网络上时滞相关脉冲非线性时滞耦合系统的积分输入到状态稳定性
The Integral Input-to-State Stability of Networked Delay-Related Impulse Nonlinear Time-Delay Coupling Systems
DOI: 10.12677/pm.2024.145206, PDF, HTML, XML, 下载: 24  浏览: 59 
作者: 杨 弘, 马 楠:上海理工大学理学院,上海
关键词: 耦合系统积分输入到状态稳定性(iISS)图论Coupled Systems Integral Input-to-State Stability (iISS) Graph Theory
摘要: 本文研究了网络上时滞相关脉冲的一般非线性时滞耦合系统的积分输入到状态稳定(iISS)性质。利用图论方法和Lyapunov-Krasovskii方法,在单个顶点系统iISS的Lyapunov函数的基础上,构造了整个网络iISS的Lyapunov函数,并推导出了网络上时滞相关脉冲的一般非线性时滞耦合系统存在iISS的充分条件。这些条件表明,如果每个节点上的连续时间系统都是iISS时,网络上时滞相关脉冲非线性时滞耦合系统在不稳定的脉冲出现的频率不太高的情况下仍能保证iISS的性质。
Abstract: This paper investigates the integral-input-to-state stability (iISS) of general nonlinear delayed impulsive coupled systems on networks with delay-dependent impulses. With the assistance of graph theory and the Lyapunov-Krasovskii method, an iISS Lyapunov function for the total network is constructed based on the iISS Lyapunov functions of individual vertex systems, and sufficient conditions for iISS for general nonlinear delayed impulsive coupled systems on networks are derived. It is demonstrated that, when every continuous vertex system is iISS, the nonlinear delayed impulsive coupled systems on networks can still maintain iISS property provided the destabilizing impulses do not occur too frequently.
文章引用:杨弘, 马楠. 网络上时滞相关脉冲非线性时滞耦合系统的积分输入到状态稳定性[J]. 理论数学, 2024, 14(5): 525-536. https://doi.org/10.12677/pm.2024.145206

1. 引言

在过去的几十年里,CSNs在生物系统 [1] [2] 、神经网络 [3] [4] 、化学系统 [5] 、流行病模型 [6] [7] 等领域广泛应用。特别的,在系统稳定性的动态特性上,相关结果已在 [6] [8] - [24] 及其参考文献中提出。

对于CSNs,假设每个顶点系统都有一个全局稳定的平衡点,那么顶点按照有向图连接时,研究CSNs是否也有一个全局稳定的平衡点就具有重要意义。自从Li和Shuai [17] [18] 利用图论中的基尔霍夫矩阵树定理提出了一种构造CSNs全局Lyapunov函数来解决全局稳定性问题的一般方法以来,这种方法的各种扩展已经应用于各种不同类型的动力系统。例如,Chen和Sun [19] 利用图论构造网络上的时滞耦合系统全局Lyapunov泛函的系统方法,得到了系统的稳定性的一些充分条件;Li和Su [9] [20] 等人研究了同时存在白噪声和色噪声的CSNs的随机稳定性;Su等人 [21] 同样利用这些技术推导出了保证网络上离散时间耦合系统全局稳定的充分条件;Suo等人 [22] 得到了网络上的脉冲耦合系统渐近稳定和指数稳定的一些充分条件。因此,图论方法是研究耦合系统的一个很好的方法。此外,对于CSNs的稳定性,人们还提出了许多方法,如Lyapunov方法 [23] 、线性矩阵不等式方法 [24] 等。

控制系统的ISS和iISS在物理学、生物学、社会科学、神经网络、工程领域以及人工复杂动力系统中有着广泛的应用。输入到状态稳定性ISS的概念由Sontag [25] 首次提出,随后在文献 [26] 中引入了ISS的一个积分变体,称为iISS。由于它们在描述外源干扰对动力系统的影响方面的广泛应用,已吸引了越来越多的研究关注。在Sontag的研究之后,许多学者将ISS概念推广到了许多不同类型的动力系统,如离散时间系统 [27] 、脉冲系统 [28] 、切换系统 [29] 等。而iISS的应用可以在小增益定理 [30] 、跟踪问题 [31] 、干扰衰减 [32] 等方面找到。特别的,脉冲系统的ISS和iISS问题在文献 [33] [34] 中受到了广泛关注,并在文献 [28] 中扩展到具有时滞的混合脉冲和切换系统。在文献 [35] 中,对于带脉冲的切换时滞系统,ISS和iISS的结果得到了进一步改进。随后,Liu和Zhang [36] [37] 的研究将ISS和iISS的结果发展到了脉冲带时滞的一般非线性时滞系统。然而,关于单个控制系统的ISS和iISS的论文虽然有很多,但是关于脉冲带时滞的网络上时滞耦合系统的ISS和iISS问题的研究还很少。

我们研究网络上时滞相关脉冲非线性时滞耦合系统的iISS性质有助于丰富和完善非线性系统的稳定性理论。由于此类系统通常具有复杂的动态行为,因此对其进行深入研究可以提供新的数学工具和理论框架,以更好地理解和预测系统的行为。在实际应用中,了解系统的iISS性质对于设计鲁棒控制器至关重要。这些性质保证了即使在外部扰动或不确定性的影响下,系统仍然能保持稳定运行。对于提高系统的性能和可靠性具有显著意义。此外,随着技术的发展,对于高速、高精度控制系统的需求日益增长。在这种情况下,时滞和非线性效应变得更加显著,因此研究此类系统的ISS和iISS性质对于指导现代控制工程设计具有重要意义,为各种实际问题提供解决方案。

