Traveling Waves of Modified Leslie-Gower

Predator-prey Systems
Hongliang Li∗, Min Zhao, Rong Yuan
arXiv:2306.00701v1 [math.AP] 1 Jun 2023

Laboratory of Mathematics and Complex Systems (Ministry of Education)

School of Mathematical Sciences, Beijing Normal university
Beijing 100875, People’s Republic of China
E-mail: [email protected], [email protected],
[email protected]

Abstract
The spreading phenomena in modified Leslie-Gower reaction-diffusion predator-prey sys-
tems are the topic of this paper. We mainly study the existence of two different types of
traveling waves. Be specific, with the aid of the upper and lower solutions method, we estab-
lish the existence of traveling wave connecting the prey-present state and the coexistence state
or the prey-present state and the prey-free state by constructing different and appropriate
Lyapunov functions. Moreover, for traveling wave connecting the prey-present state and the
prey-free state, we gain more monotonicity information on wave profile based on the asymp-
totic behavior at negative infinite. Finally, our results are applied to modified Leslie-Gower
system with Holling II type or Lotka-Volterra type, and then a novel Lyapunov function is
constructed for the latter, which further enhances our results. Meanwhile, some numerical
simulations are carried to support our results.
Key words: Traveling waves; Upper and lower solutions; LaSalle’s Invariance Principle

1 Introdcution
Due to the complexity of ecosystems, long-term coexisting species frequently exhibit diversified
interspecific interactions, such as cooperation, competition, and predation. One of the most famous
systems is Lotka-Volterra predator-prey system derived by Lotka [28] and Volterra [37] in the
1920s, and there has been a substantial amount of research devoted to predator-prey systems,
seeing [1, 3, 8, 21, 30, 31]. However, the majority of biological species live in spatially heterogeneous
natural habitats, and it is fair to anticipate that the variability of the habitats will have an
impact on population density. Consequently, the response terms should be expanded to include
the diffusion.
What we are curious about is the change in ecological processes when alien species are intro-
duced into new habitats. One way to study this topic is to look at the so-called traveling waves,
∗
Corresponding author.

1
seeing [12–14, 23–25], another way is to characterize spreading speeds of the predator by solving
the Cauchy problem, seeing [4, 5, 10]. It should be emphasized that the functional response, the
rate of prey consumption by an average predator, and the predator growth are identical in most
two species predator-prey systems. For example,

∂u(x, t)



 ∂t = ∆u(x, t) + g(u(x, t)) − f (u(x, t))v(x, t),

 ∂v(x, t) = d∆v(x, t) + (βf (u(x, t)) − d) v(x, t).



∂t
While some scholars have also studied predator-prey systems whose predator growth function is
different from the functional response. For example, Leslie in [26, 27] proposed a predator-prey
system in which the predator’s environmental carrying capacity is proportional to the abundance
of the prey  
∂u(x, t) av(x, t)



 ∂t = ∆u(x, t) + u(x, t) 1 − u(x, t) − ,
u(x, t) + e1

 
 ∂v(x, t) v(x, t)
= d∆v(x, t) + sv(x, t) 1 − ,



∂t u(x, t)
where the term v(x, t)/u(x, t) is called the Leslie-Gower term. Such predators are called special-
ist predators [35] since the population density of a favored prey is small, the predator should
concentrate on other prey species, and vice versa. Recently, the existence of traveling waves to
the above system has attracted increasing interest, including Lotka-Volterra type, Holling-III and
Beddington-DeAngelis functional responses and so on, seeing [2, 9, 11, 38, 41].
Based on the above system, the authors in [6, 7] proposed a modified Leslie-Gower predator-
prey system by adding the positive constant µ to the Leslie-Gower term. Such predators can be
called generalist predators [19] since they explore other alternative food sources in the absence of
their favorite prey. More precisely,
 
∂u(x, t) av(x, t)



 ∂t = ∆u(x, t) + u(x, t) 1 − u(x, t) − ,
u(x, t) + e1

  (1)
 ∂v(x, t) v(x, t)
= d∆v(x, t) + sv(x, t) 1 − ,



∂t u(x, t) + µ

where parameters a, s, e1 and µ are positive, u(x, t) and v(x, t) represent population densities of
prey and predator species respectively, and d denotes the ratio of the diffusion of the predator to
that of the prey. For more detailed descriptions, one can refer to [6, 7].
When the prey is stationary, Tian et al. in [33, 34] studied traveling waves connecting two
appropriate equilibria by the shoot argument and the upper and lower solutions method respec-
tively. However, when the prey diffuses slower than the predator, Hsu and Lin in [22] obtained
the existence of traveling waves connecting (0, 0) and the coexistence state by the upper and lower
solutions method with the aid of contracting rectangles. For system (1), to the best of the author’s
knowledge, there are no results on traveling waves connecting the prey-present state and the co-
existence state or the prey-present state and the prey-free state, even for a modified Leslie-Gower
predator-prey system with Lotka-Volterra type [16, 18]. Based on these, we consider a generalized

2
Leslie-Gower system

∂u(x, t)
= ∆u(x, t) + f (u(x, t)) (p(u(x, t)) − v(x, t)) ,


 ∂t

  (2)
 ∂v(x, t) v(x, t)
= d∆v(x, t) + sv(x, t) 1 − .


∂t u(x, t)h(u(x, t)) + µ


where parameters d, s, µ are positive constants. In this paper, we assume that f (u), p(u) and h(u)
are C 1 functions satisfying the following assumption:
Assumption 1.1
(H1) f (0) = 0 and f ′ (u) > 0 for u ∈ [0, +∞).
(H2) p(1) = 0 and p′ (u) < 0 for u ∈ (0, +∞).
(H3) h(0) > 0 and h′ (u) ≥ 0 for u ∈ [0, +∞).
Some classical examples satisfying Assumption 1.1 as follows
(1) Lotka-Volterra type system: for given a > 0,
(1 − u)
f (u) = au, p(u) = and h(u) = 1.
a
(2) Holling II type system: for given a > 0 and e1 ≥ 1,
au (1 − u)(u + e1 )
f (u) = , p(u) = and h(u) = 1.
u + e1 a

(3) Ivlev-type type system: for given a > 0 and 2 > m > 0,
u(1 − u)
f (u) = a(1 − e−mu ), p(u) = and h(u) = 1.
a(1 − e−mu )

In view of Assumption 1.1, as for the corresponding kinetic system, there are always three
boundary equilibria (0, 0), (1, 0) and (0, µ). Clearly, (1, 0) is unstable. Moreover, one can readily
verify that (0, µ) is stable if µ > p(0), whereas (0, µ) is unstable if µ < p(0). On the other
hand, we also conclude that function Q(u) = p(u) − uh(u) − µ satisfies Q(1) < 0, Q(0) > 0
and Q′ (u) < 0 for u ∈ (0, 1) if p(0) > µ, which implies that system (2) has a unique positive
equilibrium (u∗ , v ∗ ) := (u∗ , u∗h(u∗ ) + µ) with u∗ ∈ (0, 1). Consequently, we may expect traveling
waves connecting (1, 0) and E of system (2), where
(
(0, µ), if µ > p(0),
E=
(u∗ , v ∗ ), if µ < p(0).

We are interested in the spreading phenomena of system (2). First of all, let us focus on
spreading speeds of v(x, t). Constant s∗ is called spreading speeds of species w(x, t) if
( )
lim sup w(x, t) = 0, for ǫ > 0,
t→∞ |x|>(s∗ +ǫ)t

3
  
lim inf inf∗ w(x, t) > 0, for ǫ ∈ (0, s∗ ).
t→∞ |x|<(s −ǫ)t

We assume that u(x, 0) = 1 and v(x, 0) ∈ [0, h(1) + µ] is continuous function with nonempty
compact support, by the theory of reaction-diffusion system, then [0, 1] × [0, h(1) + µ] is a invariant
region of system (2). Hence, we obtain (u(x, t), v(x, t)) ∈ [0, 1] × [0, h(1) + µ] for x ∈ R and t > 0.
Meanwhile, one can directly observe that
 
∂v(x, t) v(x, t)
≤ d∆v(x, t) + sv(x, t) 1 − ,
∂t h(1) + µ
and  
∂v(x, t) v(x, t)
≥ d∆v(x, t) + sv(x, t) 1 − ,
∂t µ
thus, the spreading theory and the comparison principle of [4, 5] give
( )
lim sup
√
v(x, t) = 0, for ǫ > 0,
t→∞
|x|>(2 ds+ǫ)t

√
 
lim inf inf
√ v(x, t) > µ, for ǫ ∈ (0, 2 ds),
t→∞ |x|<(2 ds−ǫ)t
√
which implies that spreading speeds of v(x, t) is c∗ = 2 ds. Therefore, in this paper, we mainly
establish the existence of traveling waves connecting (1, 0) and E of system (2). A positive solution
is called a traveling wave of system (2) if it has the form

(u(x, t), v(x, t)) = (ũ(z), ṽ(z)), z = x + ct,

where c is the wave speed. With the tilde removed, (u(z), v(z)) should satisfy
 ′′ ′
u (z) − cu (z) + f (u(z)) (p(u(z)) − v(z)) = 0,


v(z)
 (3)
′′ ′
dv (z) − cv (z) + sv(z) 1 −
 = 0.
u(z)h(u(z)) + µ
Further, if (u(z), v(z)) also satisfies the boundary condition

(u(−∞), v(−∞)) = (1, 0), (u(+∞), v(+∞)) = E, (4)

then it is referred to as strong-traveling wave of system (2). In addition, one that just meets the
left-hand tail limit is referred to as semi-traveling wave of system (2).
In the sequence, strong or semi-traveling waves as mentioned are the above waves. Moreover,
for the sake of convenience, we assume that

q(u) = uh(u) + µ,

clearly, q(0) = µ, q(1) = h(1) + µ and q(u) is monotone increasing function in [0, +∞).
We are now in the position to state the existence results. The first theorem states necessary
and sufficient condition on the existence of semi-traveling wave.

