Skip to content
BY 4.0 license Open Access Published by De Gruyter Open Access May 23, 2024

Almost periodic dynamics for a delayed differential neoclassical growth model with discontinuous control strategy

  • Qian Wang , Wei Wang and Qian Zhan EMAIL logo
From the journal Open Mathematics

Abstract

In this study, we are concerned with the existence and exponential stability issue of a delayed differential neoclassical growth model with discontinuous control strategy. By employing the Filippov’s theory and dichotomy theory, together with the Lyapunov functional method, novel criteria on existence and exponential stability are established for the addressed model. The established theoretical results extend and supplement the related results in the existing literature. Moreover, a simulation example is presented to verify the practicability of the proposed results.

MSC 2010: 34K14; 34K20; 34K39

1 Introduction

Day [1,2] used the nonlinear delay differential equation of the following form:

(1.1) x ( t ) = α x ( t ) + β x n ( t τ ) e δ x ( t τ ) ,

where α , β , δ , and τ are positive constants, to describe the appearance of complex dynamics even though the economic structure is very simple. Economically, x ( t ) is the capital per capita at time t , α stands for the depreciation ratio of capital, δ suggests an intensity of a “negative effect” given rise by increasing condensation of capital, and τ means the delay in the production process. The dynamical behaviors of equation (1.1) have been widely investigated in [35], where this model was called as the differential neoclassical growth model.

The theory and application of the differential neoclassical growth model have made prominent progress in the last two decades and drawn considerable attentions among researchers (e.g., [68] and references therein). Many meaningful and important results on the dynamical properties including the well-posedness, positive (almost) periodic solutions, persistence, attractivity, and stability of equation (1.1) and its generalizations (in particular, to variable parameters and stochastic equations) have been obtained. For example, Wang and Chen [9] studied the existence and ultimate boundedness in mean for a delayed differential neoclassical growth model in random environments. Long and Wang [10] established a criterion ensuring the existence and global exponential stability of an almost periodic non-autonomous delayed differential neoclassical growth model with an oscillating death rate. For more interesting dynamical results for differential neoclassical growth models, one can refer [1115] and references therein.

However, in real economic activities, under the influence of inventory changes, capital flows, financial fluctuations, and even financial crises, issues such as commodity pricing decisions and monetary policy regulation in modern economic and financial markets are highly susceptible to instantaneous impacts from external reality factors, it is necessary to implement some economic regulation policies under such circumstances. Therefore, the research on economic dynamics with control strategy is an interesting and important research topic in mathematical economics, which is in accordance with the optimal management of economic activities, according to the economic situation of different stages, the relevant departments will adopt different control strategies, even discontinuous. On the other hand, for most successful applications of various models in the real world, stability is usually a prerequisite and the convergence dynamics of models depends mainly on their parametric configuration (e.g., [16,17]). Until now, fewer results have been reported on the dynamics of positive almost periodic solutions of delayed differential neoclassical growth model with discontinuous control strategy. Inspired by the aforementioned discussion, in this work, we shall investigate the existence and exponential stability problems of positive almost periodic solutions for a discontinuous differential neoclassical growth model of the following type:

(1.2) x ( t ) = α ( t ) x ( t ) + β ( t ) x n ( t τ ( t ) ) e δ ( t ) x ( t τ ( t ) ) b ( t ) C ( x ( t ) ) ,

where C ( x ( t ) ) is the discontinuous control function and b ( t ) stands for the technological coefficient like control strength at time t . In order to ensure the sustainability of the economy, the economy is prohibited from controlling, which can also be made to be zero to reflect this fact. The other variables and parameters admit the same economic significance as described in (1.1) but it needs to consider the influence of time factor.

In consideration of the above analysis, the main purpose of this work is to study the almost periodic dynamics of delayed differential neoclassical growth model (1.2) with discontinuous control strategy. Compared with the results of previous literature, the distinct characteristics of this work are summarized as follows:

  1. This work first considers the exponential stability problem of solutions for a delayed almost periodic differential neoclassical growth model with discontinuous control strategy, which expands the scope of the study.

  2. Since the control strategy employed herein is discontinuous, by employing the concept of Filippov solutions and the set-valued mapping theory, a unified criterion ensuring the existence and exponential stability of almost periodic solutions for the considered discontinuous model is established.

  3. The exponential stability criteria are independent of time delay, which enhanced the anti-interference ability and robustness against time delays.

