 1 ( x , t ) , x [ r , s ] , t = 0 , ε e ( λ 1 + θ ) τ ζ 1 ( x , t ) , x [ r , s ] , t = 0 + , ε e ( λ 1 + θ ) ( τ t ) ζ 1 ( x , t ) , x [ r , s ] , t ( 0 + , τ ] ,

and

V ̲ ( x , t ) = ε η 1 ( x , t ) , x [ r , s ] , t = 0 , ε e ( λ 1 + θ ) τ η 1 ( x , t ) , x [ r , s ] , t = 0 + , ε e ( λ 1 + θ ) ( τ t ) η 1 ( x , t ) , x [ r , s ] , t ( 0 + , τ ] ,

where ( ζ 1 , η 1 ) is the unique eigenfunction pair corresponding to the principal eigenvalue λ 1 , and ε is a positive number such that

e ( λ 1 + θ ) τ H ( 0 ) ε ζ 1 ( x , 0 ) H ( ε ζ 1 ( x , 0 ) ) 0

and

G ( 0 ) ε e ( λ 1 + θ ) ( τ t ) ζ 1 ( x , t ) G ( ε e ( λ 1 + θ ) ( τ t ) ζ 1 ( x , t ) ) ε θ min Ω 0 η 1 ( x , t ) .

A standard calculation shows that the constructed ( U ¯ , V ¯ ) and ( U ̲ , V ̲ ) are the ordered upper and lower solutions of the steady-state problem (3.25), respectively.

The two sequences of iteration { ( U ¯ ( j ) , V ¯ ( j ) ) } j = 0 + and { ( U ̲ ( j ) , V ̲ ( j ) ) } j = 0 + may then be derived from the iteration law (3.28) by choosing ( U ¯ ( 0 ) , V ¯ ( 0 ) ) = ( U ¯ , V ¯ ) and ( U ̲ ( 0 ) , V ̲ ( 0 ) ) = ( U ̲ , V ̲ ) , respectively. Using a similar discussion as in case 2 in (1), it follows that

U ̲ U ̲ ( m 1 ) U ̲ ( m ) U ¯ ( m ) U ¯ ( m 1 ) U ¯ and V ̲ V ̲ ( m 1 ) V ̲ ( m ) V ¯ ( m ) V ¯ ( m 1 ) V ¯ ,

where m Z + . By the monotone boundedness theorem, it follows that

lim m + ( U ̲ ( m ) , V ̲ ( m ) ) = ( U 1 , V 1 ) and lim m + ( U ¯ ( m ) , V ¯ ( m ) ) = ( U 2 , V 2 ) ,

where ( U 1 , V 1 ) and ( U 2 , V 2 ) are the solutions of the steady-state problem (3.25). By standard methods (see, e.g., the proof of [45, Theorem 4.2]), it follows that ( U 1 , V 1 ) and ( U 2 , V 2 ) are the maximum and minimum solutions of the steady-state problem (3.25), respectively.

It remains to show that the positive periodic solution of problem (3.25) is unique. Arguing indirectly, suppose that ( U ˜ , V ˜ ) and ( U ¨ , V ¨ ) are different solutions of problem (3.25). Since the two different solutions are both strongly positive,

ω 0 min { ω 1 ω ( U ˜ , V ˜ ) ( U ¨ , V ¨ ) for ( x , t ) [ r , s ] × [ 0 , τ ] }

is well defined and finite. In addition, we have that ω 0 ( U ˜ , V ˜ ) ( U ¨ , V ¨ ) for ( x , t ) [ r , s ] × [ 0 , τ ] and there exists some ( x 0 , t 0 ) [ r , s ] × ( 0 + , τ ] such that

(i) ω 0 U ˜ ( x 0 , t 0 ) = U ¨ ( x 0 , t 0 ) or (ii) ω 0 V ˜ ( x 0 , t 0 ) = V ¨ ( x 0 , t 0 ) .

To simplify notation, let

P ( x , t ) = ω 0 U ˜ ( x , t ) U ¨ ( x , t ) and Q ( x , t ) = ω 0 V ˜ ( x , t ) V ¨ ( x , t ) .

If (i) holds, then

0 P t ( x 0 , t 0 ) = d 1 r s J 1 ( x 0 y ) P ( y , t 0 ) d y ( d 1 + a 11 ) P ( x 0 , t 0 ) + a 12 Q ( x 0 , t 0 ) a 12 Q ( x 0 , t 0 ) .

This implies that the aforementioned inequality can hold only if Q ( x 0 , t 0 ) = 0 , that is, (ii) holds. We claim ω 0 = 1 . If not, ω 0 > 1 . From the condition ω 0 > 1 and assumption (G), it follows that ω 0 G ( U ˜ ) > G ( ω 0 U ˜ ) . If (ii) holds, then

0 Q t ( x 0 , t 0 ) = d 1 r s J 1 ( x 0 y ) Q ( y , t 0 ) d y ( d 2 + a 22 ) Q ( x 0 , t 0 ) + ω 0 G ( U ˜ ( x 0 , t 0 ) ) G ( U ¨ ( x 0 , t 0 ) ) ω 0 G ( U ˜ ( x 0 , t 0 ) ) G ( U ¨ ( x 0 , t 0 ) ) > G ( ω 0 U ˜ ( x 0 , t 0 ) ) G ( U ¨ ( x 0 , t 0 ) ) = 0 ,

which is impossible. Hence, ω 0 = 1 , which means that ( U ˜ , V ˜ ) ( U ¨ , V ¨ ) for ( x , t ) [ r , s ] × [ 0 , τ ] . By exchanging ( U ˜ , V ˜ ) and ( U ¨ , V ¨ ) , the conclusion that ( U ¨ , V ¨ ) ( U ˜ , V ˜ ) for ( x , t ) [ r , s ] × [ 0 , τ ] may also be obtained. As a result, ( U ¨ , V ¨ ) = ( U ˜ , V ˜ ) for x [ r , s ] and t [ 0 , τ ] . This completes the proof of the first assertion.

Next, we prove the second assertion. A standard calculation yields that ( U ¯ , V ¯ ) is an upper solution of (3.1). Recall that ( u 0 ( x ) , v 0 ( x ) ) , 0 for x [ r , s ] . We assume, without loss of generality, that ( u 0 ( x ) , v 0 ( x ) ) 0 for x [ r , s ] . Otherwise, Lemma 3.6 yields that ( u ( x , τ ) , v ( x , τ ) ) 0 for x [ r , s ] , and it can be seen as a new initial function. By decreasing the ε if necessary, it follows from a standard calculation that ( U ̲ , V ̲ ) is a lower solution of (3.1). Then. Lemma 3.6 yields that the solution ( u , v ) of (3.1) satisfies the following estimate:

U ̲ u U ¯ and V ̲ v V ¯

for ( x , t ) [ r , s ] × [ 0 , + ) . By a similar discussion as in case 2 in (1), we have that

(3.35) U ̲ ( m ) ( x , t ) u ( x , t + m τ ) U ¯ ( m ) ( x , t ) and V ̲ ( m ) ( x , t ) v ( x , t + m τ ) V ¯ ( m ) ( x , t )

for ( x , t ) Ω k , where m = 0 , 1 , 2 , . With the help of the first assertion, we have that

(3.36) lim m + ( U ̲ ( m ) , V ̲ ( m ) ) = lim m + ( U ¯ ( m ) , V ¯ ( m ) ) = ( U , V ) for ( x , t ) [ r , s ] × [ 0 , τ ] .

Noting that ( U , V ) is periodic, (3.36) also holds for ( x , t ) [ r , s ] × [ 0 , + ) . Hence, it follows from (3.35) that

lim m + ( u ( x , t + m τ ) , v ( x , t + m τ ) ) = ( U ( x , t ) , V ( x , t ) ) uniformly for ( x , t ) [ r , s ] × [ 0 , + ) .

This proof is finished.□

Theorem 3.2 gives the asymptotical profiles of problem (3.1) when the infection environment [ r , s ] is fixed. Next, we will analyze the situation when the infection environment is increasing and eventually goes to infinity.

Lemma 3.7

Assume that [ r , s ] = [ l , l ] . Let λ 1 ( ) denote the eigenvalue of (3.2) as l goes to infinity. If λ 1 ( ) < 0 , then

lim l ( U l ( x , t ) , V l ( x , t ) ) = ( U 1 ( t ) , V 1 ( t ) ) locally u n i f o r m l y i n R ,

where ( U l , V l ) denotes the solution to problem (3.25) with [ r , s ] replaced by [ l , l ] , and ( U 1 , V 1 ) stands for the unique solution to problem

(3.37) ζ t = a 12 η a 11 ζ , t ( 0 + , τ ] , η t = G ( ζ ) a 22 η , t ( 0 + , τ ] , ( ζ ( 0 ) , η ( 0 ) ) = ( ζ ( τ ) , η ( τ ) ) , ( ζ ( 0 + ) , η ( 0 + ) ) = ( H ( ζ ( 0 ) ) , η ( 0 ) ) .

Proof

It is well known that the asymptotic profiles of problem (3.37) are determined by the eigenvalue μ 1 of the following eigenvalue problem:

ϕ t = a 12 φ a 11 ϕ + μ ϕ , t ( 0 + , τ ] , φ t = G ( 0 ) ϕ a 22 φ + μ φ , t ( 0 + , τ ] , ( ϕ ( 0 ) , φ ( 0 ) ) = ( ϕ ( τ ) , φ ( τ ) ) , ( ϕ ( 0 + ) , φ ( 0 + ) ) = ( H ( 0 ) ϕ ( 0 ) , φ ( 0 ) ) .