4
Theorem √ 1.1 Assume that (H1)-(H3) hold. System (2) has a semi-traveling wave if and only if
∗
c ≥ c := 2 ds.
The following theorem establishes the existence of traveling wave connecting (1, 0) and (u∗ , v ∗ )
of a class of modified Leslie-Gower predator-prey systems, that is h(u) = 1.
Theorem 1.2 Assume that (H1)-(H2) hold and h(u) = 1. For c ≥ c∗ , if p(0) > µ and
(P) (u∗ + µ)f (u) − (u − u∗ )(u + µ)f ′ (u) > 0 for u ∈ (0, 1),
then system (2) has a traveling wave connecting (1, 0) and (u∗ , v ∗ ).
Finally, we state the existence result on traveling wave connecting (1, 0) and (0, µ). Before we
can do that, we define the sets
A = v ∈ C 2 (R) | 0 < v(z) < µ and v ′ (z) > 0 over R ,


 ∃ zv such that v(zv ) = µ, v ′(z) > 0 in (−∞, zv )
 
2
B = v ∈ C (R)  .
and v(z) > µ in (zv , +∞)
Theorem 1.3 Assume that (H1)-(H3) hold. For c ≥ c∗ , if
f (u) = ug(u) and p(u) = (1 − u)/g(u),
where g(u) satisfies
h(1)(h(1) + µ)
min g(u) > ,
u∈[0,1] µ2 h(0)
then the following hold.
(a) System (2) has a traveling wave (u(z), v(z)) connecting (1, 0) and (0, µ).
(b) v(z) ∈ A ∪ B and u′(z) < 0 over R.
This paper is organized as follows. In Section 2, we introduce some preliminary knowledges.
In Section 3, based on the upper and lower solutions method and the unstable manifold theorem,
we obtain necessary and sufficient condition on the existence of semi-traveling wave. In section 4,
we establish that the existence of strong-traveling waves, and for traveling wave connecting (1, 0)
and (0, µ), we further gain more monotonicity information on wave profile based on the asymptotic
behavior at z = −∞. In the last section, we apply Theorems 1.2 and 1.3 to modified Leslie-Gower
system with Holling II type or Lotka-Volterra type, then a novel Lyapunov function is constructed
for the latter, which further enhances our results. Moreover, the numerical simulations of traveling
waves are carried to support our results.

2 Preliminary
In this section, we will introduce the upper and lower solutions method, which actually pro-
vides a constructive approach for producing traveling waves.
Let us begin with the definition of upper and lower solutions of system (3). Define sets
X = {Φ(z) = (u(z), v(z)) | Φ(z) ∈ C (R, R2 )} ,
X0 = {Φ(z) ∈ X | 0 ≤ u(z) ≤ 1, 0 ≤ v(z) ≤ q(1), z ∈ R} .

5
Definition 2.1 Function pairs (u(z), v(z)) and (u(z), v(z)) in X0 are called a pair of upper and
lower solutions of system (3) if they satisfy

(a) u(z) ≤ u(z), v(z) ≤ v(z) for z ∈ R.

(b) There is a finite set E such that for z ∈ R\E

U (u, v) = u′′ (z) − cu′ (z) + f (u(z)) (p(u(z)) − v(z)) ≤ 0,

U (u, v) = u′′ (z) − cu′ (z) + f (u(z)) (p(u(z)) − v(z)) ≥ 0,
 
′′ ′ v(z)
V (u, v) = dv (z) − cv (z) + sv(z) 1 − ≤ 0,
q(u(z))
 
′′ ′ v(z)
V (u, v) = dv (z) − cv (z) + sv(z) 1 − ≥ 0.
q(u(z))

To apply Schauder’s fixed point theorem, we define the Banach space

Bρ R, R2 = φ(z) = (u(z), v(z)) ∈ C R, R2 : |φ(z)|ρ < +∞



with the norm  

−ρ|z| −ρ|z|
|φ(z)|ρ = max sup |u(z)|e , sup |v(z)|e
z∈R z∈R

for small positive constant ρ. Then we will seek traveling waves in wave profile set

Σ = (u, v) ∈ C R, R2 : u(z) ≤ u(z) ≤ u(z), v(z) ≤ v(z) ≤ v(z) .



Obviously, Σ is bounded, closed and convex in C (R, R2 ). Furthermore, we define

F (u, v) = τ u + f (u)(p(u) − v),

 
v
G(u, v) = τ v + sv 1 − ,
q(u)

where τ satisfies
  
′ ′ q(1)
τ > max max {f (1)|p (u)| + q(1)|f (u)|} , s 2 −1 .
u∈[0,1] q(0)

At the moment, one can verify that for (u, v) ∈ [0, 1] × [0, q(1)]

∂F (u, v) ∂F (u, v) ∂G(u, v) ∂G(u, v)

≥ 0, ≤ 0, ≥ 0, ≥ 0.
∂u ∂v ∂u ∂v
Thus, system (3) can be rewritten as
( ′′
u − cu′ − τ u + F (u, v) = 0,
dv ′′ − cv ′ − τ v + G(u, v) = 0.

6
Next, we define the operator P = (P1 , P2 ) : Σ → C (R, R2 ) by
Z z Z ∞ 
1 ν1− (z−y) ν1+ (z−y)
P1 (u, v)(z) = + e + e F (u, v)(y)dy,
ν1 − ν1− −∞ z
Z z Z ∞ 
1 ν2− (z−y) ν2+ (z−y)
P2 (u, v)(z) = e + e G(u, v)(y)dy,
d ν2+ − ν2−


−∞ z

where
1 √  1  √ 
ν1± = c ± c2 + 4τ , ν2± = c ± c2 + 4dτ .
2 2d
It is easy to check that a fixed point of operator P is identical to the solution of system (3). Then
a standard argument similar to [29, 32, 39, 40] gives that P (Σ) ⊆ Σ, P is continuous and compact
with respect to the norm in Bρ (R, R2 ). Thus, solution of system (3) is found by using Schauder’s
fixed point theorem. As a consequence, we conclude the following lemma.

Lemma 2.1 Assume that system (3) has a pair of upper and lower solutions (u(z), v(z)) and
(u(z), v(z)) satisfying

u(zi− ) ≤ u(zi+ ), u(zi+ ) ≤ u(zi− ), v(zi− ) ≤ v(zi+ ), v(zi+ ) ≤ v(zi− ) for zi ∈ E.

Then it has a solution (u(z), v(z)) satisfying for z ∈ R

u(z) ≤ u(z) ≤ u(z), v(z) ≤ v(z) ≤ v(z).

3 The existence of semi-traveling wave

In this section, we employ the upper and lower solutions method to establish a positive solution
of system (3) that meets the left-hand tail limit in (4), that is, semi-traveling wave. We further
investigate the asymptotic behavior of such waves at z = −∞ by using the unstable manifold
theorem, and then provide the sufficient and necessary condition for the existence of semi-traveling
wave.
To construct an appropriate pair of upper and lower solutions, we define P (c, λ) by

P (c, λ) = dλ2 − cλ + s,

which gives two roots λ1 and λ2 as follows

1  √  1  √ 
λ1 =c − c2 − 4ds , λ2 = c + c2 − 4ds . (5)
2d 2d
√
Case 1: Assume that c > c∗ = 2 ds, then 0 < λ1 < λ2 . First, for given constants

0 < β < min {c, λ1 } ,

0 < ε < min {λ1 , λ2 − λ1 } ,

7
  
f (1)q(1)
σ > max 1, ,
β (c − β)
 
−sq(1)
r > max 1, ,
P (c, λ1 + ε)q(0)
we define functions
(
1 − σeβz , z ≤ z1 ,
u(z) = 1, u(z) =
0, z > z1 ,
( ( (6)
λ1 z λ1 z εz
q(1)e , z ≤ 0, q(1)e (1 − re ) , z ≤ z2 ,
v(z) = v(z) =
q(1), z > 0, 0, z > z2 ,
where z1 = −(1/β) ln σ < 0 and z2 = −(1/ε) ln r < 0 by the choice of σ and r.

Case 2: Assume that c = c∗ , we have λ1 = λ2 = λ := c/ (2d). For given constants

h = λe2 /2,
0 < β < min {c, λ} ,
 
2β/λ f (1)hq(1)
σ > max e , ,
(c − β) (λ − β) βe
( 7/2 )
4sh2 q(1)

p 7
r > max h 2/λ, ,
dq(0) 2eλ

we define functions
(
1 − σeβz , z ≤ z1
u(z) = 1, u(z) =
0, z > z1 ,
( ( (7)
λz λz 1/2


−hq(1)ze , z ≤ −2/λ, q(1)e −hz − r(−z) , z ≤ z2 ,
v(z) = v(z) =
q(1), z > −2/λ, 0, z > z2 .
where z1 = −(1/β) ln σ < −2/λ and z2 = −(r/h)2 < −2/λ by the choice of σ and r.
Next, we will check that (u(z), v(z)) and (u(z), v(z)) described in (6) and (7) are indeed a
pair of upper and lower solutions of system (3) for c ≥ c∗ . Apparently, u(z) ≤ u(z), v(z) ≤ v(z)
for z ∈ R, then it is sufficient to verify the requirement (b) in Definition 2.1.