The structure of this work is outlined as follows. Section 2 presents some preliminaries, Section 3 establishes the existence and exponential stability of solutions for the considered almost periodic model. Section 4 provides a numerical example to verify the feasibility of the established theoretical results.

Notations: As to a bounded function p : R R , denote

p + = sup t R p ( t ) and p = inf t R p ( t ) .

Let τ + = sup t R τ ( t ) and C = C ( [ τ + , 0 ] , R ) indicate the Banach space composed of all real-valued continuous functions mapping from [ τ + , 0 ] to R supplemented with supremum norm . Observe that C + = { φ C : φ ( θ ) R + for θ [ τ + , 0 ] } is a positive cone in C , where R + = [ 0 , ) . Generally, a continuous function x ( t ) on [ τ + + t 0 , ϱ ) with t 0 < ϱ , then, for all t [ t 0 , ϱ ) , we denote x t ( θ ) = x ( t + θ ) by x t C , for all θ [ τ + , 0 ) .

Let Q R be a set, K [ Q ] means the closure of the convex hull of Ω .

2 Preliminaries

Given that the economic explanation of model (1.2), just positive solutions are meaningful and thus admissible, and so, we give the following initial condition:

(2.1) x t 0 = φ , φ C + , φ ( 0 ) > 0 .

Indicate x t ( t 0 , φ ) or x ( t ; t 0 , φ ) for an admissible solution for initial value problem (1.2) and (2.1) with x t 0 ( t 0 , φ ) = φ C + and t 0 R . Moreover, let [ t 0 , η ( t 0 , φ ) ) or [ t 0 , η ( φ ) ) be the maximal right interval of existence of x ( t ; t 0 , φ ) .

Let 0 < n < 1 , it is readily seen that L ( v ) = e v v n 1 ( n v ) is decreasing with the range ( 0 , n + n ) and increasing with the range ( n + n , ) , and there is a unique κ 1 ( 0 , n ) such that

(2.2) L ( κ 1 ) = sup v κ 1 L ( v ) = L ( n + n ) .

In addition, one can readily see that v n e v is increasing on [ 0 , n ] and decreasing on [ n , ) , let κ 2 be the unique constant in [ n , + ) satisfying

κ 1 n e κ 1 = κ 2 n e κ 2 .

Next we always assume that the controlling function in model (1.2) admits the following discontinuous characteristics.

Definition 2.1

(Function class C ) Let us call C C , if C meets the assumptions in the following:

  1. C 0 is a monotone nondecreasing function and admits at most finite discontinuous jump points on every compact interval of R .

  2. C ( 0 ) = C ( 0 + ) = 0 , and there is a nonnegative constant N such that for any x y , for all γ K [ C ( x ) ] , η K [ C ( y ) ] ,

    (2.3) γ η x y N ,

    in which

    K [ C ( x ) ] = [ min { C ( x ) , C ( x + ) } , max { C ( x ) , C ( x + ) } ] .

Remark 2.1

Since the controlling function considered in model (1.2) is discontinuous, the usual Lipschitz condition is no longer satisfied, so the unilateral Lipschitz condition (2.3) is imposed to investigate the exponential stability of model (1.2), which has been successfully used to study the neural network and biomathematical models. Moreover, it is readily seen from Assumptions (1)–(2) that

(2.4) 0 sup γ K [ C ( x ) ] γ N x , for each x R .

In view of the proposed control strategy being discontinuous, there is a need to explain what does it mean of a solution to model (1.2). In this work, we shall employ the Filippov solution which was initially proposed by Filippov in [18].

Definition 2.2

(Filippov solution) Function x ( t ) is said to be a solution of model (1.2) on [ τ + , η ( φ ) ) ( η ( φ ) ( t 0 , ] ) if

  1. x ( t ) is continuous on [ τ + , η ( φ ) ) and absolutely continuous on each compact interval of [ t 0 , η ( φ ) ) ;

  2. there exists a function γ ( t ) : [ τ + , η ( φ ) ) R that is measurable satisfying γ ( t ) K [ C ( x ( t ) ) ] for a.a. t [ τ + , η ( φ ) ) . Moreover, it satisfies

    (2.5) x ( t ) = α ( t ) x ( t ) + β ( t ) x n ( t τ ( t ) ) e δ ( t ) x ( t τ ( t ) ) b ( t ) γ ( t ) ,

    for a.a. t [ t 0 , η ( φ ) ) .