Now, we claim that if λ 1 ( ) < 0 , then μ 1 < 0 . Otherwise, assume that μ 1 0 and its corresponding non-negative eigenfunction is ( ϕ 1 ( t ) , φ 1 ( t ) ) . Next, it follows to check that for any given l > 0 , ( ζ , η , λ ̲ ) = ( ϕ 1 , φ 1 , 0 ) satisfies (3.20). By standard computation, for x [ l , l ] ,

d 1 l l J 1 ( x y ) ζ ( y , t ) d y d 1 ζ a 11 ζ + a 12 η ζ t , a 11 ϕ 1 + a 12 φ 1 ϕ 1 t 0 ,

and

d 1 l l J 2 ( x y ) η ( y , t ) d y d 2 η a 22 η + G ( 0 ) ζ η t , a 22 φ 1 + G ( 0 ) ϕ 1 φ 1 t 0 .

Then, Lemma 3.3 yields that λ 1 ( l ) > 0 for any l > 0 , which implies that λ 1 ( ) 0 . This contradicts with the condition λ 1 ( ) < 0 . Therefore, the claim holds by contraposition.

By using a similar proof method as in the first assertion of Theorem 3.2(2), it follows that when μ 1 < 0 , (3.37) has a unique positive solution ( U 1 , V 1 ) . Finally, the desired conclusion can be obtained by standard methods (see, e.g., the proof of [33, Lemma 3.11]). This ends the proof.□

4 The asymptotic profile in a moving environment

This section first presents some basic knowledge. Based on this, the vanishing-spreading dichotomy is proved and sufficient conditions used for determining spreading or vanishing are provided. Throughout this section, we always denote the solution of (2.1) by ( u , v , g , h ) or ( u ( x , t ) , v ( x , t ) , g ( t ) , h ( t ) ) .

4.1 Some introductory results

In here, the definition of the order upper and lower solutions of model (2.1), the comparison principle applicable to model (2.1), and the definition of the spreading and vanishing of model (2.1) will be presented in turn.

Definition 4.1

If for any T > 0 , g ¯ , h ¯ C [ 0 , T ] [ C 1 ( n τ , ( n + 1 ) τ ] C 1 ( N τ , T ] ] , u ¯ , v ¯ , u ¯ t , u ¯ t C ( Ω ¯ n τ ( n + 1 ) τ ) C ( Ω ¯ N τ T ) satisfy u ¯ , v ¯ 0 and

(4.1) u ¯ t d 1 g ¯ ( t ) h ¯ ( t ) J 1 ( x y ) u ¯ ( y , t ) d y d 1 u ¯ a 11 u ¯ + a 12 v ¯ , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , v ¯ t d 2 g ¯ ( t ) h ¯ ( t ) J 2 ( x y ) v ¯ ( y , t ) d y d 2 v ¯ a 22 v ¯ + G ( u ¯ ) , ( x , t ) Ω n τ ( n + 1 ) τ Ω N τ T , g ¯ ( t ) μ 1 g ¯ ( t ) h ¯ ( t ) g ¯ ( t ) J 1 ( x y ) u ¯ ( x , t ) d y d x μ 2 g ¯ ( t ) h ¯ ( t ) g ¯ ( t ) J 2 ( x y ) v ¯ ( x , t ) d y d x , t ( 0 , T ] , h ¯ ( t ) μ 1 g ¯ ( t ) h ¯ ( t ) h ¯ ( t ) + J 1 ( x y ) u ¯ ( x , t ) d y d x + μ 2 g ¯ ( t ) h ¯ ( t ) h ¯ ( t ) + J 2 ( x y ) v ¯ ( x , t ) d y d x , t ( 0 , T ] , u ¯ ( g ¯ ( t ) , t ) , u ¯ ( h ¯ ( t ) , t ) 0 , v ¯ ( g ¯ ( t ) , t ) , v ¯ ( h ¯ ( t ) , t ) 0 , t ( 0 , T ] , u ¯ ( x , 0 ) u 0 ( x ) , v ¯ ( x , 0 ) v 0 ( x ) , g ¯ ( 0 ) , h ¯ ( 0 ) h 0 , x [ h 0 , h 0 ] , u ¯ ( x , ( n τ ) + ) H ( u ¯ ( x , n τ ) ) , v ¯ ( x , ( n τ ) + ) v ¯ ( x , n τ ) , x ( g ¯ ( n τ ) , h ¯ ( n τ ) ) ,

then the quadruple ( u ¯ , v ¯ , g ¯ , h ¯ ) is called an upper solution of (2.1).

Remark 4.1

By reversing all the inequalities in (4.1), a lower solution of model (2.1) can also be defined. In this section, denote the upper and lower solutions of model (2.1) by the quadruples ( u ¯ , v ¯ , g ¯ , h ¯ ) and ( u ̲ , v ̲ , g ̲ , h ̲ ) , respectively.

Lemma 4.1

Denote the unique solution of model (2.1) by ( u , v , g , h ) . If the quadruples ( u ¯ , v ¯ , g ¯ , h ¯ ) and ( u ̲ , v ̲ , g ̲ , h ̲ ) are the ordered upper and lower solutions of model (2.1), respectively, then

g ¯ ( t ) g ( t ) g ̲ ( t ) , h ̲ ( t ) h ( t ) h ¯ ( t ) for t R + , u ̲ ( x , t ) u ( x , t ) u ¯ ( x , t ) , v ̲ ( x , t ) v ( x , t ) v ¯ ( x , t ) for x [ g ( t ) , h ( t ) ] , t R + .

Proof

Here, we only need to prove that the conclusion holds for the upper solution, due to the fact that a similar analysis yields that the conclusion also holds for the lower solution.

By using assumption ( H ), at the pulse moment t = 0 + , we have that

u ( x , 0 + ) C [ h 0 , h 0 ] , u ( ± h 0 , 0 + ) = 0 and u ( x , 0 + ) > 0 in ( h 0 , h 0 ) , v ( x , 0 + ) C [ h 0 , h 0 ] , v ( ± h 0 , 0 + ) = 0 and v ( x , 0 + ) > 0 in ( h 0 , h 0 ) .

By viewing ( u ( x , 0 + ) , v ( x , 0 + ) ) as an initial function, the standard discussion (see, e.g., the proof of [5, Theorem 3.1]) yields that the conclusion holds for t ( 0 + , τ ] , that is, the conclusion holds for t ( 0 , τ ] . From mathematical induction, it follows that the conclusion holds for t ( 0 , ) . The proof is completed.□

Theorem 2.1 and Lemma 2.2 show that ( u ( x , t ) , v ( x , t ) ) 0 for x ( g ( t ) , h ( t ) ) and t [ 0 , ) . This implies that h ( t ) , g ( t ) > 0 for t [ 0 , ) . Therefore,

lim t g ( t ) = g [ , h 0 ) and lim t h ( t ) = h ( h 0 , + ]

are always well-defined.

Definition 4.2

The model (2.1) is called as vanishing if

h g < and lim t u ( x , t ) + v ( x , t ) C ( R ) = 0 ,

and spreading if

h g = and lim t ( u ( x , t ) , v ( x , t ) ) = ( U 1 ( t ) , V 1 ( t ) ) locally uniformly in R ,

where ( U 1 , V 1 ) denotes the solution of (3.37).

4.2 Spreading-vanishing dichotomy

This subsection will prove that when the time t tends to infinity, model (2.1) is either vanishing or spreading. This conclusion will be discussed separately for the three cases: λ 1 ( , + ) > 0 , = 0 , and < 0 .

Lemma 4.2

Let λ 1 ( ) = λ 1 ( , + ) . If λ 1 ( ) > 0 , then a constant k 0 > 0 , which depends only the eigenvalue λ 1 ( ) , can be found such that

lim t e k 0 t u ( x , t ) + v ( x , t ) C ( R ) = 0 .

Moreover, we have that ( g , h ) is a finite interval.

Proof

The first conclusion is proved first. Define, for x [ g ( t ) , h ( t ) ] and t [ 0 , + ) ,

u ¯ ( x , t ) M e σ t ζ ( x , t ) , v ¯ ( x , t ) M e σ t η ( x , t ) ,

where M and σ are positive constants to be determined later, and ( ζ ( x , t ) , η ( x , t ) ) is the eigenfunction of (3.2), in which [ r , s ] is substituted by ( , + ) . Recalling that ( ζ ( x , t ) , η ( x , t ) ) X τ + + , M can be determined such that M ( ζ ( x , 0 ) , η ( x , 0 ) ) ( u 0 ( x ) , v 0 ( x ) ) for x [ h 0 , h 0 ] . Then, check the impulsive conditions. Clearly, assumption ( H ) gives that

u ¯ ( x , ( k τ ) + ) = M e σ k τ ζ ( x , ( k τ ) + ) = M e σ k τ H ( 0 ) ζ ( x , k τ ) > H ( u ¯ ( x , k τ ) ) , x R , v ¯ ( x , ( k τ ) + ) = M e σ k τ η ( x , ( k τ ) + ) = M e σ k τ ζ ( x , k τ ) = v ¯ ( x , k τ ) , x R .

Finally, assumption ( G ) yields that

u ¯ t d 1 g ( t ) h ( t ) J 1 ( x y ) u ¯ ( y , t ) d y + d 1 u ¯ + a 11 u ¯ a 12 v ¯ M e σ t σ ζ + ζ t d 1 + J 1 ( x y ) [ ζ ( t , y ) ζ ( t , x ) ] d y + a 11 ζ a 12 η M e σ t [ λ 1 ( ) ζ σ ζ ] 0 ,

and

v ¯ t d 2 g ( t ) h ( t ) J 2 ( x y ) v ¯ ( y , t ) d y + d 2 v ¯ + a 22 v ¯ G ( u ¯ ) M e σ t σ η + η t d 2 + J 2 ( x y ) [ η ( y , t ) η ( x , t ) ] d y + a 22 η G ( 0 ) ζ M e σ t [ λ 1 ( ) η σ η ] 0 ,

for x ( g ( t ) , h ( t ) ) and t ( ( k τ ) + , ( k + 1 ) τ ] provided that σ λ 1 ( ) 2 . Now, applying Lemma 3.6 to ( u ¯ u , v ¯ v ) yields that

(4.2) u u ¯ and v v ¯ for ( t , x ) R + × R .

