Next Article in Journal
Correlating Morphology and Multifractal Spatial Patterns of the Leaf Surface Architecture of Anacardium occidentale L.
Next Article in Special Issue
Particle Swarm Optimization Fractional Slope Entropy: A New Time Series Complexity Indicator for Bearing Fault Diagnosis
Previous Article in Journal
Neutrosophic Double Controlled Metric Spaces and Related Results with Application
Previous Article in Special Issue
Hidden and Coexisting Attractors in a Novel 4D Hyperchaotic System with No Equilibrium Point
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Sensitivity of Uniformly Convergent Mapping Sequences in Non-Autonomous Discrete Dynamical Systems

1
College of Mathematics and Statistics, Sichuan University of Science and Engineering, Zigong 643000, China
2
The Key Laboratory of Higher Education of Sichuan Province for Enterprise Informationalization and Internet of Things, Zigong 643000, China
*
Authors to whom correspondence should be addressed.
Fractal Fract. 2022, 6(6), 319; https://doi.org/10.3390/fractalfract6060319
Submission received: 16 April 2022 / Revised: 30 May 2022 / Accepted: 30 May 2022 / Published: 7 June 2022

Abstract

:
Let H be a compact metric space. The metric of H is denoted by d. And let ( H , f 1 , ) be a non-autonomous discrete system where f 1 , = { f n } n = 1 is a mapping sequence. This paper discusses infinite sensitivity, m-sensitivity, and m-cofinitely sensitivity of f 1 , . It is proved that, if f n ( n N ) are feebly open and uniformly converge to f : H H , f i f = f f i for any i { 1 , 2 , } , and i = 1 D ( f i , f ) < , then ( H , f ) has the above sensitive property if and only if ( H , f 1 , ) has the same property where D ( · , · ) is the supremum metric.

1. Introduction

Chaos, as a universal motion form of topological dynamical systems, is one of the core contents of the research for dynamical systems. At present, fruitful results of chaos theory have been obtained in autonomous discrete dynamical systems. However, many complex systems in real life, such as medicine, biology, and physics, are difficult to describe by autonomous systems. Therefore, it is necessary to use other models (for example, non-autonomous discrete systems). Since 1996, chaos of non-autonomous discrete dynamical systems (for convenience, we abbreviate it to NDDS) has began to be studied [1]. In recent years, the discussion about the chaotic properties in NDDS has been active. Si [2] gives some sufficient conditions for NDDS to have asymptotically stable sets. Lan and Peris [3] showed the relation between the weak stability of an NDDS and its induced set-valued system. Li, Zhao, and Wang [4] studied stronger forms of sensitivity and transitivity for NDDS by using the Furstenberg family. Meanwhile, under the condition lim n d ( g m m , g m ) = 0 , a necessary and sufficient condition for g to be F -mixing is established in [5]. Vasisht and Das [6] discussed the difference between F -sensitivity and some other stronger forms of sensitivity by some examples. Salman and Das [7] proved that on a compact metric space, every finitely generated NDDS which is topologically transitive and has a dense set of periodic points is thickly syndetically sensitive. Vasisht and Das [8] proved that if the rate of convergence at which ( f n ) converges to f is “sufficiently fast”, then various forms of sensitivity for the autonomous system ( X , f ) and the NDDS ( X , f 1 , ) coincide. For the chaoticity of other maps in NDDS, see [9,10,11,12] and other literature.
This paper further studies the chaotic properties in the sense of sensitivity. The basic definitions of chaos are given in Section 2. In Section 3, under the conditions of that, f n : H H ( n N ) are feebly open and uniformly converge to f : H H , f i f = f f i for any i { 1 , 2 , } , and i = 1 D ( f i , f ) < . This paper proves that ( H , f ) is Q -sensitive if and only if ( H , f 1 , ) is Q -sensitive where, D ( · , · ) is the supremum metric (see Section 3), Q -sensitive denotes one of the four properties: accessible, infinitely sensitive, m-sensitive, and m-cofinitely sensitive.

