Home Mathematics Non-existence, radial symmetry, monotonicity, and Liouville theorem of master equations with fractional p-Laplacian
Article Open Access

Non-existence, radial symmetry, monotonicity, and Liouville theorem of master equations with fractional p-Laplacian

  • Mengru Liu and Lihong Zhang EMAIL logo
Published/Copyright: November 15, 2025

Abstract

In this article, first, we introduce a new operator

( t Δ p ) s u ( z , t ) = C n , s p t R n u ( z , t ) u ( ζ , ϱ ) p 2 ( u ( z , t ) u ( ζ , ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ ,

where 0 < s < 1 , 2 p < . Second, for the purpose of overcoming the challenge brought by the non-locality of the operator ( t Δ p ) s in space-time, the narrow region principle and the maximum principle are studied, and the direct method of moving planes suitable for the operator ( t Δ p ) s is proposed. Based on this, we establish the radial symmetry, monotonicity and non-existence of solutions for the master equations with fractional p -Laplacian and the Liouville theorem for the homogeneous master equation with fractional p -Laplacian. Finally, we are confident that the ideas in the proof, as well as those involving perturbation techniques, limit arguments, and Fourier transforms will be useful tools in exploring qualitative properties and the Liouville theorem for solutions to various nonlocal parabolic problems.

MSC 2020: 35R11; 47G30; 35K05; 35B50; 35B53

1 Introduction

The primary aim of this article is to study

(1) The radial symmetry and strict monotonicity of the solution of the master equation with fractional p -Laplacian in the unit ball

(1.1) ( t Δ p ) s u ( z , t ) = f ( u ( z , t ) ) in B 1 ( 0 ) × R , u ( z , t ) = 0 in B 1 c ( 0 ) × R .

(2) The Liouville theorem of the following homogeneous master equation with fractional p -Laplacian

(1.2) ( t Δ p ) s u ( z , t ) = 0 in R n × R .

(3) The non-existence of the solution for the master equation with fractional p -Laplacian

(1.3) ( t Δ p ) s u ( z , t ) = f ( u ( z , t ) ) , in R n × R .

Here, the operator ( t Δ p ) s is defined later.

The new operator proposed in this article originates from the famous fully fractional heat operator ( t Δ ) s introduced by Riesz [25], which is defined as follows:

( t Δ ) s u ( z , t ) = C n , s t R n u ( z , t ) u ( ζ , ϱ ) ( t ϱ ) n 2 + 1 + s e z ζ 2 4 ( t ϱ ) d ζ d ϱ .

Because the value of ( t Δ ) s at a point ( z , t ) is related to the value of u in the entire R n and all past times up to t , this operator is non-local in both space and time. Interestingly, when u depends only on z , then it can be expressed as

( t Δ ) s u ( z ) = ( Δ ) s u ( z ) ,

where ( Δ ) s represents the famous fractional Laplacian. When u depends only on t , then

( t Δ ) s u ( t ) = t s u ( t ) ,

where t s represents the Marchaud fractional derivative [20], commonly denoted as D left s , which is the Marchaud left fractional derivative, defined by

D right s u ( t ) = 1 Γ ( s ) t u ( t ) u ( ϱ ) ( t ϱ ) 1 + s d ϱ .

The Marchaud fractional derivative has application in describing numerous physical phenomena, including plasma turbulence [4], wave propagation [8] and magneto-thermoelastic heat conduction [9]. Furthermore, when s 1 , the operator ( t Δ ) s becomes the local operator t Δ .

The master equation has been applied in physical phenomena to anomalous diffusion [3] and chaotic dynamics [14]. In the biological field, it is used to study biological invasions [1]. In the financial field, it can simulate the relationship between waiting time and the increase in trading prices [24]. From a probabilistic perspective, the master equation plays a crucial role in the theory of continuous-time random walks, where u represents the distribution of particles that perform random jumps while experiencing stochastic time delays [21]. Unlike the nonlocal parabolic equation

(1.4) t u + ( Δ ) s u = f ,

and the dual fractional parabolic equation

(1.5) t α u + ( Δ ) s u = f ,

the master equation accounts for the strong correlation between waiting times and particle jumps, which is absent in the nonlocal parabolic equation (1.4) and the dual fractional parabolic equation (1.5). Obviously, the master equation is of significant value in different fields. Therefore, further study of the master equation will enhance our understanding of complex phenomena. At present, the radial symmetry, existence, regularity, uniqueness, Liouville theorem and Gibbons’ conjecture of solutions for the master equation have obtained a series of outstanding results [2,5,7,17,19,26].

The fractional p -Laplacian appears in various fields such as engineering, chemistry, physics, image processing, and biological modeling. In recent years, many scholars have shown great interest in problems with the fractional p -Laplacian. For instance, in 2020, Wang and Yang investigated the non-existence of positive solutions of the fractional p -Laplacian equation with indefinite nonlinearities [31]. In 2021, Mi et al. studied a class of fractional p -Laplacian equations involving asymptotically periodic conditions and established the existence results of two new ground state solutions by variational methods [22]. In 2023, Hamdani et al. studied a novel tri-nonlocal Kirchhoff problem, which involves the p ( x ) -fractional Laplacian equations of variable order [11]. In 2024, Liao established the local Hölder regularity of weak solutions of a fractional p -Laplacian parabolic equation involving only measurable kernels [16]. The other results regarding the fractional p -Laplacian equation can be referred in [6,23,28,29,34,36]. In the particular case when p = 2 , ( Δ ) p s transforms into the well-known fractional Laplacian. For the fractional Laplacian, Zhao et al. combined the minimax method, bifurcation theory, and Morse theory to study the multiplicity of nontrivial solutions for a class of nonlinear fractional Laplace problems with a parameter and superlinear nonlinearity [35]. Kossowski and Przeradzki established the existence theory of solutions for the heat equations with generalized fractional Laplacian [12]. Another special case is that when s = 1 , the fractional p -Laplacian degenerates into the classical p -Laplacian. For the p -Laplacian, Wang and Sun studied the normalized solutions of the p -Laplacian equation with a trapping potential and L r -supercritical growth, where r = p or r = 2 [30]. Wu et al. studied the local solvability of a singular parabolic p -Laplacian equation with logarithmic nonlinearity by the technique of cut-off combining with the method of Faedo-Galerkin approximation. Based on the potential well method and Hardy-Sobolev inequality, they derived the global existence of solutions [33]. Many fruitful results can be found in [10,18,37].

Inspired by the above articles, we introduce a new operator that combines the fractional p -Laplacian and the fully fractional heat operator, which is defined as follows:

(1.6) ( t Δ p ) s u ( z , t ) = C n , s p t R n u ( z , t ) u ( ζ , ϱ ) p 2 ( u ( z , t ) u ( ζ , ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ ,

where 0 < s < 1 , 2 p < , C n , s p = s p 2 ( 1 s ) 2 2 s 1 n s p π n 1 2 Γ ( p + 1 2 ) Γ ( 2 s ) is the normalized positive constant and the integral with respect to ζ is considered in the sense of Cauchy principal values. To make sure that the integral in (1.6) is well defined, let

u ( z , t ) C z , t , loc 2 s + ε , s + ε ( R n × R ) L ( R n × R ) ,

where L ( R n × R ) is a class of slowly increasing functions defined as

L ( R n × R ) u ( z , t ) L loc p 1 ( R n × R ) t R n u ( z , ϱ ) p 1 e z 2 4 ( t ϱ ) 1 + ( t ϱ ) n 2 + 1 + s p 2 d z d ϱ < for any t R ,

and the local parabolic Hölder space C z , t , loc 2 s + ε , s + ε ( R n × R ) is defined in [19]. In the following, we recall the definition of the parabolic Hölder space C z , t 2 α , α ( R n × R ) (cf. [13]).

(i) When 0 < α 1 2 , if u ( z , t ) C z , t 2 α , α ( R n × R ) , then there exists a constant C > 0 such that

u ( z , t ) u ( ζ , ϱ ) C ( z ζ + t ϱ 1 2 ) 2 α

for any z , ζ R n and t , ϱ R .

(ii) When 1 2 < α 1 , we say that

u ( z , t ) C z , t 2 α , α ( R n × R ) C z , t 1 + ( 2 α 1 ) , α ( R n × R ) ,

if u is α -Hölder continuous in t uniformly with respect to z and its gradient z u is ( 2 α 1 ) -Hölder continuous in z uniformly with respect to t and ( α 1 2 ) -Hölder continuous in t uniformly with respect to z .

(iii) When α > 1 , if u ( z , t ) C z , t 2 α , α ( R n × R ) , then it means that

t u , D z 2 u C z , t 2 α 2 , α 1 ( R n × R ) .

Then, we can analogously define the local parabolic Hölder space C z , t , loc 2 α , α ( R n × R ) . Interestingly, when u depends only on z , then

( t Δ p ) s u ( z ) = ( Δ ) p s u ( z ) ,

where ( Δ ) p s represents the famous fractional p -Laplacian. When u depends only on t , then

( t Δ p ) s u ( t ) = t s u ( t ) ,

where t s represents the Marchaud fractional derivative [20]. Moreover, when p = 2 , the operator ( t Δ p ) s transforms into the famous fully fractional heat operator ( t Δ ) s .

