This paper is devoted to the dynamics of a stochastic modified Bazykin predator-prey system with Lévy jumps. First, we show that the system has a unique global positive solution and give some properties of solutions. Then, some sufficient conditions for persistence and extinction are derived by Itô formula and some inequalities on stochastic analysis. At last, some simulations are provided to check the main results.
National Natural Science Foundation of China117710011147101511401002Nature Science Foundation of Anhui Province1508085QA011508085MA10Natural Science Foundation of Anhui CollegesKJ2014A010Program for Excellent Young Talents in University of Anhui Provincegxyq2017092Anhui Province Workshop of Prestigious Teacher2016msgzs0061. Introduction
In population biology, construction of models and relevant qualitative analysis are popular fields [1, 2]. In the last few decades, many predator-prey models with functional and numerical responses have been proposed and studied extensively. Particularly, ratio-dependent functional response is common in some classical literature [3–5].
Alexeev and Bazykin firstly proposed a Bazykin system [1] (1)x˙t=xta-bxt-cyt1+Axt,y˙t=yt-g-hyt+fxt1+Axt,where x(t) and y(t) are the size of prey and predator at time t. a,b,c,f,g, and m are positive constants (some details refer to [1]).
Considering the ratio effect in hunting process of predation, Haque built a modified Bazykin system [2] (2)x˙t=xta-bxt-cytyt+Axt,y˙t=yt-g-hyt+fxtyt+Axt,of which dynamical behavior near equilibrium point and bifurcation are observed. System (2) is more reasonable and many researchers began to pay more attention on it. Its generalized forms have been investigated and a lot of results were obtained [6–8].
However, in the natural world, environmental noise is everywhere, and the growth rate of the populations is not constants. In this case, many systems are described by stochastic differential equations driven by Brownian motion of the actual research. For example, [9] studied the following stochastic modified Bazykin system: (3)dxt=xta-bxt-cytyt+Axtdt+αxtdw1t,dyt=yt-g-hyt+fxtyt+Axtdt+βytdw2t,where w1(t) and w2(t) are two independent Wiener processes defined on a complete probability space Ω,F,Ftt≥0,P and α and β stand for the level of the white noises. Some sufficient conditions for persistence and extinction are obtained.
Recently, the stochastic differential equation driven by jump has drawn more and more researchers’ attention [10–17]. This is mainly according to sudden perturbation of environment, such as toxic contamination of water, torrential flood, and hurricane. Motivated by the above, this paper considers the following stochastic modified Bazykin system with Lévy jumps: (4)dxt=xt-a-bxt-cytyt+Axtdt+αxtdw1t+∫Zγuxt-N~dt,du,dyt=yt--g-hyt+fxtyt+Axtdt+βytdw2t+∫Zηuyt-N~dt,du,where x(t-), y(t-) represent the left limit of x(t), y(t). N~(dt,du)=N(dt,du)-λ(du)dt, N is a Poisson random measure, and λ is the characteristic measure of N on a measurable subset Z⊂R+=(0,+∞) with λ(Z)<+∞.
The rest of this paper is organized as follows. In Section 2, some properties of positive solutions to system (4) are discussed. In Section 3, the main results for persistence and extinction are given. Finally, the simulation results show the validity of our results.
2. Properties of Positive Solutions
Throughout this paper, we require that w1(t), w2(t), and N are independent and(5)H1max∫Zγupλdu,∫Zηupλdu<+∞,p>0H2max∫Zln1+γu2λdu,∫Zln1+ηu2λdu<+∞with min1+γu,1+ηu>0,u∈Z.
First, we present the global existence of positive solutions.
Lemma 1.
For any given value (x0,y0)∈R+2, system (4) has a unique positive solution X(t)=(x(t),y(t)) on t≥0 and the solution will be in R+2 a.s. (almost surely).
Proof.
