
It is interesting and challenging to study chaotic phenomena in partial differential equations. In this paper, we mainly study the problems for oscillations governed by 1D wave equation with general nonlinear feedback control law and energy-conserving or energy-injecting effects at the boundaries. We show that i) energy-injecting effect at the boundary is the necessary condition for the onset of chaos when the nonlinear feedback law is an odd function; ii) chaos never occurs if the nonlinear feedback law is an even function; iii) when one of the two ends is fixed, only the effect of self-regulation at the other end can still cause the onset of chaos; whereas if one of the two ends is free, there will never be chaos for any feedback control law at the other end. In addition, we give a sufficient condition about the general feedback law at one of two ends to ensure the occurrence of chaos. Numerical simulations are provided to demonstrate the effectiveness of the theoretical outcomes.
Citation: Liangliang Li. Chaotic oscillations of 1D wave equation with or without energy-injections[J]. Electronic Research Archive, 2022, 30(7): 2600-2617. doi: 10.3934/era.2022133
[1] | Xiaoqiang Dai, Chao Yang, Shaobin Huang, Tao Yu, Yuanran Zhu . Finite time blow-up for a wave equation with dynamic boundary condition at critical and high energy levels in control systems. Electronic Research Archive, 2020, 28(1): 91-102. doi: 10.3934/era.2020006 |
[2] | Min Li . Long-wavelength limit for the Green–Naghdi equations. Electronic Research Archive, 2022, 30(7): 2700-2718. doi: 10.3934/era.2022138 |
[3] | Shasha Bian, Yitong Pei, Boling Guo . Numerical simulation of a generalized nonlinear derivative Schrödinger equation. Electronic Research Archive, 2022, 30(8): 3130-3152. doi: 10.3934/era.2022159 |
[4] | Xuerong Shi, Zuolei Wang, Lizhou Zhuang . Spatiotemporal pattern in a neural network with non-smooth memristor. Electronic Research Archive, 2022, 30(2): 715-731. doi: 10.3934/era.2022038 |
[5] | Yang Yang, Yun-Rui Yang, Xin-Jun Jiao . Traveling waves for a nonlocal dispersal SIR model equipped delay and generalized incidence. Electronic Research Archive, 2020, 28(1): 1-13. doi: 10.3934/era.2020001 |
[6] | Hyungyeong Jung, Sunghwan Moon . Reconstruction of the initial function from the solution of the fractional wave equation measured in two geometric settings. Electronic Research Archive, 2022, 30(12): 4436-4446. doi: 10.3934/era.2022225 |
[7] | José Luis Díaz Palencia, Abraham Otero . Oscillatory solutions and smoothing of a higher-order p-Laplacian operator. Electronic Research Archive, 2022, 30(9): 3527-3547. doi: 10.3934/era.2022180 |
[8] | Yang Liu, Wenke Li . A family of potential wells for a wave equation. Electronic Research Archive, 2020, 28(2): 807-820. doi: 10.3934/era.2020041 |
[9] | Yi Wei . The Riccati-Bernoulli subsidiary ordinary differential equation method to the coupled Higgs field equation. Electronic Research Archive, 2023, 31(11): 6790-6802. doi: 10.3934/era.2023342 |
[10] | Mohammad Kafini, Maher Noor . Delayed wave equation with logarithmic variable-exponent nonlinearity. Electronic Research Archive, 2023, 31(5): 2974-2993. doi: 10.3934/era.2023150 |
It is interesting and challenging to study chaotic phenomena in partial differential equations. In this paper, we mainly study the problems for oscillations governed by 1D wave equation with general nonlinear feedback control law and energy-conserving or energy-injecting effects at the boundaries. We show that i) energy-injecting effect at the boundary is the necessary condition for the onset of chaos when the nonlinear feedback law is an odd function; ii) chaos never occurs if the nonlinear feedback law is an even function; iii) when one of the two ends is fixed, only the effect of self-regulation at the other end can still cause the onset of chaos; whereas if one of the two ends is free, there will never be chaos for any feedback control law at the other end. In addition, we give a sufficient condition about the general feedback law at one of two ends to ensure the occurrence of chaos. Numerical simulations are provided to demonstrate the effectiveness of the theoretical outcomes.