By the above definition, it indicates that x ( t ) is a solution of model (1.2) in the sense of Filippov, since it satisfies

(2.6) x ( t ) α ( t ) x ( t ) + β ( t ) x n ( t τ ( t ) ) e δ ( t ) x ( t τ ( t ) ) b ( t ) K [ C ( x ( t ) ) ] ,

for a.a. t [ t 0 , η ( φ ) ) .

Remark 2.2

As is shown in [19], for all C C , the following set-valued map

x ( t ) α ( t ) x ( t ) + β ( t ) x n ( t τ ( t ) ) e δ ( t ) x ( t τ ( t ) ) b ( t ) K [ C ( x ( t ) ) ]

has nonempty compact convex values. Besides, it is upper semicontinuous and hence is measurable. It follows from the measurable selection theorem that there is a bounded measurable function γ ( t ) K [ C ( x ( t ) ) ] such that (2.5) holds.

Now, we present some definitions and dichotomy theory of almost periodic functions from [20].

Definition 2.3

Assume that f ( t ) : R R is continuous. f ( t ) is said to be almost periodic on R , if, for arbitrary ε > 0 , the set T ( f , ε ) = { w : f ( t + w ) f ( t ) < ε , t R } is relatively dense, that is, for arbitrary ε > 0 , one can find a real number l = l ( ε ) > 0 such that, for each interval with length l ( ε ) , there is a number w = w ( ε ) belonging to this interval such that f ( t + w ) f ( t ) < ε for all t R . The set of all almost periodic functions is represented by AP ( R , R ) .

Definition 2.4

Suppose x R n and Q ( t ) is a n -th order continuous matrix defined on R . The linear system

(2.7) x ( t ) = Q ( t ) x ( t )

is said to admit an exponential dichotomy on R , if there are positive constants k , α , and projection P such that the fundamental solution matrix X ( t ) of (2.7) meets

X ( t ) P X 1 ( s ) k e α ( t s ) , for t s ,

X ( t ) ( I P ) X 1 ( s ) k e α ( t s ) , for t s ,

in which I is the identity matrix.

Lemma 2.1

If system (2.7) has an exponential dichotomy, then the following almost periodic system

x ( t ) = Q ( t ) x ( t ) + q ( t )

admits a unique almost periodic solution x ( t ) , and

x ( t ) = t X ( t ) P X 1 ( s ) q ( s ) d s t X ( t ) ( I P ) X 1 ( s ) q ( s ) d s .

Lemma 2.2

Suppose c i is almost periodic, which is defined on R and

M [ c i ] = lim t 1 T t t + T c i ( s ) d s > 0 , i = 1 , 2 , , n .

Then, the linear system

x ( t ) = diag ( c 1 ( t ) , c 2 ( t ) , , c n ( t ) ) x ( t )

satisfies exponential dichotomy on R .

In order to obtain exponential stability criteria on almost periodic equation (1.2), the following Halanay inequality will be employed.

Lemma 2.3

[21,22] If

V ( t ) β V ( t ) + ζ sup t τ + s t V ( s ) for t t 0 ,

and β > ζ > 0 , then as t t 0 , the following inequality holds:

V ( t ) sup t 0 τ + s t 0 V ( s ) e μ ( t t 0 ) ,

where μ is the unique positive root to the following algebra equation:

μ = β ζ e μ τ + .

We are in a position to establish the criteria ensuring the global existence of positive solutions to model (1.2) with admissible initial conditions (2.1).

Lemma 2.4

Let C C , then the maximal right existence interval of each solution x ( t ; t 0 , φ ) for problem (1.2)–(2.1) is [ t 0 τ + , ) and x ( t ; t 0 , φ ) for t [ t 0 , ) . Moreover, limsup t x ( t ; t 0 , φ ) n n e n α β + ( δ ) n .

Proof

For convenience, let us, respectively, denote x ( t ; t 0 , φ ) and η ( t 0 , φ ) by x ( t ) and η ( φ ) .

We first prove that

(2.8) x ( t ) > 0 , for t ( t 0 , η ( φ ) ) .