Let ut=lnxt,vt=lnyt, then we consider the following system: (6)dut=a-beut-cevtevt+Aeut-α22-∫Zγu-ln1+γuλdudt+αdw1t+∫Zln1+γuN~dt,du,dvt=-g-hevt+feutevt+Aeut-β22-∫Zηu-ln1+ηuλdudt+βtdw2t+∫Zln1+ηuN~dt,duwith initial date u(0)=lnx0,v(0)=lny0 on t≥0. Since the coefficients of system (6) are locally Lipschitz continuous, then there is a unique local solution on [0,T] a.s., where T is blow-up time. Then x=eu(t),y=ev(t) is the unique positive local solution to system (4) with initial data x0>0,y0>0. We will show that T=+∞, and this mean the solution is global.
Consider the following equations:(7)dX1t=X1ta-c-bX1tdt+αX1tdw1t+∫ZX1tγuN~dt,du,X10=x0,dX2t=X2ta-bX2tdt+αX2tdw1t+∫ZX2tγuN~dt,du,X20=x0,dY1t=Y1t-g+fA-h+fA2X1tY1tdt+βY1tdw2t+∫ZY1tηuN~dt,du,Y10=y0,dY2t=Y2t-g+fA-hY2tdt-βY2tdw2t+∫ZY2tηuN~dt,du,Y20=y0.
By the comparison theorem of stochastic differential equation, we conclude that (8)X1t≤xt≤X2t,Y1t≤yt≤Y2ta.s.
According to [10], we can give(9)X1t=expa-c-ρ1t+αw1t+κ1tx0-1+b∫0texpa-c-ρ1s+αw1s+κ1sds,X2t=expa-ρ1t+αw1t+κ1tx0-1+b∫0texpa-ρ1s+αw1s+κ1sds,Y1t=exp-g+f/A-ρ2t+βw2t+κ2ty0-1+∫0th+f/A2X1texpf/A-g-ρ2s+βw2s+κ2sds,Y2t=exp-g+f/A-ρ2t+βw2t+κ2ty0-1+h∫0texpf/A-g-ρ2s+βw2s+κ2sds,where (10)ρ1=0.5α2+∫Zγu-ln1+γuλdu,ρ2=0.5β2+∫Zηu-ln1+ηuλdu,κ1t=∫0t∫Zln1+γuN~dt,du,κ2t=∫0t∫Zln1+ηuN~dt,du.
Because t≥0 is the existence range of solutions X1(t),X2(t),Y1(t),Y2(t), that means T=+∞.
Next, we will show the asymptotic property of the solution to system (4).
Theorem 2.
The solutions of system (4) are bounded in mean.
Proof.
Let V1≔V1(t,x)=etxp(p>0). Direct computation, by the formula EV1(t,x(t))-EV1(0,x0)=E∫0tLV(s)ds [18], now leads to(11)Eetxp=xp0+E∫0tLV1ds,where (12)LV1=etxp+etpxpa-bx+0.5p-1α2-cyy+Ax+∫Z1+γup-1λdu≤etxp+etpxpa-bx+0.5p-1α2+∫Z1+γup-1λdu.From actual meanings of parameters and assumption H1, we get that(13)xp+pxpa-bx+0.5p-1α2+∫Z1+γup-1λduis bounded. There exists a constant K1>0, such that (14)Eetxpt≤x0p+∫0tK1esdswhen 1+xp+pxp[a-bx+0.5(p-1)α2∫Z[(1+γ(u))p-1]λ(du)]>0; otherwise Eetxp(t)≤x0p. Let K2=max{K1,0}; we have Eetxp(t)≤x0p+K2et. Denote K=K2+x0p, then Exp(t)≤K<+∞. Similarly, Eyp(t)≤K.
3. Persistence and Extinction
In this section, some properties of the solutions of system (4) are investigated. Some sufficient conditions for persistence and extinction are shown. To proceed, some definitions and lemma are as follows.
Definition 3 (see [19]).
(1) The population x is called to be extinct if x(t)→0(t→∞)a.s.
(2) The population x is called to be persistent if liminft→∞1/t∫0tx(s)ds>0 a.s.
(3) System (4) is called to be persistent if populations x and y are all persistent a.s.
Lemma 4 (see [20]).
Under H1 and H2, suppose Y(t)∈C(Ω×[0,+∞),R+).