2. Preliminaries

For any initial value x 0 H , the orbit of x under f 1 , is denoted by { f n f n 1 f 1 ( x 0 ) : n N } .
A subset K of N is cofinite [4,5] if there exists a N N such that [ N , + ] K .
A system ( H , f 1 , ) (or maps sequence { f n } n N ) is called “feebly open” [4,5] if for any nonempty open subset V of H, int ( f n ( V ) ) ϕ for any n N . Where intA denotes the interior of set A.
A pair ( x , y ) is proximal [13] for ( H , f 1 , ) if for any x H , lim inf n d ( f 1 n ( x ) , f 1 n ( y ) ) = 0 .
Definition 1
([14]). A system ( H , f 1 , ) is "spatio-temporal chaotic" if for any x H and each neighborhood V of x, there is a y V such that lim sup n d ( f 1 n ( a ) , f 1 n ( b ) ) > 0 but lim inf n d ( f 1 n ( a ) , f 1 n ( b ) ) = 0 .
Definition 2
([4,5]). A system ( H , f 1 , ) is called "sensitive dependent on initial condition" if there exists an η > 0 such that for any x H and ε > 0 , there exists a y B ( x , ε ) and an n N such that d ( f 1 n ( x ) , f 1 n ( y ) ) > η .
Definition 3
([7,8]). A system ( H , f 1 , ) is called "infinitely sensitive" if there exists an η > 0 such that, for any x H and ε > 0 , one can find a y B ( x , ε ) and an n N such that lim sup n d ( f 1 n ( x ) , f 1 n ( y ) ) η .
Definition 4
([15]). A system ( H , f 1 , ) is called "accessible" if for any ε > 0 and any two nonempty open subsets U 1 , U 2 H , there are two points x U 1 and y U 2 such that d ( f 1 n ( x ) , f 1 n ( y ) ) ) < ε for some integer n > 0 .
For convenience, write
A ( U , m , n ) = m i n { d ( f 1 n ( x i ) , f 1 n ( y j ) ) : x i , y j U , i , j { 1 , 2 , , m } , i j }
and
S f 1 , , m ( U , λ ) = { n N : there is x i , y j U ( i , j { 1 , 2 , , m } , i j ) such that A ( U , m , n ) λ } ,
where m , n N , U is an arbitrary nonempty open subset in X.
Definition 5
([16]). Given an integer m with m 2 . The system ( H , f 1 , ) is called "m-sensitive", if there is a real number λ > 0 such that for any nonempty open subset U of H, there are 2 m points x 1 , x 2 , , x m ; y 1 , y 2 , , y m U such that S f 1 , , m ( U , λ ) is nonempty.
Definition 6
([16]). Given an integer m with m 2 . The system ( H , f 1 , ) is called "m-cofinitely sensitive", if there is a real number λ > 0 such that for any nonempty open subset U of H, there are 2 m points x 1 , x 2 , , x m ; y 1 , y 2 , , y m U such that S f 1 , , m ( U , λ ) is a cofinite set.