Verification of upper and lower solutions for c > c∗ : First, since u ≡ 1, v ≥ 0 and
p(1) = 0, we have for z ∈ R
U (u, v) = −f (1)v ≤ 0.
Next, for z > z1 , it follows from u = 0 that U (u, v) = 0. For z < z1 < 0, since u = 1 − σeβz and
v = q(1)eλ1 z , we have
U (u, v) =eβz βσ (c − β) − f (1 − σeβz )q(1)e(λ1 −β)z + f (1 − σeβz )p(1 − σeβz )
 

8
 ≥eβz [βσ (c − β) − f (1)q(1)] > 0

by the choice of β and σ. For V (u, v), we know that v = q(1) for z > 0, then
 
q(1)
V (u, v) = V (1, q(1)) = sq(1) 1 − = 0.
q(1)

For z < 0, v = q(1)eλ1 z . From the definition of λ1 , we have

V (u, v) = q(1)dλ21 eλ1 z − q(1)cλ1 eλ1 z + sλ1 eλ1 z − sq(1)e2λ1 z

= P (c, λ1 )q(1)eλ1 z − sq(1)e2λ1 z

= −sq(1)e2λ1 z ≤ 0.

Finally, for z > z2 , V (u, v) = 0 since v = 0. Note that v = q(1)eλ1 z (1 − reεz ) and v = q(1)eλ1 z for
z < z2 < 0, we have for z < z2

v = v − rq(1)e(λ1 +ε)z ,

v ′ = v ′ − rq(1) (λ1 + ε) e(λ1 +ε)z ,

v ′′ = v ′′ − rq(1) (λ1 + ε)2 e(λ1 +ε)z .

Then by the choice of ε and r, we further deduce that

V (u, v) =dv ′′ − drq(1) (λ1 + ε)2 e(λ1 +ε)z − cv ′ + crq(1) (λ1 + ε) e(λ1 +ε)z

s(v − rq(1)e(λ1 +ε)z )2

+ sv − srq(1)e(λ1 +ε)z −
q(u)
sv 2
≥q(1)eλ1 z dλ21 − cλ1 + s − rq(1)e(λ1 +ε)z d (λ1 + ε)2 − c (λ1 + ε) + s −

 
q(0)
sv2
=q(1)eλ1 z P (c, λ1) − rq(1)e(λ1 +ε)z P (c, λ1 + ε) −
q(0)
sv 2
≥ − rq(1)e(λ1 +ε)z P (c, λ1 + ε) −
q(0)
 
(λ1 +ε)z sq(1) (λ1 −ε)z
≥q(1)e −rP (c, λ1 + ε) − e
q(0)
 
(λ1 +ε)z sq(1)
≥q(1)e −rP (c, λ1 + ε) − > 0.
q(0)
Verification of upper and lower solutions for c = c∗ : First, as before, we have

U (u, v) = −f (1)v ≤ 0, for z ∈ R,

U (u, v) = 0, for z > z1 ,

9
 V (u, v) = 0, for z > −2/λ,
V (u, v) = 0, for z > z2 .

Let us go on to prove the remaining inequalities. For U (u, v), when z < z1 < −2/λ, we have
u = 1 − σeβz and v = −hq(1)zeλz . Note that for z ∈ R
1
ze(λ−β)z ≥ − ,
(λ − β)e
after a simple computation, we arrive at

U (u, v) =eβz βσ (c − β) + f (1 − σeβz )hq(1)ze(λ−β)z + f (1 − σeβz )p(1 − σeβz )

 

≥eβz βσ (c − β) + f (1 − σeβz )hq(1)ze(λ−β)z

 

≥eβz βσ (c − β) + f (1)hq(1)ze(λ−β)z
 

 
βz f (1)hq(1)
≥e βσ (c − β) − >0
(λ − β)e
by the choice of β and σ. Next, for V (u, v), it is easy to see that if z < −2/λ, then

v′ = −hq(1) (1 + λz) eλz , v ′′ = −hq(1)λ (2 + λz) eλz .

Then we deduce that for z < −2/λ

V (u, v) ≤dv′′ − cv ′ + sv

= − dhq(1)λ (2 + λz) eλz + chq(1) (1 + λz) eλz − shq(1)zeλz (8)

= − hq(1)eλz [P (c, λ)z + (2dλ − c)] = 0.

Finally, for z < z2 < −2/λ, v = −hq(1)zeλz , we have

v = q(1)eλz −hz − r(−z)1/2 = v − rq(1)(−z)1/2 eλz .

Direct calculation gives

 
′ ′ 1λz −1/2 1/2
v = v + rq(1)e (−z) − λ(−z) ,
2
 
′′ ′′ λz −1/2 1 −3/2 2 1/2
v = v + rq(1)e λ(−z) + (−z) − λ (−z) .
4
Substituting the above equalities into V (u, v), we obtain
 
′′ λz −1/2 1 −3/2 2 1/2
V (u, v) =dv + drq(1)e λ(−z) + (−z) − λ (−z)
4
sv 2
 
′ λz 1 −1/2 1/2
− cv − crq(1)e (−z) − λ(−z) + sv − srq(1)(−z)1/2 eλz − .
2 q(u)

10
From (8), we have for z < −2/λ
dv ′′ − cv ′ + sv = 0.
Since q(u) is increasing function for u > 0, then by the choice of r, we further deduce that
 
λz −1/2 1 −3/2 2 1/2
V (u, v) =drq(1)e λ(−z) + (−z) − λ (−z)
4
sv2
 
λz 1 −1/2 1/2
− crq(1)e (−z) − λ(−z) − srq(1)(−z)1/2 eλz − ,
2 q(u)
sv2
 
λz 1 c

−3/2 −1/2 1/2
=q(1)e dr(−z) + r(−z) dλ − − r(−z) P (c, λ) −
4 2 q(u)
sh2 q(1) 2 λz
 
1
≥q(1)eλz dr(−z)−3/2 − z e
4 q(0)
sh2 q(1)
 
λz −3/2 1 7/2 λz
=q(1)e (−z) dr − (−z) e
4 q(0)
" 7/2 #
2

1 sh q(1) 7
≥q(1)eλz (−z)−3/2 dr − > 0,
4 q(0) 2eλ

by taking advantage of the fact that for z ≤ 0

 7/2
7/2 λz 7
(−z) e ≤ .
2eλ
Summarizing the above discussion, we further conclude the following lemma.
Lemma 3.1 For c ≥ c∗ , system (3) has a non-negative solution (u(z), v(z)).
Proof. For c ≥ c∗ , function pairs (u(z), v(z)) and (u(z), v(z)) described in (6) or (7) are a pair of
upper and lower solutions of system (3), respectively. Moreover, for c > c∗ , we have
u′ (z1− ) = −β < 0 = u′ (z1+ ),
v ′ (0+ ) = 0 < q(1)λ1 = v ′ (0− ),

v ′ (z2− ) = −q(1)εeλ1 z2 < 0 = v ′ (z2+ ).

While, for c = c∗ , we have
u′ (z1− ) = −β < 0 = u′ (z1+ ),

v ′ (z2− ) = −q(1)heλz2 /2 < 0 = v′ (z2+ ),

v ′ ((−2/λ)+ ) = 0 < q(1)he−2 = v ′ ((−2/λ)− ).
Hence, we complete the proof with Lemma 2.1.

Actually, the above non-negative solution is a semi-traveling wave of system (2).

11
 √
Theorem 3.1 For c ≥ c∗ := 2 ds, system (3) has a solution (u(z), v(z)) satisfying

(u(−∞), u′(−∞), v(−∞), v ′(−∞)) = (1, 0, 0, 0)

and 0 < u(z) < 1, 0 < v(z) < q(1) for z ∈ R.

Proof. From Lemma 2.1, we can easily get that for c ≥ c∗

0 ≤ u(z) ≤ u(z) ≤ u(z) = 1 and 0 ≤ v(z) ≤ v(z) ≤ v(z) ≤ q(1) for z ∈ R,

And it is clear that

1 = lim u(z) ≤ lim inf u(z) ≤ lim sup u(z) ≤ lim u(z) = 1,
z→−∞ z→−∞ z→−∞ z→−∞

0 = lim v(z) ≤ lim inf v(z) ≤ lim sup v(z) ≤ lim v(z) = 0,
z→−∞ z→−∞ z→−∞ z→−∞

which implies that (u(−∞), v(−∞)) = (1, 0).

We further claim that 0 < u(z) < 1, 0 < v(z) < q(1) for z ∈ R. For contradiction, we
assume that u(z0 ) = 0 for some z0 ∈ R, then u′ (z0 ) = 0 owing to u(z) ≥ 0 over R. The
uniqueness of solution yields u(z) ≡ 0, which contradicts u(z) ≥ u(z) > 0 for z < z1 . Hence,
we have u(z) > 0 over R. And similar reasons lead to v(z) > 0 over R. In order to prove
u(z) < 1 over R, we also assume that u(z0 ) = 1 for some z0 ∈ R. It follows that u′ (z0 ) = 0
and u′′ (z0 ) ≤ 0. Using the first equation of system (3) and the assumption (H1)-(H2), we have
0 = u′′ (z0 ) − f (u(z0 ))v(z0 ) ≤ −f (u(z0 ))v(z0 ) < 0. Hence, this contradiction yields u(z) < 1 over
R. Similarly, we also get v(z) < q(1) over R by using u(z) < 1 over R.
Finally, we show that (u′ (−∞), v ′ (−∞)) = (0, 0). Indeed, for the first equation of system (3),
we utilize the variation constants formula to deduce that
Z ς
′ c(z−ς) ′ cz
u (z) = e u (ς) − e e−cτ [f (u(τ )) (v(τ ) − p(u(τ )))] dτ,
z

thus, it is easy to check that for z ≤ ς

 
′ c(z−ς) ′
c|u (z)| ≤ ce |u (ς)| + max f (u(τ )) (v(τ ) − p(u(τ ))) .
 