(1) If there exist three positive T,k,k0 such that (15)lnYt≤kt-k0∫0tYsds+∑i=1nkiwit+∑i=1nci∫0t∫Zln1+γuN~ds,dua.s.for all t>T, where ki,ci are constants, then(16)limsupt→∞1t∫0tYsds≤kk0a.s.If k<0 and other conditions are the same, then limt→∞Y(t)=0 a.s.
(2) If there exist three positive T,k,k0 such that (17)lnYt≥kt-k0∫0tYsds+∑i=1nkiwit+∑i=1nci∫0t∫Zln1+γuN~ds,dua.s.for all t>T, where ki,ci are constants, then (18)liminft→∞1t∫0tYsds≥kk0a.s.
Now, the main results about persistence and extinction of system (4) are as follows.
Theorem 5.
(1) The populations x and y are extinct if a-ρ1<0 a.s.
(2) The population x is persistent and y is extinct if a-c-ρ1>0,f/A-g-ρ2<0 a.s.
(3) The populations x and y are persistent if a-c-ρ1>0,f/A-g-ρ2>0 a.s.
Proof.
(1) By using the Itô formula, the following formulas are hold: (19)dlnxt=a-bxt-cytyt+Axt-ρ1dt+αdw1t+∫Zln1+γuN~dt,du,dlnyt=-g-hyt+fxtyt+Axt-ρ2dt+βdw2t+∫Zln1+ηuN~dt,du.Calculated by integral, we have the following form: (20)lnxt-lnx0=a-ρ1t+αw1t-∫0tbxsds-∫0tcysys+Axsds+κ1t≤a-ρ1t-b∫0txsds+αw1t+κ1t.
Then by Lemma 4, note that a<ρ1, and therefore x(t)→0(t→∞) (population x is extinct) a.s. Then, for arbitrarily small ε>0, there exist a sufficiently large T, when t>T, and we have x(t)<ε and (21)dlnyt≤-g-hyt+ε-ρ2dt+βdw2t+∫Zln1+ηuN~dt,du,where -g-ε-ρ2<0. It is clear that y(t)→0(t→∞) (population y is extinct) a.s. Thus (1) is correct.
(2) From (1), the following forms are clear. (22)lnxt-lnx0=a-ρ1t+αw1t-∫0tbxsds-∫0tcysys+Axsds+κ1≤a-ρ1t-b∫0txsds+αw1t+κ1,lnxt-lnx0=a-ρ1t+αw1t-∫0tbxsds-∫0tcysys+Axsds+κ1≥a-c-ρ1t-b∫0txsds+αw1t+κ1.Then by Lemma 4, we have (23)a-c-ρ1b≤liminft→∞1t∫0txsds≤limsupt→∞1t∫0txsds≤a-ρ1ba.s.Therefore the population x is persistent in mean. For population y, we have (24)dlnyt≤-g-hyt+fA-ρ2dt+βdw2t+∫Zln1+ηuN~dt,du.
From Lemma 4 and condition f/A-g-ρ2<0, we know that y(t)→0(t→∞) (i.e., population y is extinct) a.s.
(3) In view of above, we can conclude the following when a-c-ρ1>0,g/A-g-ρ2>0, and(25)lnyt-lny0≥-g+fA-ρ2t-h+fA2X1∗∫0tysds+βw2t+κ2t.Then, we have (26)f/A-g-ρ2h+f/A2X1∗≤liminft→∞1t∫0tysds≤limsupt→∞1t∫0tysds≤f/A-g-ρ2h,a.s.,where X1∗ is the minimum of X1(t)>0.
It is clear that population y is persistent.
Remark 6.
For (1) in Theorem 5, it implies that the population x of extinction a.s. leads to the extinction of population y a.s. As shown in Figure 2, the simulations also affirm this point. In Figure 2, we can see that population x becomes extinct, and after a while, population y becomes extinct.
4. Numerical Simulations
We will demonstrate our results with the help of numerical simulations by using the Euler-Maruyama scheme [21, 22]. In numerical simulation, randomly selected parameters are as follows a=0.55,b=0.21,c=0.15,f=0.93,h=0.11,A=1.23,α=0.75,β=0.73,Z=(0,+∞), and λ(Z)=1, with simulation time span T=50 and step size Δt=T/N, where N=212. The initial data is (x0,y0)=(0.3,0.11) in Figure 2, and others are (3,1).