3. The Relation of Chaoticity between f 1 , and Its Limit Map f

Let C ( H ) be the set of all continuous self-maps on ( H , d ) . For any f , g C ( H ) , the supremum metric (see [4]) is defined by D ( f , g ) = sup x H d ( f ( x ) , g ( x ) ) . This section will give equivalence of chaotic properties between ( H , f 1 , ) and ( H , f ) .
Lemma 1
([5]). Let ( H , f 1 , ) be an NDDS on a nontrivial compact metric space ( H , d ) and f C ( H ) . If f i f = f f i for any i { 1 , 2 , } , then for any x H , any integer q 1 and any integer p 1 one has
d ( f 1 q + p ( x ) , f q ( f 1 p ( x ) ) ) j = p + 1 q + p D ( f j , f ) .
Theorem 1.
If f n ( n N ) are a feebly open mapping sequence which uniformly converges to f, f i f = f f i for any i { 1 , 2 , } , and i = 1 D ( f i , f ) < , then ( H , f ) is accessible if and only if ( H , f 1 , ) is accessible.
Proof. 
Suppose that ( H , f ) is accessible. Given ε > 0 , let U , V are two nonempty open subsets in H. Because f i f = f f i for any i { 1 , 2 , } , by Lemma 1, for the above ε > 0 , d ( f 1 p 0 + q ( x ) , f q ( f 1 p 0 ( x ) ) ) < j = p 0 + 1 q + p 0 D ( f j , f ) for any x H and any integer p 0 , q 1 . Moreover, because i = 1 D ( f i , f ) < , then there is an integer S 0 1 such that j = s D ( f j , f ) < 1 3 ε for any s s 0 . Combine with the arbitrariness of p 0 , q , one can get that d ( f 1 p 0 + q ( x ) , f q ( f 1 p 0 ( x ) ) ) < ε 3 . Because f i ( i { 1 , 2 , } ) are feebly open, the interiors of f 1 p 0 ( U ) and f 1 p 0 ( V ) are nonempty sets. Let U , V be the interiors of f 1 p 0 ( U ) and f 1 p 0 ( V ) , respectively.
Because ( H , f ) is accessible, for the above ε > 0 , there are x U and y V such that d ( f q ( x ) , f q ( y ) ) < ε 3 for some q > 0 . Then, there exist x U , y V satisfying x = f 1 p 0 ( x ) , y = f 1 p 0 ( y ) . Thus, d ( f q ( f 1 p 0 ( x ) ) , f q ( f 1 p 0 ( y ) ) ) < ε 3 . Noting that d ( f 1 q + p 0 ( x ) , f q ( f 1 p 0 ( x ) ) ) < ε 3 for x H , by triangle inequality, one has that
d ( f 1 p 0 + q ( x ) , f 1 p 0 + q ( y ) ) d ( f 1 p 0 + q ( x ) , f q ( f 1 p 0 ) ( x ) ) + d ( f q ( f 1 p 0 ) ( x ) , f q ( f 1 p 0 ) ( y ) ) + d ( f 1 p 0 + q ( y ) , f q ( f 1 p 0 ) ( y ) ) ε 3 + ε 3 + ε 3 = ε .
Hence, ( H , f 1 , ) is accessible.
Now, suppose that ( H , f 1 , ) is accessible. For a given ε > 0 , let U , V H be nonempty and open. Because i = 1 D ( f i , f ) < , by Lemma 1, there is an integer p 0 1 such that for the above ε > 0 , x H , d ( f 1 p 0 + q ( x ) , f q ( f 1 p 0 ( x ) ) ) < ε 3 for any integer q 1 . Because f i ( i { 1 , 2 , } ) are feebly open, then the interiors of f 1 p 0 ( U ) and f 1 p 0 ( V ) are nonempty sets. Let U , V be the interiors of f 1 p 0 ( U ) and f 1 p 0 ( V ) , respectively.
Because ( H , f 1 , ) is accessible for the above ε > 0 , there are x U and y V such that d ( f 1 q + p 0 ( x ) , f 1 q + p 0 ( y ) ) < ε 3 for some q > 0 . Then, there exist x U , y V satisfying x = f 1 p 0 ( x ) , y = f 1 p 0 ( y ) . Noted that d ( f 1 q + p 0 ( x ) , f q ( x ) < ε 3 , by triangle inequality,
d ( f q ( x ) , f q ( y ) ) d ( f q ( x ) , f q + p 0 ( x ) ) + d ( f q + p 0 ( x ) , f q + p 0 ( y ) ) + d ( f q ( y ) , f q + p 0 ( y ) ) < ε .
Hence, ( H , f ) is accessible. □
Theorem 2.
If f n ( n N ) is a feebly open mapping sequence which uniformly converges to f, f i f = f f i for any i { 1 , 2 , } , and i = 1 D ( f i , f ) < , then ( H , f ) is infinitely sensitive if and only if ( H , f 1 , ) is infinitely sensitive.
Proof. 
Suppose that ( H , f ) is infinitely sensitive with λ > 0 as an infinitely sensitive constant. Let ε > 0 , U H is a nonempty open set. Because i = 1 D ( f i , f ) < , by Lemma 1, there is an integer p 1 such that d ( f 1 p + q ( x ) , f q ( f 1 p ( x ) ) ) < ε for any integer q 1 , x H and the above ε > 0 . Taking an integer k { 1 , 2 , } satisfying k > 4 λ . Then, there is an integer p 0 1 such that d ( f 1 p 0 + q ( x ) , f q ( f 1 p 0 ( x ) ) ) < 1 k for any integer q 1 and x H . Because f i is feebly open ( i { 1 , 2 , } ), then the interior of f 1 p 0 ( U ) is nonempty. Let U be the interior of f 1 p 0 ( U ) . Because ( H , f ) is infinitely sensitive with infinitely sensitive constant λ > 0 , then there is a y U such that lim sup q d ( f q ( x ) , f q ( y ) ) > λ . Because
x = f 1 p 0 ( x ) , y = f 1 p 0 ( y ) , lim sup q d ( f q ( f 1 p 0 ( x ) ) , f q ( f 1 p 0 ( y ) ) ) > λ ,