τ ≤ς

Consequently, for fixed ς, we have

 
′ 1 
lim sup |u (z)| ≤ max f (u(τ )) (v(τ ) − p(u(τ )))  .
 
z→−∞ c τ ≤ς

It follows from (H1) and (H2) that

f (u(−∞)) (p(u(−∞)) − v(−∞)) = 0,

then we have w(−∞) = 0 by arbitrary of ς. Similar reasons give v ′ (−∞) = 0.



12
 In order to obtain more detailed information on semi-traveling wave, it is essential for us to
study the asymptotic behavior of semi-traveling wave at z = −∞. Hence, we set w(z) = u′ (z),
y(z) = v ′ (z) and rewrite system (3) as a system of first order ODEs in R4
 ′

 u (z) = w(z),


w ′(z) = cw(z) − f (u(z))(p(u(z)) − v(z)),




v ′ (z) = y(z), (9)



  

 ′ v(z)
dy (z) = cy(z) − sv(z) 1 − .


q(u(z))

The eigenvalues of the linearization of (9) at e0 = (1, 0, 0, 0) are

1  √  1 p 
λ1 = c − c2 − 4ds , λ3 = c + c2 − 4f (1)p′ (1) ,
2d 2
1  √  1 p 
λ2 = 2
c + c − 4ds , λ4 = 2 ′
c − c − 4f (1)p (1) .
2d 2
Obviously, for c ≥ c∗ , we have λ4 < 0 < λ1 , λ2 , λ3 due to p′ (1) < 0 < f (1).

Lemma 3.2 For c ≥ c∗ , the positive solution (u(z), v(z)) obtained by Theorem 3.1 satisfies

y(z) w(z)
lim = λ1 and lim = Λ ∈ {λ1 , λ3 } .
z→−∞ v(z) z→−∞ u(z) − 1

Proof. Based on the choice of upper and lower solutions, from Lemma 2.1, we have
(
O(eλ1 z ), c > c∗ ,
v(z) =
O(−zeλ1 z ), c = c∗ ,

as z → −∞, which implies that

y(z)
lim = λ1 .
z→−∞ v(z)

Now, we show the second equality, which is quite complicated and tedious. We only consider the
case c > c∗ , since when c = c∗ and λ1 = λ2 < (>, =)λ3 , it can be similarly treated by using
generalized eigenvectors and the unstable manifold theorem.
(1) If λ1 , λ2 6= λ3 , then the corresponding eigenvectors are given by
     
−1 −1 −1
 −λ1 
 , e2 =  −λ2  , e3 = −λ3  ,
   
e1 =   −ψ (λ1 )   −ψ (λ2 )   0 
−λ1 ψ (λ1 ) −λ2 ψ (λ2 ) 0

where
λ2 − cλ + f (1)p′ (1)
ψ(λ) = .
f (1)

13
Thus, every solution W (z) of the corresponding linearized system, which converges to e0 as z →
−∞, is given by
W (z) = e0 + C1 e1 eλ1 z + C2 e2 eλ2 z + C3 e3 eλ3 z
for some constants Ci , i = 1, 2, 3. Applying the unstable manifold theorem yields that as z → −∞,
there are α, β and γ such that

u(z) = 1 − αeλ1 z − βeλ2 z − γeλ3 z + h.o.t.,

w(z) = 0 − αλ1 eλ1 z − βλ2 eλ2 z − γλ3 eλ3 z + h.o.t.,

v(z) = 0 − αψ (λ1 ) eλ1 z − βψ (λ2 ) eλ2 z + h.o.t..

Note that v(z) = O(eλ1 z ) as z → −∞, we must have α 6= 0. Then the following hold.

(i) If λ1 < λ3 , then ψ(λ1 ) < 0. Hence, α > 0 due to v(z) > 0 for z ∈ R.

(ii) If λ1 > λ3 , then ψ(λ1 ) > 0. Hence, α < 0 < γ due to v(z) > 0 and u(z) < 1 for z ∈ R.

Consequently, one can easily verify that

w(z)
lim = min {λ1 , λ3 } .
z→−∞ u(z) − 1

(2) If λ1 = λ3 < λ2 , we have a generalized eigenvector

 
0
−f (1)/(2λ1 − c)
e4 =  .
 −1 
−λ1

Thus, every solution W (z) of the corresponding linearized system, which converges to e0 as z →
−∞, is given by
W (z) = e0 + C1 (e1 z + e4 ) eλ1 z + C2 e2 eλ2 z
for some constants Ci , i = 1, 2. Applying the unstable manifold theorem yields that as z → −∞,
there are α and β such that

u(z) = 1 − αzeλ1 z − βeλ2 z + h.o.t.,

w(z) = 0 − α [λ1 z + f (1)/(2λ1 − c)] eλ1 z − βλ2eλ2 z + h.o.t.,

v(z) = 0 − αeλ1 z − βψ (λ2 ) eλ2 z + h.o.t..

Note that v(z) = O(eλ1 z ) as z → −∞ and v(z) > 0, we must have α < 0. Moreover, one can easily
verify that
w(z)
lim = λ1 . (10)
z→−∞ u(z) − 1

14
 (3) If λ1 < λ2 = λ3 , we have a generalized eigenvector
 
0
−f (1)/(2λ2 − c)
e5 =  .
 −1 
−λ2

Thus, every solution W (z) of the corresponding linearized system, which converges to e0 as z →
−∞, is given by
W (z) = e0 + C1 e1 eλ1 z + C2 (e2 z + e5 )eλ2 z
for some constants Ci , i = 1, 2. Applying the unstable manifold theorem yields that as z → −∞,
there are α and β such that

u(z) = 1 − αeλ1 z − βzeλ2 z + h.o.t.,

w(z) = 0 − αλ1 eλ1 z − β [λ2 z + f (1)/(2λ2 − c)] eλ2 z + h.o.t.,

v(z) = 0 − αψ (λ1 ) eλ1 z − βeλ2 z + h.o.t..

Similarly, we have α > 0 and (10) still hold. Therefore, we complete the proof.


Remark 3.1 From Theorem 3.1, we know that v(z) > 0 and u(z) < 1 over R. Combining
with Lemma 3.2, one can infer that y(z) > 0 > w(z) for z ≪ −1, which is used to prove the
monotonicity of traveling wave connecting (1, 0) and (0, µ) in Theorem 4.2.

Finally, let us end this section by proving Theorem 1.1.

The proof of Theorem 1.1: We investigate the linearization of the second equation of
system (3) at (1, 0), which provides two eigenvalues λ1 and λ2 defined in (5). If |c| < c∗ , then
λ1 and λ2 are a pair of complex eigenvalues. If c < −c∗ , then λ1 < λ2 < 0. The former implies
that v(z) cannot have the same sign as z → −∞, while the latter implies that v(z) is unbounded
as z → −∞. Since neither is feasible, system (2) has no semi-traveling wave if c < c∗ , which
concludes the result by coupling Theorem 3.1.


4 The existence of strong-traveling wave

In this section, our main purpose is to prove the existence of strong-traveling waves of system
(2). Firstly, inspired by [17, 23, 36, 40], we shall develop two derivative estimates to produce an
invariant set of system (9), which is a required and crucial effort. Then semi-traveling wave
obtained in Theorem 3.1 can be viewed as a good candidate for strong-traveling waves. Finally,
we deduce the right-hand tail limit of such a solution, which establishes the existence of strong-
traveling waves. In the sequence, we always assume that c ≥ c∗ .
We start by recalling LaSalle’s Invariance Principle.

15
Proposition 4.1 ( [20]) Consider the following initial value problem
( ′
y = f (y),
(11)
y(0) = y0 ,

where f : Rn → Rn is continuous and satisfies Lipschitz condition. Let Σ ⊆ Rn be an open set.

Suppose y(t, y0) is a solution of equation (11) which is positively invariant in Σ. If there is a
continuous and bounded below function V (y) : Σ → Rn such that the orbital derivative of V (y)
along y(t, y0) is non-positive, i.e.,

dV
= grad V (y) · f (y) ≤ 0,
dt
then the ω-limit set of y(t, y0) is contained in the largest invariant set

I = {y ∈ Σ : grad V (y) · f (y) = 0} .

4.1 Derivative estimates

In this subsection, we shall develop two derivative estimates of u(z) and v(z).

Lemma 4.1 There is a large enough K > q(1)/c such that for z ∈ R,

−Kf (u(z)) < w(z) < Kf (u(z)).

Proof. To prove the above inequalities, we define functions

φ1 (z) = w(z) − Kf (u(z)),

φ2 (z) = w(z) + Kf (u(z)).

It is sufficient to show that φ1 (z) < 0 and φ2 (z) > 0 over R. By recalling Theorem 3.1, we have
0 < u(z) < 1 and (u(−∞), w(−∞)) = (1, 0). Thus, there are K > 0 and z0 ≪ −1 such that

φ1 (z) < 0 and φ2 (z) > 0 for z ≤ z0 .

(1) We show that φ1 (z) < 0 for z > z0 by contradiction. Assume that there is a z1 > z0 such
that φ1 (z1 ) = 0 and φ′1 (z1 ) ≥ 0. Since u(z) and v(z) are bounded, and w(z1 ) = Kf (u(z1 )), then
for large enough K > 0

0 ≤ φ′1 (z1 ) = w ′(z1 ) − Kf ′ (u(z1 ))w(z1 )

= cw(z1 ) − f (u(z1 ))p(u(z1 )) + f (u(z1 ))v(z1 ) − Kf ′ (u(z1 ))w(z1 )

= cK − p(u(z1 )) + v(z1 ) − K 2 f ′ (u(z1 )) f (u(z1)) < 0.

 

Hence, the contradiction yields φ1 (z) < 0 for z > z0 .