基于上述讨论,我们研究了网络上时滞相关脉冲的一般非线性时滞耦合系统的iISS性质。利用图论方法和Lyapunov-krasovskii方法,在单个顶点系统iISS的Lyapunov函数的基础上,构造了整个网络iISS的Lyapunov函数,并推导出了网络上时滞相关脉冲的一般非线性时滞耦合系统存在iISS的充分条件。这些条件表明,如果每个节点上的连续时间系统都是iISS时,网络上时滞相关脉冲非线性时滞耦合系统在不稳定的脉冲出现的频率不太高的情况下仍能保iISS的性质。

本文的其余部分组织如下,第2节介绍了一些基本定义和符号;第3节给出了主要结果;第4节给出一个例子来证明结果的有效性;第5节给出了结论。

2. 预备知识

+ n 分别表示非负实数和n维欧氏空间, | · | 为n维空间欧氏范数; 为正整数集;上标‘T’表示向量或矩阵的转置; C ( X , Y ) 是从空间 X 到空间 Y 的连续映射族。当 n = 0 , 1 , 2 , 时,令 k = n + 1 n a k = 0 k = n + 1 n a k = 1 。设 P C ( [ a , b ] , n ) 为分段右连续函数 ϕ : [ a , b ] n 的集合,其范数定义为 ϕ = sup s [ a , b ] | ϕ ( s ) | (当 b = 时, [ a , b ] [ a , b ) 代替)。对于任意 v : ,定义 [ v ( t ) ] τ = sup τ s 0 v ( t + s )

如果函数 α : + + 是连续的,严格递增的函数且 α ( 0 ) = 0 则称函数 α 为K类函数。特别的,如果 α K α 是⽆界的,则称 α K 类函数。如果 β ( , t ) 对于任意固定的 t > 0 是K类函数,并且 β ( s , t ) 随着 t 对于每个固定的 s > 0 减少到0,则称函数 β : + × + + 为KL类函数。

下面回顾图论的一些知识。有向图 G = ( L , E ) 包含一个由顶点组成的集合 L 和从初始顶点i到终端顶点j的弧 ( i , j ) 组成的集合E。如果每个弧 ( i , j ) 被赋一个正权重 m i j ,则有向图 G 被加权。当且仅当 G 中存在从顶点j到顶点i的弧时 m i j > 0 ,称 M = ( m i j ) l × l 为权重矩阵。如果对于任何一对不同的顶点,存在从一个顶点到另一个顶点的有向路径,则有向图 G 是强连通的。将具有权重矩阵M的有向图记为 ( G , M ) ( G , M ) 的拉普拉斯矩阵定义为

L = ( i 1 m 1 i m 12 m 1 l m 21 i 2 m 2 i m 2 l m l 1 m l 2 i l m l i ) .

以下引理是图论中有用的结论,将用于证明。通常被称为广义基尔霍夫矩阵树定理。

引理1 [17] :假设 l 2 c i ( G , M ) 的拉普拉斯矩阵的第i个对角元素的余子式,则下列恒等式成立:

i , j l c i m i j T ( x i , x j ) = Q W ( Q ) ( h , k ) E ( F Q ) T k h ( x k , x h ) . (1)

其中 T ( x i , x j ) 是一个任意的连续函数, ( G , M ) 的所有生成无圈图的集合, W ( G ) 是Q的权重, F Q 表示Q的有向圈, E ( F Q ) F Q 中的弧合。特别地,如果 ( G , M ) 是强连接的,那么 c i > 0 1 i l

考虑以下形式的网络上时滞相关脉冲非线性时滞耦合系统:

{ x ˙ i ( t ) = f i ( t , x i t , ω i ( t ) ) + j = 1 l F i j ( x i ( t ) , x j ( t ) ) , i L , t t k , k , Δ x i ( t ) = I i k ( t , x i t , ω i ( t ) ) , i L , t = t k , k , x i t 0 = ϕ i , (2)

其中 L = { 1 , 2 , , l } x i ( t ) n i 是第i个动态节点在时间t时的状态变量; x i t ( s ) = x i ( t + s ) , s [ τ , 0 ] 其中时滞 τ > 0 ω i ( t ) P C ( [ t 0 , ) , m i ) 是输入函数; f i , I i k : × P C ( [ τ , 0 ] , n i ) × m i n i 满足

f i ( t , 0 , 0 ) = I i k ( t , 0 , 0 ) = 0 , t , k ;耦合函数 F i j : n i × n j n i 是连续的,表示第j个分量对第i个分量的影响,当且仅当系统中不存在第j个分量对第i个分量的影响时 F i j = 0 t 0 是初始时间, { t k }

( t 0 , + ) 中严格递增的脉冲时间序列,满足 lim k t k = Δ x ( t ) : = x i ( t + ) x i ( t ) 。这里我们假设 n = i = 1 l n i m = i = 1 l m i n i , m i