and because
d ( f 1 p 0 + q ( x ) , f q ( f 1 p 0 ( x ) ) ) < 1 k   and   d ( f 1 p 0 + q ( y ) , f q ( f 1 p 0 ( y ) ) ) < 1 k
for any integer q 1 . By triangle inequality,
d ( f 1 p 0 + q ( x ) , f 1 p 0 + q ( y ) ) > λ 2 k > 1 2 λ .
Taking the upper limit of both sides of the inequality, one has that
lim sup q d ( f q + p 0 ( x ) , f q + p 0 ( y ) ) > 1 2 λ .
Therefore, ( H , f 1 , ) is infinitely sensitive.
Conversely, let ( H , f 1 , ) be infinitely sensitive with λ > 0 as an infinitely sensitive constant. Let ε > 0 , U H be a nonempty open set. Because i = 1 D ( f i , f ) < , by Lemma 1, there is an integer p 1 such that d ( f 1 p + q ( x ) , f q ( f 1 p ( x ) ) ) < ε for any integer q 1 , x H , and the above ε > 0 . Taking an integer k { 1 , 2 , } satisfying k > 4 λ . Then, there is an integer p 0 1 such that d ( f 1 p 0 + q ( x ) , f q ( f 1 p 0 ( x ) ) ) < 1 k for any integer q 1 and x H . Because f i is feebly open ( i { 1 , 2 , } ), the interior of f 1 p 0 ( U ) is nonempty. Let U be the interior of f 1 p 0 ( U ) . Because ( H , f 1 , ) is infinitely sensitive with λ > 0 as a sensitive constant, then there is a y U such that lim sup q d ( f 1 q + p 0 ( x ) ) , f 1 q + p 0 ( y ) ) ) > λ . So, there exist x , y U such that x = f 1 p 0 ( x ) , y = f 1 p 0 ( y ) . Noted that
d ( f 1 p 0 + q ( x ) , f q ( f 1 p 0 ( x ) ) ) < 1 k   and   d ( f 1 p 0 + q ( y ) , f q ( f 1 p 0 ( y ) ) ) < 1 k
for any integer q 1 , then
d ( f 1 p 0 + q ( x ) , f q ( x ) ) < 1 k   and   d ( f 1 p 0 + q ( y ) , f q ( y ) ) < 1 k
for any integer q 1 . By triangle inequality, one has that
d ( f q ( x ) , f q ( y ) ) > λ 2 k > 1 2 λ .
Taking the upper limit of both sides of the inequality, one has that lim sup q d ( f q ( x ) ) , f q ( y ) ) ) > 1 2 λ . Consequently, ( H , f ) is infinitely sensitive. □
Theorem 3.
If f n ( n N ) is a feebly open mapping sequence which uniformly converges to f, f i f = f f i for any i { 1 , 2 , } , and i = 1 D ( f i , f ) < , then ( H , f ) is m-sensitive if and only if ( H , f 1 , ) is m-sensitive.
Proof. 
Suppose that ( H , f ) is m-sensitive with m-sensitive constant λ > 0 . Let ε > 0 and a open set U H : U ϕ . Because i = 1 D ( f i , f ) < , by Lemma 1, there is an integer p 1 such that d ( f 1 p + q ( x ) , f q ( f 1 p ( x ) ) ) < ε for any integer q 1 , x H , and the above ε > 0 . Taking m { 1 , 2 , } with m > 4 λ . Then, there is an integer p 0 1 such that d ( f 1 p 0 + q ( x ) , f q ( f 1 p 0 ( x ) ) ) < 1 m for any integer q 1 and x H . Because f i is feebly open for all i { 1 , 2 , } , the interior of f 1 p 0 ( U ) is nonempty. Let U be the interior of f 1 p 0 ( U ) . Because ( H , f ) is m-sensitive with m-sensitive constant λ > 0 , there are m points x 1 , x 2 , , x m U and a q N such that
m i n { d ( f q ( x i ) , f q ( x j ) ) : i , j { 1 , 2 , , m } i j } λ .
Because x 1 , x 2 , , x m f 1 p 0 ( U ) , there are x 1 , x 2 , , x m U satisfying x 1 = f 1 p 0 ( x 1 ) , x 2 = f 1 p 0 ( x 2 ) , , x m = f 1 p 0 ( x m ) and
m i n { d ( f q ( f 1 p 0 ( x i ) ) , f q ( f 1 p 0 ( x j ) ) ) : i , j { 1 , 2 , , m } i j } λ .
And because d ( f 1 p 0 + q ( x i ) , f q ( f 1 p 0 ( x i ) ) ) < 1 m for any i = 1 , 2 , , m . By triangle inequality,
m i n { d ( f 1 p 0 + q ( x i ) , f 1 p 0 + q ( x j ) ) : i , j { 1 , 2 , , m } i j } λ 2 m > 1 2 λ .
This implies ( H , f 1 , ) is m-sensitive.
Conversely, let ε > 0 and U H : U ϕ be an open set. Because i = 1 D ( f i , f ) < , by Lemma 1, there is an integer p 1 such that d ( f 1 p + q ( x ) , f q ( f 1 p ( x ) ) ) < ε for any integer q 1 x H , and the above ε > 0 . Taking m { 1 , 2 , } with m > 4 λ . Then, there is an integer p 0 1 such that d ( f 1 p 0 + q ( x ) , f q ( f 1 p 0 ( x ) ) ) < 1 m for any integer q 1 and x H . Because f i is feebly open for all i { 1 , 2 , } , the interior of f 1 p 0 ( U ) is nonempty. Let U be the interior of f 1 p 0 ( U ) . Because ( H , f 1 , ) is m-sensitive with λ > 0 as a sensitive constant, there are m points x 1 , x 2 , , x m U and p 0 > 0 such that m i n { d ( f 1 q ( x i ) , f 1 q ( x j ) ) : i j { 1 , 2 , , m } } > λ for any integer q > 0 . Because x 1 , x 2 , , x m U , then there are x 1 , x 2 , , x m U satisfying x 1 = f 1 p 0 ( x 1 ) , x 2 = f 1 p 0 ( x 2 ) , , x m = f 1 p 0 ( x m ) . And because d ( f 1 p 0 + q ( x i ) , f q ( f 1 p 0 ( x i ) ) ) < 1 m for any i { 1 , 2 , , m } , then d ( f 1 p 0 + q ( x i ) , f q ( x i ) ) < 1 m for any i { 1 , 2 , , m } . By triangle inequality, one has that