First, I take this opportunity to express my great admiration toward Professor Goong Chen by dedicating this paper to him on the occasion of his 70th birthday.
In this paper, we mainly consider the problem of vibrations governed by 1D wave equation wtt−c2wxx=0, where c denotes wave propagation speed, associated with a generalized boundary condition. Recall the definition of chaos for this kind of system, which is firstly introduced in [1], as below:
Definition 1.1. Consider an initial-boundary problem (S) governed by 1D wave equation wtt−c2wxx=0 defined on a segment I, where c>0 denotes the propagation speed of the wave. The system is said to be chaotic if there exists a large class of initial data (w0,w1) such that
(i) |wx|+|wt| is uniformly bounded,
(ii) V(t)def=VI(wx(⋅,t))+VI(wt(⋅,t))<+∞ for all t≥0,
(iii) lim inft→+∞lnV(t)t>0.
Remark 1.1. When chaos occurs, the function x↦(|wx|+|wt|)(x,t) is uniformly bounded, whereas the length of the curve {(x,(|wx|+|wt|)(x,t)),x∈[0,1]} grows exponentially w.r.t. time t. Therefore, the system (S) must undergo extremely complex oscillations as time t increasing.
Remark 1.2. There are lots of work about chaos studies, see e.g., [2,3,4] and references therein. As we know, there is no a common mathematical definition for chaos, which is actually a challenge to give. However, Li-Yorke chaos is probably one of the most popular and acceptable notions of chaos. We will study the relationship between Li-Yorke chaos and chaos in the sense of Definition 1.1 in our future work. Particularly, if chaos happens and the solution (wx,wt) can be represented by two interval maps, denoted by K1 and K2, respectively, then K1 and K2 have positive entropy, which implies K1 and K2 are chaotic in the sense of Li-Yorke.
Let us take a classical model to introduce the research background and our motivation, as below:
{wtt−wxx=0, x∈(0,1), t>0,wx(0,t)=−ηwt(0,t), η≠1, t>0,wx(1,t)=αwt(1,t)−βw3t(1,t), 0<α<1, β≥0, t>0,w(x,0)=w0(x),wt(x,0)=w1(x),0≤x≤1, | (1.1) |
where α,β, and η are given constants. If η=1, the system (1.1) is not well-posed. Thus throughout this paper we assume η≠1. The wave equation itself is linear and represents the infinite-dimensional harmonic oscillator. Let
E(t)=12∫10|∇w(x,t)|2dx=12∫10[w2x(x,t)+w2t(x,t)]dx, |
be the energy function of this system. And assume that (1.1) admits a C2 solution, then the boundary conditions show
ddtE(t)=ηw2t(0,t)+w2t(1,t)[α−βw2t(1,t)]. |
The right-handed side boundary condition (at x=1) is nonlinear when β≠0, which is usually called a van der Pol type boundary condition (see, e.g., [1,5,6,7,8,9]). The left-handed side boundary condition (at x=0) is linear, where η>0 indicates that energy is being injected into the system at x=0. Thus if η>0, the system (1.1) has a self-excited mechanism that supplies energy to the system itself, which induces irregular vibrations [1,6]. In particular, when η=0, the free end at x=0 has no effect to the energy, which is also called an energy-conserving boundary condition.
The existence and uniqueness of the classical solution of (1.1) can be found in [5,6]. Furthermore, the system (1.1) has a smooth solution w∈C2 if the initial data satisfy
w0∈C20([0,1]),w1∈C10([0,1]), | (1.2) |
where
Ck0([0,1])={f∈Ck([0,1]) | f(i)(0)=f(i)(1)=0, 0≤i≤k }, k∈N, | (1.3) |
see Theorem 6.1 in [6]. The weak solution as well as its numerical approximation are discussed in [9].