定义1 对于给定序列 { t k } ,如果存在函数 β K L γ K ,使得对任意 ϕ P C ( [ τ , 0 ] , n ) 和输入函数 ω P C ( [ t 0 , ) , m ) ,系统(2.2)的解 x ( t , t 0 , ϕ ) 全局存在并满⾜以下条件,

| x ( t , t 0 , ϕ ) | β ( ϕ τ , t t 0 ) + t 0 t γ ( | ω ( s ) | ) d s + t k ( t 0 , t ] γ ( | ω ( t k ) | ) , t t 0

则称其是iISS的。

对于一个局部Lipschitz连续函数 V i C ( + × n i , + ) ,对系统(3.1)的Dini导数定义如下:

D + V i ( t , x i ( t ) ) = V i ( t , x i ( t ) ) t + V i ( t , x i ( t ) ) x i [ f i ( t , x i t , w i ( t ) ) + j = 1 l F i j ( x i ( t ) , x j ( t ) ) ] ,

其中

V i ( t , x i ) x i = ( V i ( t , x i ) x i 1 , V i ( t , x i ) x i 2 , , V i ( t , x i ) x i n i ) , x i n i .

3. 理论结果

在本节中,研究了系统(2)的iISS性,通过图论和顶点Lyapunov函数构造了系统(2)的整个网络iISS的Lyapunov函数,建立了系统(2) iISS的一些充分条件。

定理1假设加权有向图 ( G , M ) 是强连通的。对任意 i , j L ,存在一个Lyapunov函数 V i C ( + × n i , + ) ,函数 χ i K θ i ( s ) P C ( [ τ , 0 ] , + ) ,常数 λ 1 ( i ) < 0 r i > 0 和正常数 μ λ 2 ( i ) δ k ( i ) ρ k ( i ) k i L ,使得对于任意 ϕ = ( ϕ 1 T , ϕ 2 T , , ϕ l T ) T t t 0 ϕ i P C ( [ τ , 0 ] , n i ) 有,

(i) 当 t t k , k 时,

D + V i ( t , x i ( t ) ) [ r i χ i ( | ω i ( t ) | ) + λ 1 ( i ) ] V i ( t , x i ( t ) ) + λ 2 ( i ) τ 0 θ i ( s ) V i ( t + s , x i ( t + s ) ) d s + χ i ( | ω i ( t ) | ) + j = 1 l m i j T i j ( x i ( t ) , x j ( t ) ) ;

(ii) 当 t = t k , k 时,

V i ( t , x i ( t ) + I i k ( t , x i t , ω i ( t ) ) ) δ k ( i ) V i ( t , x i ( t ) ) + ρ k ( i ) τ 0 θ i ( s ) V i ( t + s , x i ( t + s ) ) d s + χ i ( | ω i ( t ) | ) ;

(iii) λ 1 + λ 2 e μ τ τ 0 θ ( s ) d s < μ

其中 λ 1 = max i L { λ 1 ( i ) } λ 2 = max i L { λ 2 ( i ) } θ ( s ) = max i L { θ i ( s ) }

(iv) δ ˜ k : = δ k + ρ k e μ τ τ 0 θ ( s ) d s 1 , k

其中 δ k = max i L { δ k ( i ) } ρ k = max i L { ρ k ( i ) }

(v) 在加权有向图 ( G , M ) 的每个有向圈 C D 上,有

( j , i ) E ( C D ) T i j ( x i , x j ) 0 ;

其中 x ( t ) : = x ( t , t 0 , ϕ ) = ( x 1 T ( t ) , x 2 T ( t ) , , x l T ( t ) ) T 是系统(3.1)初值 x t 0 = ϕ 的解。则对于 t [ t p , t p + 1 ) p = 0 , 1 , 2 , ,有

V ( t , x ( t ) ) : = i = 1 l c i V i ( t , x i ( t ) ) ε ˜ ( t 0 , t ) [ η 0 k = 1 p δ ˜ k e μ ( t t 0 ) + k = 0 p h = k + 1 p δ ˜ h t k t k + 1 t χ ( | ω ( s ) | ) e μ ( t s ) d s + k = 1 p h = k + 1 p δ ˜ h e μ ( t t k ) χ ( | ω ( t k ) | ) ] , (3)

其中 c i 由引理2.1定义, ε ˜ ( t 0 , t ) = exp { r t 0 t χ ( ω ( s ) ) d s } r = max i L { r i }

η 0 = sup s [ τ , 0 ] V ( t 0 + s , ϕ ( s ) ) ,

χ ( r ) = max { sup { i = 1 l c i χ i ( | ω i | ) : | ω | = r , ω n i , ω = ( ω 1 T , ω 2 T , , ω l T ) T , r 0 } , sup { χ i ( | ω i | ) : | ω | = r , ω n i , ω = ( ω 1 T , ω 2 T , , ω l T ) T , r 0 } } .