Due to the non-local and strong correlation exhibited by ( t Δ p ) s in the context of the master equation with fractional p -Laplacian, many traditional methods for solving local operators are no longer applicable in non-local environments. However, for the exploration of explosion rate, prior estimation and optimal global estimation for non-local parabolic problems with initial boundary values, the study of the qualitative properties of the solution plays a vital role. Therefore, this article mainly overcomes the challenge of non-local and strong correlation of operator ( t Δ p ) s and establishes the direct method of moving planes suitable for this operator ( t Δ p ) s . Consequently, we obtain the Liouville theorem, radial symmetry, monotonicity, and non-existence of the solution of the master equation with fractional p -Laplacian.

The primary conclusions of this article are as follows.

Radial symmetry and monotonicity: Assume that

u C z , t 2 s + ε , s + ε ( B 1 ( 0 ) × R )

is a positive bounded solution of equation (1.1). If f C 1 ( [ 0 , + ) ) with f ( 0 ) 0 , f ( 0 ) 0 , the solution u of equation (1.1) is radially symmetric and strictly decreasing with respect to the origin in z B 1 ( 0 ) for any t R .

Liouville theorem: Assume that

u C z , t , loc 2 s + ε , s + ε ( R n × R )

is a bounded solution of equation (1.2). Then, u ( z , t ) must be a constant.

Non-existence: Assume that

u C z , t , loc 2 s + ε , s + ε ( R n × R ) L ( R n × R )

is a positive bounded solution of equation (1.3). Suppose f satisfies

  1. f ( u ) = z 1 ψ ( u ) ,

  2. ψ is monotonically increasing with respect to u ,

  3. ψ > C , where C is a positive constant.

If conditions ( a ) , ( b ) or ( a ) , ( c ) hold, then equation (1.3) has no positive bounded solution.

The structure of this article is as follows. In Section 2, we establish the narrow region principle. On this basis, using the direct method of moving planes, we obtain equation (1.1), which is radially symmetric and strictly decreasing from the origin within the unit ball for any t R . In Section 3, we introduce the maximum principle in unbounded domains. By applying the Fourier transform, the Liouville theorem of the homogeneous equation (1.2) is obtained, that is, all bounded solutions must be constants. In Section 4, we prove that equation (1.3) has no bounded positive solution by utilizing the maximum principle in bounded domains.

Notations. Let z 1 be any given direction in R n . For any λ R , let us take

T λ { z = ( z 1 , z 2 , , z n ) R n z 1 = λ }

as the moving planes, and it’s perpendicular to z 1 -axis. Let

Σ λ { z R n z 1 < λ }

and

Ω λ { z B 1 ( 0 ) z 1 < λ }

represent the left region of T λ . Let

z λ ( 2 λ z 1 , z 2 , , z n )

be reflection point of z with respect to T λ . Denote

u λ ( z , t ) = u ( z λ , t )

and

w λ ( z , t ) = u ( z λ , t ) u ( z , t ) = u λ ( z , t ) u ( z , t ) .

Then, w λ is an antisymmetric function satisfying

(1.7) w λ ( z , t ) = w λ ( z λ , t ) , in Σ λ × R .

2 Radial symmetry and monotonicity

In this section, we first establish the narrow region principle (i.e., Theorem 2.1). This principle is essential for proving the radial symmetry and monotonicity of the solution of the master equation with fractional p -Laplacian. Second, Theorem 2.1 and the direct method of moving planes are applied to prove the radial symmetry and strict monotonicity of the solution of the master equation with fractional p -Laplacian (i.e., Theorem 2.2). In the entire article, C represents a positive constant, which value may vary from line to line.

2.1 Narrow region principle for antisymmetric functions

Theorem 2.1

Assume that Ω is a bounded narrow region contained in { z Σ λ λ 2 l < z 1 < λ } , where l is small. Let

w λ C z , t , loc 2 s + ε , s + ε ( Ω × R ) L ( R n × R )

be lower semi-continuous up to the boundary Ω and bounded from below, satisfying (1.7) and

(2.1) ( t Δ p ) s u λ ( t Δ p ) s u = c ( z , t ) w λ i n Ω × R , w λ 0 i n ( Σ λ \ Ω ) × R ,

where c ( z , t ) is bounded from above.

Then

(2.2) w λ 0 , i n Σ λ × R

for sufficiently small l.

In addition, if there exists a point ( z , t ) Ω × R such that w λ ( z , t ) = 0 , then

(2.3) w λ 0 , i n R n × ( , t ] .

Proof

First, we derive (2.2) by contradiction. Since Ω is a bounded narrow region and w λ ( z , t ) is lower semi-continuous, there exists z ( t ) Ω such that for every fixed t in R ,

w λ ( z ( t ) , t ) = min z Ω w λ ( z , t ) .

If (2.2) is not true, we can assume that there exists h > 0 such that

(2.4) inf Ω × R w λ ( z , t ) = inf R w λ ( z ( t ) , t ) = h < 0 .

Because t belongs to an unbounded interval R , the infimum of w λ ( z ( t ) , t ) may not be obtained, but there definitely exists a sequence { ( z ( t k ) , t k ) } Ω × R such that

w λ ( z ( t k ) , t k ) = h k h as k .

Let δ k = h h k . Obviously, the sequence { δ k } is nonnegative and δ k 0 when k .

To solve the problem that the infimum of w λ ( z ( t ) , t ) may not exist for t in R , we adopt perturbation techniques for w λ about t . To do this, suppose the auxiliary function

v k ( z , t ) = w λ ( z , t ) δ k η k ( t ) ,

where the smooth cut-off function

η k ( t ) = η ( t t k ) C 0 ( 1 + t k , 1 + t k )

and satisfies

(2.5) η k ( t ) = 1 , t 1 2 + t k , 1 2 + t k , [ 0, 1 ] , t 1 + t k , 1 2 + t k 1 2 + t k , 1 + t k , = 0 , t ( 1 + t k , 1 + t k ) .

It follows that

v k ( z ( t k ) , t k ) = w λ ( z ( t k ) , t k ) δ k = h k h + h k = h .

By (2.4), we conclude that when ( z , t ) Ω × ( R \ ( 1 + t k , 1 + t k ) ) , then

v k ( z , t ) = w λ ( z , t ) h .

Thus, there exists ( z ¯ k , t ¯ k ) Ω × ( 1 + t k , 1 + t k ) such that

(2.6) h δ k v k ( z ¯ k , t ¯ k ) = inf Ω × R v k ( z , t ) = inf Σ λ × R v k ( z , t ) h .

From the definition of v k , we can infer that

(2.7) h w λ ( z ¯ k , t ¯ k ) h + δ k = h k < 0 .

Next, our goal is to estimate the lower and upper bounds of

(2.8) ( t Δ p ) s ( u λ ( z , t ) δ k η k ( t ) ) ( t Δ p ) s u ( z , t )

at the minimum point ( z ¯ k , t ¯ k ) of v k in Σ λ × R to derive a contradiction. For convenience, we denote G ( t ) = t p 2 t , then G ( t ) = ( p 1 ) t p 2 0 . On the one hand, we have

( t Δ p ) s ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) ) ( t Δ p ) s u ( z ¯ k , t ¯ k ) = C n , s p t ¯ k R n G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 e z ¯ k ζ 2 4 ( t ¯ k ϱ ) d ζ d ϱ C n , s p t ¯ k R n G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 e z ¯ k ζ 2 4 ( t ¯ k ϱ ) d ζ d ϱ = C n , s p t ¯ k Σ λ e z ¯ k ζ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) ] d ζ d ϱ + C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u λ ( ζ , ϱ ) ) ] d ζ d ϱ = C n , s p t ¯ k Σ λ ( e z ¯ k ζ 2 4 ( t ¯ k ϱ ) e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) ] d ζ d ϱ + C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 { [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u λ ( ζ , ϱ ) ) ] + [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) ] } d ζ d ϱ = C n , s p t ¯ k Σ λ ( e z ¯ k ζ 2 4 ( t ¯ k ϱ ) e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) ] d ζ d ϱ + C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 { [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) ] + [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u λ ( ζ , ϱ ) ) ] } d ζ d ϱ I 1 + I 2 .

We first estimate I 1 . Since ( z ¯ k , t ¯ k ) is the minimum point of v k in Σ λ × R , for any ζ Σ λ , we have

( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ) ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) = w λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) ( w λ ( ζ , ϱ ) δ k η k ( ϱ ) ) = v k ( z ¯ k , t ¯ k ) v k ( ζ , ϱ ) 0 .

By the strict monotonicity of G , it follows

(2.9) G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) 0 .

Combine (2.9) with the fact that

e z ¯ k ζ 2 4 ( t ¯ k ϱ ) e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) > 0 ,

then

(2.10) I 1 = C n , s p t ¯ k Σ λ ( e z ¯ k ζ 2 4 ( t ¯ k ϱ ) e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) ] d ζ d ϱ 0 .

Next, we estimate I 2 .

(2.11) I 2 = C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 { [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) ] + [ G ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u λ ( ζ , ϱ ) ) ] } d ζ d ϱ = C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( ξ 1 ( z , ζ ) ) ( w λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) + δ k η k ( ϱ ) ) + G ( ξ 2 ( z , ζ ) ) ( w λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) + δ k η k ( ϱ ) ) ] d ζ d ϱ C t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 ( v k ( z ¯ k , t ¯ k ) + δ k η k ( ϱ ) ) d ζ d ϱ = C v k ( z ¯ k , t ¯ k ) t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 d ζ d ϱ + C δ k t ¯ k Σ λ η k ( ϱ ) e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 d ζ d ϱ ,

where

u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u ( ζ , ϱ ) + δ k η k ( ϱ ) ξ 1 ( z , ζ ) u ( z ¯ k , t ¯ k ) u ( ζ , ϱ )

and

u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) u λ ( ζ , ϱ ) + δ k η k ( ϱ ) ξ 2 ( z , ζ ) u ( z ¯ k , t ¯ k ) u λ ( ζ , ϱ ) .