Therefore, taking k 0 σ 2 , it follows from (4.2) that

lim t e k 0 t u ( x , t ) + v ( x , t ) C ( R ) = 0 .

This proves the first assertion.

Next, we prove the second assertion. The first assertion tells us that a constant T 0 can be found such that for any t > T 0 , it follows that

e k 0 t u ( x , t ) 1 4 ( μ 1 + μ 2 ) and e k 0 t v ( x , t ) 1 4 ( μ 1 + μ 2 )

for ( x , t ) [ g ( t ) , h ( t ) ] × ( T 0 , ) . This combined with assumption ( J ) yields that

h ( t ) g ( t ) = μ 1 g ( t ) h ( t ) h ( t ) + J 1 ( x y ) u ( x , t ) d y d x + μ 2 g ( t ) h ( t ) h ( t ) + J 2 ( x y ) u ( x , t ) d y d x + μ 1 g ( t ) h ( t ) g ( t ) J 1 ( x y ) u ( x , t ) d y d x + μ 2 g ( t ) h ( t ) g ( t ) J 2 ( x y ) v ( x , t ) d y d x 2 ( μ 1 + μ 2 ) g ( t ) h ( t ) u ( x , t ) d x + 2 ( μ 1 + μ 2 ) g ( t ) h ( t ) v ( x , t ) d x [ h ( t ) g ( t ) ] e k 0 t

for t > T 0 . Then, a standard calculation gives us follows:

h ( t ) g ( t ) [ h ( T ) g ( T ) ] e e k 0 T k 0 e k 0 t k 0 ,

which implies that

h g [ h ( T ) g ( T ) ] e e k 0 T k 0 < .

Therefore, the second assertion is valid. The proof is complected.□

Lemma 4.2 shows that model (2.1) is vanishing when λ 1 ( ) > 0 . Subsequently, we will discuss the case λ 1 ( ) = 0 .

Lemma 4.3

Assume that the region at infinity is particularly unfavourable for the survival of the infectious agents, that is, h g < . If λ 1 ( ) = 0 , then we have that

lim t u ( x , t ) + v ( x , t ) C [ g ( t ) , h ( t ) ] = 0 .

Proof

Let ( u ˜ , v ˜ ) be the unique solution of the problem

u ˜ t = a 11 u ˜ + a 12 v ˜ , t ( ( k τ ) + , ( k + 1 ) τ ] , v ˜ t = a 22 v ˜ + G ( u ˜ ) , t ( ( k τ ) + , ( k + 1 ) τ ] , u ˜ ( 0 ) = C 1 , v ˜ ( 0 ) = C 2 , u ˜ ( ( k τ ) + ) = H ( u ˜ ( k τ ) ) , v ˜ ( ( k τ ) + ) = v ˜ ( k τ ) ,

where C 1 and C 2 are defined in (2.3). Then, applying Lemma 3.6 to ( u ˜ u , v ˜ v ) yields that

(4.3) u u ˜ and v v ˜ for ( t , x ) R + × R .

From the condition λ 1 ( ) = 0 , it follows from Lemma 3.5 that μ 1 = 0 , where μ 1 is the principal eigenvalue of (3.24). Since μ 1 = 0 , (0, 0) is the only solution of problem (3.37). Finally, the upper and lower solutions method show that

lim t u ˜ ( t ) = lim t v ˜ ( t ) = 0 .

This together with (4.3) yields that

lim t u ( x , t ) + v ( x , t ) C ( R ) = 0 .

This ends the proof.□

Next, we will discuss the dynamic behaviour of model (2.1) when λ 1 ( ) < 0 . To do this, we start by proving the following two lemmas.

Lemma 4.4

If ( g , h ) is a finite interval, then

lim t u ( x , t ) = lim t v ( x , t ) = 0 uniformly f o r x [ g ( t ) , h ( t ) ]

and λ 1 ( g , h ) 0 , where λ 1 ( g , h ) is the eigenvalue of problem (3.2) in which [ r , s ] is substituted by [ g , h ] .

Proof

Begin by showing the second assertion. Arguing indirectly, suppose that λ 1 ( g , h ) < 0 . Lemma 3.4(2) gives that a sufficiently small positive constant ε can be found such that

λ 1 ( g + ε , h ε ) < 0 .

The assumption that J C ( R ) and J ( 0 ) > 0 in (J) yields that there exist sufficiently small positive constants ε 0 , δ 0 h 0 2 such that

J i ( x ) > δ 0 for x ( 3 ε 0 , 3 ε 0 ) , i = 1 , 2 .

Then, for given ε 1 < min { h 0 , 2 ε 0 } , a sufficiently large positive integer k 0 can be found such that

(4.4) ( g , h ε 1 ) ( g ( t ) , h ( t ) ) ( g + ε 1 , h ) for t k 0 τ .

Now, we consider the following problem:

u ˜ t = d 1 g + ε 1 h ε 1 J 1 ( x y ) u ˜ ( y , t ) d y d 1 u ˜ a 11 u ˜ + a 12 v ˜ , x Γ , t ( ( k τ ) + , ( k + 1 ) τ ] , v ˜ t = d 2 g + ε 1 h ε 1 J 2 ( x y ) v ˜ ( y , t ) d y d 2 v ˜ a 22 v ˜ + G ( u ˜ ) , x Γ , t ( ( k τ ) + , ( k + 1 ) τ ] , u ˜ ( x , ( k τ ) + ) = H ( u ˜ ( x , k τ ) ) , v ˜ ( x , ( k τ ) + ) = v ˜ ( x , k τ ) , x Γ , u ˜ ( x , k 0 τ ) = u ( x , k 0 τ ) , v ˜ ( x , k 0 τ ) = v ( x , k 0 τ ) , x Γ , k = k 0 , k 0 + 1 , ,

where Γ = [ g + ε 1 , h ε 1 ] . It follows from (4.4) that

g + ε 1 h ε 1 J 1 ( x y ) u ( y , t ) d y g ( t ) h ( t ) J 1 ( x y ) u ( y , t ) d y for x Γ , t [ k 0 τ , ) .

Therefore, from Lemma 3.6, we have that

u ˜ ( x , t ) u ( x , t ) and v ˜ ( x , t ) v ( x , t ) for x Γ and t [ k 0 τ , ) .

Noting that λ 1 ( g + ε 1 , h ε 1 ) < 0 , it follows from Theorem 3.2(2) that

lim m + ( u ˜ ( x , t + m τ ) , v ˜ ( x , t + m τ ) ) = ( U ε 1 ( x , t ) , V ε 1 ( x , t ) ) uniformly for ( x , t ) Γ × [ 0 , τ ] ,

where ( U ε 1 , V ε 1 ) is the only non-negative periodic solution of (3.25) in which [ r , s ] is substituted by [ g + ε 1 , h ε 1 ] . Then, an integer k 1 k 0 can be found such that for any k k 1 , we have that

1 4 ( U ε 1 ( x , t ) , V ε 1 ( x , t ) ) ( u ˜ ( x , t + k τ ) , v ˜ ( x , t + k τ ) ) ( u ( x , t + k τ ) , v ( x , t + k τ ) )

for x Γ and t [ 0 , τ ] . To simplify notation, let

ϑ < min { min Γ × [ 0 , τ ] U ε 1 ( x , t ) , min Γ × [ 0 , τ ] V ε 1 ( x , t ) } .

For t k 1 τ , we have that

h ( t ) = μ 1 g ( t ) h ( t ) h ( t ) + J 1 ( x y ) u ( x , t ) d y d x + μ 2 g ( t ) h ( t ) h ( t ) + J 2 ( x y ) v ( x , t ) d y d x μ 1 h ( t ) 2 ε 0 h ( t ) h ( t ) h ( t ) + ε 0 J 1 ( x y ) u ( x , t ) d y d x + μ 2 h ( t ) 2 ε 0 h ( t ) h ( t ) h ( t ) + ε 0 J 2 ( x y ) v ( x , t ) d y d x μ 1 ε 0 δ 1 h ( t ) 2 ε 0 h ( t ) u ( x , t ) d x + μ 2 ε 0 δ 1 h ( t ) 2 ε 0 h ( t ) v ( x , t ) d x μ 1 ε 0 δ 1 h 2 ε 0 h ε 1 u ( x , t ) d x + μ 2 ε 0 δ 1 h 2 ε 0 h ε 1 v ( x , t ) d x ( 2 ε 0 ε 1 ) ( μ 1 + μ 2 ) ε 0 δ 1 ϑ > 0 ,

which implies that h = + . A similar process yields that g = . Therefore, we have that ( h , g ) = R . This contradicts the assumption that ( g , h ) is a finite interval. The proof of the second assertion is now completed.

The first assertion is shown next. Let ( u ¯ ( x , t ) , v ¯ ( x , t ) ) be the solution of (3.1) in which [ r , s ] and ( u ¯ ( x , 0 ) , v ¯ ( x , 0 ) ) are substituted by [ g , h ] and ( u 0 , v 0 ) , respectively. Lemma 3.6 yields that

0 ( u ( x , t ) , v ( x , t ) ) ( u ¯ ( x , t ) , v ¯ ( x , t ) ) for x [ g , h ] , t [ 0 , ) .

With the help of Theorem 3.2(1), it follows that

lim t + u ¯ ( x , t ) + v ¯ ( x , t ) C [ g , h ] = 0 ,

and hence the first assertion holds. The proof is finished.□

Lemma 4.4 implies that if λ 1 ( ) < 0 , then h g = . The following Lemma 4.5 will show us that if λ 1 ( ) < 0 and h g = , then h = g = .

Lemma 4.5

If λ 1 ( ) < 0 , then g < + iff h < + .

Proof

Without loss of generality, we suppose by contradiction that g < and h = . Noting that the principal eigenvalue λ 1 has translation invariance with respect to the space variable x , it follows from λ 1 ( ) < 0 that for any given sufficiently small positive number ε , a sufficiently large l ˜ can be found such that λ 1 ( g + ε , l ˜ ) < 0 . For such ε and l ˜ , a sufficiently large k 0 can be found such that

h ( t ) > l ˜ , g ( t ) < g + ε , t k 0 τ .