证明 设 v ( t ) : = V ( t , x ( t ) ) ,由条件(i),(v)和引理1可知

v ˙ ( t ) [ r χ ( | ω ( t ) | ) + λ 1 ] v ( t ) + λ 2 [ v ( t ) ] τ τ 0 θ ( s ) d s + χ ( | ω ( t ) | ) + i = 1 l j = 1 l c i m i j T i j ( x i ( t ) , x j ( t ) ) [ r χ ( | ω ( t ) | ) + λ 1 ] v ( t ) + λ 2 [ v ( t ) ] τ τ 0 θ ( s ) d s + χ ( | ω ( t ) | ) . (4)

首先证明以下不等式

v ( t ) ε ˜ ( t 0 , t ) [ η 0 e μ ( t t 0 ) + t 0 t χ ( | ω ( s ) | ) e μ ( t s ) d s ] , t [ t 0 , t 1 ) . (5)

h ( t ) = { v ( t ) , t [ t 0 τ , t 0 ] , v ( t ) ε ˜ ( t 0 , t ) t 0 t χ ( | ω ( s ) | ) e μ ( t s ) d s , t [ t 0 , t 1 ] .

显然, h ( t ) t = t 0 时是连续的。然后,由(iii),(4)和 η 0 的定义,有

D + h ( t ) [ r χ ( | ω ( t ) | ) + λ 1 ] [ h ( t ) + ε ˜ ( t 0 , t ) t 0 t χ ( | ω ( s ) | ) e μ ( t s ) d s ] + λ 2 τ 0 θ ( s ) d s [ h ( t ) + ε ˜ ( t 0 , t ) t 0 t χ ( | ω ( s ) | ) e μ ( t s ) d s ] τ + χ ( | ω ( t ) | ) ε ˜ ( t 0 , t ) [ r χ ( | ω ( t ) | ) t 0 t χ ( | ω ( s ) | ) e μ ( t s ) d s + χ ( | ω ( t ) | ) μ t 0 t χ ( | ω ( s ) | ) e μ ( t s ) d s ] [ r χ ( | ω ( t ) | ) + λ 1 ] h ( t ) + λ 2 [ h ( t ) ] τ τ 0 θ ( s ) d s + [ λ 1 + λ 2 e μ τ τ 0 θ ( s ) d s + μ ] ε ˜ ( t 0 , t ) t 0 t χ ( | ω ( s ) | ) e μ ( t s ) d s + [ 1 ε ˜ ( t 0 , t ) ] χ ( | ω ( t ) | ) [ r χ ( | ω ( t ) | ) + λ 1 ] h ( t ) + λ 2 [ h ( t ) ] τ τ 0 θ ( s ) d s , t [ t 0 , t 1 ] , (6)

h ( t ) η 0 e μ ( t t 0 ) , t [ t 0 τ , t 0 ] . (7)

对于任意 ε > 0 ,设

h ˜ ( t ) = { ( η 0 + ε ) e μ ( t t 0 ) , t [ t 0 τ , t 0 ] , ( η 0 + ε ) ε ˜ ( t 0 , t ) e μ ( t t 0 ) , t [ t 0 , t 1 ) .

接下来证明

h ( t ) h ˜ ( t ) , t [ t 0 , t 1 ) . (8)

如果不等式(8)不成立,则存在 t * ( t 0 , t 1 ) ,使得

h ( t * ) = h ˜ ( t * ) , D + h ( t * ) D + h ˜ ( t * ) , (9)

h ( t ) h ˜ ( t ) , t [ t 0 τ , t * ] . (10)

由(iii)、(6)、(8)~(10),可以推导出

D + h ( t * ) [ λ 1 + r χ ( | ω ( t ) | ) ] h ( t * ) + λ 2 [ h ( t * ) ] τ τ 0 θ ( s ) d s ( η 0 + ε ) [ r χ ( | ω ( t ) | ) + λ 1 + λ 2 e μ τ τ 0 θ ( s ) d s ] ε ˜ ( t 0 , t ) e μ ( t * t 0 ) < ( η 0 + ε ) [ r χ ( | ω ( t ) | ) μ ] ε ˜ ( t 0 , t ) e μ ( t * t 0 ) = D + h ˜ ( t * ) ,

这与(9)中的不等式矛盾。因此,不等式(8)成立。在(8)中令 ε 0 + ,得到(5)成立。

Z ( t ) = { v ( t ) , t [ t 0 τ , t 0 ] , e μ ( t t 0 ) ( ε ˜ ( t 0 , t ) ) 1 v ( t ) , t [ t 0 , + ) .

接下来证明:

Z ( t ) η 0 k = 1 n δ ˜ k + k = 0 n h = k + 1 n δ ˜ h t k t k + 1 t χ ( | ω ( s ) | ) e μ ( s t 0 ) d s + k = 1 n h = k + 1 n δ ˜ h e μ ( t k t 0 ) χ ( | ω ( t k ) | ) : = H ( t ) , t [ t n , t n + 1 ) . (11)

显然, Z ( t ) η 0 , t [ t 0 τ , t 0 ] 。由(5)式可得

Z ( t ) η 0 + t 0 t χ ( | ω ( s ) | ) e μ ( s t 0 ) d s , t [ t 0 , t 1 ) . (12)

n = 0 时(11)成立。根据(12),可得

Z ( t 1 ) η 0 + t 0 t 1 χ ( | ω ( s ) | ) e μ ( s t 0 ) d s . (13)

s [ τ , 0 )

Z ( t 1 + s ) { η 0 + t 0 t 1 + s χ ( | ω ( z ) | ) e μ ( z t 0 ) d z , t 1 + s [ t 0 , t 1 ) , η 0 , t 1 + s [ t 0 τ , t 0 ] , η 0 + t 0 t 1 χ ( | ω ( z ) | ) e μ ( z t 0 ) d z . (14)