The PDE system (1.1) has received considerable attention since it exhibits many interesting and complicated dynamical phenomena, such as limit cycles and chaotic behavior of (wt,wx) when the parameters α,β and η assume certain values [1,6]. Different from dynamics of a system of ODEs, this is a simple and useful infinite-dimensional model for the study of spatiotemporal behaviors as time evolutes. For instance, the propagation of acoustic waves in a pipe satisfies the linear wave equation: wtt−wxx=0. As we know, the solution of 1D wave equation describes a superposition of two traveling wave with arbitrary profiles, one propagating with unit speed to the left, the other with unit speed to the right. The boundary conditions appeared in (1.1) can create irregularly acoustical vibrations ([1,6,7]). This type of vibrations, for example, can be generated by noise signals radiated from underwater vehicles, and there are intensive research for the properties of acoustical vibrations in current literature (see e.g., [10] and references therein). Hence the study of this type of vibration is not only important but also may lead to a better understanding of the dynamics of acoustic systems.
In this paper, we mainly consider the oscillation problems described by the following models:
{wtt(x,t)−c2wxx(x,t)=0, x∈(0,1), t>0,wx(0,t)=−ηwt(0,t), η≠c−1, t>0,wx(1,t)=h(wt(1,t)), t>0,w0(x)=w(x,0),w1(x)=wt(x,0),0≤x≤1, | (1.4) |
{ztt(x,t)−c2zxx(x,t)=0, x∈(0,1), t>0,z(0,t)=0, t>0,zx(1,t)=h(zt(1,t)), t>0,z0(x)=z(x,0),z1(x)=zt(x,0),0≤x≤1, | (1.5) |
where the function h∈C0(R) satisfies
(A1) ϕ:t↦12(h(t)−1ct) strictly monotonically decreases on R,
(A2) ϕ(R)=R.
It is clear that ϕ−1:R→R is well-defined and strictly decreases on R. In fact, the model (1.4) is a generalized case of the van del Pol type boundary condition, h(x)=αx−βx3, say. Denote the total energy of system (1.4) as
E(t)=12c2∫10[c2w2x(x,t)+w2t(x,t)]dx, |
then the boundary conditions show
ddtE(t)=ηw2t(0,t)+w2t(1,t)⋅h(wt(1,t))wt(1,t). |
Hence η>0 can cause the energy of system to increase. Moreover, the system (1.1) has a generalized self-excited mechanism if we assume that, roughly speaking, h(y)y>0 if |y| is small and h(y)y<0 if |y| is large. Thus if η>0, the system (1.4) has a self-excited mechanism that supplies energy to the system itself, which could induce irregular oscillations. The interesting part is that the system (1.5) may have chaos even though there is no energy supplier at the boundary.
We can treat the initial-boundary problems (1.4) and (1.5) by using wave propagation method. Let's take (1.4) as a start. It is well-known that w, a solution of 1D wave equation, has the following form:
w(x,t)=L(x+ct)+R(x−ct), | (1.6) |
where L and R are two C1 functions. Then the gradient of w can be represented as follows:
wt(x,t)=cL′(x+ct)−cR′(x−ct),wx(x,t)=L′(x+ct)+R′(x−ct), | (1.7) |
for x∈[0,1],t≥0. Introduce two new variables
u(x,t)=L′(x+ct),v(x,t)=R′(x−ct), | (1.8) |
which are called the Riemann invariants. It is evident that u and v keeps constants along the lines x+ct=const. and x−ct=const., respectively, which are referred to as characteristics.
When t>0 and x=0, from the boundary condition at the left end x=0,
v(0,t)=cwx(0,t)−wt(0,t)2c=−cηwt(0,t)−wt(0,t)2c=−1+cη2cwt(0,t)=−1+cη2(u(0,t)−v(0,t)), |
that is
v(0,t)=cη+1cη−1u(0,t)def=γ(η)⋅u(0,t). | (1.9) |
Without confusion, we also take γ(η) as a linear map.
When t>0 and x=1, from the boundary condition at the right end x=1, it follows:
h(c(u(1,t)−v(1,t)))−1c⋅c(u(1,t)−v(1,t))−2v(1,t)=0, |
that is
v(1,t)=ϕ(c(u(1,t)−v(1,t))), |
which determines a reflection relationship between u(1,t) and v(1,t) as follow
u(1,t)=1c⋅ϕ−1(v(1,t))+v(1,t)def=φ(v(1,t)), | (1.10) |
where
φ(x)=1c⋅ϕ−1(x)+x. | (1.11) |
When t=0, from the initial conditions,
u0(x)def=u(x,0)=cw′0(x)+w1(x)2c,v0(x)def=v(x,0)=cw′0(x)−w1(x)2c, | (1.12) |
which are referred to as the initial date of (u,v).