After that, we make a variable substitution to estimate the integral in (2.11). Let

ϑ = z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) .

Then,

t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 d ζ d ϱ = Σ λ 0 + e ϑ ( z ¯ k ζ λ 2 4 ϑ ) n 2 + 1 + s p 2 z ¯ k ζ λ 2 4 ϑ 2 d ϑ d ζ = C Σ λ 1 z ¯ k ζ λ n + s p 0 + e ϑ ϑ n 2 + s p 2 1 d ϑ d ζ = C Σ λ 1 z ¯ k ζ λ n + s p d ζ .

Substituting the above equation into (2.11) and combining with η k [ 0 , 1 ] and (2.6), we obtain

(2.12) I 2 C v k ( z ¯ k , t ¯ k ) Σ λ 1 z ¯ k ζ λ n + s p d ζ + C δ k Σ λ 1 z ¯ k ζ λ n + s p d ζ C v k ( z ¯ k , t ¯ k ) l s p + C δ k l s p C h l s p + C δ k l s p .

Combining (2.10) and (2.12), we have

(2.13) ( t Δ p ) s ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) ) ( t Δ p ) s u ( z ¯ k , t ¯ k ) C h l s p + C δ k l s p .

In order to continue with the proof, the following lemma is required.

Lemma 2.1

Assume that u ( z , t ) C z , t , loc 2 s + ε , s + ε ( R n × R ) and smooth cut-off function η ( t ) C 0 ( 1 , 1 ) , where η ( t ) is defined in (2.5). Then, there exists a positive constant C 0 such that

( t Δ p ) s ( u ( z , t ) + δ k η ( t ) ) ( t Δ p ) s u ( z , t ) δ k C 0 for ( z , t ) R n × ( 1 , 1 ) .

Proof

By using the definitions of ( t Δ p ) s and η ( t ) , combined with the fact that

e x 2 4 ϱ ϱ n 2 + 1 + s p 2 C x n + 2 + s p + ϱ n 2 + 1 + s p 2 .

It follows that

( t Δ p ) s ( u ( z , t ) + δ k η ( t ) ) ( t Δ p ) s u ( z , t ) = C n , s p t R n [ G ( ( u ( z , t ) + δ k η ( t ) ) ( u ( ζ , ϱ ) + δ k η ( ϱ ) ) ) G ( u ( z , t ) u ( ζ , ϱ ) ) ] ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ = C n , s p t R n δ k G ( ξ ( z , t ) ) ( η ( t ) η ( ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ C δ k t 1 R n η ( t ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ + C δ k t 1 t η ( t ) η ( ϱ ) ( t ϱ ) 1 + s p 2 R n e z ζ 2 4 ( t ϱ ) ( t ϱ ) n 2 d ζ d ϱ C δ k t 1 R n 1 z ζ n + s p 2 + ( t ϱ ) n 2 + 1 + s p 2 d ζ d ϱ + C δ k t 1 t η ( t ) η ( ϱ ) ( t ϱ ) 1 + s p 2 d ϱ C δ k t 1 R n 1 z ζ n + s p 2 + ( t ϱ ) n 2 + 1 + s p 2 d ζ d ϱ + C δ k t 1 t t ϱ ( t ϱ ) 1 + s p 2 d ϱ δ k C 0 ,

where

( u ( z , t ) + δ k η ( t ) ) ( u ( ζ , ϱ ) + δ k η ( ϱ ) ) ξ ( z , t ) u ( z , t ) u ( ζ , ϱ ) .

This completes the proof of Lemma 2.1.□

From Lemma 2.1, it is easy to derive the following corollary.

Corollary 2.1

For some t 0 R and r > 0 , let u ( z , t ) C z , t , loc 2 s + ε , s + ε ( R n × R ) and

η r ( t ) C 0 ( r 2 + t 0 , r 2 + t 0 ) ,

then

( t Δ p ) s ( u ( z , t ) + δ k η ( t ) ) ( t Δ p ) s u ( z , t ) δ k C r 2 s for ( z , t ) R n × ( 1 , 1 ) ,

where constant C > 0 .

Now, we continue our proof.

On the other hand, it follows from (2.1), (2.7), and Lemma 2.1 that

(2.14) ( t Δ p ) s ( u λ ( z ¯ k , t ¯ k ) δ k η k ( t ¯ k ) ) ( t Δ p ) s u ( z ¯ k , t ¯ k ) ( t Δ p ) s u λ ( z ¯ k , t ¯ k ) ( t Δ p ) s u ( z ¯ k , t ¯ k ) δ k C C 0 h δ k C .

Here, C 0 is the upper bound of c ( z , t ) . Combining (2.13) and (2.14), we obtain

C 0 h C h l s p + C δ k l s p + C δ k C h l s p ,

as k .

Thus, for small enough l , we derive a contradiction and conclude that (2.2) is true.

Finally, we prove that (2.3) is valid. From (2.2), we know

w λ ( z , t ) = min Σ λ × R w λ ( z , t ) = 0 .

Obviously, according to the equation in (2.1), we have

(2.15) ( t Δ p ) s u λ ( z , t ) ( t Δ p ) s u ( z , t ) = 0 .

By (2.2) and

e z ζ λ 2 4 ( t ϱ ) < e z ζ 2 4 ( t ϱ ) ,

we have

( t Δ p ) s u λ ( z , t ) ( t Δ p ) s u ( z , t ) = C n , s p t R n G ( u λ ( z , t ) u λ ( ζ , ϱ ) ) G ( u ( z , t ) u ( ζ , ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ = C n , s p t R n G ( ξ 3 ( z , t ) ) ( w λ ( z , t ) w λ ( ζ , ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ = C n , s p t R n G ( ξ 3 ( z , t ) ) w λ ( ζ , ϱ ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ = C n , s p t Σ λ G ( ξ 3 ( z , t ) ) w λ ( ζ , ϱ ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ λ 2 4 ( t ϱ ) e z ζ 2 4 ( t ϱ ) d ζ d ϱ < 0 ,

where

u λ ( z , t ) u λ ( ζ , ϱ ) ξ 3 ( z , t ) u ( z , t ) u ( ζ , ϱ ) .

Then, by (2.15), we have

w λ ( z , t ) 0 for ( z , t ) Σ λ × ( , t ] .

Since w λ is an antisymmetric function, we have □

w λ 0 in R n × ( , t ] .

2.2 Radial symmetry of solutions in a unit ball

Theorem 2.2

Assume that

u C z , t 2 s + ε , s + ε ( B 1 ( 0 ) × R )

is a positive bounded solution of

(2.16) ( t Δ p ) s u = f ( u ) i n B 1 ( 0 ) × R , u = 0 i n B 1 c ( 0 ) × R .

If f C 1 ( [ 0 , + ) ) with f ( 0 ) 0 , f ( 0 ) 0 , the solution u of equation (2.16) is radially symmetric and strictly decreasing with respect to the origin in z B 1 ( 0 ) for any t R .

Proof

We choose z 1 as any direction and T λ , Σ λ , Ω λ , z λ , u λ , w λ are defined as shown before. By the simple calculation, it follows w λ ( z , t ) = w λ ( z λ , t ) in Σ λ × R and

(2.17) ( t Δ p ) s u λ ( t Δ p ) s u = C λ ( z , t ) w λ in Ω λ × R , w λ 0 in ( Σ λ \ Ω λ ) × R ,

where

C λ ( z , t ) = 0 1 f ( ς u λ ( z , t ) + ( 1 ς ) u ( z , t ) ) d ς .

Combining the boundedness of u and f C 1 ( [ 0 , + ) ) , we obtain that C λ ( z , t ) is bounded. Next, we will break the proof into three steps.

Step 1 We first move the plane T λ from z 1 = 1 along the z 1 -axis to the right. Therefore, when λ is close enough to 1 , we can observe that Ω λ becomes a narrow region. By applying Theorem 2.1, it holds

(2.18) w λ 0 in Ω λ × R .

Here, the inequality (2.18) offers a starting point for moving the plane T λ .

Step 2 The plane T λ continues to move to the right, ensuring that (2.18) remains valid up to its limit position. Denote

λ 0 = sup { λ < 0 w ν ( z , t ) 0 , ( z , t ) Σ ν × R , ν λ } .

Now, our goal is to prove that

(2.19) λ 0 = 0 .

If λ 0 < 0 , according to the definition of λ 0 and w λ k ( z , t ) 0 in ( Σ λ k \ Ω λ k ) × R , there exists a negative sequence { λ k } with { λ k } λ 0 such that

inf Σ λ k × R w λ k ( z , t ) = inf Ω λ k × R w λ k ( z , t ) = h k < 0

and

h k 0 when k .

Because t R , the infimum of w λ k about t may not exist. However, there must exist a sequence { ( z k , t k ) } Ω λ k × R and a nonnegative sequence { δ k } 0 as k , which satisfy

w λ k ( z k , t k ) = inf z Σ λ k w λ k ( z , t k ) h k + δ k h k < 0 .