m i n { d ( f q ( x i ) , f q ( x j ) ) : i j { 1 , 2 , , m } } > λ 2 m > 1 2 λ .
Hence, ( H , f ) is m-sensitive with 1 2 λ as an m-sensitive constant. □
Theorem 4.
If f n ( n N ) is a feebly open mapping sequence which uniformly converge to f, f i f = f f i for any i { 1 , 2 , } , and i = 1 D ( f i , f ) < , then ( H , f ) is m-cofinitely sensitive if and only if ( H , f 1 , ) is m-cofinitely sensitive.
Proof. 
This proof is similar to that of Theorem 1, and hence is omitted. □
Example 1.
Let H be the compact interval [0,1] and g , h be defined by g ( x ) = x for any x [ 0 , 1 ] and
h ( x ) = 2 x + 1 3 f o r x [ 0 , 1 3 ] 3 x + 2 f o r x [ 1 3 , 2 3 ] x 2 3 f o r x [ 2 3 , 1 ] .
In fact, for any nonempty open subset V of H, int ( h ( V ) ) ϕ . Then h ( x ) is feeble open. It is easy to know that, for any x 1 , x 2 [ 0 , 1 ] : x 1 x 2 (without loss of generality, x 1 < x 2 ), the following conclusions are held.
If x 1 , x 2 [ 0 , 1 3 ] or x 1 , x 2 [ 1 3 , 2 3 ] or x 1 , x 2 [ 2 3 , 1 ] , one can get that
h ( x 1 ) h ( x 2 ) x 1 x 2 .
If x 1 [ 0 , 1 3 ] , x 2 [ 1 3 , 2 3 ] , one has
h ( x 1 ) h ( x 2 ) = 2 x 1 + 1 3 ( 3 x 2 + 2 ) = 2 x 1 + 3 x 2 5 3 > x 1 + 3 2 x 2 5 6 .
If x 1 [ 1 3 , 2 3 ] , x 2 [ 2 3 , 1 ] , one has
h ( x 1 ) h ( x 2 ) = 3 x 1 + 2 ( x 2 2 3 ) = 3 x 1 + x 2 8 3 > x 1 + 1 3 x 2 8 9 .
If x 1 [ 0 , 1 3 ] , x 2 [ 2 3 , 1 ] , one has
h ( x 1 ) h ( x 2 ) = 2 x 1 + 1 3 ( x 2 2 3 ) = 2 x 1 x 2 + 1 > x 1 1 2 x 2 + 1 2 .
Write
Δ 1 = { x 1 x 2 : x 1 , x 2 [ 0 , 1 3 ] } ; Δ 2 = { x 1 x 2 : x 1 , x 2 [ 1 3 , 2 3 ] } ;
Δ 3 = { x 1 x 2 : x 1 , x 2 [ 2 3 , 1 ] } ; Δ 4 = { x 1 + 3 2 x 2 5 6 : x 1 [ 0 , 1 3 ] , x 2 [ 1 3 , 2 3 ] }
Δ 5 = { x 1 + 1 3 x 2 8 9 : x 1 [ 1 3 , 2 3 ] , x 2 [ 2 3 , 1 ] } ; Δ 6 = { x 1 1 2 x 2 + 1 2 : x 1 [ 0 , 1 3 ] , x 2 [ 2 3 , 1 ] } .
Taking δ = i n f ( i = 1 6 Δ i ). Then, for any n N , h n ( x 1 ) h n ( x 2 ) δ . This implies that the map h : [ 0 , 1 ] [ 0 , 1 ] is sensitive-dependent on initial condition. The computer simulation with explanation of chaotic behavior is provided in Figure 1. The red dots and the green dots represent the trajectories of initial value x 1 = 0.3556 and x 2 = 0.3557 iterate for 3000 times, respectively. It can be seen that, after iteration, the orbit of x 1 (or x 2 ) is ergodic and disorder (see red dots or green dots). And with little difference between initial values x 1 and x 2 , there is a big gap between the iterative values after 1995 times (see h 1995 ( x 1 ) = 0.0803 , h 1995 ( x 2 ) = 0.9032 ). This means that h is sensitive-dependent on initial condition.
Then, it can be proved that the system ( H , h ) is infinitely sensitive, m-sensitive, and m-cofinitely sensitive.
Now, let f n ( x ) = g ( x ) ( n = 2 k + 1 , k N ) and f n ( x ) = h ( x ) ( n = 2 k , k N ) . Then the family ( f n ) consists of feebly open mappings converging uniformly to h. Obviously, ( H , f 1 , ) is also infinitely sensitive, m-sensitive, and m-cofinitely sensitive. Thus, the system ( H , f 1 , ) is conform to the assumption of Theorems 1–4.
Example 2.
Defining
p ( x ) = 25 s a w ( x ) + c o s ( x 2 ( 1 x ) ) , x H = R ,
where, s a w ( x ) is the sawtooth function defined by
s a w ( x ) = ( 1 ) m ( x 2 m ) , 2 m 1 x 2 m + 1 , m Z .
One can prove that the map p ( x ) satisfies the definitions of chaos in Section 2. The computer simulation with explanation of chaotic behavior is provided in Figure 2. The red dots and the green dots represent the trajectories of initial value x 1 = 0.3556 and x 2 = 0.3557 iterate for 6000 times, respectively. And with little difference between initial values x 1 and x 2 , there is a big gap between the iterative values after 4123 times (see p n ( x 1 ) = 18.3449 , p n ( x 2 ) = 24.1185 ).
Now, let f n ( x ) = p ( x ) ( n N ) . Then f n ( n N ) are feebly open mappings which uniformly converge to p. Similar to Example 1, ( H , f 1 , ) is infinitely sensitive, m-sensitive, and m-cofinitely sensitive.
Remark 1.
The above discussion tells us that under some conditions, studying the effect of a series of disturbances on the system can be simplified to studying the effect of a single map (i.e., the limit map) on the system.