For x∈[0,1] and τ≥0, it follows from the boundary reflections (1.9) and (1.10) that
u(x,τ+2c)=L′(x+cτ+2)=u(1,1cx+τ+1c)=φ(v(1,1cx+τ+1c))=φ(R′(−x−cτ))=φ(v(0,1cx+τ))=φ(γ(η)⋅u(x,τ)), |
and
v(x,τ+2c)=γ(η)∘φ(v(x,τ)); |
inductively,
u(x,τ+2nc)=(φ∘γ(η))n(u(x,τ)),v(x,τ+2nc)=(γ(η)∘φ)n(v(x,τ)), | (1.13) |
where the superscript n∈N denotes the nth iteration of a function. Analogously, for x∈[0,1] and τ∈[0,2c),
u(x,τ)={u0(x+cτ),0≤cτ≤1−x,φ(v0(2−x−cτ)),1−x<cτ≤2−x,φ∘γ(η)(u0(x+cτ−2)),2−x<cτ<2, | (1.14) |
and
v(x,τ)={v0(x−cτ),0≤cτ≤x,γ(η)(u0(cτ−x)),x<cτ≤x+1,γ(η)∘φ(v0(2+x−cτ)),x+1<cτ<2. | (1.15) |
Equations (1.13)–(1.15) show that the system (1.4) is solvable and the dynamics of the solution (w,wx,wt) to the equation (1.4) can be uniquely determined by the initial data and the following two functions:
ψη=γ(η)∘φ,g=φ∘γ(η). | (1.16) |
Note that g=γ−1(η)∘ψη∘γ(η), that is to say there is topological conjugacy between ψη and g. Therefore one only needs to consider one of them, say ψη.
The paper will be organized as follows. In the next two sections, we will present the necessary and sufficient conditions to cause the onset of chaos, respectively. Section 4 shows, as a special case, in which the wave equation has a fixed end, that chaos can occur with the effects of self-regulations and energy-conservation. In Section 5, there are some applications of the theoretical outcomes. In the last section, it is the numerical simulations.
In this section, we firstly give a necessary condition for the onset of chaos in the following system:
{wtt(x,t)−c2wxx(x,t)=0, x∈(0,1), t>0,wx(0,t)=−ηwt(0,t), η≠c−1, t>0,wx(1,t)=h(wt(1,t)), t>0,w0(x)=w(x,0),w1(x)=wt(x,0),0≤x≤1, | (2.1) |
where h satisfies hypotheses (A1) and (A2). In addition, assume the function t↦(h(t)+1ct) is piecewise monotone.
Theorem 2.1. Suppose the system (2.1) is chaotic in the sense of Definition 1.1. Then h is not even and η≠0.
Proof. It is equivalent to prove that there is no chaos in the system (2.1) if η=0 or h is an even function.
Firstly, assume h is an even function. Let η∈R and η≠c−1. Recall the function ψη given by (1.16):
ψη=γ(η)∘φ,φ(x)=1c⋅ϕ−1(x)+x. |
Introduce a new map Q from R to R as follow:
Q(y)=12(h(y)+1cy). | (2.2) |
For x∈R, let y=ϕ−1(x), then
ψη(x)=γ(η)(1cy+ϕ(y))=12γ(η)(1cy+h(y))=γ(η)⋅Q∘ϕ−1(x). |
Since ϕ−1 is strictly decreasing and γ(η)∈R∗, ψ is monotonic if and only if Q is monotonic. Let y1,y2∈R with y1<y2. From the hypothesis h being an even function, it follows
Q(y2)−Q(y1)=12(h(−y2)−1c(−y2))−12(h(−y1)−1c(−y1))=ϕ(−y2)−ϕ(−y1)>0, |
which implies Q strictly monotonically increases. Therefore ψ strictly monotonically increases (decreases) if γ(η)<0(γ(η)>0). It is easily seen that
wt(x,t)=c[γ−1(η)∘ψn(γ(η)u(x,τ))−ψn(v(x,τ))],wx(x,t)=γ−1(η)∘ψn(γ(η)u(x,τ))+ψn(v(x,τ)). |
Hence the dynamics of (wx,wt) is simple, and the chaos doesn't occur in the system.