(2) We show that φ2 (z) > 0 for z > z0 by contradiction. Assume that there is a z1 > z0
such that φ2 (z1 ) = 0 and φ′2 (z1 ) ≤ 0. We claim that there is a z2 > z1 such that φ2 (z2 ) = 0 and

16
φ′2 (z2 ) ≥ 0. To see this, we further assume that φ2 (z) < 0 for all z > z1 , that is, w(z) < −Kf (u(z))
for all z > z1 . Note that 0 < u(z) < 1, 0 < v(z) < q(1) over R and p(u) > 0 for u ∈ (0, 1), then
we have the estimate for z > z1

w ′(z) =cw(z) + f (u(z))v(z) − f (u(z))p(u(z))

<[−cK + v(z) − p(u(z))]f (u(z))
<(q(1) − cK)f (u(z)) < 0

as long as K > q(1)/c. Thus, it is easy to check that for z > z1

w(z) < w(z1 ) = −Kf (u(z1 )) < 0,

which contradicts the boundedness of u(z). Hence, the claim is valid. On the other hand, at
z = z2 , w(z2 ) = −Kf (u(z2 )), we also get for large enough K > 0

0 ≤ φ′2 (z2 ) = w ′(z2 ) + Kf ′ (u(z2 ))w(z2 )

= cw(z2 ) − f (u(z2 ))p(u(z2 )) + f (u(z2))v(z2 ) + Kf ′ (u(z2 ))w(z2 )

= −(K 2 f ′ (u(z2 )) + cK) + v(z2 ) − p(u(z2 )) f (u(z2 )) < 0,

 

due to the boundedness of u(z) and v(z). Hence, the contradiction yields φ2 (z) > 0 for z > z0 .

√
Lemma 4.2 If c ≥ c∗ = 2 ds, then for z ∈ R,
sq(1) c
− v(z) ≤ y(z) ≤ v(z).
cq(0) 2d

Proof. To prove the above inequalities, we define function

c
ψ1 (z) = y(z) − v(z),
2d
sq(1)
ψ2 (z) = y(z) + v(z).
cq(0)
It is sufficient to show that ψ1 (z) ≤ 0 and ψ2 (z) ≥ 0 over R.
(1) We show that ψ1 (z) ≤ 0 over R by contradiction. Assume√that there is a z0 such that
ψ1 (z0 ) > 0, then the v(z) equation of system (3) gives for c ≥ c∗ = 2 ds

c c2
dψ1′ (z) ′
− ψ1 (z) = dy (z) − cy(z) + v(z)
2 4d
2
 
sv(z) c
= + − s v(z)
q(u(z)) 4d
sv 2 (z)
≥ > 0.
q(u(z))

17
Hence, by the comparison principle, we have ψ1 (z) > ec(z−z0 )/2d ψ1 (z0 ) > 0 for z > z0 . Using the
expression of ψ1 (z) and the comparison principle again, we obtain
v(z) > ec(z−z0 )/2d v(z0 ) for z > z0 .
Hence, we have v(z) → +∞ as z → +∞, which contradicts the boundedness of v(z).
(2) We show that ψ2 (z) ≥ 0 over R. From Lemma 3.2, we have for c ≥ c∗
y(z)
lim = λ1 > 0,
z→−∞ v(z)

then there is a z0 ≪ −1 such that for z ≤ z0

 
y(z) sq(1)
ψ2 (z) = v(z) + > 0.
v(z) cq(0)
For contradiction, we assume that there is a z1 > z0 such that ψ2 (z1 ) = 0 and ψ2′ (z1 ) ≤ 0. Similar to
the proof of Lemma 4.1, ψ2 (z2 ) = 0 and ψ2′ (z2 ) ≥ 0 for some z2 > z1 . Actually, since 0 < u(z) < 1,
0 < v(z) < q(1) over R and q(u(z)) > q(0), then as long as y ′(z) ≤ −sq(1)v(z)/cq(0), we have
 
′ v(z)
dy (z) = cy(z) − sv(z) 1 −
q(u(z))
  (12)
v(z) q(1)
≤ sv(z) − − 1 < 0.
q(u(z)) q(0)
Therefore, if ψ2 (z) < 0 for all z > z1 , that is, for z > z1
sq(1)
y(z) < − v(z),
cq(0)
then (12) yields that for z > z1 ,
sq(1)
y(z) < y(z1 ) = − v(z1 ) < 0,
cq(0)
which contradicts the positivity of v(z). Thanks to ψ2′ (z2 ) ≥ 0 and y(z2 ) = −sq(1)v(z2 )/cq(0), we
have  2
′ sq(1) sq(1)
y (z2 ) ≥ − y(z2) = v(z2 ) > 0.
cq(0) cq(0)
Thus, from 0 < v(z2 ) < q(1) and q(u(z2 )) > q(0), we deduce that
 
′ v(z2 )
0 = dy (z2 ) − cy(z2 ) + sv(z2 ) 1 −
q(u(z2))
 
q(1) v(z2 )
≥ sv(z2 ) − + 1 > 0,
q(0) q(u(z2 ))
which leads to a contradiction. Hence, we complete the proof.

Then according to Lemmas 4.1 and 4.2, the set
 
 2sq(1) c
D = χ(z)0 < u < 1, 0 < v < q(1), −Kf (u) < w < Kf (u), − v<y< v

cq(0) d
is an invariant set of system (9), where χ(z) := (u, w, v, y)(z).

18
4.2 Convergence of semi-traveling wave
In this subsection, we will deduce the right-hand tail limit of semi-traveling wave obtained in
Theorem 3.1. To be specific, based on the open set D, the appropriate Lyapunov functions are
constructed for system (9), such that the candidate is a strong-traveling wave.
We firstly prove Theorem 1.2. For convenience, let us recall it again.

Theorem 4.1 Assume that (H1)-(H2) hold and h(u) = 1. For c ≥ c∗ , if p(0) > µ and
(P) (u∗ + µ)f (u) − (u − u∗ )(u + µ)f ′ (u) > 0 for u ∈ (0, 1),
then system (2) has a traveling wave connecting (1, 0) and (u∗ , v ∗ ).

Proof. Let u v
η − u∗ 1 η − v∗
Z Z
H(u, v) = dη + dη.
u∗ q(η)f (η) s v∗ η
Obviously, H(u, v) ≥ 0 for u, v > 0. Then we define Lyapunov function L : R4 → R by

L(u, w, v, y) = cH(u, v) − w∂u H − dy∂v H,

thus, L(u, w, v, y) is continuous function with a lower bound since for (u, w, v, y) ∈ D
w(u − u∗ ) w(u + u∗ ) 2w 2K
w∂u H = ≤ ≤ < ,
f (u)q(u) q(0)f (u) q(0)f (u) q(0)
y(v − v ∗ ) 2sv ∗
 
1 c y ∗  q(1) c
y∂v H = ≤ v− v ≤ + .
sv s d v s d cq(0)

From (9), the orbital derivative of L along χ(z) is

L′ (χ(z)) = (cw − w ′)∂u H + (cy − dy ′)∂v H − w 2 ∂uu H − dy 2 ∂vv H

= f (u)(p(u) − v)∂u H + sv (1 − v/q(u)) ∂v H − w 2 ∂uu H − dy 2∂vv H
= [(u − u∗ ) (p(u) − v) + (v − v ∗ ) (q(u) − v)] /q(u) − w 2 ∂uu H − dy 2∂vv H.

Using the expression of p(u∗) = q(u∗) = v ∗ , we further obtain

(u − u∗ ) (p(u) − v) = (u − u∗ ) (p(u) − p(u∗ )) + (u − u∗ ) (v ∗ − v),

(v − v ∗ ) (q(u) − v) = (v − v ∗ ) (q(u) − q(u∗ )) − (v − v ∗ )2 .

Hence,

L′ (χ(z)) = (u − u∗ ) (p(u) − p(u∗ )) − (v − v ∗ )2 /q(u) − w 2 ∂uu H − dy 2∂vv H + G(u, v),

 

where

G(u, v) = (v − v ∗ ) [(q(u) − q(u∗ )) − (u − u∗ )] /q(u)

= (v − v ∗ ) [(u − u∗ ) − (u − u∗ )] /q(u) = 0,

19
due to q(u) = uh(u) + µ = u + µ. Moreover, a simple computation gives

v∗ (u∗ + µ)f (u) − (u − u∗ )(u + µ)f ′(u)

∂vv H = and ∂uu H = .
sv 2 (u + µ)2 f 2 (u)

Thus, (H2) and (P) imply that the first and third terms are non-positive in L′ (χ(z)). Hence, the
orbital derivative of L along χ(z) is non-positive, and it is zero if and only if χ(z) = (u∗ , 0, v ∗, 0).
On the other hand, p(0) > µ implies that there is a unique positive equilibrium (u∗ , v ∗ ) in D.
Therefore, we complete the proof by using LaSalle’s Invariance Principle.

Finally, we show Theorem 1.3. By recalling that

A = v ∈ C 2 (R) | 0 < v(z) < µ and y(z) > 0 over R ,



  
2
 ∃ zv such that v(zv ) = µ, y(z) > 0 in (−∞, zv )
B = v ∈ C (R)   ,
and v(z) > µ in (zv , +∞)

then Theorem 1.3 is repeated as below.

Theorem 4.2 Assume that (H1)-(H3) hold. For c ≥ c∗ , if

f (u) = ug(u) and p(u) = (1 − u)/g(u),

where g(u) satisfies

h(1)(h(1) + µ)
min g(u) > , (13)
u∈[0,1] µ2 h(0)
then the following hold.

(a) System (2) has a traveling wave (u(z), v(z)) connecting (1, 0) and (0, µ).

(b) v(z) ∈ A ∪ B and w(z) < 0 over R.