根据条件(ii),得到

v ( t k ) δ k v ( t k ) + ρ k τ 0 θ ( s ) d s [ v ( t k ) ] τ + χ ( | ω ( t k ) | ) . (15)

由(13)、(14)和条件(ii)得到

Z ( t 1 ) e μ ( t 1 t 0 ) ( ε ˜ ( t 0 , t 1 ) ) 1 [ δ 1 v ( t 1 ) + ρ 1 τ 0 θ ( s ) d s [ v ( t 1 ) ] τ + χ ( | ω ( t 1 ) | ) ] δ ˜ 1 [ η 0 + t 0 t 1 χ ( | ω ( z ) | ) e μ ( z t 0 ) d z ] + e μ ( t 1 t 0 ) ( ε ˜ ( t 0 , t 1 ) ) 1 χ ( | ω ( t 1 ) | ) = H ( t 1 ) , (16)

这表明(11)在 t = t 1 时成立。假设(11)对 t [ t 1 , t 2 ) 不成立,则有

t * : = inf { t [ t 1 , t 2 ) : Z ( t ) > H ( t ) } ( t 1 , t 2 ) .

此外,有 Z ( t 1 * ) = H ( t 1 * ) , Z ( t ) H ( t ) , t [ t 1 , t 1 * ] D + Z ( t 1 * ) > D + H ( t 1 * )

由(4)可得

D + Z ( t 1 * ) μ Z ( t 1 * ) + λ 1 Z ( t 1 * ) + λ 2 e μ τ [ Z ( t 1 * ) ] τ τ 0 θ ( s ) d s + e μ ( t 1 * t 0 ) ( ε ˜ ( t 0 , t 1 ) ) 1 χ ( | ω ( t 1 * ) | ) [ μ + λ 1 + λ 2 e μ τ τ 0 θ ( s ) d s ] H ( t 1 * ) + e μ ( t 1 * t 0 ) χ ( | ω ( t 1 * ) | ) < e μ ( t 1 * t 0 ) χ ( | ω ( t 1 * ) | ) = D + H ( t 1 * ) .

D + Z ( t 1 * ) > D + H ( t 1 * ) 相矛盾。因此, t [ t 1 , t 2 ) 时,(11)成立。假设当 n 1 时,(11)在 t [ t 0 , t n ) 上成立,则证明(11)在 [ t n , t n + 1 ) 上成立。由(11)在 t = t n 时,有

Z ( t n ) η 0 k = 1 n 1 δ ˜ k + k = 0 n 1 h = k + 1 n 1 δ ˜ h t k t k + 1 χ ( | ω ( s ) | ) e μ ( s t 0 ) d s + k = 1 n 1 h = k + 1 n 1 δ ˜ h e μ ( t k t 0 ) χ ( | ω ( t k ) | ) = H ( t n 1 ) . (17)

不失一般性,假设 s [ τ , 0 ] , t n + s t 0 ,则对于任意固定的 s [ τ , 0 ] ,存在一个整数 m [ 0 , n 1 ] ,使得 t n + s [ t m , t m + 1 ) 。根据条件(iv),和(11)在 t = t n + s 时,得到

Z ( t n + s ) η 0 k = 1 m δ ˜ k + k = 0 m h = k + 1 m δ ˜ h t k t k + 1 ( t n + s ) χ ( | ω ( s ) | ) e μ ( s t 0 ) d s + k = 1 m h = k + 1 m δ ˜ h e μ ( t k t 0 ) χ ( | ω ( t k ) | ) η 0 k = 1 n 1 δ ˜ k + k = 0 n 1 h = k + 1 n 1 δ ˜ h t k t k + 1 χ ( | ω ( s ) | ) e μ ( s t 0 ) d s + k = 1 n 1 h = k + 1 n 1 δ ˜ h e μ ( t k t 0 ) χ ( | ω ( t k ) | ) = H ( t n 1 ) , s [ τ , 0 ) . (18)

因此,由不等式(15)、(17)、(18),我们得到

Z ( t n ) e μ ( t n t 0 ) ( ε ˜ ( t 0 , t n ) 1 ) [ δ n v ( t n ) + ρ n [ v ( t n ) ] τ τ 0 θ ( s ) d s + χ ( | ω ( t n ) | ) ] δ n Z ( t n ) + ρ n e μ τ [ Z ( t n ) ] τ τ 0 θ ( s ) d s + e μ ( t n t 0 ) χ ( | ω ( t n ) | ) δ ˜ n H ( t n 1 ) + e μ ( t n t 0 ) χ ( | ω ( t n ) | ) = η 0 k = 1 n δ ˜ k + k = 0 n 1 h = k + 1 n δ ˜ h t k t k + 1 χ ( | ω ( s ) | ) e μ ( s t 0 ) d s + k = 1 n 1 h = k + 1 n δ ˜ h e μ ( t k t 0 ) χ ( | ω ( t k ) | ) + e μ ( t n t 0 ) χ ( | ω ( t n ) | ) = H ( t n ) , (19)

这意味着(11)在 t = t n 时成立。假设(11)在 t [ t n , t n + 1 ) 上不成立,则

t n * : = inf { t [ t n , t n + 1 ) : Z ( t ) > H ( t ) } ( t n , t n + 1 ) .