To solve the problem that w λ k does not reach the infimum. Suppose the auxiliary function

V k ( z , t ) = w λ k ( z , t ) δ k h k η k ( t ) ,

where smooth cut-off function η k ( t ) = η ( t t k ) that satisfies (2.5). By the simple calculation, it holds

V k ( z k , t k ) h k .

For ( z , t ) Ω λ k × ( R n \ ( 1 + t k , 1 + t k ) ) , we obtain

V k ( z , t ) = w λ k ( z , t ) h k .

Thus, there exists ( z ¯ k , t ¯ k ) Ω λ k × ( 1 + t k , 1 + t k ) such that

(2.20) h k δ k h k V k ( z ¯ k , t ¯ k ) = inf Σ λ k × R V k ( z , t ) h k .

By the definition of V k , we have

(2.21) h k w λ k ( z ¯ k , t ¯ k ) h k + δ k h k < 0 .

Next, we focus on the estimate of

( t Δ p ) s ( u λ k ( z , t ) δ k h k η k ( t ) ) ( t Δ p ) s u ( z , t ) ,

at the minimum point ( z ¯ k , t ¯ k ) of V k in Σ λ k × R . For convenience, we denote G ( t ) = t p 2 t , then G ( t ) = ( p 1 ) t p 2 0 . First of all, similar to the estimate of (2.12), using the antisymmetry of w λ about z , (2.20), η k [ 0, 1 ] and z ¯ k ζ < z ¯ k ζ λ k , we have

(2.22) ( t Δ p ) s ( u λ k ( z ¯ k , t ¯ k ) δ k h k η k ( t ¯ k ) ) ( t Δ p ) s u ( z ¯ k , t ¯ k ) = C n , s p t ¯ k Σ λ k ( e z ¯ k ζ 2 4 ( t ¯ k ϱ ) e z ¯ k ζ λ k 2 4 ( t ¯ k ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ k ( z ¯ k , t ¯ k ) δ k h k η k ( t ¯ k ) u λ k ( ζ , ϱ ) + δ k h k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) ] d ζ d ϱ + C n , s p t ¯ k Σ λ k e z ¯ k ζ λ k 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 { [ G ( u λ k ( z ¯ k , t ¯ k ) δ k h k η k ( t ¯ k ) u ( ζ , ϱ ) + δ k h k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u ( ζ , ϱ ) ) ] + [ G ( u λ k ( z ¯ k , t ¯ k ) δ k h k η k ( t ¯ k ) u λ k ( ζ , ϱ ) + δ k h k η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) u λ k ( ζ , ϱ ) ) ] } d ζ d ϱ C n , s p t ¯ k Σ λ k e z ¯ k ζ λ k 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( ξ 1 ( z , ζ ) ) ( w λ k ( z ¯ k , t ¯ k ) δ k h k η k ( t ¯ k ) + δ k h k η k ( ϱ ) ) + G ( ξ 2 ( z , ζ ) ) ( w λ k ( z ¯ k , t ¯ k ) δ k h k η k ( t ¯ k ) + δ k h k η k ( ϱ ) ) ] d ζ d ϱ C t ¯ k Σ λ k e z ¯ k ζ λ k 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 ( V k ( z ¯ k , t ¯ k ) + δ k h k η k ( ϱ ) ) d ζ d ϱ = C V k ( z ¯ k , t ¯ k ) t ¯ k Σ λ k e z ¯ k ζ λ k 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 d ζ d ϱ + C δ k h k t ¯ k Σ λ k η k ( ϱ ) e z ¯ k ζ λ k 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 d ζ d ϱ C V k ( z ¯ k , t ¯ k ) Σ λ k 1 z ¯ k ζ λ k n + s p d ζ + C δ k h k Σ λ k 1 z ¯ k ζ λ k n + s p d ζ C V k ( z ¯ k , t ¯ k ) dist ( z ¯ k , T λ k ) s p + C δ k h k dist ( z ¯ k , T λ k ) s p C h k dist ( z ¯ k , T λ k ) s p + C δ k h k dist ( z ¯ k , T λ k ) s p C h k ( 1 δ k ) 2 s p ,

where

u λ k ( z ¯ k , t ¯ k ) δ k h k η k ( t ¯ k ) u ( ζ , ϱ ) + δ k h k η k ( ϱ ) ξ 1 ( z , ζ ) u ( z ¯ k , t ¯ k ) u ( ζ , ϱ )

and

u λ k ( z ¯ k , t ¯ k ) δ k h k η k ( t ¯ k ) u λ k ( ζ , ϱ ) + δ k h k η k ( ϱ ) ξ 2 ( z , ζ ) u ( z ¯ k , t ¯ k ) u λ k ( ζ , ϱ ) .

Second, due to the boundedness of C λ k ( z ¯ k , t ¯ k ) , combining (2.17), (2.21), and Lemma 2.1, we have

(2.23) ( t Δ p ) s ( u λ k ( z ¯ k , t ¯ k ) δ k h k η k ( t ¯ k ) ) ( t Δ p ) s u ( z ¯ k , t ¯ k ) ( t Δ p ) s u λ k ( z ¯ k , t ¯ k ) ( t Δ p ) s u ( z ¯ k , t ¯ k ) δ k h k C = C λ k ( z ¯ k , t ¯ k ) w λ k ( z ¯ k , t ¯ k ) C δ k h k C λ k ( z ¯ k , t ¯ k ) h k C δ k h k .

Here C > 0 . It can be inferred from (2.22) and (2.23) that

C λ k ( z ¯ k , t ¯ k ) C δ k + C ( C δ k ) 2 s p C ( 1 δ k ) .

Applying δ k 0 when k , it follows that for sufficiently large k

C λ k ( z ¯ k , t ¯ k ) C 0 > 0 .

In addition, because

C λ k ( z ¯ k , t ¯ k ) = f ( ξ )

and f ( 0 ) 0 , where ξ ( u λ k ( z ¯ k , t ¯ k ) , u ( z ¯ k , t ¯ k ) ) , then there exists a subsequence of { ( z ¯ k , t ¯ k ) } , which we will continue to represent by { ( z ¯ k , t ¯ k ) } , such that

(2.24) u ( z ¯ k , t ¯ k ) C 1 > 0 .

Denote

u k ( z , t ) = u ( z , t + t ¯ k ) , w ¯ k ( z , t ) = w λ k ( z , t + t ¯ k ) and C ¯ k ( z , t ) = C λ k ( z , t + t ¯ k ) .

According to the Arzelà-Ascoli theorem, there exist u ¯ ( z , t ) , w ¯ ( z , t ) , and C ¯ ( z , t ) satisfying

lim k u k ( z , t ) = u ¯ ( z , t ) , lim k w ¯ k ( z , t ) = w ¯ ( z , t ) and lim k C ¯ k ( z , t ) = C ¯ ( z , t ) .

Then, by equation

( t Δ p ) s u k ( z λ k , t ) ( t Δ p ) s u k ( z , t ) = C ¯ k ( z , t ) w ¯ k ( z , t ) in Ω λ k × R ,

and λ k λ 0 as k , we obtain

(2.25) ( t Δ p ) s u ¯ ( z λ 0 , t ) ( t Δ p ) s u ¯ ( z , t ) = C ¯ ( z , t ) w ¯ ( z , t ) in Ω λ 0 × R .

Since Ω λ k is a bounded domain, we can make the assumption that z ¯ k z 0 , combined with (2.21), we obtain that

w ¯ ( z 0 , 0 ) w ¯ k ( z ¯ k , 0 ) = w λ k ( z ¯ k , t ¯ k ) 0 ,

when k , that is,

w ¯ ( z 0 , 0 ) = 0 .

Obviously, according to (2.25), we have

(2.26) ( t Δ p ) s u ¯ ( ( z 0 ) λ 0 , 0 ) ( t Δ p ) s u ¯ ( z 0 , 0 ) = C ¯ ( z 0 , 0 ) w ¯ ( z 0 , 0 ) = 0 .

By w ¯ ( z , ζ ) 0 in Σ λ 0 × R , G 0 and z 0 ζ < z 0 ζ λ 0 , we obtain

( t Δ p ) s u ¯ ( ( z 0 ) λ 0 , 0 ) ( t Δ p ) s u ¯ ( z 0 , 0 ) = C n , s p 0 R n G ( u ¯ ( ( z 0 ) λ 0 , 0 ) u ¯ ( ζ λ 0 , ϱ ) ) G ( u ¯ ( z 0 , 0 ) u ¯ ( ζ , ϱ ) ) ( ϱ ) n 2 + 1 + s p 2 e z 0 ζ 2 4 ϱ d ζ d ϱ = C n , s p 0 R n G ( ξ 3 ( z , t ) ) ( w ¯ ( z 0 , 0 ) w ¯ ( ζ , ϱ ) ) ( ϱ ) n 2 + 1 + s p 2 e z 0 ζ 2 4 ϱ d ζ d ϱ = C n , s p 0 R n G ( ξ 3 ( z , t ) ) w ¯ ( ζ , ϱ ) ( ϱ ) n 2 + 1 + s p 2 e z 0 ζ 2 4 ϱ d ζ d ϱ = C n , s p 0 Σ λ 0 G ( ξ 3 ( z , t ) ) w ¯ ( ζ , ϱ ) e z 0 ζ λ 0 2 4 ϱ ( ϱ ) n 2 + 1 + s p 2 e z 0 ζ 2 4 ϱ ( ϱ ) n 2 + 1 + s p 2 d ζ d ϱ > 0 ,

where

u ¯ ( ( z 0 ) λ 0 , 0 ) u ¯ ( ζ λ 0 , ϱ ) ξ 3 ( z , t ) u ¯ ( z 0 , 0 ) u ¯ ( ζ , ϱ ) .