Consider the following problem:

u ¯ t = d 1 g + ε l ˜ J 1 ( x y ) u ¯ ( y , t ) d y d 1 u ¯ a 11 u ¯ + a 12 v ¯ , x ( g + ε , l ˜ ) , t ( ( k τ ) + , ( k + 1 ) τ ] , v ¯ t = d 2 g + ε l ˜ J 2 ( x y ) v ¯ ( y , t ) d y d 2 v ¯ a 22 v ¯ + G ( u ¯ ) , x ( g + ε , l ˜ ) , t ( ( k τ ) + , ( k + 1 ) τ ] , u ¯ ( x , ( k τ ) + ) = H ( u ¯ ( x , k τ ) ) , v ¯ ( x , ( k τ ) + ) = v ¯ ( x , k τ ) , x ( g + ε , l ˜ ) , u ¯ ( x , k 0 τ ) = u ( x , k 0 τ ) , v ¯ ( x , k 0 τ ) = v ( x , k 0 τ ) , x ( g + ε , l ˜ ) , k = k 0 , k 0 + 1 , .

Then, ( u ( x , t ) , v ( x , t ) ) ( u ¯ ( x , t ) , u ¯ ( x , t ) ) for x [ g + ε , l ˜ ] and t k 0 τ by Lemma 3.6. Using methods similar to the proof of Lemma 4.4, there exist small constant c * and large integer k 1 > k 0 such that g ( t ) c * < 0 for t k 1 τ . Therefore, we have that g = , which shows a contradiction with the assumption g < . The proof is completed.□

Now, we begin to analyze the dynamical behaviour of model (2.1) when λ 1 ( ) < 0 .

Lemma 4.6

If λ 1 ( g , h ) < 0 , then

lim m ( u ( x , t + m τ ) , v ( x , t + m τ ) ) = ( U 1 ( t ) , V 1 ( t ) ) in [ C loc ( R × [ 0 , τ ] ) ] 2

where ( U 1 , V 1 ) denotes the unique solution to problem (3.37), and ( h , g ) = R .

Proof

In virtue of Lemma 4.4, it follows from λ 1 ( ) < 0 that ( h , g ) is an infinite interval. Then, Lemma 4.5 yields that ( h , g ) = R .

Next verify the first claim by showing

(4.5) limsup m ( u ( x , t + m τ ) , v ( x , t + m τ ) ) ( U 1 ( t ) , V 1 ( t ) )

uniformly in R × [ 0 , τ ] and

(4.6) liminf m ( u ( x , t + m τ ) , v ( x , t + m τ ) ) ( U 1 ( t ) , V 1 ( t ) )

locally uniformly in R × [ 0 , τ ] . We first prove that (4.5) holds. From Lemma 3.6, it follows that

(4.7) u u ˜ and v v ˜ for ( x , t ) R × R + ,

where ( u ˜ , v ˜ ) is defined in the proof of Lemma 4.3. Therefore, it follows from (4.7) that

(4.8) limsup m ( u ( x , t + m τ ) , v ( x , t + m τ ) ) lim m ( u ˜ ( t + m τ ) , v ˜ ( t + m τ ) )

uniformly for x [ g ( t + m τ ) , h ( t + m τ ) ] and t R + . By using methods similar to the proof of Theorem 3.2(2), we find that

(4.9) lim m ( u ˜ ( t + m τ ) , v ˜ ( t + m τ ) ) = ( U 1 ( t ) , V 1 ( t ) )

uniformly in [ 0 , τ ] . Hence, (4.8) and (4.9) yield that (4.5) holds.

We then prove that (4.6) also holds. Since λ 1 ( ) < 0 , a sufficiently large l ˜ can be found such that λ 1 ( l ˜ , l ˜ ) < 0 . Noticing that h = g = , there exists a large k 0 such that ( h ( t ) , g ( t ) ) ( l ˜ , l ˜ ) for t k 0 τ . Consider the problem

(4.10) u ¯ t = d 1 l ˜ l ˜ J 1 ( x y ) u ¯ ( y , t ) d y d 1 u ¯ a 11 u ¯ + a 12 v ¯ , x [ l ˜ , l ˜ ] , t ( ( k τ ) + , ( k + 1 ) τ ] , v ¯ t = d 2 l ˜ l ˜ J 2 ( x y ) v ¯ ( y , t ) d y d 2 v ¯ a 22 v ¯ + G ( u ¯ ) , x [ l ˜ , l ˜ ] , t ( ( k τ ) + , ( k + 1 ) τ ] , u ¯ ( x , ( k τ ) + ) = H ( u ¯ ( x , k τ ) ) , v ¯ ( x , ( k τ ) + ) = v ¯ ( x , k τ ) , x [ l ˜ , l ˜ ] , u ¯ ( x , k 0 τ ) = u ( x , k 0 τ ) , v ¯ ( x , k 0 τ ) = v ( x , k 0 τ ) , x [ l ˜ , l ˜ ] , k = k 0 , k 0 + 1 , .

Then, it follows from Lemma 3.6 that

(4.11) u ( x , t ) u ¯ ( x , t ) and v ( x , t ) v ¯ ( x , t ) for x [ l ˜ , l ˜ ] and t k 0 τ .

Since λ 1 ( l ˜ , l ˜ ) < 0 , Theorem 3.2(2) gives us that (4.10) has a unique positive steady-state ( U l ˜ ( x , t ) , V l ˜ ( x , t ) ) defined in problem (3.25) with [ r , s ] replaced by [ l ˜ , l ˜ ] , and

(4.12) lim m + ( u ¯ ( x , t + m τ ) , v ¯ ( x , t + m τ ) ) = ( U l ˜ ( x , t ) , V l ˜ ( x , t ) )

uniformly in [ 0 , τ ] × [ l ˜ , l ˜ ] . From (4.11) and (4.12), it follows that

(4.13) liminf m ( u ( x , t + m τ ) , v ( x , t + m τ ) ) ( U l ˜ ( x , t ) , V l ˜ ( x , t ) )

uniformly in [ 0 , τ ] × [ l ˜ , l ˜ ] . Finally, when l ˜ , it follows from Lemma 3.7 that

liminf m ( u ( x , t + m τ ) , v ( x , t + m τ ) ) ( U 1 ( t ) , V 1 ( t ) )

locally uniformly in R × [ 0 , τ ] . This completes the proof.□

On the basis of Lemmas 4.2,4.3 and 4.6, we now present the main result of this subsection.

Theorem 4.1

The model (2.1) is either vanishing or spreading.

4.3 Spreading-vanishing criteria

In this subsection, we will look for criteria guaranteeing vanishing or spreading for (2.1). For convenience, let λ 1 ( h 0 ) and λ 1 ( ) denote the principal eigenvalues of (3.2) with [ r , s ] replaced by [ h 0 , h 0 ] and ( , + ) , respectively.

Based on Lemmas 4.2 and 4.3, the following conclusion can be obtained directly.

Theorem 4.2

If λ 1 ( ) 0 , then extinction happens, that is,

lim t u ( x , t ) + v ( x , t ) C [ g ( t ) , h ( t ) ] = 0 .

In addition, if λ 1 ( ) > 0 , then vanishing happens.

Next, we will consider the case of λ 1 ( ) < 0 . Its proof will be divided into two cases: λ 1 ( h 0 ) 0 and λ 1 ( h 0 ) > 0 .

Theorem 4.3

Assume that λ 1 ( ) < 0 . If λ 1 ( h 0 ) 0 , then spreading happens.

Proof

Suppose by contradiction that h < + or g > . Then, Lemma 4.5 yields that ( g , h ) is a finite interval. By using Lemma 4.4, it follows that λ 1 ( g , h ) 0 . Noticing that both g ( t ) and h ( t ) are strictly monotonically increasing, Lemma 3.4(2) gives that

λ 1 ( h 0 ) > λ 1 ( g , h ) 0 ,

which shows a contradiction with the condition λ 1 ( h 0 ) 0 . This proof is completed.□

Next, we will explore the case of λ 1 ( h 0 ) > 0 . To do this, we first provide the following lemma.

Lemma 4.7

Assume that λ 1 ( ) < 0 . If λ 1 ( h 0 ) > 0 and the initial value ( u 0 ( x ) , v 0 ( x ) ) of model (2.1) is sufficiently small, then vanishing happens.

Proof

A suitable upper solution of model (2.1) is constructed to prove this conclusion. Since λ 1 ( h 0 ) > 0 , Lemma 3.4(2) gives that a positive number ε can be found such that

λ 1 ( h 1 ) λ 1 ( h 1 , h 1 ) λ 1 ( h 0 ε , h 0 + ε ) > 0 .

Denote by ( ζ , η ) the eigenfunction of (3.2) in which [ r , s ] is substituted by [ h 1 , h 1 ] . For x [ h 1 , h 1 ] and t > 0 , define

h ¯ ( t ) h 1 ( h 1 h 0 ) e σ t , g ¯ ( t ) h ¯ ( t ) , u ¯ ( x , t ) M e σ t ζ ( x , t ) , v ¯ ( x , t ) M e σ t η ( x , t ) ,

where the non-negative constants M and σ are determined later.

We start by examining some conditions and determining M and σ before applying Lemma 4.1. Clearly,

( u ¯ ( x , t ) , v ¯ ( x , t ) ) 0 for ( x , t ) { g ¯ ( t ) , h ¯ ( t ) } × ( 0 , ) .