(1) As illustrated in Figure 1, we choose γ(u)=0.99,η(u)=0.85, then ρ1=a-0.5α2-∫Z(γ(u)-ln(1+γ(u)))λ(du)=0.58312, ρ2=a-0.5β2-∫Z(η(u)-ln(1+η(u)))λ(du)=0.50126, and a-ρ1=-0.03312<0,f/A-g-ρ2=-0.09517<0. By Theorem 5, the populations x and y are extinct a.s. Numerical experiments verify the correctness of (1) in Theorem 5.
(2) As illustrated in Figure 2, we choose γ(u)=0.99,η(u)=0.05, then ρ1=a-0.5α2-∫Z(γ(u)-ln(1+γ(u)))λ(du)=0.58312, ρ2=a-0.5β2-∫Z(η(u)-ln(1+η(u)))λ(du)=0.26766, and a-ρ1=-0.03312<0,f/A-g-ρ2=0.13844>0. By Theorem 5, the populations x and y are extinct a.s. Numerical experiments also verify the correctness of (1) in Theorem 5.
(3) As shown in Figure 3, we choose γ(u)=0.15,η(u)=0.91, then ρ1=a-0.5α2-∫Z(γ(u)-ln(1+γ(u)))λ(du)=0.29149, ρ2=a-0.5β2-∫Z(η(u)-ln(1+η(u)))λ(du)=0.52935, and a-c-ρ1=0.048512>0,f/A-g-ρ2=-0.12325<0. By Theorem 5, the population x is persistent a.s. and population y is extinct a.s. The correctness of (2) in Theorem 5 is verified.
(4) As shown in Figure 4, we choose γ(u)=0.15,η(u)=0.05, then ρ1=a-0.5α2-∫Z(γ(u)-ln(1+γ(u)))λ(du)=0.29149, ρ2=a-0.5β2-∫Z(η(u)-ln(1+η(u)))λ(du)=0.52935, and a-c-ρ1=0.048512>0,f/A-g-ρ2=0.13844>0. By Theorem 5, the populations x and y are persistent a.s. The validity of correctness of (3) in Theorem 5 is verified.
5. Conclusions
In this paper, a stochastic modified Bazykin system with Lévy jumps is discussed. Firstly, the existence of global positive solutions is investigated, and boundedness in mean is also given. Further, under related assumption, we obtain the sufficient conditions for the stochastic permanence and extinction of system (4). There are some interesting topics which deserve further discussion. For example, many authors consider the stationary distribution for stochastic predator-prey model with harvesting and delays [23–25], epidemic model with regime switching [26], impulsive stochastic model [27], and budworm growth model with Markovian switching [28]. The above investigations are left for future work.
Conflicts of Interest
The authors declare that there are no conflicts of interest regarding the publication of this manuscript.
Acknowledgments
This work is supported by National Natural Science Foundation of China (nos. 11771001, 11471015, and 11401002), Nature Science Foundation of Anhui Province (no. 1508085QA01 and no. 1508085MA10), Natural Science Foundation of Anhui Colleges (no. KJ2014A010), Program for Excellent Young Talents in University of Anhui Province (no. gxyq2017092), and Anhui Province Workshop of Prestigious Teacher (no. 2016msgzs006).