Next, assume η=0, in other words, the system is free at the left end. In this case,
ψ0(x)=−x−ϕ−1(x). |
We will show that there is no period point of ψ0 with period 2. Let x0∈R satisfy x0<ψ0(x0). Define two functions as follows:
q:x↦−x+ψ0(x0)+x0,k=ψ0−q. |
Since ϕ−1(⋅) strictly monotonically decreases and
k(x)=−x−ϕ−1(x)−q(x)=ϕ−1(x0)−ϕ−1(x), |
k strictly monotonically increases in [x0,ψ0(x0)]. Hence
k(ψ0(x0))=ψ20(x0)−x0>k(x0)=0. |
That implies there are no period points of ψ0 with period 2. By Sharkovsky's Theorem, there are no periods of ψ0 with period lager than 2. By virtue of the Main Theorem 6 in [12], there is still no chaos in the system when η=0.
Next, we consider the following system governed by 1D wave equation with a fixed end:
{ztt(x,t)−c2zxx(x,t)=0, x∈(0,1), t>0,z(0,t)=0, t>0,zx(1,t)=h(zt(1,t)), t>0,z0(x)=z(x,0),z1(x)=zt(x,0),0≤x≤1, | (2.3) |
where h satisfies hypotheses (A1) and (A2).
Theorem 2.2. Suppose the system (2.3) is chaotic in the sense of Definition 1.1. Thenh is neither an even function nor an odd function.
Proof. Put
ψ∞(x)=x+ϕ−1(x). |
By the analysis in Section 1, it is clear that (zx,zt) can be represented by iterations of ψ∞ and initial data. Hence, the dynamics of (zx,zt) is completely determined by ψ∞.
If h is even, ψ∞ is monotonically monotone. If h is odd, −ψ∞ has no periodic points of period larger than 2. Therefore, chaos never occurs in the system 2.3 if h is either an even function or an odd function.
Remark 2.1. Chaos can definitely happen in the system 2.3 for a special kind of h. One can find more details about that in the later section.
In this section, we mainly try to determine some sufficient conditions to ensure the onset of chaotic oscillations in the following system:
{wtt(x,t)−c2wxx(x,t)=0, x∈(0,1), t>0,wx(0,t)=−ηwt(0,t), η≠c−1, t>0,wx(1,t)=h(wt(1,t)), t>0,w0(x)=w(x,0),w1(x)=wt(x,0),0≤x≤1, | (3.1) |
where h∈C0(R) satisfies h(0)=0, hypotheses (A1) and (A2). When wt(1,t)≡0, there should be no signals feedback to wx(1,t), that's to say the right end should be free. Therefore it is reasonable to let h(0)=0.
According to the analysis in the first section, we have known that the function ψη given by (1.16) plays a vital role in studying the dynamics of system (3.1). We firstly give two lemmas that are useful in analyzing dynamics of ψη, as follow:
Lemma 3.1. Let I be a non-degenerate closed interval, J⊆R and F:I→J, G:J→R. Assume that VJG<+∞ and F is piecewise monotone. Then
VF(I)G≤VIG∘F<+∞. |
Proof. Let I1=[a,b] be a monotone interval of F. Without loss of generality, let F monotonically increase in I1. Let
P:a=t0≤t1≤⋯≤tn=b, |
be a partition of I1. Then
VI1[G∘F,P]=VF(I1)[G,P′]≤VF(I1)G, |
where
P′:F(a)=F(t0)≤F(t1)≤⋯≤F(tn)=F(b), |
is a partition of F(I1). Consequently, VI1G∘F≤VF(I1)G. Conversely, if
P′:F(a)=q0≤q1≤⋯≤qn=F(b), |
is a partition of of F(I1), then
P:a=˜F−1(q0)≤˜F−1(q1)≤⋯≤˜F−1(qn)≤qn+1=b, |
where ˜F−1(q)=min{x∈I1,F(x)=q}, is a partition of I1. That implies
VF(I1)[G,P′]=VI1[G∘F,P]≤VI1G∘F. |
Therefore, VI1G∘F=VF(I1)G.