A standard calculation yields that, for x ( g ¯ ( t ) , h ¯ ( t ) ) and t ( ( k τ ) + , ( k + 1 ) τ ] ,

u ¯ t d 1 g ( t ) h ( t ) J 1 ( x y ) u ¯ ( y , t ) d y + d 1 u ¯ + a 11 u ¯ a 12 v ¯ M e σ t σ ζ + ζ t d 1 h 1 h 1 J 1 ( x y ) ζ ( y , t ) d y + d 1 ζ + a 11 ζ a 12 η M e σ t [ λ 1 ( h 1 ) ζ σ ζ ] 0 ,

and

v ¯ t d 2 g ( t ) h ( t ) J 2 ( x y ) v ¯ ( y , t ) d y + d 2 v ¯ + a 22 v ¯ G ( u ¯ ) M e σ t σ η + η t d 2 h 1 h 1 J 2 ( x y ) η ( y , t ) d y + d 2 η + a 22 η G ( 0 ) ζ M e σ t [ λ 1 ( h 1 ) η σ η ] 0 ,

provided that σ λ 1 ( h 1 ) 3 . On the basis of assumption ( H ), we have that for x ( g ¯ ( k τ ) , h ¯ ( k τ ) )

u ¯ ( x , ( k τ ) + ) = M e σ k τ ζ ( x , ( k τ ) + ) = M e σ k τ H ( 0 ) ζ ( x , k τ ) H ( u ¯ ( x , k τ ) ) ,

and

v ¯ ( x , ( k τ ) + ) = M e σ k τ η ( x , ( k τ ) + ) = M e σ k τ ζ ( x , k τ ) = v ¯ ( x , k τ ) .

In addition, assumption (J) yields that

μ 1 g ¯ ( t ) h ¯ ( t ) h ¯ ( t ) + J 1 ( x y ) u ¯ ( x , t ) d y d x + μ 2 g ¯ ( t ) h ¯ ( t ) h ¯ ( t ) + J 2 ( x y ) v ¯ ( x , t ) d y d x μ 1 g ¯ ( t ) h ¯ ( t ) u ¯ ( x , t ) d x + μ 2 g ¯ ( t ) h ¯ ( t ) v ¯ ( x , t ) d x M e σ t μ 1 h 1 h 1 max [ h 1 , h 1 ] × [ 0 , τ ] ζ ( x , t ) d x + μ 2 h 1 h 1 max [ h 1 , h 1 ] × [ 0 , τ ] η ( x , t ) d x ( h 1 h 0 ) σ e σ t = h ¯ ( t ) for t > 0 ,

provided that

M ( h 1 h 0 ) σ μ 1 h 1 h 1 max [ h 1 , h 1 ] × [ 0 , τ ] ζ ( x , t ) d x + μ 2 h 1 h 1 max [ h 1 , h 1 ] × [ 0 , τ ] η ( x , t ) d x .

Similarly, we have that

μ 1 g ( t ) h ( t ) g ( t ) J 1 ( x y ) u ¯ ( x , t ) d y d x μ 2 g ( t ) h ( t ) g ( t ) J 2 ( x y ) v ¯ ( x , t ) d y d x g ¯ ( t ) for t > 0 .

For the initial value, if we choose ( u 0 ( x ) , v 0 ( x ) ) sufficiently small such that

( u 0 ( x ) C [ h 0 , h 0 ] , v 0 ( x ) C [ h 0 , h 0 ] ) M ( min [ h 0 , h 0 ] ζ ( x , 0 ) , min [ h 0 , h 0 ] η ( x , 0 ) ) ,

then

( u 0 ( x ) , v 0 ( x ) ) ( u ¯ ( x , 0 ) , v ¯ ( x , 0 ) ) for x [ h 0 , h 0 ] .

Now it is clear that the quadruple ( u ¯ , v ¯ , g ¯ , h ¯ ) is an upper solution of (2.1). Therefore, by Lemma 4.1, it follows that

( g , h ) ( lim t + g ¯ ( t ) , lim t + h ¯ ( t ) ) ( h 1 , h 1 ) ,

which implies that ( g , h ) is a finite interval. Therefore, Lemma 4.4 yields that vanishing happens. This ends the proof.□

Remark 4.2

In the proof of Lemma 4.7, it follows from observing the definition of M that when μ 2 = ρ μ 1 , where ρ is a positive constant, we have that M as μ 1 0 . It means that, if λ 1 ( ) < 0 , λ 1 ( h 0 ) > 0 , and μ 2 = ρ μ 1 , for any ( u 0 ( x ) , v 0 ( x ) ) , a constant μ ̲ > 0 can be found such that when μ 1 ( 0 , μ ̲ ] vanishing happens.

Lemma 4.8

Assume that λ 1 ( ) < 0 . If λ 1 ( h 0 ) > 0 , then there exists l * > 0 such that λ 1 ( g , h ) < 0 if h g > 2 l * .

Proof

Since λ 1 ( ) < 0 and λ 1 ( h 0 ) > 0 , it follows from Lemma 3.4(2) that there exists a constant l * > 0 such that λ 1 ( l * , l * ) = 0 . Again using Lemma 3.4(2) yields that when h g > 2 l * , λ 1 ( g , h ) < 0 . The proof is completed.□

Lemma 4.9

Suppose that λ 1 ( ) < 0 . If λ 1 ( h 0 ) > 0 , then a positive constant μ ¯ can be found such that if μ 1 + μ 2 > μ ¯ spreading occurs.

Proof

Denote the unique solution of model (2.1) by ( u μ , v μ , g μ , h μ ) to emphasise the dependency on μ ( μ 1 , μ 2 ) . First, we claim that μ ˜ 0 can be found such that for some large t 0 > 0 ,

(4.14) λ 1 ( g μ ˜ ( t 0 ) , h μ ˜ ( t 0 ) ) < 0 .

If not, suppose that for any μ 0 and t > 0 , λ 1 ( g μ ( t ) , h μ ( t ) ) 0 . We will derive a contradiction.

By Theorem 2.1, we have that g μ ( t ) and h μ ( t ) are strongly increasing with respect to t > 0 . In addition, it follows from Lemma 4.1 that g μ ( t ) and h μ ( t ) are non-decreasing with respect to μ 0 . Hence,

G lim μ lim t g μ ( t ) and H lim μ lim t h μ ( t )

are well defined. By Lemma 4.8, it follows that H G 2 l * . Assumption (J) yields that the sufficiently small ε 0 > 0 and δ 0 > 0 can be found such that

J i ( x ) δ 0 > 0 for x [ ε 0 , ε 0 ] and i = 1 , 2 .

Then, for the fixed ε 0 , we can find μ 0 0 and k 0 N + such that

h μ ( t ) + ε 0 4 > H for μ μ 0 and t k 0 τ .

By integrating both sides of the fourth equation in the model over [ k 0 τ , ( k 0 + 1 ) τ ] , it follows that

h μ ( ( k 0 + 1 ) τ ) h μ ( k 0 τ ) = k 0 τ ( k 0 + 1 ) τ μ 1 g μ ( t ) h μ ( t ) h μ ( t ) + J 1 ( x y ) u ( x , t ) d y d x + μ 2 g μ ( t ) h μ ( t ) h μ ( t ) + J 2 ( x y ) v ( x , t ) d y d x d t k 0 τ ( k 0 + 1 ) τ μ 1 g μ 0 ( t ) h μ 0 ( t ) h μ 0 ( t ) + ε 0 4 + J 1 ( x y ) u ( x , t ) d y d x + μ 2 g μ 0 ( t ) h μ 0 ( t ) h μ 0 ( t ) + ε 0 4 + J 2 ( x y ) v ( x , t ) d y d x d t k 0 τ ( k 0 + 1 ) τ μ 1 h μ 0 ( t ) ε 0 2 h μ 0 ( t ) h μ 0 ( t ) + ε 0 4 h μ 0 ( t ) + ε 0 2 J 1 ( x y ) u ( x , t ) d y d x + μ 2 h μ 0 ( t ) ε 0 2 h μ 0 ( t ) h μ 0 ( t ) + ε 0 4 h μ 0 ( t ) + ε 0 2 J 2 ( x y ) v ( x , t ) d y d x d t ε 0 δ 0 4 k 0 τ ( k 0 + 1 ) τ μ 1 h μ 0 ( t ) ε 0 2 h μ 0 ( t ) u ( x , t ) d x + μ 2 h μ 0 ( t ) ε 0 2 h μ 0 ( t ) v ( x , t ) d x d t ( μ 1 + μ 2 ) δ 0 τ A ε 0 2 8 ,

where

A min min [ h μ 0 ( t ) ε 0 2 , h μ 0 ( t ) ] × [ k 0 τ , ( k 0 + 1 ) τ ] u ( x , t ) , min [ h μ 0 ( t ) ε 0 2 , h μ 0 ( t ) ] × [ k 0 τ , ( k 0 + 1 ) τ ] v ( x , t ) .

This implies that

μ 1 + μ 2 8 ( h μ ( ( k 0 + 1 ) τ ) h μ ( k 0 τ ) ) δ 0 τ A ε 0 2 32 l * δ 0 τ A ε 0 2 < ,

which shows a contradiction with the assumption. Therefore, (4.14) holds. Finally, it follows from Lemma 4.6 that the conclusion holds. The proof is completed.□

Now we consider the case λ 1 ( h 0 ) > 0 .

Theorem 4.4

Assume that λ 1 ( ) < 0 . If λ 1 ( h 0 ) > 0 , then there exist μ ̲ > 0 and μ ¯ > 0 such that vanishing occurs when μ 1 + μ 2 μ ̲ and spreading happens when μ 1 + μ 2 > μ ¯ . In addition, if μ 2 = ρ μ 1 where ρ is a positive constant, then a positive constant μ * can be found such that vanishing occurs when μ 1 μ * and spreading happens when μ 1 > μ * .

Proof

We first prove the first assertion. In the proof of Lemma 4.7, we have from observing the definition of M that as μ 1 + μ 2 0 M . This implies that for any given pair of initial function ( u 0 ( x ) , v 0 ( x ) ) of model (2.1), a sufficiently small positive constant μ ̲ can be found such that when μ 1 + μ 2 μ ̲ vanishing occurs. The remainder of the first conclusion is a direct result of Lemma 4.9. For the proof of the second assertion, it can be obtained by a method similar to that in [5, Theorem 3.14], and we are omitting it here. This proof is completed.□

5 Numerical simulation

In this section, we will present a numerical example to demonstrate the theoretical outcomes and further understand the effect of the impulse intervention on the spread of the faecal-oral diseases.