Otherwise, there is t 1 ( t 0 , η ( φ ) ) satisfying x ( t 1 ) = 0 and x ( t ) > 0 for t [ t 0 , t 1 ) . It follows from C C that

0 x ( t 1 ) = α ( t 1 ) x ( t 1 ) + β ( t 1 ) x n ( t 1 τ ( t 1 ) ) e δ ( t 1 ) x ( t 1 τ ( t 1 ) ) b ( t 1 ) γ ( t 1 ) = β ( t 1 ) x n ( t 1 τ ( t 1 ) ) e δ ( t 1 ) x ( t 1 τ ( t 1 ) ) > 0 .

This is a contradiction, which means that (2.8) is true. We then prove that η ( φ ) = . Considering (2.8) and sup x [ 0 , + ) x n e x = n n e n , we have

x ( t ) α ( t ) x ( t ) + β ( t ) ( δ ( t ) ) n δ n ( t ) x n ( t τ ( t ) ) e δ ( t ) x ( t τ ( t ) ) α x ( t ) + n n e n β + ( δ ) n , for t [ t 0 , η ( φ ) ) ,

which produces

(2.9) x ( t ) e α ( t t 0 ) x ( t 0 ) + n n e n α β + ( δ ) n ( 1 e α ( t t 0 ) ) , for t [ t 0 , η ( φ ) ) .

This, combined with (2.8) and the continuation theorem [18, p. 78, Theorem 2], means that η ( φ ) = .

At last, one finds from η ( φ ) = and (2.9) that limsup t x ( t ) n n e n α β + ( δ ) n . The proof is completed.□

3 Main results

In this section, we shall establish new conditions on existence and exponential stability of solutions to almost periodic model (1.2).

Lemma 3.1

(Positive invariance) Let 0 < n < 1 . If there is a positive constant M > κ 1 satisfies

(3.1) n n e n M β + ( δ ) n < α α + < κ 1 n 1 e κ 1 β ( δ + ) n b + N M κ 1

and

(3.2) 1 δ δ + κ 2 M .

Then, the set C 0 = { φ φ C , κ 1 < φ ( t ) < M , for any t [ τ + , 0 ] } is positively invariant.

Proof

Suppose φ C 0 . According to Lemma 2.3, model (1.2) with initial condition (2.1) admits a unique solution x ( t ; t 0 , φ ) on [ t 0 τ + , ) and x ( t ; t 0 , φ ) > 0 for t [ t 0 τ + , ) . For simplicity, we denote x ( t ; t 0 , φ ) by x ( t ) . It is required to testify that

κ 1 < x ( t ) < M , for all t [ t 0 τ + , ) .

First, we claim that

(3.3) x ( t ) < M , for all t [ t 0 , ) .

Argued by contradiction, there is t 1 ( t 0 , η ( φ ) ) such that

(3.4) x ( t 1 ) = M and x ( t ) < M , for all t [ t 0 τ + , t 1 ) .

It follows from (2.4), (2.5), (3.1), (3.4), and the fact sup x R + x n e x = n n e n that

0 x ( t 1 ) = α ( t 1 ) x ( t 1 ) + β ( t 1 ) δ n ( t 1 ) δ n ( t 1 ) x n ( t 1 τ ( t 1 ) ) e δ ( t 1 ) x ( t 1 τ ( t 1 ) ) b ( t 1 ) γ ( t 1 ) α ( t 1 ) M + β ( t 1 ) δ n ( t 1 ) n n e n < 0 ,

which is impossible. This proves (3.3).

Second, we show that

(3.5) x ( t ) > κ 1 , for each t [ t 0 , ) .

If (3.5) does not hold, there is t 2 ( t 0 , ) satisfying

(3.6) x ( t 2 ) = κ 1 and x ( t ) > κ 1 , for all t [ t 0 τ + , t 2 ) .

Then,

κ 1 δ ( t 2 ) x ( t 2 τ ( t 2 ) ) δ ( t 2 ) M κ 2 ,

accordingly,

(3.7) ( δ ( t 2 ) x ( t 2 τ ( t 2 ) ) ) n e δ ( t 2 ) x ( t 2 τ ( t 2 ) ) min { κ 1 n e κ 1 , κ 2 n e κ 2 } = κ 1 n e κ 1 .