Therefore, combining (2.26) and the antisymmetry of w ¯ ( z , ζ ) about z , it follows

(2.27) w ¯ ( z , t ) 0 for ( z , t ) R n × ( , 0 ] .

Furthermore, according to (2.16), we can obtain that u ¯ ( z , t ) satisfies

(2.28) ( t Δ p ) s u ¯ ( z , t ) = f ( u ¯ ( z , t ) ) , ( z , t ) B 1 ( 0 ) × R .

Then, by (2.24), we have

(2.29) u ¯ ( z 0 , 0 ) = lim k u k ( z ¯ k , 0 ) = lim k u ( z ¯ k , t ¯ k ) > 0 .

Next, we demonstrate that

(2.30) u ¯ ( z , 0 ) > 0 for z B 1 ( 0 ) .

If (2.30) does not hold, because of the external condition and internal positivity of u , there exists a point z ¯ B 1 ( 0 ) such that

u ¯ ( z ¯ , 0 ) = 0 = min R n × R u ¯ ( z , 0 ) .

Combining (2.28) and f ( 0 ) 0 , we obtain

( t Δ p ) s u ¯ ( z ¯ , 0 ) = f ( u ¯ ( z ¯ , 0 ) ) = f ( 0 ) 0 .

But, by the direct calculation, we arrive at

( t Δ p ) s u ¯ ( z ¯ , 0 ) = C n , s p 0 R n u ¯ ( ζ , ϱ ) p 2 u ¯ ( ζ , ϱ ) ( ϱ ) n 2 + 1 + s p 2 e z 0 ζ 2 4 ϱ d ζ d ϱ 0 .

So, we obtain

( t Δ p ) s u ¯ ( z ¯ , 0 ) = 0 .

Considering u ¯ 0 , it holds

u ¯ ( z , t ) = 0 in R n × ( , 0 ] ,

which is contradictory to (2.29). Thus, (2.30) is valid. Through u ¯ ( z , 0 ) = 0 in B 1 c ( 0 ) , (2.30), and λ 0 < 0 , then when z B 1 c ( 0 ) , we have z λ 0 B 1 ( 0 ) and

w ¯ ( z , 0 ) = u ¯ ( z λ 0 , 0 ) u ¯ ( z , 0 ) = u ¯ ( z λ 0 , 0 ) > 0 .

This conflicts with (2.27). Therefore, we prove that the limit position is T 0 , that is, λ 0 = 0 .

Finally, since the direction of z 1 is arbitrary, according to the definition of λ 0 . Consequently, for any t R , we have u ( z , t ) is radially symmetric and monotonically decreasing with respect to the origin in z B 1 ( 0 ) .

Step 3 We show that this decrease is strict. It is enough to prove that for any 1 < λ < 0 ,

(2.31) w λ ( z , t ) > 0 in Ω λ × R .

Assume that (2.31) is invalid, then there exist some λ 0 ( 1 , 0 ) and ( z 0 , t 0 ) Ω λ 0 × R such that

w λ 0 ( z 0 , t 0 ) = min Σ λ 0 × R w λ 0 ( z , t ) = 0 .

From w λ 0 0 in Σ λ 0 × R , G 0 and

e z 0 ζ λ 2 4 ( t 0 ϱ ) < e z 0 ζ 2 4 ( t 0 ϱ ) ,

it follows

( t Δ p ) s u λ 0 ( z 0 , t 0 ) ( t Δ p ) s u ( z 0 , t 0 ) = C n , s p t 0 R n G ( u λ 0 ( z 0 , t 0 ) u λ 0 ( ζ , ϱ ) ) G ( u ( z 0 , t 0 ) u ( ζ , ϱ ) ) ( t 0 ϱ ) n 2 + 1 + s p 2 e z 0 ζ 2 4 ( t 0 ϱ ) d ζ d ϱ = C n , s p t 0 R n G ( ξ 4 ( z , t ) ) ( w λ 0 ( z 0 , t 0 ) w λ 0 ( ζ , ϱ ) ) ( t 0 ϱ ) n 2 + 1 + s p 2 e z 0 ζ 2 4 ( t 0 ϱ ) d ζ d ϱ = C n , s p t 0 R n G ( ξ 4 ( z , t ) ) w λ 0 ( ζ , ϱ ) ( t 0 ϱ ) n 2 + 1 + s p 2 e z 0 ζ 2 4 ( t 0 ϱ ) d ζ d ϱ = C n , s p t 0 Σ λ 0 G ( ξ 4 ( z , t ) ) w λ 0 ( ζ , ϱ ) ( t 0 ϱ ) n 2 + 1 + s p 2 e z 0 ζ λ 0 2 4 ( t 0 ϱ ) e z 0 ζ 2 4 ( t 0 ϱ ) d ζ d ϱ < 0 ,

where

u λ 0 ( z 0 , t 0 ) u λ 0 ( ζ , ϱ ) ξ 4 ( z , t ) u ( z 0 , t 0 ) u ( ζ , ϱ ) .

This is contradictory to

( t Δ p ) s u λ 0 ( z 0 , t 0 ) ( t Δ p ) s u ( z 0 , t 0 ) = C λ 0 ( z 0 , t 0 ) w λ 0 ( z 0 , t 0 ) = 0 .

Thus

(2.32) w λ 0 ( z , t ) 0 in Σ λ 0 × ( , t 0 ] .

However, because of the external conditions and internal positivity of u ( z , t ) , it follows that for any fixed t ( , t 0 ] ,

w λ 0 ( z , t ) 0 in Σ λ 0 ,

which contradicts (2.32). Then (2.31) is valid.□

3 Liouville theorem

In this section, we first prove the maximum principle in unbounded domains (i.e., Theorem 3.1). This principle is an essential element of the Liouville theorem for proving homogeneous master equations with fractional p -Laplacian. Second, applying Theorem 3.1 and Fourier transform, we prove the Liouville theorem of the master equation with fractional p -Laplacian (i.e., Theorem 3.2).

3.1 Maximum principle in unbounded domains

Theorem 3.1

Assume that

w λ C z , t , loc 2 s + ε , s + ε ( Σ λ × R ) L ( R n × R )

is upper semi-continuous up to the boundary T λ and bounded from above, satisfying (1.7) and

(3.1) ( t Δ p ) s u λ ( t Δ p ) s u 0 at t h e p o i n t s i n Σ λ × R ( u λ > u ) .

Then

(3.2) w λ 0 in Σ λ × R .

Proof

If (3.2) is invalid, since w λ ( z , t ) is bounded in Σ λ × R , there exists a constant h > 0 such that

(3.3) sup Σ λ × R w λ ( z , t ) = h > 0 .

Because Σ λ × R is unbounded, the supremum of w λ ( z , t ) about z and t may not exist, but from (3.3), we know that there must be a sequence { ( z k , t k ) } Σ λ × R such that

0 < w λ ( z k , t k ) = h k h as k .

Let δ k = h h k , obviously the sequence { δ k } is nonnegative and δ k 0 when k .

To solve the situation where the supremum of w λ for z and t cannot be reached. We perturb the function w λ for z and t . This allows the perturbation function to reach its supremum and maintains the antisymmetry of z . To do this, let φ ( z ) C 0 ( R n ) satisfy

(3.4) φ ( z ) = e 1 + 1 z 2 1 , z B 1 ( 0 ) , 0 , z B 1 ( 0 )

and

φ λ ( z ) = φ ( z λ ) = e 1 + 1 z λ 2 1 z B 1 ( 0 λ ) , 0 z B 1 ( 0 λ ) .

Define

Φ k ( z ) = φ 2 ( z z k ) r k φ λ 2 ( z z k ) r k C 0 ( B r k 2 ( z k ) B r k 2 ( ( z k ) λ ) ) ,

where r k = dist ( z k , T λ ) , it can be seen that Φ k ( z ) is an antisymmetric function about T λ .

We have

max z R n Φ k ( z ) = Φ k ( z k ) = φ ( 0 ) = 1 .

For convenience, we denote

φ k ( z ) = φ 2 ( z z k ) r k , ( φ k ) λ ( z ) = φ λ 2 ( z z k ) r k ,

then Φ k ( z ) = φ k ( z ) ( φ k ) λ ( z ) .

Let the smooth cut-off function

η k ( t ) = η t t k ( r k 2 ) 2 C 0 t k r k 2 2 , t k + r k 2 2

satisfy (2.5).

Next, assume that

V k ( z , t ) = w λ ( z , t ) + δ k Φ k ( z ) η k ( t )

and

Q r k 2 ( z k , t k ) = B r k 2 ( z k ) × t k r k 2 2 , t k + r k 2 2 .

Therefore, by a direct calculation, we obtain

V k ( z k , t k ) = h k + h h k = h

and

V k ( z , t ) = w λ ( z , t ) h in Q r k 2 c ( z k , t k ) ( Σ λ × R ) .

Thus, there exists ( z ¯ k , t ¯ k ) Q r k 2 ( z k , t k ) such that

(3.5) h + δ k V k ( z ¯ k , t ¯ k ) = sup Σ λ × R V k ( z , t ) h > 0 .