Since the relationship between the infectious agents and the infected individuals is cooperative, we only give the numerical simulation of the infectious agents to save space. Now we choose the dispersal kernel function as follows:

(5.1) J i ( x ) = k e 1 x 3 2 1 , x < 3 , 0 , x 3 ,

where i = 1 , 2 and 1 k = + e 1 x 3 2 1 d x . The diagram for J i ( x ) ( i = 1 , 2 ) is shown as Figure 1.

Figure 1 
               Dispersal kernel function 
                     
                        
                        
                           
                              
                                 J
                              
                              
                                 i
                              
                           
                           
                              (
                              
                                 x
                              
                              )
                           
                           
                              (
                              
                                 i
                                 =
                                 1
                                 ,
                                 2
                              
                              )
                           
                        
                        {J}_{i}\left(x)\left(i=1,2)
                     
                   defined by (5.1).
Figure 1

Dispersal kernel function J i ( x ) ( i = 1 , 2 ) defined by (5.1).

The initial infection region is chosen as [ 2 , 2 ] and the initial functions are as follows

u 0 ( x ) = 3 cos π x 4 , v 0 ( x ) = cos π x 4 , x [ 2 , 2 ] .

The other parameters of model (2.1) are given in the following example.

Example 5.1

Fix d 1 = 0.10 , d 2 = 0.10 , τ = 1 , a 11 = 0.35 , a 12 = 0.11 , a 22 = 0.10 , μ 1 = 20 , μ 2 = 200 , and G ( u ) = 0.5 u ( 10 + u ) . H ( u ) is chosen as u and 0.1 u 10 + u , respectively.

When there is no the impulsive intervention, it can be seen from Figure 2(c) that h > h ( 24 ) > 9.75 . By using the theory in [10] and the method in [21,35], we obtain Figure 2(d). This combined with Lemma 3.4(2) yields that

λ 1 ( g , h ) < λ 1 ( 9.75 , 9.75 ) 0.22 < 0 .

Then, it follows from Lemma 4.6 that ( U 1 ( t ) , V 1 ( t ) ) is globally asymptotically stable. Actually, it follows from Figure 2(a) that the infectious agents indeed converge to a heterogeneous stable state, which is consistent with the conclusion of Lemma 4.6.

Figure 2 
               When 
                     
                        
                        
                           H
                           
                              (
                              
                                 u
                              
                              )
                           
                           =
                           u
                        
                        H\left(u)=u
                     
                  , which implies no pulse, 
                     
                        
                        
                           u
                        
                        u
                     
                   converges to a steady-state and 
                     
                        
                        
                           
                              
                                 λ
                              
                              
                                 1
                              
                           
                           <
                           0
                        
                        {\lambda }_{1}\lt 0
                     
                  . (a) 3D graph of 
                     
                        
                        
                           u
                           
                              (
                              
                                 x
                                 ,
                                 t
                              
                              )
                           
                        
                        u\left(x,t)
                     
                  , (b) distribution of the infectious agents 
                     
                        
                        
                           u
                           
                              (
                              
                                 x
                                 ,
                                 t
                              
                              )
                           
                        
                        u\left(x,t)
                     
                   at different times, (c) change of the right boundary 
                     
                        
                        
                           h
                           
                              (
                              
                                 t
                              
                              )
                           
                        
                        h\left(t)
                     
                   with time, (d) approximated principal eigenvalue 
                     
                        
                        
                           
                              
                                 λ
                              
                              
                                 1
                                 ,
                                 n
                              
                           
                        
                        {\lambda }_{1,n}
                     
                   (
                     
                        
                        
                           
                              
                                 λ
                              
                              
                                 1
                              
                           
                           ,
                           80
                           ≈
                           −
                           0.22
                        
                        {\lambda }_{1},80\approx -0.22
                     
                  ).
Figure 2

When H ( u ) = u , which implies no pulse, u converges to a steady-state and λ 1 < 0 . (a) 3D graph of u ( x , t ) , (b) distribution of the infectious agents u ( x , t ) at different times, (c) change of the right boundary h ( t ) with time, (d) approximated principal eigenvalue λ 1 , n ( λ 1 , 80 0.22 ).

When the impulse function is taken as H ( u ) = 0.1 u 10 + u , one can obtain from Figure 3(c) that h < 7.8 . Similarly, Figure 3(d) can be obtained. This and Lemma 3.4 tell us that

λ 1 ( H ( 0 ) , ( g , h ) ) = λ 1 ( 0.01 , ( g , h ) ) > λ 1 ( 1 , ( 7.8 , 7.8 ) ) 0.08 > 0 .

Then, we have from Lemma 4.2 that the disease-free equilibrium point ( 0 , 0 ) is globally asymptotically stable. Actually, it is not hard to see from Figure 3(a) that the infectious agents continuously converge to zero as time t increases, which is consistent with the conclusion obtained.

Figure 3 
               When 
                     
                        
                        
                           H
                           
                              (
                              
                                 u
                              
                              )
                           
                           =
                           
                              
                                 0.1
                                 u
                              
                              
                                 10
                                 +
                                 u
                              
                           
                        
                        H\left(u)=\frac{0.1u}{10+u}
                     
                  , 
                     
                        
                        
                           u
                        
                        u
                     
                   decays to 0 and 
                     
                        
                        
                           
                              
                                 λ
                              
                              
                                 1
                              
                           
                           >
                           0
                        
                        {\lambda }_{1}\gt 0
                     
                  . (a) 3D graph of 
                     
                        
                        
                           u
                           
                              (
                              
                                 x
                                 ,
                                 t
                              
                              )
                           
                        
                        u\left(x,t)
                     
                  , (b) distribution of the infectious agents 
                     
                        
                        
                           u
                           
                              (
                              
                                 x
                                 ,
                                 t
                              
                              )
                           
                        
                        u\left(x,t)
                     
                   at different times, (c) change of the right boundary 
                     
                        
                        
                           h
                           
                              (
                              
                                 t
                              
                              )
                           
                        
                        h\left(t)
                     
                   with time, (d) approximated principal eigenvalue 
                     
                        
                        
                           
                              
                                 λ
                              
                              
                                 1
                                 ,
                                 n
                              
                           
                        
                        {\lambda }_{1,n}
                     
                   (
                     
                        
                        
                           
                              
                                 λ
                              
                              
                                 1
                              
                           
                           ,
                           80
                           ≈
                           0.88
                        
                        {\lambda }_{1},80\approx 0.88
                     
                  ).
Figure 3

When H ( u ) = 0.1 u 10 + u , u decays to 0 and λ 1 > 0 . (a) 3D graph of u ( x , t ) , (b) distribution of the infectious agents u ( x , t ) at different times, (c) change of the right boundary h ( t ) with time, (d) approximated principal eigenvalue λ 1 , n ( λ 1 , 80 0.88 ).

It can be obtained from comparing Figures 2 and 3 that the implementation of the impulsive intervention can reduce the number of the infectious agents, slow down the expansion speed of the diseases, decrease the infected region of the diseases, and extinguish the diseases which are originally persistent. Therefore, the impulsive intervention is beneficial in combating the diseases transmitted by the faecal-oral route.

6 Conclusion and future work

In this article, we develop an impulsive non-local model to capture the effect of the combination of periodic spraying of disinfectant and non-local diffusion of individuals on the diseases transmitted by the faecal-oral route. Our model extends the model of Chang and Du [9], which does not take into account the pulse intervention. With the introduction of the pulse intervention, the model is not continuous at the pulse point and the corresponding steady-state is governed by a periodic parabolic problem, hence the techniques used here are rather different from those in [9].

Based on the work of Wang and Du [33], Theorem 2.1 shows that the model has a unique non-negative global classical solution. Then, the corresponding periodic eigenvalue problem (3.2) is presented to study the asymptotical profiles of the model. The existence and uniqueness of the principal eigenvalue of problem (3.2) is proved in Theorem 3.1 by using the theory of resolvent positive operators with their perturbations. With the help of this eigenvalue, Theorem 4.1 shows that model (2.1) is either vanishing or spreading. To provide the criteria for determining spreading and vanishing, the principal eigenvalues of the eigenvalue problem (3.2) with [ r , s ] replaced by [ h 0 , h 0 ] and ( , + ) are next defined in λ 1 ( h 0 ) and λ 1 ( ) , respectively. We obtain the following criteria:

  • If λ 1 ( ) > 0 , then the diseases are vanishing, see Lemma 4.2;

  • If λ 1 ( ) = 0 , then the diseases are vanishing, see Lemma 4.3;

  • If λ 1 ( ) < 0 and λ 1 ( h 0 ) 0 , then the diseases are spreading, see Theorem 4.3;

  • If λ 1 ( ) < 0 and λ 1 ( h 0 ) > 0 , then

    • there exist two positive constants μ ¯ and μ ̲ such that vanishing occurs when μ 1 + μ 2 μ ̲ and spreading happens when μ 1 + μ 2 > μ ¯ , see Theorem 4.4,

    • or if μ 2 = ρ μ 1 where ρ is a positive constant, then a positive constant μ * can be found such that vanishing occurs when μ 1 μ * and spreading happens when μ 1 > μ * , see Theorem 4.4.

Our theoretical results suggest that the implementation of the impulse intervention is beneficial in combating the diseases transmitted by the faecal-oral route. Numerical simulation is used to further understand this finding. In addition, the initial infected region also has a significant impact on the evolution of the diseases. Specifically, the smaller the initial infected area, the more advantageous for human beings to fight against the diseases. The effect of the non-local diffusion of individuals on the evolution of the diseases depends on the choice of the diffusion kernel functions.

When λ 1 ( ) = 0 , the conclusion that the diseases are vanishing is obtained under the assumption that the length of the infected region is finite. We strongly believe that this assumption can be removed. This will be the direction of our future research.

Acknowledgments

The authors are grateful to the anonymous reviewers for careful reading and valuable comments that helped to improve the article.