Combining (2.4), (2.5), (3.1) with (3.7) yields

0 x ( t 2 ) = α ( t 2 ) x ( t 2 ) + β ( t 2 ) δ n ( t 2 ) δ n ( t 2 ) x n ( t 2 τ ( t 2 ) ) e δ ( t 2 ) x ( t 2 τ ( t 2 ) ) b ( t 2 ) γ ( t 2 ) x ( t 2 ) x ( t 2 ) α ( t 2 ) κ 1 + β ( t 2 ) δ n ( t 2 ) κ 1 n e κ 1 b ( t 2 ) N κ 1 κ 1 α + + β ( δ + ) n κ 1 n 1 e κ 1 b + N > 0 ,

which is a contradiction. Thus, (3.5) holds. (3.3) together with (3.5) implies that the main results of Lemma 3.1 hold.□

Theorem 3.2

If the assumptions in Lemma 3.1hold, assume further that

(3.8) ρ = 1 α ( β + ( δ + ) 1 n L ( n + n ) + b + N ) < 1 ,

in which L ( u ) = e u u n 1 ( n u ) . Then, model (1.2) has a unique almost periodic solution in

Ω = { x x AP ( R , R ) , κ 1 x ( t ) M , t R } .

Proof

For each ϕ AP ( R , R ) , consider the following equation:

(3.9) x ( t ) = α ( t ) x ( t ) + β ( t ) ϕ n ( t τ ( t ) ) e δ ( t ) ϕ ( t τ ( t ) ) b ( t ) γ ϕ ( t ) ,

where γ ϕ ( t ) K [ C ( ϕ ( t ) ) ] is measurable. It is easy to see that M [ α ] > 0 , one finds from Lemma 2.2 that the following linear equation

x ( t ) = α ( t ) x ( t )

has exponential dichotomy on R . By Lemma 2.1, equation (3.9) just possesses an almost periodic solution, which is represented as

(3.10) x ϕ ( t ) = t e s t α ( u ) d u [ β ( s ) ϕ n ( s τ ( s ) ) e δ ( s ) ϕ ( s τ ( s ) ) b ( s ) γ ϕ ( s ) ] d s .

Set

Ω = { x x AP ( R , R ) , κ 1 x ( t ) M , t R } ,

evidently, Ω AP ( R , R ) is a closed set.

Define

(3.11) ( Γ ϕ ) ( t ) = t e s t α ( u ) d u [ β ( s ) ϕ n ( s τ ( s ) ) e δ ( s ) ϕ ( s τ ( s ) ) b ( s ) γ ϕ ( s ) ] d s , for ϕ Ω .

Apparently, to show that model (1.2) has a unique almost periodic solution, one only needs to prove that Γ admits a fixed point in Ω .

Let us first testify that the operator Γ is a self-mapping from Ω to itself. Actually, for each ϕ Ω , by (3.1), (3.11), and sup x R + x n e x = n n e n , one has

(3.12) ( Γ ϕ ) ( t ) t e α ( t s ) β ( s ) δ n ( s ) δ n ( s ) ϕ n ( s τ ( s ) ) e δ ( s ) ϕ ( s τ ( s ) ) d s t e α ( t s ) β ( s ) δ n ( s ) n n e n d s n n e n α β + ( δ ) n M , for every t R .

Meanwhile, in regard to (2.4) and (3.1), one has

( Γ ϕ ) ( t ) t e α + ( t s ) β ( s ) δ n ( s ) δ n ( s ) ϕ n ( s τ ( s ) ) e δ ( s ) ϕ ( s τ ( s ) ) b ( s ) γ ϕ ( s ) d s t e α + ( t s ) β ( s ) δ n ( s ) κ 1 n e κ 1 b + N M d s 1 α + κ 1 κ 1 n 1 e κ 1 β ( δ + ) n b + N M κ 1 κ 1 , for all t R ,

which, and (3.12), implies that the mapping Γ is a self-mapping from Ω to Ω .

Next we demonstrate that the mapping Γ is contractive on Ω . For every ϕ 1 , ϕ 2 Ω , we obtain