Proof. Firstly, we show the existence by LaSalle’s Invariance Principle. It follows from (13) that
there is a ̺ satisfying
sh(1) sµh(0)

<̺< . (14)
µ minu∈[0,1] g(u) h(1) + µ
Then we define Lyapunov function L : R4 → R by
Z v
η − q(0) dq(0)y
L(χ(z)) = ̺(cu − w) + c dη − dy + .
q(0) η v

Since q(u) − q(0) = q(u) − µ = uh(u) and (9), after a direct computation, we have

L′ (χ(z)) =̺ (cw − w ′ ) + (1 − q(0)/v) (cy − dy ′) − dq(0) (y/v)2

=̺u (1 − u − g(u)v) + s(v − q(0)) (1 − v/q(u)) − dq(0) (y/v)2

= − ̺u2 + ̺u (1 − g(u)v) − s(v − q(0))2 /q(u)

20
 + s(v − q(0))(q(u) − q(0))/q(u) − dq(0) (y/v)2

= − ̺u2 − s(v − q(0))2 /q(u) − dq(0) (y/v)2

+ ̺u [1 − g(u)v + s(v − q(0))h(u)/̺q(u)]

= − ̺u2 − s(v − q(0))2 /q(u) − dq(0) (y/v)2 + ̺uC(u),

where

C(u) = 1 − g(u)v + s(v − q(0))h(u)/̺q(u).

Through the monotonicity of h(u) and 0 < u < 1, we have

C(u) = [(̺q(u) − sq(0)h(u)) + v(sh(u) − ̺q(u)g(u))] /̺q(u)

< [(̺q(1) − sq(0)h(0)) + v(sh(1) − ̺q(0)g(u))] /̺q(u)
= {[̺(h(1) + µ) − sµh(0)] + v(sh(1) − ̺µg(u))} /̺q(u).

Automatically, C(u) < 0 for u ∈ [0, 1] by the choice of ̺. Then the orbital derivative of L along
χ(z) is non-positive and it is zero if and only if u = 0, v = µ, y = 0. Thus, LaSalle’s Invariance
Principle yields χ(z) → (0, w(+∞), µ, 0) as z → +∞ if w(+∞) exists. Since u(z), v(z) and w(z)
are bounded for z ∈ R, then the u(z) equation gives that w(z) is uniformly continuous. And we
further infer that w(+∞) = 0 by the Barbalat Lemma. Hence, the existence of traveling wave is
proved.
Next, we show v(z) ∈ A ∪ B. We claim that y(z) 6= 0 as long as z satisfies 0 < v(z) ≤ µ.
Actually, we assume that there is a z0 ∈ R such that y(z0) = 0 and 0 < v(z0 ) ≤ µ. Since
0 < u(z0 ) < 1, µ = q(0) < q(u(z0 )) by the monotonicity of q(u), we have
   
′ v(z0 ) µ
dy (z0 ) = −sv(z0 ) 1 − < −sv(z0 ) 1 − = 0,
q(u(z0 )) q(0)

which gives that v(z) attains a local maximum at z0 . Thus, 0 < v(z1 ) < µ and y(z1 ) < 0 for some
z1 > z0 . Observe that  
′ v(z)
dy (z) = cy(z) − sv(z) 1 − <0
q(u(z))
as long as z satisfies y(z) ≤ 0 and 0 < v(z) ≤ µ, we have y ′ (z) < 0 and y(z) < y(z1 ) < 0 for
z > z1 , which contradicts the positive of v(z). Hence, we prove the assertion of the claim. Since
v(−∞) = 0 and v(+∞) = µ, y(z) > 0 as long as z satisfies 0 < v(z) ≤ µ. Meanwhile, it holds that
either v(z) ∈ (0, µ) over R or there exists a zv such that v(zv ) = µ. For the former case, clearly,
v(z) ∈ A. For the latter case, y(z) > 0 for z ≤ zv . On the other hand, there is no point z > zv
with v(z) = µ and y(z) ≤ 0, which implies that v(z) ∈ B.
Finally, we show that w(z) < 0 over R, which be considered in two different cases.
For the case v ∈ A. For contradiction, we assume that there is z0 ∈ R such that w(z0 ) = 0.
From Remark 3.1, we know that w(z) < 0 for z ≪ −1, then we can reasonably set

z̄ = sup {z ∈ (−∞, z0 ] : w(ς) < 0 for ς ∈ (−∞, z)} .

21
It follows from the definition of z̄ that w(z̄) = 0 and w ′(z̄) ≥ 0. Thus, the u(z) equation of system
(3) gives
v(z̄) ≥ p(u(z̄)). (15)
Using u(+∞) = 0 and the definition of z̄, we can define

ẑ = inf {z ∈ (z̄, +∞) : w(ς) > 0 for ς ∈ (z̄, z)} .

Obviously, u(ẑ) > u(z̄), w(ẑ) = 0 and w ′ (ẑ) ≤ 0. Then the u(z) equation gives

v(ẑ) ≤ p(u(ẑ)). (16)

Since p′ (u) < 0 for u > 0, we see that

v(z̄) ≥ p(u(z̄)) > p(u(ẑ)) ≥ v(ẑ),

which contradicts with y(z) > 0 over R. Therefore, we have w(z) < 0 over R if v ∈ A.
For the another case v ∈ B, we first claim that w(z) < 0 if z ≥ zv . Otherwise, we assume that
w(z0 ) ≥ 0 for some z0 ≥ zv . Since v(z) ≥ µ for z ≥ zv , we deduce that for z ≥ zv

w ′(z) − cw(z) = f (u(z)) (v(z) − p(u(z)))

(17)
≥ f (u(z)) (µ − p(0)) > 0,

here we have used the fact that

1 1 µ2 h(0)
p(0) = < < < µ.
g(0) min g(u) h(1)(h(1) + µ)
u∈[0,1]

For the case w(z0 ) > 0, (17) and the comparison principle give for all z > z0

w(z) > w(z0 )ec(z−z0 ) , (18)

which contradicts the boundedness of u(z). For the case w(z0 ) = 0, (18) yields that w(z1 ) > 0 for
some z1 > z0 , using (17) and the comparison principle again, we obtain all z > z1

w(z) > w(z1 )ec(z−z1 ) ,

which contradicts the boundedness of u(z). Thus, we prove the assertion of the claim. Therefore,
it is sufficient to prove that w(z) < 0 for z < zv . Similar to the case v(z) ∈ A, we assume that
there is z0 < zv such that w(z0 ) = 0 and define

z̄ = sup {z ∈ (−∞, z0 ] : w(ς) < 0 for ς ∈ (−∞, z)} .

Obviously, from the definition of z̄, we have w(z̄) = 0 and w ′ (z̄) ≥ 0. Then inequality (15) is
still valid. Note that w(zv ) < 0 from the above claim, then there is a ẑ ∈ (z̄, zv ) such that u(z)
attains a local maximum at ẑ, and inequality (16) also remains valid. These again contradict the
monotonicity of v(z). Above all, we complete the proof.


22
5 Application
In this section, we apply Theorems 1.2 and 1.3 to modified Leslie-Gower system with Holling
II or Lotka-Volterra type, then some numerical simulations are carried by using MATLAB. It
should be pointed out that we achieve a better result on existence of traveling wave connecting
(1, 0) and (u∗ , v ∗ ) by constructing a novel Lyapunov function for system with Lotka-Volterra type.

5.1 Modified Leslie-Gower system with Holling II type

Let us now consider system
 
∂u(x, t) av(x, t)



 ∂t = ∆u(x, t) + u(x, t) 1 − u(x, t) − ,
u(x, t) + e1

  (19)
 ∂v(x, t) v(x, t)
= d∆v(x, t) + sv(x, t) 1 − .



∂t u(x, t) + e2
where a, e1 , d, s and e2 are positive constants. Then
au 1
f (u) = , p(u) = (1 − u)(u + e1 ), q(u) = u + e2 , h(u) = 1, µ = e2 .
u + e1 a
For convenience, we define

γ(x) = (1 − x)(x + e1 )/(x + e2 ), x ∈ (0, 1).

By Theorem 1.2, we have the following result on existence of traveling wave connecting (1, 0) and
(u∗ , v ∗ ).
Theorem 5.1 For c ≥ c∗ , there is a constant

 e1 /e2 ,
 if e1 ∈ [1, +∞) ∩ (0, e2 ] ,
ā =
  
e1 − e2
 min e1 /e2 , γ
 , if e1 ∈ [1, +∞) ∩ (e2 , +∞) ,
1 + 2e1 + e1 e2

such that if a < ā, then system (19) has a traveling wave connecting (1, 0) and (u∗ , v ∗ ).
Proof. Clearly, Assumption 1.1 hold if e1 ≥ 1. If a < e1 /e2 , there is a u∗ ∈ (0, 1) satisfying
K(u) = u2 − (1 − a − e1 )u + (ae2 − e1 ) = 0, such that system has a unique positive equilibrium
(u∗ , v ∗ ). Then we only check the assumption (P), that is, for u ∈ (0, 1)

B(u) = (u∗ + e2 − e1 )u2 + 2u∗e1 u + e1 e2 u∗ > 0.

(1) When u∗ = e1 − e2 ∈ (0, 1), that is, e1 ∈ (e2 , e2 + 1) and a = γ(e1 − e2 ). It is obvious that
B(u) > 0 for u ∈ (0, 1).
(2) When u∗ ∈ (e1 − e2 , 1), we have B ′ (u) > 0 for u ∈ (0, 1) and B(u) > B(0) = e1 e2 u∗ > 0.
(a) If e1 ∈ (0, e2 ], we have u∗ ∈ (0, 1) ⊆ (e1 − e2 , 1).
(b) If e1 ∈ (e2 , e2 + 1), then u∗ ∈ (e1 − e2 , 1) is equivalent to K(e1 − e2 ) < 0 from the graph
of K(u), that is, a < γ(e1 − e2 ).