By the definition of V k , it holds

(3.6) h w λ ( z ¯ k , t ¯ k ) h δ k = h k > 0 .

For the maximum point ( z ¯ k , t ¯ k ) of V k ( z , t ) in Σ λ × R , we have

( t Δ p ) s ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) ) ( t Δ p ) s ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) ) = C n , s p t ¯ k R n G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 e z ¯ k ζ 2 4 ( t ¯ k ϱ ) d ζ d ϱ C n , s p t ¯ k R n G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 e z ¯ k ζ 2 4 ( t ¯ k ϱ ) d ζ d ϱ = C n , s p t ¯ k Σ λ e z ¯ k ζ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) ] d ζ d ϱ + C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) ] d ζ d ϱ = C n , s p t ¯ k Σ λ ( e z ¯ k ζ 2 4 ( t ¯ k ϱ ) e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) ] d ζ d ϱ + C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 { [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) ] + [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) ] } d ζ d ϱ = C n , s p t ¯ k Σ λ ( e z ¯ k ζ 2 4 ( t ¯ k ϱ ) e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) ] d ζ d ϱ + C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 { [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) ] + [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) ] } d ζ d ϱ I 1 + I 2 .

We first estimate I 1 . By the strict monotonicity of G and the fact that

( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) = ( w λ ( z ¯ k , t ¯ k ) + δ k Φ k ( z ¯ k ) η k ( t ¯ k ) ) ( w λ ( ζ , ϱ ) + δ k Φ k ( ζ ) η k ( ϱ ) ) = V k ( z ¯ k , t ¯ k ) V k ( ζ , ϱ ) 0 ,

we have

(3.7) G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) 0 .

Combining (3.7) and

e z ¯ k ζ 2 4 ( t ¯ k ϱ ) e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) > 0 ,

then

(3.8) I 1 = C n , s p t ¯ k Σ λ ( e z ¯ k ζ 2 4 ( t ¯ k ϱ ) e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) ] d ζ d ϱ 0 .

Next, we estimate I 2 . By (3.5) and Lemma 4.1 in [32], we have

(3.9) I 2 = C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 { [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u ( ζ , ϱ ) δ k ( φ k ) λ ( ζ ) η k ( ϱ ) ) ] + [ G ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) G ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) u λ ( ζ , ϱ ) δ k φ k ( ζ ) η k ( ϱ ) ) ] } d ζ d ϱ 2 C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 ( w λ ( z ¯ k , t ¯ k ) + δ k Φ k ( z ¯ k ) η k ( t ¯ k ) ) p 1 d ζ d ϱ 2 ( V k ( z ¯ k , t ¯ k ) ) p 1 C n , s p t ¯ k Σ λ e z ¯ k ζ λ 2 4 ( t ¯ k ϱ ) ( t ¯ k ϱ ) n 2 + 1 + s p 2 d ζ d ϱ = C ( V k ( z ¯ k , t ¯ k ) ) p 1 Σ λ 1 z ¯ k ζ λ n + s p d ζ C h p 1 r k s p .

Combining (3.8) and (3.9), we have

(3.10) ( t Δ p ) s ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) ) ( t Δ p ) s ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) ) = I 1 + I 2 C h p 1 r k s p .

In order to continue with the proof, the following lemma is required.

Lemma 3.1

Assume that u ( z , t ) C z , t , loc 2 s + ε , s + ε ( R n × R ) and ϕ ( z , t ) = φ ( z ) η ( t ) C 0 ( B 1 ( 0 ) × ( 1 , 1 ) ) , where φ ( z ) is defined in (3.4) and η ( t ) is defined in (2.5). Then, there exists a positive constant C 0 such that

( t Δ p ) s ( u + δ k ϕ ) ( z , t ) ( t Δ p ) s u ( z , t ) δ k C 0 f o r ( z , t ) B 1 ( 0 ) × ( 1 , 1 ) .

Proof

By using the definition of ( t Δ p ) s and G ( t ) 0 . For ( z , t ) B 1 ( 0 ) × ( 1 , 1 ) , we have

( t Δ p ) s ( u + δ k ϕ ) ( z , t ) ( t Δ p ) s u ( z , t ) = C n , s p t R n [ G ( ( u + δ k ϕ ) ( z , t ) ( u + δ k ϕ ) ( ζ , ϱ ) ) G ( u ( z , t ) u ( ζ , ϱ ) ) ] ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ = C n , s p t R n δ k G ( ξ ( z , t ) ) ( ϕ ( z , t ) ϕ ( ζ , ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ = C δ k t 1 t B 1 c ( z ) + t 1 B 1 c ( z ) + t B 1 ( z ) ϕ ( z , t ) ϕ ( ζ , ϱ ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ I 1 + I 2 + I 3 ,

where

( u + δ k ϕ ) ( z , t ) ( u + δ k ϕ ) ( ζ , ϱ ) ξ ( z , t ) u ( z , t ) u ( ζ , ϱ ) .

Similar to the proof of Lemma 2.1 in [7], it follows

( t Δ p ) s ( u + δ k ϕ ) ( z , t ) ( t Δ p ) s u ( z , t ) δ k C 0 for ( z , t ) B 1 ( 0 ) × ( 1 , 1 ) .

From Lemma 3.1, it is easy to derive the following corollary.□

Corollary 3.1

For some ( z 0 , t 0 ) R n × R and r > 0 , let u ( z , t ) C z , t , loc 2 s + ε , s + ε ( R n × R ) and

ϕ r ( z , t ) C 0 ( B r ( z 0 ) × ( r 2 + t 0 , r 2 + t 0 ) ) ,

then

( t Δ p ) s ( u + δ k ϕ r ) ( z , t ) ( t Δ p ) s u ( z , t ) δ k C 0 r s p ,

where ϕ and the constant C 0 > 0 are defined in Lemma 3.1.

Now we continue the proof of Theorem 3.1

From (3.1) and Corollary 3.1, we can conclude that

(3.11) ( t Δ p ) s ( u λ ( z ¯ k , t ¯ k ) + δ k φ k ( z ¯ k ) η k ( t ¯ k ) ) ( t Δ p ) s ( u ( z ¯ k , t ¯ k ) + δ k ( φ k ) λ ( z ¯ k ) η k ( t ¯ k ) ) ( t Δ p ) s u λ ( z ¯ k , t ¯ k ) ( t Δ p ) s u ( z ¯ k , t ¯ k ) + C δ k r k s p C δ k r k s p .

It follows from (3.10) and (3.11) that

( h ) p 1 C δ k .

Therefore, for sufficiently large k , we derive a contradiction. It follows that (3.2) is valid.□

3.2 Liouville theorem in the whole space

Theorem 3.2

Assume that

u C z , t , loc 2 s + ε , s + ε ( R n × R )

is a bounded solution of

(3.12) ( t Δ p ) s u = 0 , in R n × R .

Then, u ( z , t ) must be a constant.

Proof

We choose z 1 as any direction and T λ , Σ λ , z λ , u λ , w λ are defined as shown before. According to (3.12), we have

( t Δ p ) s u λ ( t Δ p ) s u = 0 , in Σ λ × R .

and

w λ ( z , t ) = w λ ( z λ , t ) , in Σ λ × R .

At the same time, since u is bounded, w λ is also bounded. Applying Theorem 3.1, we have

w λ 0 in Σ λ × R .

By substituting w λ with w λ , it holds

w λ 0 in Σ λ × R .

Therefore, we derive

w λ = 0 in Σ λ × R .

For any hyperplane perpendicular to the z 1 -axis, u ( z , t ) is symmetric due to the arbitrariness of λ . Further, combined with the fact that the direction of z 1 is arbitrary, for t R , it follows that u ( z , t ) is symmetric. Consequently, u depends on t , that is,

u ( z , t ) = u ( t ) , ( z , t ) R n × R .

The proof of Theorem 3.2 can be simplified to proving that the bounded solution of

(3.13) D left s u ( t ) = 0 in R

must be a constant.

The fact that u is bounded means that u ( t ) s ( R ) , where

s ( R ) = u L loc 1 ( R ) t u ( ϱ ) 1 + ϱ 1 + s d ϱ < + for any t R

is a one-side distributional space.

Thus, let D left s u be a distribution [27]

+ ( D left s u ( t ) ) ψ ( t ) d t = + u ( t ) D right s ψ ( t ) d t , for any ψ S ,

where S represents the class of Schwartz functions and D right s represents the Marchaud right fractional derivative given by

D right s ψ ( t ) = 1 Γ ( s ) t + ψ ( t ) ψ ( ϱ ) ( t ϱ ) 1 + s d ϱ .

Moreover, define its Fourier transform and the inverse Fourier transform with respect to distribution, as u is a tempered distribution. Combining

F ( D right s ψ ) ( ρ ) = ( i ρ ) s F ( ψ ) ( ρ ) for any ψ S

in [27] and (3.13), for any ψ S , we have

(3.14) 0 = + ( D left s u ( t ) ) ψ ( t ) d t = + u ( t ) D right s ψ ( t ) d t = + u ( t ) F 1 ( ( i ρ ) s F ( ρ ) ) ( t ) d t .

Next, we prove that

(3.15) F u , ϕ = 0 , ϕ C 0 ( R \ { 0 } ) .