(3.13) ( Γ ϕ 1 ) ( t ) ( Γ ϕ 2 ) ( t ) = sup t R ( Γ ϕ 1 ) ( t ) ( Γ ϕ 2 ) ( t ) = sup t R t e s t α ( u ) d u β ( s ) δ n ( s ) [ δ n ( s ) ϕ 1 n ( s τ ( s ) ) e δ ( s ) ϕ 1 ( s τ ( s ) ) δ n ( s ) ϕ 2 n ( s τ ( s ) ) e δ ( s ) ϕ 2 ( s τ ( s ) ) ] b ( s ) ( γ ϕ 1 ( s ) γ ϕ 2 ( s ) ) d s sup t R t e s t α ( u ) d u β ( s ) δ n ( s ) δ n ( s ) ϕ 1 n ( s τ ( s ) ) e δ ( s ) ϕ 1 ( s τ ( s ) ) δ n ( s ) ϕ 2 n ( s τ ( s ) ) e δ ( s ) ϕ 2 ( s τ ( s ) ) + b ( s ) γ ϕ 1 ( s ) γ ϕ 2 ( s ) d s sup t R t e s t α ( u ) d u β ( s ) δ 1 n ( s ) d s L ( n + n ) ϕ 1 ϕ 2 β + ( δ + ) 1 n α L ( n + n ) + b + N α ϕ 1 ϕ 2 = ρ ϕ 1 ϕ 2 ,

that is,

(3.14) ( Γ ϕ 1 ) ( t ) ( Γ ϕ 2 ) ( t ) ρ ϕ 1 ϕ 2 .

In (3.13), we have used the following fact: Denote L ( u ) = e u u n 1 ( n u ) , it is easy to see that

sup u κ 1 L ( u ) = L ( n + n ) = L ( κ 1 ) , for 0 < n < 1 ,

one has

x n e x y n e y = L ( x + θ ( y x ) ) x y L ( n + n ) x y , where x , y [ κ 1 , ) , 0 < θ < 1 .

Since ρ < 1 in (3.14), it means that Γ is contractive. We then conclude from the Banach fixed point theorem that Γ admits a unique fixed point which is exactly the solution to model (1.2) in Ω AP ( R , R ) . This completes the proof of Theorem 3.2.□

Theorem 3.3

Under the assumptions of Theorem 3.2, then the unique almost periodic solution of model (1.2) is globally exponentially stable.

Proof

Let x ( t ) be any solution of model (1.2) and x * ( t ) be an almost periodic solution of model (1.2), consider the following Lyapunov function:

V ( e ) ( t ) = 1 2 e 2 ( t ) ,

in which e ( t ) = x ( t ) x * ( t ) . In view of the monotonicity of controlling function, one obtains

d d t V ( e ) ( t ) = e ( t ) { α ( t ) e ( t ) + β ( t ) x n ( t τ ( t ) ) e δ ( t ) x ( t τ ( t ) ) β ( t ) ( x * ( t τ ( t ) ) ) n e δ ( t ) x * ( t τ ( t ) ) b ( t ) ( γ ( t ) γ * ( t ) ) } α e 2 ( t ) + β + ( δ + ) 1 n 2 L ( n + n ) ( e 2 ( t ) + e 2 ( t τ ( t ) ) ) α β + ( δ + ) 1 n 2 L ( n + n ) e 2 ( t ) + β + ( δ + ) 1 n L ( n + n ) sup t τ + s t V ( e ) ( s ) = λ V ( e ) ( t ) + ζ sup t τ + s t V ( e ) ( s ) ,

where

λ 2 α β + ( δ + ) 1 n L ( n + n ) , ζ β + ( δ + ) 1 n L ( n + n ) .

By (3.8), one sees λ > ζ > 0 . It follows from Lemma 2.2 that

V ( e ) ( t ) ( sup τ + s 0 V ( e ) ( s ) ) e ψ t ,

for each t 0 , in which ψ is the unique root of ψ = λ ζ e ψ τ + . It is not difficult to obtain that

x ( t ) x * ( t ) sup τ + s 0 x ( s ) x * ( s ) e 1 2 ψ t .

This says that the almost periodic solution x * ( t ) is globally exponentially stable. This completes the proof.□

Corollary 3.4

Suppose that the delay τ ( t ) is a constant and all conditions in Theorem 3.2 hold, then the unique almost periodic solution of model (1.2) is globally exponential stable.

Remark 3.1

It is interesting to note that the periodic or equilibrium dynamics would follow, if the economic parameters are periodic or constant in Theorems 3.1 and 3.2.

Remark 3.2

The exponential stability issue of model (1.2) has been well addressed in [5,12,14], however, one finds that the exponential stability results obtained by the Halanay inequality are different from these references, and the dynamics of discontinuous model (1.2) has seldom been considered up to now, which means that the established theoretical results are novel and extend some existing results.

4 Numerical example

In order to illustrate the validity of the obtained results, a numerical example is now given below.