  1. Funding information: The first author is supported by Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX24_3711) and the third author acknowledges the support of the National Natural Science Foundation of China (No. 12271470).

  2. Author contributions: All the authors Qi Zhou, Michael Pedersen, and Zhigui Lin contribute equally.

  3. Conflict of interest: The authors state no conflict of interest.

  4. Data availability statement: The authors confirm that the data supporting the findings of this study are available within the manuscript.

References

[1] I. Ahn, S. Baek, and Z. G. Lin, The spreading fronts of an infective environment in a man-environment-man epidemic model, Appl. Math. Model. 40 (2016), 7082–7101, https://doi.org/10.1016/j.apm.2016.02.038. Search in Google Scholar

[2] M. Ali, A. R. Nelson, A. L. Lopez, and D. A. Sack, Updated global burden of cholera in endemic countries, PLoS Negl Trop Dis. 9 (2015), 0003832, https://doi.org/10.1371/journal.pntd.0003832. Search in Google Scholar PubMed PubMed Central

[3] X. Bao and W. Shen, Criteria for the existence of principal eigenvalues of time periodic cooperative linear systems with nonlocal dispersal, Proc. Amer. Math. Soc. 145 (2017), 2881–2894, https://doi.org/10.1090/proc/13602. Search in Google Scholar

[4] H. Brezis, Functional Analysis, Sobolev Spaces and Partial Differential Equations, Springer, New York, 2011. 10.1007/978-0-387-70914-7Search in Google Scholar

[5] J. F. Cao, Y. H. Du, F. Li, and W. T. Li, The dynamics of a Fisher-KPP nonlocal diffusion model with free boundaries, J. Funct. Anal. 277 (2019), 2772–2814, https://doi.org/10.1016/j.jfa.2019.02.013. Search in Google Scholar

[6] V. Capasso, Asymptotic stability for an integro-differential reaction–diffusion system, J. Math. Anal. Appl. 103 (1984), 575–588, https://doi.org/10.1016/0022-247X(84)90147-1. Search in Google Scholar

[7] V. Capasso and L. Maddalena, Convergence to equilibrium states for a reaction–diffusion system modeling the spatial spread of a class of bacterial and viral diseases, J. Math. Biol. 13 (1981), 173–184, https://doi.org/10.1007/BF00275212. Search in Google Scholar PubMed

[8] V. Capasso and S. L. Paveri-Fontana, A mathematical model for the 1973 cholera epidemic in the European Mediterranean region, Rev. Epidemiol. Sante. Publique. 27 (1979), 121–132, https://europepmc.org/article/MED/538301. Search in Google Scholar

[9] T. Y. Chang and Y. H. Du, Long-time dynamics of an epidemic model with nonlocal diffusion and free boundaries, Electron Res Arch. 30 (2021), 289–313, https://doi.org/10.3934/era.2022016. Search in Google Scholar

[10] F. Chatelin, The spectral approximation of linear operators with applications to the computation of eigenelements of differential and integral operators, SIAM Rev. 23 (1981), 495–522, https://doi.org/10.1137/1023099. Search in Google Scholar

[11] Q. L. Chen, S. Y. Tang, Z. D. Teng, and F. Wang, Long-time dynamics and semi-wave of a delayed nonlocal epidemic model with free boundaries, Proc. Roy. Soc. Edinburgh Sect. A. 155 (2025), 564–610, https://doi.org/10.1017/prm.2023.100. Search in Google Scholar

[12] Y. H. Du and Z. G. Lin, Spreading-vanishing dichotomy in the diffusive logistic model with a free boundary, SIAM J. Math. Anal. 42 (2010), 377–405, https://doi.org/10.1137/090771089. Search in Google Scholar

[13] Y. H. Du and W. J. Ni, Spreading speed for monostable cooperative systems with nonlocal diffusion and free boundaries, part 1: semi-wave and a threshold condition, J. Differential Equations 308 (2022), 369–420, https://doi.org/10.1016/j.jde.2021.11.014. Search in Google Scholar

[14] M. Fazly, M. Lewis, and H. Wang, On impulsive reaction–diffusion models in higher dimensions, SIAM J. Appl. Math. 77 (2017), 224–246, https://doi.org/10.1137/15M1046666. Search in Google Scholar

[15] M. Fazly, M. Lewis, and H. Wang, Analysis of propagation for impulsive reaction–diffusion models, SIAM J. Appl. Math. 80 (2020), 521–542, https://doi.org/10.1137/19M1246481. Search in Google Scholar

[16] Y. X. Feng, W. T. Li, S. G. Ruan, and M. Z. Xin, Principal spectral theory of time-periodic nonlocal dispersal cooperative systems and applications, SIAM J. Math. Anal. 56 (2024), 4040–4083, https://doi.org/10.1137/22M1543902. Search in Google Scholar

[17] S. Ghosh, M. Kumar, M. Santiana, A. Mishra, M. Zhang, H. Labayo, et al., Enteric viruses replicate in salivary glands and infect through saliva, Nature. 607 (2022), 345–350, https://doi.org/10.1038/s41586-022-04895-8. Search in Google Scholar PubMed PubMed Central

[18] A. E. Schumacher, H. H. Hyu, A. Aali, C. Abbafati, J. Abbas, R. Abbasgholizadeh, et al., Global age-sex-specific mortality, life expectancy, and population estimates in 204 countries and territories and 811 subnational locations, 1950-2021, and the impact of the COVID-19 pandemic: a comprehensive demographic analysis for the Global Burden of Disease Study 2021, Lancet. 403 (2024), 1989–2056, https://doi.org/10.1016/S0140-6736(24)00476-8. Search in Google Scholar PubMed PubMed Central

[19] L. Heller, C. R. Mota, and D. B. Greco, COVID-19 faecal-oral transmission: Are we asking the right questions? Sci. Total Environ. 729 (2020), 138919, https://doi.org/10.1016/j.scitotenv.2020.138919. Search in Google Scholar PubMed PubMed Central

[20] C. H. Hsu and T. S. Yang, Existence, uniqueness, monotonicity and asymptotic behaviour of travelling waves for epidemic models, Nonlinearity. 26 (2013), 121–139, https://doi.org/10.1088/0951-7715/26/1/121. Search in Google Scholar

[21] T. Kuniya and J. L. Wang, Global dynamics of an SIR epidemic model with nonlocal diffusion, Nonlinear Anal. Real World Appl. 43 (2018), 262–282, https://doi.org/10.1016/j.nonrwa.2018.03.001. Search in Google Scholar

[22] M. A. Lewis and B. T. Li, Spreading speed, traveling waves, and minimal domain size in impulsive reaction–diffusion models, Bull. Math. Biol. 74 (2012), 2383–2402, https://doi.org/10.1007/s11538-012-9757-6. Search in Google Scholar PubMed

[23] J. H. Liang, Q. Yan, C. C. Xiang, and S. Y. Tang, A reaction–diffusion population growth equation with multiple pulse perturbations, Commun. Nonlinear Sci Numer. Simul. 74 (2019), 122–137, https://doi.org/10.1016/j.cnsns.2019.02.015. Search in Google Scholar

[24] X. Liang, L. Zhang, and X. Q. Zhao, The principal eigenvalue for degenerate periodic reaction–diffusion systems, SIAM J. Math. Anal. 49 (2017), 3603–3636, https://doi.org/10.1137/16M1108832. Search in Google Scholar

[25] Z. Ling, L. Zhang, Y. Zhou, Long time behaviour of a river-field population model with nonlocal reaction–diffusion, Math. Methods Appl. Sci. Early Access (2025), 1–10, https://doi.org/10.1002/mma.70001. Search in Google Scholar

[26] P. Meyre-Nieberg, Banach Lattices, Springer-Verlag, New York, 1991. 10.1007/978-3-642-76724-1Search in Google Scholar

[27] V. Mogasale, B. Maskery, R. L. Ochiai, J. S. Lee, V. V. Mogasale, E. Ramani, et al., Burden of typhoid fever in low-income and middle-income countries: a systematic, literature-based update with risk-factor adjustment, Lancet Glob Health. 2 (2014), 570–580, https://doi.org/10.1016/s2214-109x(14)70301-8. Search in Google Scholar

[28] J. D. Murray, Mathematical Biology II: Spatial Models and Biomedical Applications, 3rd ed., Springer-Verlag, New York, 2003. 10.1007/b98869Search in Google Scholar

[29] H. R. Thieme, Spectral bound and reproduction number for infinite-dimensional population structure and time heterogeneity, SIAM J. Appl. Math. 70 (2009), 188–211, https://doi.org/10.1137/080732870. Search in Google Scholar

[30] H. R. Thieme, Remarks on resolvent positive operators and their perturbation, Discrete Contin. Dynam. Systems 4 (1998), 73–90, https://doi.org/10.3934/dcds.1998.4.73. Search in Google Scholar

[31] H. R. Thieme and X. Q. Zhao, Asymptotic speeds of spread and traveling waves for integral equations and delayed reaction–diffusion models, J. Differential Equations 195 (2003), 430–470, https://doi.org/10.1016/S0022-0396(03)00175-X. Search in Google Scholar

[32] R. Wang and Y. H. Du, Long-time dynamics of a diffusive epidemic model with free boundaries, Discrete Contin. Dyn. Syst. Ser. B. 26 (2021), 2201–2238, https://doi.org/10.3934/dcdsb.2020360. Search in Google Scholar

[33] R. Wang and Y. H. Du, Long-time dynamics of a nonlocal epidemic model with free boundaries: Spreading-vanishing dichotomy, J. Differential Equations 327 (2022), 322–381, https://doi.org/10.1016/j.jde.2022.04.031. Search in Google Scholar

[34] R. Wang and Y. H. Du, Long-time dynamics of an epidemic model with nonlocal diffusion and free boundaries: Spreading speed, Discrete Contin. Dynam. Systems 43 (2023), 121–161, https://doi.org/10.3934/era.2022016. Search in Google Scholar