此外,可得 Z ( t n * ) = H ( t n * ) Z ( t ) H ( t ) , t [ t n , t n * ]

D + Z ( t n * ) D + H ( t n * ) . (20)

注意到

D + Z ( t n * ) μ Z ( t n * ) + λ 1 Z ( t n * ) + λ 2 e μ τ [ Z ( t n * ) ] τ τ 0 θ ( s ) d s + e μ ( t n * t 0 ) ( ε ˜ ( t 0 , t n ) ) 1 χ ( | ω ( t n * ) | ) [ μ + λ 1 + λ 2 e μ τ τ 0 θ ( s ) d s ] H ( t n * ) + e μ ( t n * t 0 ) χ ( | ω ( t n * ) | ) < e μ ( t n * t 0 ) χ ( | ω ( t n * ) | ) = D + H ( t n * ) ,

这与(20)矛盾。因此,当 t [ t n , t n + 1 ) , n = 0 , 1 , 2 , 时,(11)成立。因此,通过在(11)式两边除 e μ ( t t 0 ) 可得(3)成立。这就完成了证明。

接下来,建立了系统(2)在 l ( ε 1 , ε 2 ) 型时间序列(即 ε 1 t k + 1 t k ε 2 , k ε 1 , ε 2 > 0 )上iISS的结果。

定理2假设定理1中的条件都满足,若存在正常数 d ( 0 , μ ) α 1 ( i ) , α 2 ( i ) > 0 p 2 ,使得

(vi)

(vii) ln δ ˜ k d ( t k t k 1 ) , k

则系统(2)在 l ( ε 1 , ε 2 ) 上是iISS的。

证明 设 v ( t ) : = V ( t , x ( t ) ) 。由定理1和条件(vii),对任意 t [ t n , t n + 1 )

v ( t ) ε ˜ ( t 0 , t ) [ η 0 e d ( t n t 0 ) e μ ( t t 0 ) + k = 0 n e d ( t n t k ) t k t k + 1 t χ ( | ω ( s ) | ) e μ ( t s ) d s + k = 1 n e d ( t n t k ) e μ ( t t k ) χ ( | ω ( t k ) | ) ] ε ˜ ( t 0 , t ) [ η 0 e ( μ d ) ( t t 0 ) + k = 0 n e ( μ d ) ( t t k ) t k t k + 1 t χ ( | ω ( s ) | ) e μ ( s t k ) d s + k = 1 n e ( μ d ) ( t t k ) χ ( | ω ( t k ) | ) ] ε ˜ ( t 0 , t ) [ η 0 e ( μ d ) ( t t 0 ) + e μ ε 2 t 0 t χ ( | ω ( s ) | ) d s + k = 1 n χ ( | ω ( t k ) | ) ] , (21)

为了证明(2)的iISS性,注意到

ε ˜ ( t 0 , t ) η 0 = η 0 + [ ε ˜ ( t 0 , t ) 1 ] η 0 η 0 + 1 2 η 0 2 + 1 2 [ ε ˜ ( t 0 , t ) 1 ] 2 (22)

ε ˜ ( t 0 , t ) k = 1 n χ ( | ω ( t k ) | ) 1 2 ε ˜ 2 ( t 0 , t ) + 1 2 [ k = 1 n χ ( | ω ( t k ) | ) ] 2 1 2 ε ˜ 2 ( t 0 , t ) + n 2 k = 1 n χ 2 ( | ω ( t k ) | ) (23)

定义 γ 1 ( s ) = s + 1 2 s 2 γ 2 ( s ) = 1 2 ( e r s 1 ) 2 + s e μ ε 2 + r s + 1 2 e 2 r s χ ˜ ( s ) = n 2 χ 2 ( s ) s 0 。很明显, γ 1 γ 2 都是 K 类, γ 2 是一个凸函数。将(22)和(23)应用于(21),得到

v ( t ) γ 1 ( η 0 ) e ( μ d ) ( t t 0 ) + t 0 t γ 2 ( χ ( | ω ( s ) | ) ) d s + k = 1 n χ ˜ ( | ω ( t k ) | ) , t [ t n , t n + 1 ) (24)

由条件(vi)易知

α 1 | x ( t ) | p v ( t ) α 2 | x ( t ) | p , (25)

其中 α 1 = ( min i L { c i α 1 ( i ) } ) p 2 ( i = 1 l c i α 1 ( i ) ) 1 p 2 α 2 = i = 1 l c i α 2 ( i )

因此,根据条件(v) (vii)和(24),可得系统(2)在 l ( ε 1 , ε 2 ) 上是iISS的。

4. 实例分析

在本节中,将通过讨论网络上的非线性时滞脉冲耦合系统的iISS性来证明我们结果的有效性。

例4.1考虑以下双线性系统:

{ x ˙ i ( t ) = A i x i ( t ) + m = 1 q ω i m ( t ) [ B i m x i ( t ) + C i m τ 0 e s x i ( t + s ) d s ] + D i ω i ( t ) j = 1 l β i j ( x i ( t ) x j ( t ) ) , t t k , k , i L , Δ x i ( t ) = E k x i ( t ) + F k τ 0 e s x i ( t + s ) d s + G k ω i ( t ) , t = t k , k , i L , (26)