Example 4.1

Consider the following delayed neoclassical growth model with discontinuous controlling:

(4.1) x ( t ) = ( 0.46 + 0.01 sin 3 t ) x ( t ) + ( 1.4 + 0.03 cos 3 t ) x 0.88 ( t 0.03 ) e ( 0.23 + 0.01 cos 2 t ) x ( t 0.03 ) ( 0.12 + 0.02 sin 2 t ) C ( x ( t ) ) ,

where

(4.2) C ( x ) = 0 , if 0 x < 1 , 1 , if x 1 .

Obviously, n = 0.88 , α + = 0.47 , α = 0.45 , β + = 1.43 , β = 1.37 , δ + = 0.24 , δ = 0.22 , τ = 0.03 , b + = 0.14 , and C C with N = 1 .

A simple calculation shows that κ 1 0.326534 and κ 2 1.436471 . Let M = 5.5276 , it is not difficult to verify that

n n e n M β + ( δ ) n 0.3634 , κ 1 n 1 e κ 1 β ( δ + ) n b + N M κ 1 1.5988 , κ 2 M 0.2599 .

In addition,

ρ = 1 α ( β + ( δ + ) 1 n L ( n + n ) + b + N ) 0.6906 < 1 .

Then, model (4.1) has a unique positive almost periodic solution which is globally exponentially stable. This fact is verified by numerical simulation with different initial values in Figure 1.

Figure 1 
               Exponential stability of almost periodic solution for equation (4.1) with different initial values.
Figure 1

Exponential stability of almost periodic solution for equation (4.1) with different initial values.

Acknowledgement

The authors would like to thank the editor and anonymous reviewers for their valuable comments which have helped them improve this paper greatly.

  1. Funding information: This work was jointly supported by the Anhui Provincial Natural Science Foundation (2208085MA04), University Outstanding Young Talents Support Program of Anhui Province (gxyqZD2022035 and gxyq2021181).

  2. Author contributions: This work was completed in collaboration between the three authors. Q.W. performed the analysis and wrote the main manuscript, W.W. designed the study and guided the research, and Q.Z. prepared the example and numerical simulations. All authors read and approved the final manuscript.

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

References

[1] R. Day, Irregular growth cycles, Am. Econ. Rev. 72 (1982), no. 3, 406–414, https://www.jstor.org/stable/1831540. 10.2105/AJPH.72.4.406Search in Google Scholar PubMed PubMed Central

[2] R. Day, The emergence of chaos from classical economic growth, Q. J. Econ. 98 (1983), no. 2, 203–213, DOI: https://doi.org/10.2307/1885621. 10.2307/1885621Search in Google Scholar

[3] A. Matsumoto and F. Szidarovszky, Asymptotic behavior of a delay differential neoclassical growth model, Sustain. 5 (2013), no. 2, 440–455, DOI: https://doi.org/10.3390/su5020440. 10.3390/su5020440Search in Google Scholar

[4] A. Matsumoto and F. Szidarovszky, Delay differential neoclassical growth model, J. Econ. Behav. Organ. 78 (2011), no. 3, 272–289, DOI: https://doi.org/10.1016/j.jebo.2011.01.014. 10.1016/j.jebo.2011.01.014Search in Google Scholar

[5] W. Chen and W. Wang, Global exponential stability for a delay differential neoclassical growth model, Adv. Difference Equations 2014 (2014), no. 1, 325, DOI: https://doi.org/10.1186/1687-1847-2014-325. 10.1186/1687-1847-2014-325Search in Google Scholar

[6] C. Huang, X. Zhao, J. Cao, and F. E. Alsaadi, Global dynamics of neoclassical growth model with multiple pairs of variable delays, Nonlinearity 33 (2020), 6819, DOI: https://doi.org/10.1088/1361-6544/abab4e. 10.1088/1361-6544/abab4eSearch in Google Scholar

[7] L. Duan and C. Huang, Existence and global attractivity of almost periodic solutions for a delayed differential neoclassical growth model, Math. Methods Appl. Sci. 40 (2017), no. 3, 814–822, DOI: https://doi.org/10.1002/mma.4019. 10.1002/mma.4019Search in Google Scholar

[8] L. Shaikhet, Stability of equilibriums of stochastically perturbed delay differential neoclassical growth model, Discrete Contin. Dyn. Syst. Ser. B 22 (2017), no. 4, 1565–1573, DOI: https://doi.org/10.3934/dcdsb.2017075. 10.3934/dcdsb.2017075Search in Google Scholar