[35] X.Y. Wang and J.Y. Yang, Dynamics of a nonlocal dispersal foot-and-mouth disease model in a spatially heterogeneous environment, Acta Math. Sci. 41 (2021), 552–572, https://doi.org/10.1007/s10473-021-0217-y. Search in Google Scholar

[36] S. L. Wu, and C. H. Hsu, Existence of entire solutions for delayed monostable epidemic models, Trans. Am. Math. Soc. 368 (2016), 6033–6062, https://doi.org/10.1090/tran/6526. Search in Google Scholar

[37] R. W. Wu and X. Q. Zhao, Spatial invasion of a birth pulse population with nonlocal dispersal, SIAM J. Appl. Math. 79 (2019), 1075–1097, https://doi.org/10.1137/18M1209805. Search in Google Scholar

[38] R. W. Wu and X. Q. Zhao, The evolution dynamics of an impulsive hybrid population model with spatial heterogeneity, Commun. Nonlinear Sci. Numer. Simul. 107 (2022), 106181, https://doi.org/10.1016/j.cnsns.2021.106181. Search in Google Scholar

[39] J. Y. Yang, M. J. Gong, and G. Q. Sun, Asymptotical profiles of an age-structured foot-and-mouth disease with nonlocal diffusion on a spatially heterogeneous environment, J. Differential Equations 377 (2023), 71–112, https://doi.org/10.1016/j.jde.2023.09.001. Search in Google Scholar

[40] Y. R. Zhang, T. S. Yi, and Y. M. Chen, Spreading dynamics of an impulsive reaction–diffusion model with shifting environments, J. Differential Equations 381 (2024), 1–19, https://doi.org/10.1016/j.jde.2023.10.048. Search in Google Scholar

[41] M. Zhao, W. T. Li, and Y. H. Du, The effect of nonlocal reaction in an epidemic model with nonlocal diffusion and free boundaries, Commun. Pure Appl. Anal. 19 (2020), 4599–4620, https://doi.org/10.3934/cpaa.2020208. Search in Google Scholar

[42] G. Y. Zhao and S. G. Ruan, Spatiotemporal dynamics in epidemic models with Lévy flights: A fractional diffusion approach, J. Math. Pures Appl. 173 (2023), 243–277, https://doi.org/10.1016/j.matpur.2023.02.011. Search in Google Scholar

[43] M. Zhao, Y. Zhang, W. T. Li, and Y. H. Du, The dynamics of a degenerate epidemic model with nonlocal diffusion and free boundaries, J. Differential Equations 269 (2020), 3347–3386, https://doi.org/10.1016/j.jde.2020.02.029. Search in Google Scholar

[44] Q. Zhou, Z. G. Lin, and M. Pedersen, On an impulsive faecal-oral model in a moving infected environment, 2024, https://doi.org/10.48550/arXiv.2410.15798. Search in Google Scholar

[45] Q. Zhou, Z. G. Lin, and C. A. Santos, On an impulsive faecal-oral model in a periodically evolving environment, Chaos Solitons Fractals 191 (2025), 115825, https://doi.org/10.1016/j.chaos.2024.115825. Search in Google Scholar

Received: 2025-03-12
Revised: 2025-07-07
Accepted: 2025-08-12
Published Online: 2025-09-18

© 2025 the author(s), published by De Gruyter

This work is licensed under the Creative Commons Attribution 4.0 International License.

Articles in the same Issue

  1. Research Articles
  2. Incompressible limit for the compressible viscoelastic fluids in critical space
  3. Concentrating solutions for double critical fractional Schrödinger-Poisson system with p-Laplacian in ℝ3
  4. Intervals of bifurcation points for semilinear elliptic problems
  5. On multiplicity of solutions to nonlinear Dirac equation with local super-quadratic growth
  6. Entire radial bounded solutions for Leray-Lions equations of (p, q)-type
  7. Combinatorial pth Calabi flows for total geodesic curvatures in spherical background geometry
  8. Sharp upper bounds for the capacity in the hyperbolic and Euclidean spaces
  9. Positive solutions for asymptotically linear Schrödinger equation on hyperbolic space
  10. Existence and non-degeneracy of the normalized spike solutions to the fractional Schrödinger equations
  11. Existence results for non-coercive problems
  12. Ground states for Schrödinger-Poisson system with zero mass and the Coulomb critical exponent
  13. Geometric characterization of generalized Hajłasz-Sobolev embedding domains
  14. Subharmonic solutions of first-order Hamiltonian systems
  15. Sharp asymptotic expansions of entire large solutions to a class of k-Hessian equations with weights
  16. Stability of pyramidal traveling fronts in time-periodic reaction–diffusion equations with degenerate monostable and ignition nonlinearities
  17. Ground-state solutions for fractional Kirchhoff-Choquard equations with critical growth
  18. Existence results of variable exponent double-phase multivalued elliptic inequalities with logarithmic perturbation and convections
  19. Homoclinic solutions in periodic partial difference equations
  20. Classical solution for compressible Navier-Stokes-Korteweg equations with zero sound speed
  21. Properties of minimizers for L2-subcritical Kirchhoff energy functionals
  22. Asymptotic behavior of global mild solutions to the Keller-Segel-Navier-Stokes system in Lorentz spaces
  23. Blow-up solutions to the Hartree-Fock type Schrödinger equation with the critical rotational speed
  24. Qualitative properties of solutions for dual fractional parabolic equations involving nonlocal Monge-Ampère operator
  25. Regularity for double-phase functionals with nearly linear growth and two modulating coefficients
  26. Uniform boundedness and compactness for the commutator of an extension of Riesz transform on stratified Lie groups
  27. Normalized solutions to nonlinear Schrödinger equations with mixed fractional Laplacians
  28. Existence of positive radial solutions of general quasilinear elliptic systems
  29. Low Mach number and non-resistive limit of magnetohydrodynamic equations with large temperature variations in general bounded domains
  30. Sharp viscous shock waves for relaxation model with degeneracy
  31. Bifurcation and multiplicity results for critical problems involving the p-Grushin operator
  32. Asymptotic behavior of solutions of a free boundary model with seasonal succession and impulsive harvesting
  33. Blowing-up solutions concentrated along minimal submanifolds for some supercritical Hamiltonian systems on Riemannian manifolds
  34. Stability of rarefaction wave for relaxed compressible Navier-Stokes equations with density-dependent viscosity
  35. Singularity for the macroscopic production model with Chaplygin gas
  36. Global strong solution of compressible flow with spherically symmetric data and density-dependent viscosities
  37. Global dynamics of population-toxicant models with nonlocal dispersals
  38. α-Mean curvature flow of non-compact complete convex hypersurfaces and the evolution of level sets
  39. High-energy solutions for coupled Schrödinger systems with critical growth and lack of compactness
  40. On the structure and lifespan of smooth solutions for the two-dimensional hyperbolic geometric flow equation
  41. Well-posedness for physical vacuum free boundary problem of compressible Euler equations with time-dependent damping
  42. On the existence of solutions of infinite systems of Volterra-Hammerstein-Stieltjes integral equations
  43. Remark on the analyticity of the fractional Fokker-Planck equation
  44. Continuous dependence on initial data for damped fourth-order wave equation with strain term
  45. Unilateral problems for quasilinear operators with fractional Riesz gradients
  46. Boundedness of solutions to quasilinear elliptic systems
  47. Existence of positive solutions for critical p-Laplacian equation with critical Neumann boundary condition
  48. Non-local diffusion and pulse intervention in a faecal-oral model with moving infected fronts
  49. Nonsmooth analysis of doubly nonlinear second-order evolution equations with nonconvex energy functionals
  50. Qualitative properties of solutions to the viscoelastic beam equation with damping and logarithmic nonlinear source terms
  51. Shape of extremal functions for weighted Sobolev-type inequalities
  52. One-dimensional boundary blow up problem with a nonlocal term
  53. Doubling measure and regularity to K-quasiminimizers of double-phase energy
  54. General solutions of weakly delayed discrete systems in 3D
  55. Global well-posedness of a nonlinear Boussinesq-fluid-structure interaction system with large initial data
  56. Optimal large time behavior of the 3D rate type viscoelastic fluids
  57. Local well-posedness for the two-component Benjamin-Ono equation
  58. Self-similar blow-up solutions of the four-dimensional Schrödinger-Wave system
  59. Existence and stability of traveling waves in a Keller-Segel system with nonlinear stimulation
  60. Existence and multiplicity of solutions for a class of superlinear p-Laplacian equations
  61. On the global large regular solutions of the 1D degenerate compressible Navier-Stokes equations
  62. Normal forms of piecewise-smooth monodromic systems
  63. Fractional Dirichlet problems with singular and non-locally convective reaction
  64. Sharp forced waves of degenerate diffusion equations in shifting environments
  65. Global boundedness and stability of a quasilinear two-species chemotaxis-competition model with nonlinear sensitivities
  66. Non-existence, radial symmetry, monotonicity, and Liouville theorem of master equations with fractional p-Laplacian
  67. Global existence, asymptotic behavior, and finite time blow up of solutions for a class of generalized thermoelastic system with p-Laplacian
  68. Formation of singularities for a linearly degenerate hyperbolic system arising in magnetohydrodynamics
  69. Linear stability and bifurcation analysis for a free boundary problem arising in a double-layered tumor model
  70. Hopf's lemma, asymptotic radial symmetry, and monotonicity of solutions to the logarithmic Laplacian parabolic system
  71. Generalized quasi-linear fractional Wentzell problems
  72. Existence, symmetry, and regularity of ground states of a nonlinear Choquard equation in the hyperbolic space
  73. Normalized solutions for NLS equations with general nonlinearity on compact metric graphs
  74. Boundedness and global stability in a predator-prey chemotaxis system with indirect pursuit-evasion interaction and nonlocal kinetics
  75. Review Article
  76. Existence and stability of contact discontinuities to piston problems
Downloaded on 29.1.2026 from https://www.degruyterbrill.com/document/doi/10.1515/anona-2025-0111/html
Scroll to top button