4. Some Supplements

In NDDS, is there any connection between the chaos in the sense of proximity and sensitivity? The following theorem answers this question in part.
Theorem 5.
Let H be a compact metric space and ( H , f 1 , ) be a proximal non-autonomous system, then ( H , f 1 , ) is spatio-temporal chaotic if and only if ( H , f 1 , ) is sensitive.
Proof. 
(Sufficiency) ( H , f 1 , ) be a proximal system, i.e., for any x , y H , lim inf n d ( f 1 n ( x ) , f 1 n ( y ) ) = 0 . Because ( H , f 1 , ) is sensitive with sensitive constant δ > 0 , then for any x H and any neighborhood U of x, there exist a y U and an n N such that d ( f 1 n ( x ) , f 1 n ( y ) ) > δ .
First, we prove that ( H , f 1 , ) is infinitely sensitive. This is similar to the proof of Theorem 2.1 in Ref. [17].
Given any N N , set D N = { ( x , y ) : ρ ( f 1 n ( x ) , f 1 n ( y ) ) η 4 } for an η > 0 . It is clear that D N is a closed set. And we can claim that int D N = ϕ for any N N . In fact, if there are some N N such that int D N ϕ , then there exist nonempty open sets U , V H such that U × V D N . Thus, for any pair ( x , y ) U × V , ρ ( f 1 n ( x ) , f 1 n ( y ) ) η 4 holds for any n > N . So for arbitrary two points x 1 , x 2 U and any n > N ,
ρ ( f 1 n ( x 1 ) , f 1 n ( x 2 ) ) ρ ( f 1 n ( x 1 ) , f 1 n ( y ) ) + ρ ( f 1 n ( y ) , f 1 n ( x 2 ) ) η 2 .
It is easy to prove that, there exists a nonempty open set U * U such that for any points pair x 1 , x 2 U * and any 0 m N , ρ ( f 1 m ( x 1 ) , f 1 m ( x 2 ) ) η 2 . Hence, for any points pair x 1 , x 2 U * and any n N , ρ ( f 1 m ( x 1 ) , f 1 m ( x 2 ) ) η 2 , which contradicts the sensitivity of ( H , f 1 , ) . So int D N = ϕ for any N N . It follows that set D = N N D N is a first category set in H × H . Then, the set
( H × H ) D = { ( x , y ) : N N , n > N s u c h t h a t ρ ( f 1 n ( x ) , f 1 n ( y ) ) > η 4 }
is residual in X × X .
Assume that ( H , f 1 , ) is not infinitely sensitive, then there exist an x 0 H and a ξ > 0 such that lim sup n ρ ( f 1 n ( x 0 ) , f 1 n ( y ) ) η 16 for any y B ( x 0 , ξ ) . Noting the fact that ( H × H ) D is residual in H × H , it follows that there exists a pair ( y 1 , y 2 ) [ B ( x 0 , ξ ) × B ( x 0 , ξ ) ] [ ( H × H ) D ] . Then for any n N ,
ρ ( f 1 n ( y 1 ) , f 1 n ( y 2 ) ) ρ ( f 1 n ( y 1 ) , f 1 n ( x 0 ) ) + ρ ( f 1 n ( x 0 ) , f 1 n ( y 2 ) ) η 8 .
So,
lim sup n ρ ( f 1 n ( y 1 ) , f 1 n ( y 2 ) ) η 8 ,
which contradicts to ( y 1 , y 2 ) H × H D .
Hence, ( H , f 1 , ) is infinitely sensitive. That is to say, there exists an η * > 0 such that lim sup n ρ ( f 1 n ( x ) , f 1 n ( y ) ) η * . Then, it is easy to get that ( H , f ) is spatio-temporal chaotic.
(Necessity) It is clearly held, and hence is omitted.
The proof is completed. □
Corollary 1.
Let H be a compact metric space and ( H , f ) be a proximal system, then ( H , f ) is spatio-temporal chaotic if and only if ( H , f ) is sensitive.
Remark 2.
In fact, there are some other relationships among chaotic properties in non-autonomous discrete systems. For example, topologically weak mixing implies sensitive, dense δ-chaos implies sensitive, generic δ-chaos implies sensitive, and Li-Yorke sensitive is equivalent to sensitive under the condition that k = 1 n = 1 f 1 n ( { y H : d ( f 1 n ( x ) , y ) < 1 k } ) ¯ = H . These results are in [18,19,20,21].