[9] W. Wang and W. Chen, Stochastic delay differential neoclassical growth model, Adv. Difference Equations 2019 (2019), 355, DOI: https://doi.org/10.1186/s13662-019-2292-0. 10.1186/s13662-019-2292-0Search in Google Scholar

[10] Z. Long and W. Wang, Positive pseudo almost periodic solutions for a delayed differential neoclassical growth model, J. Difference Equations Appl. 22 (2016), no. 12, 1893–1905, DOI: https://doi.org/10.1080/10236198.2016.1253688. 10.1080/10236198.2016.1253688Search in Google Scholar

[11] S. Brianzoni, C. Mammana, and E. Michetti, Complex dynamics in the neoclassical growth model with differential savings and non-constant labor force growth, Stud. Nonlinear Dyn. Econom. 11 (2007), no. 3, 3, DOI: https://doi.org/10.2202/1558-3708.1407. 10.2202/1558-3708.1407Search in Google Scholar

[12] Y. Xu, New result on the global attractivity of a delay differential neoclassical growth model, Nonlinear Dynam. 89 (2017), no. 1, 281–288, DOI: https://doi.org/10.1007/s11071-017-3453-x. 10.1007/s11071-017-3453-xSearch in Google Scholar

[13] W. Wang, The exponential convergence for a delay differential neoclassical growth model with variable delay, Nonlinear Dynam. 86 (2016), no. 3, 1875–1883, DOI: https://doi.org/10.1007/s11071-016-3001-0. 10.1007/s11071-016-3001-0Search in Google Scholar

[14] W. Wang and W. Chen, Stochastic delay differential neoclassical growth system, Stoch. Models 37 (2021), no. 2, 415–425, DOI: https://doi.org/10.1080/15326349.2021.1888300. 10.1080/15326349.2021.1888300Search in Google Scholar

[15] L. Duan and F. Di, Exponential attractivity in a delayed almost periodic differential neoclassical growth system, Qual. Theory Dyn. Syst. 18 (2019), 653–665, DOI: https://doi.org/10.1007/s12346-018-0305-0. 10.1007/s12346-018-0305-0Search in Google Scholar

[16] H. Xu, Q. Zhu, and W. Zheng, Exponential stability of stochastic nonlinear delay systems subject to multiple periodic impulses, IEEE Trans. Autom. Control 69 (2024), no. 4, 2621–2628, DOI: https://doi.org/10.1109/TAC.2023.3335005. 10.1109/TAC.2023.3335005Search in Google Scholar

[17] L. Fan, Q. Zhu, and W. Zheng, Stability analysis of switched stochastic nonlinear systems with state-dependent delay, IEEE Trans. Autom. Control 69 (2024), no. 4, 2567–2574, DOI: https://doi.org/10.1109/TAC.2023.3315672. 10.1109/TAC.2023.3315672Search in Google Scholar

[18] A. Filippov, Differential Equations with Discontinuous Right-hand Sides, Kluwer Academic Publishers, Dordrecht, 1988. 10.1007/978-94-015-7793-9Search in Google Scholar

[19] W. Lu and T. Chen, Almost periodic dynamics of a class of delayed neural networks with discontinuous activations, Neural Comput. 20 (2008), no. 4, 1065–1090, DOI: https://doi.org/10.1162/neco.2008.10-06-364. 10.1162/neco.2008.10-06-364Search in Google Scholar PubMed

[20] A. Fink, Almost Periodic Differential Equations, Springer-Verlag, Berlin, 1974. 10.1007/BFb0070324Search in Google Scholar

[21] A. Halanay, Differential Equations: Stability, Oscillations, Time Lags, Academic Press, New York, 1966. Search in Google Scholar

[22] X. Huang and J. Cao, Almost periodic solution of shunting inhibitory cellular neural networks with time-varying delay, Phys. Lett. A 314 (2003), no. 3, 222–231, DOI: https://doi.org/10.1016/S0375-9601(03)00918-6. 10.1016/S0375-9601(03)00918-6Search in Google Scholar

Received: 2023-10-12
Revised: 2024-03-06
Accepted: 2024-03-13
Published Online: 2024-05-23

© 2024 the author(s), published by De Gruyter

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

Downloaded on 24.10.2024 from https://www.degruyter.com/document/doi/10.1515/math-2024-0006/html
Scroll to top button