23
 (3) When u∗ ∈ (0, e1 − e2 ), B(u) is a concave function. Due to B(0) > 0, then it is sufficient to
prove that B(1) > 0, that is
e1 − e2
e1 − e2 > u∗ > Λ := .
1 + 2e1 + e1 e2
(a) If e1 ∈ (e2 , e2 +1), then u∗ ∈ [Λ, e2 − e1 ), which is equivalent to K(Λ) 6 0 and K(e1 −e2 ) >
0 from the graph of K(u). Note that γ(x) is decreasing on x if e1 > e2 , thus, we have
γ(e1 − e2 ) < a 6 γ(Λ).
(b) If e1 ∈ [e2 + 1, +∞), then u∗ ∈ [Λ, 1), that is, a 6 γ(Λ).
Based on (1), (2)-(b) and (3), we know that B(u) > 0 for u ∈ (0, 1) if
  
e1 − e2
a < min e1 /e2 , γ for e1 ∈ (e2 , +∞).
1 + 2e1 + e1 e2
To summarize, we complete the proof.

Next, according to Theorem 1.3, we have the following result.
Theorem 5.2 Assume that c ≥ c∗ , if e1 ≥ 1 and ae22 > (1 + e1 )(1 + e2 ), then system (19) has a
traveling wave connecting (1, 0) and (0, e2 ) and satisfying v(z) ∈ A ∪ B and w(z) < 0 over R.
Proof. Clearly, if e1 ≥ 1, then Assumption 1.1 hold. And g(u) = a/(u + e1 ). From Theorem 1.3,
we complete the proof.

Below we use MATLAB to further present some numerical simulations. We consider the initial
value as follows
(
0, f or − 200 < x ≤ 100,
u(x, 0) = 1, −200 ≤ x ≤ 200 and v(x, 0) = (20)
0.1, f or 100 < x ≤ 200.
First of all, let d = 1, s = 0.5, e1 = 2 and e2 = 1.2, then ā ≈ 1.4373. At this time, we set a = 1.4,
then (u∗ , v ∗ ) ≈ (0.1266, 1.3266). On the other hand, let d = 1, s = 0.5, e1 = 1.2 and e2 = 1.4, then
ā ≈ 0.8571. We further choose a = 0.7, then (u∗ , v ∗ ) = (0.2, 1.6). Thus, the above two groups of
parameters satisfy different conditions in Theorem 5.1. Under the same initial value (20), one can
observe traveling wave connecting connecting (1, 0) and (u∗ , v ∗ ) of system (19) from Figure 1 and
Figure 2. Moreover, we also select d = 1, s = 0.5, e1 = 1.2, e2 = 0.5 and a = 15, which satisfies
the condition in Theorem 5.2, then system (19) has a traveling wave connecting (1, 0) and (0, e2 )
from Figure 3.

5.2 Modified Leslie-Gower system with Lotka-Volterra type

Let us now consider system

∂u(x, t)
= ∆u(x, t) + u(x, t) (1 − u(x, t) − av(x, t)) ,


 ∂t

  (21)
 ∂v(x, t) v(x, t)
= d∆v(x, t) + sv(x, t) 1 − .


∂t u(x, t) + e2


24
 t=0 t=20
1.5 1.5

1 1
u,v

u,v
0.5 0.5

u u
v v
0 0
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200
x x

t=40 t=60
1.5 1.5

1 1
u,v

u,v
0.5 0.5

u u
v v
0 0
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200
x x

Figure 1: Traveling wave of system (19) at different time

using the parameters d = 1, s = 0.5, e1 = 2, e2 = 1.2 and a = 1.4.

t=0 t=20
1.8 1.8

1.6 1.6

1.4 1.4

1.2 1.2

1 1
u,v

u,v

0.8 0.8

0.6 0.6

0.4 0.4

0.2 u 0.2 u
v v
0 0
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200
x x

t=40 t=60
1.8 1.8

1.6 1.6

1.4 1.4

1.2 1.2

1 1
u,v

u,v

0.8 0.8

0.6 0.6

0.4 0.4

0.2 u 0.2 u
v v
0 0
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200
x x

Figure 2: Traveling wave of system (19) at different time

using the parameters d = 1, s = 0.5, e1 = 1.2, e2 = 1.4 and a = 0.7.

25
 t=0 t=20
1 1

0.9 0.9

0.8 0.8

0.7 0.7

0.6 0.6
u,v

u,v
0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 u 0.1 u
v v
0 0
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200
x x

t=40 t=60
1 1

0.9 0.9

0.8 0.8

0.7 0.7

0.6 0.6
u,v

u,v
0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 u 0.1 u
v v
0 0
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200
x x

Figure 3: Traveling wave of system (19) at different time

using the parameters d = 1, s = 0.5, e1 = 1.2, e2 = 0.5 and a = 15.

where a, d, s and e2 are positive constants. Then

1−u
f (u) = au, p(u) = , q(u) = u + e2 , h(u) = 1, µ = e2 ,
a
and  
∗ ∗ 1 − ae2 1 + e2
(u , v ) = , .
1+a 1+a
From Theorem 1.2, one can easily verify that if a ≤ 1/(1 + e2 ), then system (21) has a traveling
wave connecting (1, 0) and (u∗ , v ∗). However, inspired by [15, 38], we obtain a better result by
constructing a novel Lyapunov function.
We need the following lemma to proceed.
Lemma 5.1 The polynomial S(a) = a3 + a2 − 16a(1 + e2 ) − 16(1 + e2)(2 + e2 ) has a unique positive
root ã satisfying ã > 1/(1 + e2 ).
Proof. Note that the coefficient sign of S(a) changes only once, it follows from Descartes’ Rule of
Signs that S(a) has a unique positive root ã. To show that ã > 1/(1 + e2 ), we just need to prove
S(1/(1 + e2 )) < 0. Indeed, we have
 
3 1
−(1 + e2 ) S = 16(1 + e2 )5 + 16(1 + e2 )4 + 16(1 + e2 )3 − (1 + e2 ) − 1 > 0.
1 + e2
Hence, we complete the proof.

With the aid of Lemma 5.1, we begin to prove the following theorem.

26
Theorem 5.3 For c ≥ c∗ , there is a constant ã > 1/(1 + e2 ) satisfying S(ã) = 0 such that if
a < ā = min {ã, 1/e2}, then system (21) has a traveling wave connecting (1, 0) and (u∗, v ∗ ).

Proof. Define u v
(η + e2 )η − (u∗ + e2 )u∗ η − v∗
Z Z
H(u, v) = dη + ̺ dη,
u∗ (η + e2 )η v∗ η
where ̺ is a positive constant to be determined later. It is clear that

∂u H =[(u + u∗ + e2 )(u − u∗ )]/[(u + e2 )u], ∂v H = ̺(v − v ∗ )/v,

(22)
∂uu H = [u∗ (u∗ + e2 )(2u + e2 )] /[(u + e2 )u]2 , ∂vv H = ̺v ∗ /v 2 .

Then we define Lyapunov function L : R4 → R by

L(u, w, v, y) = cH(u, v) − w∂u H − dy∂v H,

which is clearly continuous function with lower bound in D. A simple calculation yields

L′ (χ(z)) =(cw − w ′ )∂u H + (cy − dy ′)∂v H − w 2∂uu H − y 2 ∂vv H

=(u + u∗ + e2 )(u − u∗ )(1 − u − av)/(u + e2 )
+ ̺s(u + e2 − v)(v − v ∗ )/(u + e2 ) − w 2 ∂uu H − y 2∂vv H
=H(u, v)/(u + e2 ) − w 2 ∂uu H − y 2 ∂vv H

where
H(u, v) = (u + u∗ + e2 )(u − u∗ ) (1 − u − av) + ̺s(v − v ∗ ) (u + e2 − v) .
To show that L′ (χ(z)) ≤ 0 for u, v > 0, it is sufficient to prove H(u, v) ≤ 0 owing to ∂uu H > 0
and ∂vv H > 0 for u, v > 0 from (22). Note that

1 − u − av = −(u − u∗ ) − a(v − v ∗ ),
u + e2 − v = (u − u∗ ) − (v − v ∗ ),

by substituting these into H(u, v), we arrive at

H(u, v) = − (u + u∗ + e2 )(u − u∗ )2 + [̺s − a(u + u∗ + e2 )](u − u∗ )(v − v ∗ ) − ̺s(v − v ∗ )2

" #2
∗
p ̺s − a(u + u + e2 )
=− (u + u∗ + e2 )(u − u∗ ) − p (v − v ∗ )
2 (u + u∗ + e2 )

[̺s − a(u + u∗ + e2 )]2

 
− ̺s − (v − v ∗ )2 .
4(u + u∗ + e2 )

If [̺s − a(u + u∗ + e2 )]2 − 4̺s(u + u∗ + e2 ) < 0, that is

R(̺s) := (̺s)2 − 2(a + 2)(u + u∗ + e2 )̺s + a2 (u + u∗ + e2 )2 < 0, (23)

27
then H(u, v) ≤ 0 for u, v > 0. Obviously, R(x) has two roots 0 < r− (u) < r+ (u), where
√ 2
r− (u) = (u + u∗ + e2 ) a+1−1 ,
√ 2
r+ (u) = (u + u∗ + e2 ) a+1+1 .

And (23) holds if and only if r− (u) < ̺s < r+ (u) for u ∈ (0, 1), which is equivalent to

sup r− (u) < inf r+ (u)

u∈(0,1) u∈(0,1)

by the arbitrariness of u. More precisely,

√ 2 √ 2
(1 + u∗ + e2 ) a + 1 − 1 < (u∗ + e2 ) a+1+1 .

Simplifying the above inequality, we have

√ 2 √
a + 1 − 1 < 4(u∗ + e2 ) a + 1.