5. Conclusions

For a mapping sequence f 1 , = ( f n ) n = 1 , this paper gives four hypotheses. That is, (1) f n ( n N ) are feebly open; (2) f n ( n N ) uniformly converge to f; (3) f i f = f f i for any i { 1 , 2 , } ; and (4) i = 1 D ( f i , f ) < . It is proved that, under the conditions of (1)–(4), accessible or sensitivity between f 1 , and its limit map f is coincide. Then, the natural problems rise. Can the above (1)–(4) be reduced? Do other chaotic properties, such as transitive, mixing, or distributional chaos, have similar conclusions? These are topics worth studying in the future.

Author Contributions

Conceptualization, Y.J.; validation, Y.J., X.Y. and T.L.; formal analysis, Y.J. and X.Y.; investigation, Y.J.; writing original draft, Y.J.; writing review and editing, X.Y. and T.L.; supervision, T.L.; funding acquisition, T.L. All authors have read and agreed to the published version of the manuscript.

Funding

This work was funded by the Project of the Department of Science and Technology of Sichuan Provincial (No. 2021ZYD0005), the Opening Project of the Key Laboratory of Higher Education of Sichuan Province for Enterprise Informationalization and Internet of Things (No. 2020WZJ01), the Scientific Research Project of Sichuan University of Science and Engineering (No. 2020RC24), and the Graduate Student Innovation Fund (No. y2021100).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Data will be made available on reasonable request.

Acknowledgments

Many thanks to experts.

Conflicts of Interest

The authors declare no conflict of interest regarding the publication of this paper.