Let (Ij)1≤j≤n be a finite sequence consisting of monotone intervals of F. Assume ⊔1≤j≤nIj=I and #(Ii∩Ij)≤1 provided i≠j. Then,
VIG∘F=n∑j=1VIjG∘F=n∑j=1VF(Ij)G≥V⊔1≤j≤nF(Ij)G=VF(I)G. |
Lemma 3.2. Let I be a closed interval and F∈C0(I,I) be piecewise monotone. If there exist non-degenerate subintervals I1,I2⊆A with Card(I1∩I2)≤1 such that I2⊆F(I1) and I1∪I2⊆F(I2), then
lim infn→+∞lnVI1∪I2Fnn≥ln1+√52. | (3.2) |
Proof. Let two subintervals I1,I2⊆A satisfy the hypothesis. Let n∈N. Put
xn=VI2Fn,yn=VI1∪I2Fn, |
where F0=IdR if n=0. Note that F(⋅) is continuous and piecewise monotone, it follows from Lemma 3.1 that
xn+1=VI2Fn∘F≥VF(I2)Fn≥VI1∪I2Fn=yn, |
and
yn+2=VI1Fn+1∘F+VI2Fn+1∘F≥VF(I1)Fn+1+VF(I2)Fn+1≥VI2Fn+1+VI1∪I2Fn+1=xn+1+yn+1≥yn+yn+1. |
Let (zn)n∈N be Fibonacci sequence, i.e., zn+2=zn+1+zn, with the initial data z0=0,z1=1. It is well known that
zn=1√5[(1+√52)n−(1−√52)n],n∈N. |
It is clear that y1>0 and ∀n∈N,yn≥y1zn. Therefore,
lim infn→+∞lnVI1∪I2Fnn=lim infn→+∞lnynn≥limn→+∞lny1znn=ln1+√52. |
This competes the proof.
Proposition 3.1. Let h∈C0(R) satisfy the hypothesises (A1)−(A2) and ψη be given by (1.16). In addition, assume h(0)=0 and
(i) non-constant function Q:t↦12(h(t)+1ct) is piecewise monotone,
(ii) there is at least one solution to the equation h(y)+c−1y=0 in R∗.
Then there exist A>0 and a non-degenerate interval I⊆(0,+∞)∖{c−1} such that for any η∈I, [0,A] is an invariant set of ψη and
lim infn→+∞lnV[0,A]ψnηn≥ln1+√52. | (3.3) |
Proof. We need to finish the following two steps:
(1) Determine an invariant interval [0,A] of ψη.
(2) To make sure the hypothesises of Lemma 3.2 holds.
It has been known that
ψη=γ(η)⋅φ=γ(η)⋅Q∘ϕ−1. |
Put
S−={y<0∣Q(y)=0 and ∃ t∈(y,0),Q(t)≠0} |
and
S+={y>0∣Q(y)=0 and ∃ t∈(0,y),Q(t)≠0}. |
By the hypothesises (i)−(ii), S−≠∅ or S+≠∅. Without loss of generality, assume S−≠∅. We take
ˉy=maxS−,A=−c−1ˉy, | (3.4) |
then
ˉy<0, A=ϕ(ˉy)>0, ψη(A)=0. | (3.5) |
Moreover, let
M=max{Q(t)∣t∈[ˉy,0]},m=min{Q(t)∣t∈[ˉy,0]}. |
By the definition of S−, it is evident that M=0 or m=0.