其中设 x i ω i = [ t 2 e 2 t ] D i = [ 1 2 1 2 ] i = 1 , 2 A 1 = 2 A 2 = 3 B 11 = 1 2 B 12 = 1 4 B 21 = 1 6 B 22 = 1 8 C 11 = 1 4 C 12 = 1 5 C 21 = 1 7 C 22 = 1 9 ( β i j ) 2 × 2 = [ 0 1 1 0 ] E k = 1 4 F k = 1 5 G k = [ 1 3 1 3 ] τ = 1 t k t k 1 = 1.4 ,满足定义1和定理2的所有条件,因此系统(26)是iISS的(见图1)。

Figure 1. Simulation results of system (26)

图1. 系统(26)的仿真结果

5. 结论

本文聚焦于网络上时滞相关脉冲非线性时滞耦合系统的iISS问题。借助图论方法和Lyapunov-krasovskii方法的强大工具,在单个顶点系统iISS的Lyapunov函数的基础上,进一步构造了整个网络iISS的Lyapunov函数。通过深入分析和推导,我们成功地得出了网络上一般非线性时滞脉冲耦合系统存在iISS的充分条件。这些条件表明,如果每个节点上的连续时间系统都是iISS时,只要网络上不稳定的脉冲出现的频率不太高的情况下,整个网络上的非线性时滞脉冲耦合系统仍能保iISS的性质。