By using the expression u∗ + e2 = (1 + e2 )/(1 + a), we further get

√
(a + 2) a + 1 < 2(a + 1) + 4(1 + e2 ),

which is equivalent to

S(a) = a3 + a2 − 16a(1 + e2 ) − 16(1 + e2 )(2 + e2 ) < 0.

From Lemma 5.1, we complete the proof.


The last theorem establishes the existence of traveling wave connecting (1, 0) and (0, e2 ) of
system (21).
Theorem 5.4 Assume that c ≥ c∗ , if ae22 > 1 + e2 , then system (21) has a traveling wave con-
necting (1, 0) and (0, e2 ) and satisfying v(z) ∈ A ∪ B and w(z) < 0 over R.
Proof. Clearly, Assumptions 1.1 hold and g(u) = a. From Theorem 1.3, we complete the proof.

We continue to illustrate our results by numerical simulations under the initial value (20).
Let d = 1, s = 0.5, e2 = 0.1 and a = 15, Figure 4 depicts that u fades out. Moreover, let d = 1,
s = 0.5, e2 = 0.1, then ā ≈ 4.5895. We further choose a = 4.5, then (u∗ , v ∗ ) = (0.1, 0.2). As show
in Figure 5, system (21) has a traveling wave connecting (1, 0) and (u∗ , v ∗ ).

Acknowledgments
This work is supported by the National Natural Science Foundation of China (No. 12171039 and
12271044).

28
 t=0 t=20
1 1

0.9 0.9

0.8 0.8

0.7 0.7

0.6 0.6
u,v

u,v
0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 u 0.1 u
v v
0 0
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200
x x

t=40 t=60
1 1

0.9 0.9

0.8 0.8

0.7 0.7

0.6 0.6
u,v

u,v
0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 u 0.1 u
v v
0 0
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200
x x

Figure 4: Traveling wave of system (19) at different time

using the parameters d = 1, s = 0.5, a = 15 and e2 = 0.1.

t=0 t=20
1 1

0.9 0.9

0.8 0.8

0.7 0.7

0.6 0.6
u,v

u,v

0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 u 0.1 u
v v
0 0
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200
x x

t=40 t=60
1 1

0.9 0.9

0.8 0.8

0.7 0.7

0.6 0.6
u,v

u,v

0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 u 0.1 u
v v
0 0
-200 -150 -100 -50 0 50 100 150 200 -200 -150 -100 -50 0 50 100 150 200
x x

Figure 5: Traveling wave of system (21) at different time

using the parameters d = 1, s = 0.5, a = 4.5 and e2 = 0.1.

29
References
[1] Abrams P A, Ginzburg L R. The nature of predation: prey dependent, ratio dependent or
neither? Trends in Ecology & Evolution, 2000, 15(8): 337-341.

[2] Ai S, Du Y, Peng R. Traveling waves for a generalized Holling-Tanner predator-prey model,

Journal of Differential Equations, 2017, 263(11): 7782-7814.

[3] Arditi R, Ginzburg L R. Coupling in predator-prey dynamics: ratio-dependence, Journal of

Theoretical Biology, 1989, 139(3): 311-326.

[4] Aronson D G, Weinberger H F. Multidimensional nonlinear diffusion arising in population

genetics, Advances in Mathematics, 1978, 30(1): 33-76.

[5] Aronson D G, Weinberger H F. Nonlinear diffusion in population genetics, combustion, and

nerve pulse propagation, Partial differential equations and related topics. Springer, Berlin,
Heidelberg, 1975: 5-49.

[6] Aziz-Alaoui M A, Okiye M D. Boundedness and global stability for a predator-prey model
with modified Leslie-Gower and Holling-type II schemes, Applied Mathematics Letters, 2003,
16(7): 1069-1075.

[7] Aziz-Alaoui M A. Study of a Leslie-Gower-type tritrophic population model, Chaos, Solitons

& Fractals, 2002, 14(8): 1275-1293.

[8] Cantrell R S, Cosner C, Ruan S. Spatial Ecology, CRC Press, 2010.

[9] Chen Y Y, Guo J S, Yao C H. Traveling wave solutions for a continuous and discrete diffusive
predator-prey model, Journal of Mathematical Analysis and Applications, 2017, 445(1): 212-
239.

[10] Ducrot A, Giletti T, Matano H. Spreading speeds for multidimensional reaction-diffusion sys-
tems of the prey-predator type, Calculus of Variations and Partial Differential Equations, 2019,
58(4): 1-34.

[11] Ducrot A, Langlais M. A singular reaction-diffusion system modelling prey-predator inter-

actions: Invasion and co-extinction waves, Journal of Differential Equations, 2012, 253(2):
502-532.

[12] Ducrot A, Liu Z, Magal P. Large speed traveling waves for the Rosenzweig-MacArthur predator-
prey model with spatial diffusion, Physica D: Nonlinear Phenomena, 2021, 415: 132730.

[13] Dunbar S R. Travelling wave solutions of diffusive Lotka-Volterra equations, Journal of Math-
ematical Biology, 1983, 17(1): 11-32.

[14] Dunbar S R. Traveling wave solutions of diffusive Lotka-Volterra equations: a heteroclinic

connection in R4 , Transactions of the American Mathematical Society, 1984: 557-594.

[15] Du Y, Hsu S B. A diffusive predator-prey model in heterogeneous environment, Journal of

Differential Equations, 2004, 203(2): 331-364.

30
[16] Freedman H I. Deterministic Mathematical Models in Population Ecology, Marcel Dekker
Incorporated, 1980.
[17] Fu S C, Tsai J C. Wave propagation in predator-prey systems, Nonlinearity, 2015, 28(12):
4389.
[18] Goh B S. Management and Analysis of Biological Populations, Elsevier, 2012.

[19] Gonźalez-Olivares E, Arancibia-Ibarra C, Rojas-Palma A, et al. Bifurcations and multistability

on the May-Holling-Tanner predation model considering alternative food for the predators,
Mathematical Biosciences and Engineering, 2019, 16(5): 4274-4298.

[20] Hale J K. Ordinary Differential Equations, Robert E. Krieger Publishing Company, 1980.

[21] Holling C S. The functional response of invertebrate predators to prey density, The Memoirs
of the Entomological Society of Canada, 1966, 98(S48): 5-86.

[22] Hsu C H, Lin J J. Existence and non-monotonicity of traveling wave solutions for general
diffusive predator-prey models, Communications on Pure and Applied Analysis, 2019, 18(3):
1483-1508.

[23] Hsu C H, Yang C R, Yang T H, et al. Existence of traveling wave solutions for diffusive
predator-prey type systems, Journal of Differential Equations, 2012, 252(4): 3040-3075.

[24] Huang J, Lu G, Ruan S. Existence of traveling wave solutions in a diffusive predator-prey

model, Journal of Mathematical Biology, 2003, 46(2): 132-152.

[25] Huang W. Traveling wave solutions for a class of predator-prey systems, Journal of Dynamics
and Differential Equations, 2012, 24(3): 633-644.

[26] Leslie P H, Gower J C. The properties of a stochastic model for the predator-prey type of
interaction between two species, Biometrika, 1960, 47(3/4): 219-234.

[27] Leslie P H. Some further notes on the use of matrices in population mathematics, Biometrika,
1948, 35(3/4): 213-245.

[28] Lotka A J. Elements of Physical Biology, Williams & Wilkins, 1925.

[29] Ma S. Traveling wavefronts for delayed reaction-diffusion systems via a fixed point theorem,
Journal of Differential Equations, 2001, 171(2): 294-314.

[30] Murray J D. Mathematical Biology, Springer-verlag, New York, 1989.

[31] Rosenzweig M L, MacArthur R H. Graphical representation and stability conditions of
predator-prey interactions, The American Naturalist, 1963, 97(895): 209-223.

[32] Ruan W, Feng W, Lu X. On traveling wave solutions in general reaction-diffusion systems

with time delays, Journal of Mathematical Analysis and Applications, 2017, 448(1): 376-400.

[33] Tian Y, Wu C, Liu Z. Traveling wave solutions of an ordinary–parabolic system in R2 and a

2D-strip. Applicable Analysis and Discrete Mathematics, 2016, 10(1): 208-230.

31
[34] Tian Y, Wu C. Traveling wave solutions of a diffusive predator-prey model with modified
Leslie-Gower and Holling-type II schemes, Proceedings-Mathematical Sciences, 2018, 128(3):
1-18.

[35] Turchin P. Complex Population Dynamics, Complex Population Dynamics. Princeton Univer-
sity Press, 2013.

[36] Tsai J C, Kabir M H, Mimura M. Travelling waves in a reaction-diffusion system modelling

farmer and hunter-gatherer interaction in the Neolithic transition in Europe, European Jour-
nal of Applied Mathematics, 2020, 31(3): 470-510.

[37] Volterra V. Fluctuations in the abundance of a species considered mathematically, Nature,

1926, 118(2972): 558-560.

[38] Wang C H, Fu S C. Traveling wave solutions to diffusive Holling-Tanner predator-prey models,

Discrete & Continuous Dynamical Systems-B, 2021, 26(4): 2239.

[39] Wu J, Zou X. Traveling wave fronts of reaction-diffusion systems with delay, Journal of Dy-
namics and Differential Equations, 2001, 13(3): 651-687.

[40] Zhang T, Wang W, Wang K. Minimal wave speed for a class of non-cooperative diffusion-
reaction system, Journal of Differential Equations, 2016, 260(3): 2763-2791.

[41] Zhao H, Wu D. Point to point traveling wave and periodic traveling wave induced by Hopf
bifurcation for a diffusive predator-prey system, Discrete & Continuous Dynamical Systems-S,
2020, 13(11): 3271.

32