BazykinA. D.199811River Edge, NJ, USAWorld ScientificWorld Scientific Series on Nonlinear Science, Series A10.1142/9789812798725MR1635219HaqueM.Ratio-dependent predator-prey models of interacting populations200971243045210.1007/s11538-008-9368-4MR2471054Zbl1170.920272-s2.0-58349121665KuangY.BerettaE.Global qualitative analysis of a ratio-dependent predator-prey system199836438940610.1007/s002850050105MR1624192Zbl0895.920322-s2.0-0000252357XiaoD.RuanS.Global dynamics of a ratio-dependent predator-prey system200143326829010.1007/s002850100097MR1868217Zbl1007.340312-s2.0-0035464834XiaoY.ChenL.Modeling and analysis of a predator-prey model with disease in the prey20011711598210.1016/S0025-5564(01)00049-9MR1839209Zbl0978.920312-s2.0-0035016953SarwardiS.HaqueM.MandalP. .Ratio-dependent predator-prey model of interacting population with delay effect201269381783610.1007/s11071-011-0307-9MR2943348Zbl061222032-s2.0-84863980309SarwardiS.HaqueM.MandalP. .Persistence and global stability of Bazykin predator-prey model with BEDdington-DeAngelis response function201419118920910.1016/j.cnsns.2013.05.029MR3142459Zbl1344.921452-s2.0-84884143415HaqueM.A detailed study of the Beddington-DeAngelis predator-prey model2011234111610.1016/j.mbs.2011.07.003MR2894850Zbl06048788LvJ.WangK.ChenD.Analysis on a stochastic two-species ratio-dependent predator-prey model201517240341810.1007/s11009-013-9383-2MR3343413Zbl1319.601312-s2.0-84890293353BaoJ.YuanC.Stochastic population dynamics driven by Lévy noise2012391236337510.1016/j.jmaa.2012.02.043MR2903137LiuM.BaiC.Dynamics of a stochastic one-prey two-predator model with Lévy jumps201628430832110.1016/j.amc.2016.02.033MR3486371LiuQ.JiangD.ShiN.HayatT.AlsaediA.Stochastic mutualism model with Lévy jumps201743789010.1016/j.cnsns.2016.05.003MR3537304MaoW.HuL.MaoX.The existence and asymptotic estimations of solutions to stochastic pantograph equations with diffusion and Lévy jumps201526888389610.1016/j.amc.2015.06.109MR3399473ZhangX.JiangD.HayatT.AhmadB.Dynamics of a stochastic SIS model with double epidemic diseases driven by Lévy jumps201747176777710.1016/j.physa.2016.12.074MR3597480MengX.WangX.Stochastic predator-prey system subject to Lévy jumps201613574989210.1155/2016/5749892MR3484229ZhuQ.Razumikhin-type theorem for stochastic functional differential equations with Lévy noise and Markov switching20179081703171210.1080/00207179.2016.1219069MR3658475GuoL.ZhuQ.Stability analysis for stochastic Volterra-Levin equations with Poisson jumps: fixed point approach2011524042702, 1510.1063/1.3573598MR2964170Zbl1316.450022-s2.0-79955420099042702MaoX.1997Chichester, UKHorwoodMR1475218Zbl0892.60057LiuM.WangK.Dynamics of a two-prey one-predator system in random environments201323575177510.1007/s00332-013-9167-4MR3101833Zbl1279.920882-s2.0-84884672931LiuM.WangK.Stochastic Lotka-Volterra systems with Lévy noise2014410275076310.1016/j.jmaa.2013.07.078MR3111864HighamD. J.MaoX.StuartA. M.Strong convergence of Euler-type methods for nonlinear stochastic differential equations20024031041106310.1137/S0036142901389530MR1949404Zbl1026.650032-s2.0-0012279718WangL.2016LiuM.HeX.YuJ.Dynamics of a stochastic regime-switching predator-prey model with harvesting and distributed delays2018288710410.1016/j.nahs.2017.10.004MR3744970Zbl06839602QiuH.DengW.Optimal harvesting of a stochastic delay tri-trophic food-chain model with Lévy jumps20184921715172810.1016/j.physa.2017.11.092MR3735230YuJ.LiuM.Stationary distribution and ergodicity of a stochastic food-chain model with Lévy jumps2017482142810.1016/j.physa.2017.04.067MR3651404ZhangX.JiangD.AlsaediA.HayatT.Stationary distribution of stochastic SIS epidemic model with vaccination under regime switching201659879310.1016/j.aml.2016.03.010MR3494309LiuM.DuC.DengM.Persistence and extinction of a modified Leslie-Gower Holling-type II stochastic predator-prey model with impulsive toxicant input in polluted environments20182717719010.1016/j.nahs.2017.08.001MR3729565LiuM.ZhuY.Stability of a budworm growth model with random perturbations201879131910.1016/j.aml.2017.11.020MR3748605