Case 1: Assume M=0. That implies m<0. Put
y0=max{t∈[ˉy,0]∣Q(t)=m},x0=ϕ(y0). |
For any η∈[0,c−1), it is easily seen that
ψη(x0)=max{ψη(x)∣x∈[0,A]}=m⋅γ(η). |
and ψη(x)<ψη(x0) provided that x∈[0,x0). We need to prove that m⋅γ(0)<A. Proceed the proof by contradiction. Assume m⋅γ(0)≥A, which implies
ψ0([0,x0])∩ψ0([x0,A])⊇[0,ψ0(x0)]⊇[0,A]=[0,x0]∪[x0,A]. |
Thus ψ0 is turbulent. According to Lemma 3 in [11], ψ0 has periodic points of all periods. But from the proof of Theorem 2.1, ψ0 has no periodic points whose period is larger than 2. We have thus reached a contradiction. Since that m⋅γ(⋅) is strictly increasing in [0,c−1) and m⋅γ(η)→+∞ as η→(c−1)−, it is reasonable to put
ˉη=γ−1(mA)∈(0,c−1). | (3.6) |
Then for η∈[0,ˉη], we have
∀x∈[0,A],0≤ψη(x)≤A, |
that is to say [0,A] is an invariant set of ψη provided η∈[0,ˉη]. Put
η_={γ−1(mx0), if mx0<−1,0, else, | (3.7) |
then η_<ˉη and ψ(x0)>x0 if η>η_. For η∈[η_,c−1), consider the following set:
S(η,x0)={x∈(0,x0)∣γ(η)⋅φ(x)=x0}. |
It is clear that S(η,x0) is closed. Since x0 is uniquely determined by h and h is independent of η, the following function is well-defined:
α(η)=maxS(η,x0),η∈[η_,c−1). | (3.8) |
We first prove that α(⋅) is strictly monotonically decreasing in [η_,c−1). Let t1,t2∈(η_,c−1) with t1<t2. From the definition of α(⋅), it follows
γ(t1)⋅φ(α(t1))=x0,γ(t1)⋅φ(x)>x0,x∈(α(t1),x0],γ(t2)⋅φ(α(t2))=x0,γ(t2)⋅φ(x)>x0,x∈(α(t2),x0]. | (3.9) |
Since that
γ(t1)⋅φ(α(t2))<γ(t2)⋅φ(α(t2))=x0, γ(t1)⋅φ(x0)>x0, |
by applying the continuity of γ(t1)⋅φ(⋅) there exists x∈(α(t2),x0) such that γ(t1)⋅φ(x)=x0. According to the definition of α(t1), we have
α(t2)<x≤α(t1). |
Therefore, α(⋅) strictly monotonically decreases in [η_,c−1).
Next, prove that α(⋅) is continuous from the left. Let t0∈(η_,x0) be fixed. Since α is monotone, we have
α(t0)≤Ldef=limt→t−0α(t)<+∞. | (3.10) |
By the continuity of η↦γ(η) and x↦φ(x), we obtain
limε→0+γ(t0−ε)=γ(t0),limε→0+φ(α(t0−ε))=φ(L), | (3.11) |
which gives γ(t0)⋅φ(L)=γ(t0)⋅φ(α(t0))=x0. From the definition of α(⋅) and (3.10), it follows
L≤α(t0)≤L=limt→t−0α(t). |
Therefore α(⋅) is continuous from the left.
In particular, ψˉη(α(η_))=γ(η_)⋅φ(α(ˉη))=x0 and ψˉη(x0)=γ(η_)⋅φ(x0)=A. Define a function as follow:
K(s)=α(η_−s)−ψ2ˉη−s(x0). | (3.12) |
It is easily seen that K(⋅) is continuous from right at s=0. Note that
K(0)=α(η_)−ψˉη(ψˉη(x0))=α(η_)−ψˉη(A)=α(η_)>0, |
hence there exists ρ0∈(0,ˉη−η_) such that
∀s∈[0,ρ0],K(s)=α(η_−s)−ψ2ˉη−s(x0)>0. | (3.13) |
Put
I=[ˉη−ρ0,ˉη] | (3.14) |
and
J1=[α(η),x0], J2=[x0,ψη(x0)], η∈I. | (3.15) |
Let η∈I be fixed. By (3.13),
ψη(J1)⊇[ψη(α(η)),ψη(x0)]=J2,ψη(J2)⊇[ψ2η(x0),ψη(x0)]⊇[α(η),ψη(x0)]⊇J1∪J2, |
which is to say hypotheses of Lemma 3.2 holds. Therefore for any η∈I, by applying Lemma 3.2 we have
lim infn→+∞lnV[0,A]ψnηn≥lim infn→+∞lnVJ1∪J2ψnηn≥ln1+√52. | (3.16) |
For the case m=0, one just needs to consider η>c−1. The proof is similar, we omit it.
Theorem 3.1. Consider the system (3.1) with the hypotheses (A1)−(A2). Suppose h(0)=0 and
(i) non-constant function Q:t↦12(h(t)+1ct) is piecewise monotone,
(ii) there is at least one solution to the equation h(y)+c−1y=0 in R∗.