References

  1. Kolyada, S.; Snoha, L. Topological entropy of nonautonomous dynamical systems. Random Comput. Dynam. 1996, 4, 205–233. [Google Scholar]
  2. Si, H. On ω-limit sets of non-autonomous discrete dynamical system. Adv. Fixed Point Theory 2016, 6, 287–294. [Google Scholar]
  3. Lan, Y.; Peris, A. Weak stability of non-autonomous discrete dynamical systems. Topol. Appl. 2018, 1, 53–60. [Google Scholar] [CrossRef]
  4. Li, R.; Zhao, Y.; Wang, H.; Liang, H. Stronger Forms of Transitivity and Sensitivity for Nonautonomous Discrete Dynamical Systems and Furstenberg Families. J. Dyn. Control Syst. 2020, 26, 109–126. [Google Scholar] [CrossRef]
  5. Li, R.; Lu, T.; Chen, G.; Liu, G. Some stronger forms of topological transitivity and sensitivity for a sequence of uniformly convergent continuous maps. J. Math. Anal. Appl. 2020, 124443. [Google Scholar] [CrossRef]
  6. Vasisht, R.; Das, R. A note on F-sensitivity for non-autonomous systems. J. Differ. Equ. Appl. 2019, 25, 1–12. [Google Scholar] [CrossRef]
  7. Salman, M.; Das, R. Multi-sensitivity and other stronger forms of sensitivity in non-autonomous discrete systems. Chaos Soliton. Fract. 2018, 115, 341–348. [Google Scholar] [CrossRef]
  8. Radhika, V.; Das, R. On stronger forms of sensitivity in non-autonomous systems. Taiwan J. Math. 2018, 22, 230–245. [Google Scholar]
  9. Fedeli, A.; Donne, A. A note on the uniform limit of transitive dynamical systems. Bull. Belg. Math. Soc.-Sim. 2009, 16, 59–66. [Google Scholar] [CrossRef]
  10. Salman, M.; Das, R. Multi-transitivity in non-autonomous discrete systems. Topol. Appl. 2020, 278, 107237. [Google Scholar] [CrossRef]
  11. Yang, X.; Lu, T.; Anwar, W. Chaotic properties of a class of coupled mapping lattice induced by fuzzy mapping in non-autonomous discrete systems. Chaos Soliton. Fract. 2021, 148, 110979. [Google Scholar] [CrossRef]
  12. Jiang, Y.; Lu, T.; Pi, J.; Anwar, W. The retentivity of four kinds of shadowing properties in non-autonomous discrete dynamical systems. Entropy 2022, 24, 397. [Google Scholar] [CrossRef] [PubMed]
  13. Murinova, E. Generic chaos in metric space. Acta Univ. Matthiae Belii Ser. Math. 2000, 8, 43–50. [Google Scholar]
  14. Snoha, L. Dense chaos. Comment. Math. Univ. Ca. 1992, 33, 747–752. [Google Scholar]
  15. Wu, X.; Wang, J. A remark on accessibility. Chaos Soliton. Fract. 2016, 91, 115–117. [Google Scholar] [CrossRef]
  16. Zhang, C. Some Studies on M-Sensitivity Dependence in Topological Dynamical Systems. Master’s Thesis, Nanchang University, Nanchang, China, 2016. (In Chinese). [Google Scholar]
  17. Wu, X.; Zhu, P. Dense chaos and densely chaotic operators. Tsukuba J. Math. 2012, 36, 367–375. [Google Scholar] [CrossRef]
  18. He, L.; Yan, X.; Wang, L. Weak-mixing implies sensitive dependence. J. Math. Anal. Appl. 2014, 299, 300–304. [Google Scholar] [CrossRef] [Green Version]
  19. Balibrea, F.; Oprocha, P. Weak mixing and chaos in non-autonomous discrete systems. Appl. Math. Lett. 2012, 25, 1135–1141. [Google Scholar] [CrossRef] [Green Version]
  20. Shao, H.; Shi, Y.; Zhu, H. Relationships among some chaotic properties of non-autonomous discrete dynamical systems. J. Differ. Equ. Appl. 2018, 24, 1055–1064. [Google Scholar] [CrossRef]
  21. Zhu, H.; Shi, Y.; Shao, H. Devaney chaos in non-autonomous discrete systems. Int. J. Bifurcat. Chaos 2016, 26, 1650190. [Google Scholar] [CrossRef] [Green Version]
Figure 1. Chaotic behaviors of h in Example 1 with the initial data x 1 = 0.3556 , x 2 = 0.3557 and n = 3000 .
Figure 1. Chaotic behaviors of h in Example 1 with the initial data x 1 = 0.3556 , x 2 = 0.3557 and n = 3000 .
Fractalfract 06 00319 g001
Figure 2. Chaotic behaviors of p in Example 2 with the initial data x 1 = 0.3556 , x 2 = 0.3557 and n = 6000 .
Figure 2. Chaotic behaviors of p in Example 2 with the initial data x 1 = 0.3556 , x 2 = 0.3557 and n = 6000 .
Fractalfract 06 00319 g002
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Jiang, Y.; Yang, X.; Lu, T. Sensitivity of Uniformly Convergent Mapping Sequences in Non-Autonomous Discrete Dynamical Systems. Fractal Fract. 2022, 6, 319. https://doi.org/10.3390/fractalfract6060319

AMA Style

Jiang Y, Yang X, Lu T. Sensitivity of Uniformly Convergent Mapping Sequences in Non-Autonomous Discrete Dynamical Systems. Fractal and Fractional. 2022; 6(6):319. https://doi.org/10.3390/fractalfract6060319

Chicago/Turabian Style

Jiang, Yongxi, Xiaofang Yang, and Tianxiu Lu. 2022. "Sensitivity of Uniformly Convergent Mapping Sequences in Non-Autonomous Discrete Dynamical Systems" Fractal and Fractional 6, no. 6: 319. https://doi.org/10.3390/fractalfract6060319

Article Metrics

Back to TopTop