Assume that ϕ C 0 ( R \ { 0 } ) S , then ϕ ( ρ ) ( i ρ ) s also belongs to C 0 ( R \ { 0 } ) . Thus, there must be ψ S , satisfying

F ( ψ ) ( ρ ) = ϕ ( ρ ) ( i ρ ) s .

By (3.14), we have

Fu , ϕ ¯ = F u , ( i ρ ) s F ( ψ ) ( ρ ) ¯ = u , F 1 ( ( i ρ ) s F ( ψ ) ( ρ ) ) ¯ = + u ( t ) F 1 ( ( i ρ ) s F ( ψ ) ( ρ ) ) ( t ) ¯ d t = 0 .

Therefore, (3.15) holds, indicating that F ( u ) is supported at the origin. It follows that u ( t ) is a polynomial of t , then when t , we obtain

u ( t ) C ,

due to the boundedness of u .□

4 Non-existence

In this section, we obtained the non-existence of positive solutions to the master equation with fractional p -Laplacian by utilizing Lemma 4.1 (maximum principle in bounded domains).

Lemma 4.1

Let Ω R n be a bounded domain and [ a , b ] R . Suppose that the lower semi-continuous function w λ C z , t , loc 2 s + ε , s + ε ( R n × R ) L ( R n × R ) satisfies the following equation:

(4.1) ( t Δ p ) s u λ ( t Δ p ) s u 0 , i n Ω × ( a , b ] , w λ > 0 , i n ( R n × ( , t 2 ] ) \ ( Ω × ( a , b ] ) .

Then

(4.2) w λ > 0 , i n Ω × ( a , b ] .

Proof

Suppose (4.2) is incorrect, then there exists ( z 0 , t 0 ) Ω × ( a , b ] such that

w λ ( z 0 , t 0 ) = inf R n × ( , t 2 ] w λ ( z , t ) < 0 .

By a direct calculation, we obtain

( t Δ p ) s u λ ( z 0 , t 0 ) ( t Δ p ) s u ( z 0 , t 0 ) = C n , s p t 0 R n G ( u λ ( z 0 , t 0 ) u λ ( ζ , ϱ ) ) G ( u ( z 0 , t 0 ) u ( ζ , ϱ ) ) ( t 0 ϱ ) n 2 + 1 + s p 2 e z 0 ζ 2 4 ( t 0 ϱ ) d ζ d ϱ = C n , s p t 0 R n G ( ξ ( z , t ) ) ( w λ ( z 0 , t 0 ) w λ ( ζ , ϱ ) ) ( t 0 ϱ ) n 2 + 1 + s p 2 e z 0 ζ 2 4 ( t 0 ϱ ) d ζ d ϱ < 0 ,

where

u λ ( z 0 , t 0 ) u λ ( ζ , ϱ ) ξ ( z , t ) u ( z 0 , t 0 ) u ( ζ , ϱ ) .

This contradicts (4.1). So (4.2) is proved.□

Next, we prove the non-existence of positive solutions of the master equation with fractional p -Laplacian.

Theorem 4.1

Assume that u C z , t , loc 2 s + ε , s + ε ( Σ λ × R ) L ( R n × R ) is a positive bounded solution satisfying

(4.3) ( t Δ p ) s u = f ( u ) in R n × R .

Suppose f satisfies

  1. f ( u ) = z 1 ψ ( u ) ,

  2. ψ is monotonically increasing with respect to u,

  3. ψ > C , where C is a positive constant.

If conditions (a), (b) or (a), (c) hold, then equation (4.3) has no positive bounded solutions.

Proof

Case 1: Conditions ( a ) and ( b ) hold.

Assume that

(4.4) ( Δ ) s ϕ ( z ) = λ 1 ϕ ( z ) , z B 1 ( 0 ) , ϕ ( z ) = 0 z B 1 c ( 0 ) ,

where λ 1 is the first eigenvalue and ϕ ( z ) is the eigenfunction. Denote

(4.5) max R n ϕ ( z ) = 1

and

ϕ L ( z ) = ϕ ( z L e 1 ) ,

where L 1 and e 1 is a unit vector in the z 1 -direction. Let

(4.6) ν ( z , t ) = ϕ L ( z ) ρ ( t ) ,

where ρ ( t ) = t β 1 and 0 < β = 1 2 k + 1 < s for some integer k > 0 . Assume that u is a positive bounded solution of (4.3). Denote

(4.7) T = ( M + 1 ) 1 β with M sup R n × R u ( z , t ) .

By the definition of ( t Δ p ) s and Lemma 2.2 in [15], for all ( z , t ) B 1 ( L e 1 ) × [ 1 , T ] , we have

(4.8) ( t Δ p ) s ν ( z , t ) = C n , s p t R n G ( ϕ L ( z ) ρ ( t ) ϕ L ( ζ ) ρ ( ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ C n , s p t B 1 c ( L e 1 ) G ( ϕ L ( z ) ρ ( t ) ϕ L ( ζ ) ρ ( ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ + C n , s p t B 1 ( L e 1 ) G ( ϕ L ( z ) ρ ( t ) ϕ L ( ζ ) ρ ( ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ

C n , s p 1 2 2 p t B 1 c ( L e 1 ) G ( ϕ L ( z ) ρ ( t ) ) G ( ϕ L ( ζ ) ρ ( ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ + C n , s p t B 1 ( L e 1 ) G ( ϕ L ( z ) ρ ( t ) ϕ L ( ζ ) ρ ( ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ C n , s p 1 2 2 p t B 1 c ( L e 1 ) G ( ξ ( z , t ) ) ( ϕ L ( z ) ρ ( t ) ϕ L ( ζ ) ρ ( ϱ ) ) ( t ϱ ) n 2 + 1 + s p 2 e z ζ 2 4 ( t ϱ ) d ζ d ϱ + C ρ ( t ) ( Δ ) s ϕ L ( z ) + sup R n ϕ L ( z ) t s ρ ( t ) ( λ 1 ρ ( t ) + C s t γ β ) sup R n ϕ L ( z ) λ 1 ( T β 1 ) + C s C T ,

where

ϕ L ( z ) ρ ( t ) ξ ( z , t ) ϕ L ( ζ ) ρ ( ϱ ) .

Denote

m T inf B 1 ( 0 ) × [ 1 , T ] u ( z , t ) > 0 .

Combining (4.3), conditions (a) and (b), for all ( z , t ) B 1 ( L e 1 ) × [ 1 , T ] , it follows

(4.9) ( t Δ p ) s u ( z , t ) = z 1 ψ ( u ( z , t ) ) ( L 1 ) ψ ( m T ) .

By choosing a sufficiently large L , we make

(4.10) ( L 1 ) ψ ( m T ) > C T .

Let w λ ( z , t ) = u ( z , t ) ν ( z , t ) . It follows from that (4.8), (4.9), and (4.10), we have

( t Δ p ) s u ( t Δ p ) s ν 0 in B 1 ( L e 1 ) × [ 1 , T ] , w λ > 0 in ( R n × ( , T ] ) \ ( ( L e 1 ) × [ 1 , T ] ) .

Applying Lemma 4.1, it holds

u ( z , t ) > ν ( z , t ) , in B 1 ( L e 1 ) × [ 1 , T ] .

According to the definition of ν ( z , t ) , it follows

u ( z , t ) > ϕ L ( z ) ( t β 1 ) in B 1 ( L e 1 ) × [ 1 , T ] .

Then

M > max R n ϕ L ( z ) ( T β 1 ) .

This is contradictory. Thus, equation (4.3) has no positive bounded solutions.

Case 2: Conditions (a) and (b) hold.

Assume that u is a positive bounded solution of (4.3). Similar to the method in Case 1, by (4.7) and (4.8), for all ( z , t ) B 1 ( L e 1 ) × [ 1 , T ] , we have

( t Δ p ) s ν ( z , t ) C T .

Combining (4.3), conditions ( a ) and ( c ) , for all ( z , t ) B 1 ( L e 1 ) × [ 1 , T ] , it follows

( t Δ p ) s u ( z , t ) = z 1 ψ ( u ( z , t ) ) ( L 1 ) C .

By choosing a sufficiently large L , we have

( L 1 ) C > C T .

Similar to the proof above, we can derive a contradiction by applying Lemma 4.1. Therefore, we prove that equation (4.3) has no positive bounded solutions.□

  1. Funding information: The work was supported by the Fundamental Research Program of Shanxi Province, China (No. 202403021221163), and the Graduate Education Innovation Program of Shanxi, China (No. 2025JG091).

  2. Author contributions: All authors contributed equally to the writing of this article. All authors accepted the responsibility for the content of the manuscript and consented to its submission, reviewed all the results, and approved the final version of the manuscript.