参考文献

[1] Strogatz, S.H. (2001) Exploring Complex Networks. Nature, 410, 268-276.
https://doi.org/10.1038/35065725
[2] Grillner, S. (2006) Biological Pattern Generation: The Cellular and Computational Logic of Networks in Motion. Neuron, 52, 751-766.
https://doi.org/10.1016/j.neuron.2006.11.008
[3] Chandrasekar, A., Rakkiyappan, R. and Cao, J. (2015) Impulsive Synchronization of Markovian Jumping Randomly Coupled Neural Networks with Partly Unknown Transition Probabilities via Multiple Integral Approach. Neural Networks, 70, 27-38.
https://doi.org/10.1016/j.neunet.2015.07.002
[4] Li, W., Pang, L., Su, H., et al. (2012) Global Stability for Discrete Cohen-Grossberg Neural Networks with Finite and Infinite Delays. Applied Mathematics Letters, 25, 2246-2251.
https://doi.org/10.1016/j.aml.2012.06.011
[5] Wayman, J.A. and Varner, J.D. (2013) Biological Systems Modeling of Metabolic and Signaling Networks. Current Opinion in Chemical Engineering, 2, 365-372.
https://doi.org/10.1016/j.coche.2013.09.001
[6] Sun, R. (2010) Global Stability of the Endemic Equilibrium of Multigroup SIR Models with Nonlinear Incidence. Computers & Mathematics with Applications, 60, 2286-2291.
https://doi.org/10.1016/j.camwa.2010.08.020
[7] Kuniya, T. (2011) Global Stability Analysis with a Discretization Approach for an Age-Structured Multigroup SIR Epidemic Model. Nonlinear Analysis: Real World Applications, 12, 2640-2655.
https://doi.org/10.1016/j.nonrwa.2011.03.011
[8] Sun, R. and Shi, J. (2011) Global Stability of Multigroup Epidemic Model with Group Mixing and Nonlinear Incidence Rates. Applied Mathematics and Computation, 218, 280-286.
https://doi.org/10.1016/j.amc.2011.05.056
[9] Li, W., Su, H., Wei, D., et al. (2012) Global Stability of Coupled Nonlinear Systems with Markovian Switching. Communications in Nonlinear Science and Numerical Simulation, 17, 2609-2616.
https://doi.org/10.1016/j.cnsns.2011.09.039
[10] Zhang, X., Li, W. and Wang, K. (2015) The Existence of Periodic Solutions for Coupled Systems on Networks with Time Delays. Neurocomputing, 152, 287-293.
https://doi.org/10.1016/j.neucom.2014.10.067
[11] Li, C., Sun, J. and Sun, R. (2010) Stability Analysis of a Class of Stochastic Differential Delay Equations with Nonlinear Impulsive Effects. Journal of the Franklin Institute, 347, 1186-1198.
https://doi.org/10.1016/j.jfranklin.2010.04.017
[12] Wu, Q., Zhou, J. and Xiang, L. (2011) Impulses-Induced Exponential Stability in Recurrent Delayed Neural Networks. Neurocomputing, 74, 3204-3211.
https://doi.org/10.1016/j.neucom.2011.05.001
[13] Sontag, E.D. and Wang, Y. (1995) On Characterizations of the Input-to-State Stability Property. Systems & Control Letters, 24, 351-359.
https://doi.org/10.1016/0167-6911(94)00050-6
[14] Grune, L. (2002) Input-to-State Dynamical Stability and Its Lyapunov Function Characterization. IEEE Transactions on Automatic Control, 47, 1499-1504.
https://doi.org/10.1109/TAC.2002.802761
[15] Angeli, D. (2002) A Lyapunov Approach to Incremental Stability Properties. IEEE Transactions on Automatic Control, 47, 410-421.
https://doi.org/10.1109/9.989067
[16] Arcak, M. and Teel, A. (2002) Input-to-State Stability for a Class of Lurie Systems. Automatica, 38, 1945-1949.
https://doi.org/10.1016/S0005-1098(02)00100-0
[17] Li, M.Y. and Shuai, Z. (2010) Global-Stability Problem for Coupled Systems of Differential Equations on Networks. Journal of Differential Equations, 248, 1-20.
https://doi.org/10.1016/j.jde.2009.09.003
[18] Guo, H., Li, M.Y. and Shuai, Z. (2008) A Graph-Theoretic Approach to the Method of Global Lyapunov Functions. Proceedings of the American Mathematical Society, 136, 2793-2802.
https://doi.org/10.1090/S0002-9939-08-09341-6
[19] Chen, H. and Sun, J. (2012) Stability Analysis for Coupled Systems with Time Delay on Networks. Physica A: Statistical Mechanics and its Applications, 391, 528-534.
https://doi.org/10.1016/j.physa.2011.08.037
[20] Li, W., Su, H. and Wang, K. (2011) Global Stability Analysis for Stochastic Coupled Systems on Networks. Automatica, 47, 215-220.
https://doi.org/10.1016/j.automatica.2010.10.041
[21] Su, H., Li, W. and Wang, K. (2012) Global Stability Analysis of Discrete-Time Coupled Systems on Networks and Its Applications. Chaos: An Interdisciplinary Journal of Nonlinear Science, 22, Article 033135.
https://doi.org/10.1063/1.4748851
[22] Suo, J., Sun, J. and Zhang, Y. (2013) Stability Analysis for Impulsive Coupled Systems on Networks. Neurocomputing, 99, 172-177.
https://doi.org/10.1016/j.neucom.2012.06.002
[23] Cao, J., Li, P. and Wang, W. (2006) Global Synchronization in Arrays of Delayed Neural Networks with Constant and Delayed Coupling. Physics Letters A, 353, 318-325.
https://doi.org/10.1016/j.physleta.2005.12.092
[24] Amato, F., Ambrosino, R., Ariola, M., et al. (2009) Finite-Time Stability of Linear Time-Varying Systems with Jumps. Automatica, 45, 1354-1358.
https://doi.org/10.1016/j.automatica.2008.12.016
[25] Sontag, E.D. (1989) Smooth Stabilization Implies Coprime Factorization. IEEE Transactions on Automatic Control, 34, 435-443.
https://doi.org/10.1109/9.28018
[26] Sontag, E.D. (1998) Comments on Integral Variants of ISS. Systems & Control Letters, 34, 93-100.
https://doi.org/10.1016/S0167-6911(98)00003-6
[27] Lazar, M., Heemels, W.P.M.H. and Teel, A.R. (2009) Lyapunov Functions, Stability and Input-to-State Stability Subtleties for Discrete-Time Discontinuous Systems. IEEE Transactions on Automatic Control, 54, 2421-2425.
https://doi.org/10.1109/TAC.2009.2029297
[28] Liu, J., Liu, X. and Xie, W.C. (2011) Input-to-State Stability of Impulsive and Switching Hybrid Systems with Time-Delay. Automatica, 47, 899-908.
https://doi.org/10.1016/j.automatica.2011.01.061
[29] Vu, L., Chatterjee, D. and Liberzon, D. (2007) Input-to-State Stability of Switched Systems and Switching Adaptive Control. Automatica, 43, 639-646.
https://doi.org/10.1016/j.automatica.2006.10.007
[30] Ito, H. and Kellett, C.M. (2018) A Small-Gain Theorem in the Absence of Strong iISS. IEEE Transactions on Automatic Control, 64, 3897-3904.
https://doi.org/10.1109/TAC.2018.2886955
[31] Angeli, D., Sontag, E.D. and Wang, Y. (2000) A Characterization of Integral Input-to-State Stability. IEEE Transactions on Automatic Control, 45, 1082-1097.
https://doi.org/10.1109/9.863594
[32] Liberzon, D. (1999) ISS and Integral-ISS Disturbance Attenuation with Bounded Controls. Proceedings of the 38th IEEE Conference on Decision and Control, Phoenix, 7-10 December 1999, 2501-2506.
https://doi.org/10.1109/CDC.1999.831303
[33] Hespanha, J.P., Liberzon, D. and Teel, A.R. (2008) Lyapunov Conditions for Input-to-State Stability of Impulsive Systems. Automatica, 44, 2735-2744.
https://doi.org/10.1016/j.automatica.2008.03.021
[34] Chen, W.H. and Zheng, W.X. (2009) Input-to-State Stability and Integral Input-to-State Stability of Nonlinear Impulsive Systems with Delays. Automatica, 45, 1481-1488.
https://doi.org/10.1016/j.automatica.2009.02.005
[35] Sun, X.M. and Wang, W. (2012) Integral Input-to-State Stability for Hybrid Delayed Systems with Unstable Continuous Dynamics. Automatica, 48, 2359-2364.
https://doi.org/10.1016/j.automatica.2012.06.056
[36] Liu, X. and Zhang, K. (2019) Input-to-State Stability of Time-Delay Systems with Delay-Dependent Impulses. IEEE Transactions on Automatic Control, 65, 1676-1682.
https://doi.org/10.1109/TAC.2019.2930239
[37] Zhang, K. (2020) Integral Input-to-State Stability of Nonlinear Time-Delay Systems with Delay-Dependent Impulse Effects. Nonlinear Analysis: Hybrid Systems, 37, Article 100907.
https://doi.org/10.1016/j.nahs.2020.100907