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

  4. Data availability statement: No data was used for the research described in the article.

References

[1] J. Belmonte-Beitia, T. E. Woolley, J. G. Scott, P. K. Maini, and E. A. Gaffney, Modelling biological invasions: Individual to population scales at interfaces, J. Theoret. Biol. 334 (2013), 1–12, https://doi.org/10.1016/j.jtbi.2013.05.033. Search in Google Scholar PubMed

[2] C. Bertucci, J. M. Lasry, and P. L. Lions, On Lipschitz solutions of mean field games master equations, J. Funct. Anal. 287 (2024), 110486, https://doi.org/10.1016/j.jfa.2024.110486. Search in Google Scholar

[3] N. G. Bhat, and S. Balaji, Modeling and simulation of Lac-Operon using reaction-diffusion master equation on heterogeneous parallel platforms, Comput. Phys. Commun. 296 (2024), 109018, https://doi.org/10.1016/j.cpc.2023.109018. Search in Google Scholar

[4] D. del-Castillo-Negrete, B. A. Carreras, and V. E. Lynch, Fractional diffusion in plasma turbulence, Phys. Plasmas, 11 (2004), 3854–3864, https://doi.org/10.1063/1.1767097. Search in Google Scholar

[5] W. Chen, Y. Guo, and L. Ma, Liouville theorem for fully fractional master equations and its applications, Math. Z. 309 (2025), 1–22, https://doi.org/10.1007/s00209-025-03693-y. Search in Google Scholar

[6] W. Chen, C. Li, Maximum principles for the fractional p-Laplacian and symmetry of solutions, Adv. Math. 335 (2018), 735–758, https://doi.org/10.1016/j.aim.2018.07.016. Search in Google Scholar

[7] W. Chen, and L. Ma, Gibbons’ conjecture for entire solutions of master equations, Commun. Contemp. Math., 27 (2025), 2450030, https://doi.org/10.1142/S0219199724500305. Search in Google Scholar

[8] G. Cottone, M. Di Paola, and M. Zingales, Elastic waves propagation in 1D fractional non-local continuum, Physica E Low Dimens. Sys. Nanostruct. 42 (2009), 95–103, https://doi.org/10.1016/j.physe.2009.09.006. Search in Google Scholar

[9] M. A. Ezzat, and A. S. El Karamany, Fractional order heat conduction law in magneto-thermoelasticity involving two temperatures, Z. Angew. Math. Phys. 62 (2011), 937–952, https://doi.org/10.1007/s00033-011-0126-3. Search in Google Scholar

[10] T. X. Gou, and H. R. Sun, Solutions of nonlinear Schrödinger equation with fractional Laplacian without the Ambrosetti-Rabinowitz condition, Appl. Math. Comput. 257 (2015), 409–416, https://doi.org/10.1016/j.amc.2014.09.035. Search in Google Scholar

[11] M. K. Hamdani, L. Mbarki, and M. Allaoui, A new class of multiple nonlocal problems with two parameters and variable-order fractional p(⋅)-Laplacian, Commun. Anal. Mech. 15 (2023), 551–574, https://doi.org/10.3934/cam.2023027. Search in Google Scholar

[12] I. Kossowski, and B. Przeradzki, Nonlocal heat equations with generalized fractional Laplacian, Adv. Nonlinear Anal. 13 (2024), 20240053, https://doi.org/10.1515/anona-2024-0053. Search in Google Scholar

[13] N. V. Krylov, Lectures on elliptic and parabolic equations in Hölder spaces, Graduate Studies in Mathematics 12, American Mathematical Society, Providence, RI, 1996. ISBN 0-8218-0569-X.10.1090/gsm/012Search in Google Scholar

[14] V. Lecomte, C. Appert-Rolland, and F. van Wijland, Chaotic properties of systems with Markov dynamics, Phys. Rev. Lett., 95 (2005), 010601, https://doi.org/10.1103/PhysRevLett.95.010601. Search in Google Scholar PubMed

[15] Z. Li, and Q. Zhang, Sub-solutions and a point-wise Hopf’s Lemma for Fractional p-Laplacian, 2019, arXiv: http://arXiv.org/abs/arXiv:1905.00127, https://doi.org/10.48550/arXiv.1905.00127. Search in Google Scholar

[16] N. Liao, Hölder regularity for parabolic fractional p-Laplacian, Calc. Var. Partial Differ. Equ. 63 (2024), 22, https://doi.org/10.1007/s00526-023-02627-y. Search in Google Scholar PubMed PubMed Central

[17] C. S. Lin, G. Tarantello, and Y. Yang, Solutions to the master equations governing fractional vortices, J. Differ. Equ. 254 (2013), 1437–1463, https://doi.org/10.1016/j.jde.2012.10.023. Search in Google Scholar

[18] Q. Lin, X. Tian, R. Xu, and M. Zhang, Blow up and blow up time for degenerate Kirchhoff-type wave problems involving the fractional Laplacian with arbitrary positive initial energy, Dis. Cont. Dyn-S. 13 (2020), 2095–2107, https://doi.org/10.3934/dcdss.2020160. Search in Google Scholar

[19] L. Ma, Y. Guo, and Z. Zhang, Radial symmetry and Liouville theorem for master equations, Fract. Calc. Appl. Anal. 27 (2024), 2620–2643, https://doi.org/10.1007/s13540-024-00328-7. Search in Google Scholar

[20] A. Marchaud, Sur les dérivées et sur les différences des fonctions de variables réelles, J. Math. Pures Appl. 6 (1927), 337–425. http://www.numdam.org/item?id=JMPA_1927_9_6__337_0.Search in Google Scholar

[21] R. Metzler, and J. Klafter, The random walkas guide to anomalous diffusion: A fractional dynamicsapproach, Phys. Rep. 339 (2000), 1–77, https://doi.org/10.1016/S0370-1573(00)00070-3. Search in Google Scholar

[22] H. Mi, X. Deng, and W. Zhang, Ground state solution for asymptotically periodic fractional p-Laplacian equation, Appl. Math. Lett. 120 (2021), 107280, https://doi.org/10.1016/j.aml.2021.107280. Search in Google Scholar

[23] Y. Pang, V. Rǎdulescu, and R. Xu, Global existence and finite time blow-up for the m-Laplacian parabolic problem, Acta Math. Sin. 39 (2023), 1497–1524, https://doi.org/10.1007/s10114-023-1619-7. Search in Google Scholar

[24] M. Raberto, E. Scalas, and F. Mainardi, Waiting-times and returns in high-frequency fnancialdata: an empirical study, Phys. A 314 (2002), 749–755, https://doi.org/10.1016/S0378-4371(02)01048-8. Search in Google Scholar

[25] M. Riesz, Integrales de Riemann-Liouville et potentiels, Acta Sci. Math. Szeged 9 (1938), 1–42, https://doi.org/10.1007/978-3-642-37535-4_38.Search in Google Scholar

[26] P. R. Stinga, and J. L. Torrea, Regularity theory and extension problem for fractional nonlocalparabolic equations and the master equation, SIAM J. Math. Anal. 49 (2017), 3893–3924, https://doi.org/10.1137/16M1104317. Search in Google Scholar

[27] P. R. Stinga, and M. Vaughan, One-sided fractional derivatives, fractional Laplacians, and weighted Sobolev spaces, Nonlinear Anal. 193 (2020), 111505, https://doi.org/10.1016/j.na.2019.04.004. Search in Google Scholar

[28] P. Wang, Monotonicity and uniqueness of positive solutions to elliptic fractional p-equations, Fract. Calc. Appl. Anal. 26 (2023), 837–850, https://doi.org/10.1007/s13540-023-00136-5. Search in Google Scholar

[29] P. Wang, and W. Chen, Hopfas lemmas for parabolic fractional p-Laplacians, Comm. Pure Appl. Math. 21 (2022), 3055–3069, https://doi.org/10.3934/cpaa.2022089. Search in Google Scholar

[30] C. Wang, and J. Sun, Normalized solutions for the p-Laplacian equation with a trapping potential, Adv. Nonlinear Anal. 12 (2023), 20220291, https://doi.org/10.1515/anona-2022-0291. Search in Google Scholar

[31] X. Wang, and Z. Yang, Nonexistence of positive solutions for an indefinite fractional p-Laplacian equation, Nonlinear Anal. 195 (2020), 111740, https://doi.org/10.1016/j.na.2019.111740. Search in Google Scholar

[32] L. Wu, and C. Wen, The sliding methods for the fractional p-Laplacian, Adv. Math. 361 (2020), 106933, https://doi.org/10.1016/j.aim.2019.106933. Search in Google Scholar

[33] X. Wu, Y. Zhao, and X. Yang, On a singular parabolic p-Laplacian equation with logarithmic nonlinearity, Commun. Anal. Mech. 16 (2024), 528–553, https://doi.org/10.3934/cam.2024025. Search in Google Scholar

[34] L. Zhang, B. Ahmad, G. Wang, and X. Ren, Radial symmetry of solution for fractional p-Laplaciansystem, Nonlinear Anal. 196 (2020), 111801, https://doi.org/10.1016/j.na.2020.111801. Search in Google Scholar

[35] L. Zhao, H. Cai, and Y. Chen, Multiple nontrivial solutions of superlinear fractional Laplace equations without (AR) condition, Adv. Nonlinear Anal. 12 (2023), 20220281, https://doi.org/10.1515/anona-2022-0281. Search in Google Scholar

[36] L. Zhang, and W. Hou, Standing waves of nonlinear fractional p-Laplacian Schrodinger equation involving logarithmic nonlinearity, Appl. Math. Lett. 102 (2020), 106149, https://doi.org/10.1016/j.aml.2019.106149. Search in Google Scholar

[37] L. Zhang, Q. Liu, B. Ahmad, and G. Wang, Nonnegative solutions of a coupled k-Hessian system involving different fractional Laplacians, Fract. Calc. Appl. Anal 27 (2024), 1–17, https://doi.org/10.1007/s13540-024-00277-1. Search in Google Scholar

Received: 2024-06-19
Revised: 2025-09-22
Accepted: 2025-10-13
Published Online: 2025-11-15

© 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-2025-0124/html
Scroll to top button