Then there exists a non-degenerate interval I such that for all η∈I the system (3.1) is chaotic in the sense of Definition 1.1.
Proof. Without loss of generality, assume the equation h(y)+c−1y=0 has at least one solution in (−∞,0). We still use the symbols given in the proof of Proposition 3.1, such as A, x0, α(η) and so on. Let I and J1,J2 be given by (3.14) and (3.15), respectively. Take η∈I.
Let (w0,w1) be the initial data of system (3.1) and (w,wx,wt) be a solution of system (3.1). Recall the Riemann invariants as follow
u(x,t)=cwx(x,t)+wt(x,t)2c, v(x,t)=cwx(x,t)−wt(x,t)2c, x∈[0,1], t≥0, |
and u0(⋅)=u(⋅,0) and v0(⋅)=v(⋅,0) are the initial data. Let t>2c−1, n=[ct2] and τ=t−2nc−1 be fixed. It is clear that 0≤cτ<2. By using (1.13), (1.14) and (1.15), we obtain
u(x,t)={γ−1(η)∘ψnη(γ(η)⋅u0(x+cτ)),0≤cτ≤1−x,γ−1(η)∘ψn+1η(v0(2−x−cτ)),1−x<cτ≤2−x,γ−1(η)∘ψnη(γ(η)⋅u0(x+cτ−2)),2−x<cτ<2, | (3.17) |
and
v(x,t)={ψnη(v0(x−cτ)),0≤cτ≤x,ψnη(γ(η)⋅u0(cτ−x)),x<cτ≤x+1,ψn+1η(v0(2+x−cτ)),x+1<cτ<2. | (3.18) |
Since [0,A] is an invariant set of ψη, |u|+|v| is uniformly bounded if
Range(γ(η)⋅u0)∪Range(v0)⊆[0,A]. | (3.19) |
In addition, assume that u0 and v0 are piecewise monotone and
[0,α(η)]⊆Range(γ(η)⋅u0)∩Range(v0). | (3.20) |
Consider the total variations of u(⋅,t) and v(⋅,t) on [0,1]. When 0≤cτ<1, from (1.13)-(1.15) and Lemma 3.1, it follows
V[0,1]u(⋅,t)=V[0,1−cτ]u(⋅,t)+V[1−cτ,1]u(⋅,t)=V[cτ,1]γ−1(η)∘ψnη∘γ(η)∘u0+V[1−cτ,1]γ−1(η)∘ψn+1η∘v0=|γ−1(η)|(V[cτ,1]ψnη∘γ(η)∘u0+V[1−cτ,1]ψn+1η∘v0), |
V[0,1]v(⋅,t)=V[0,cτ]v(⋅,t)+V[cτ,1]v(⋅,t)=V[0,cτ]ψnη∘γ(η)∘u0+V[0,1−cτ]ψnη∘v0. |
Note that 0<|γ−1(η)|<1, by Lemma 3.1 and (3.20) we obtain
V[0,1]u(⋅,t)+V[0,1]v(⋅,t)≥|γ−1(η)|Vγ(η)⋅u0([0,1])ψnη≥|γ−1(η)|V[0,α(η)]ψnη≥|γ−1(η)|Vψ2η[0,α(η)]ψn−2η≥|γ−1(η)|VJ1∪J2ψn−2η. |
Then Proposition 3.1 and (3.16) show that
ln(V[0,1]u(⋅,t)+V[0,1]v(⋅,t))t≥ln|γ−1(η)|t+ln(VJ1∪J2ψn−2η)t≥cln(VJ1∪J2ψn−2η)2n+2+ln|γ−1(η)|t≥c2ln1+√52, as t→+∞. | (3.21) |
When 1≤cτ<2, in the same way we can obtain
ln(V[0,1]u(⋅,t)+V[0,1]v(⋅,t))t≥cln(VJ1∪J2ψn−1η)2n+2+ln|γ−1(η)|t≥c2ln1+√52, as t→+∞. | (3.22) |
It is evident that
V[0,1]u(⋅,t)+V[0,1]v(⋅,t)≤max{1,c−1}(V[0,1]wx(⋅,t)+V[0,1]wt(⋅,t)). | (3.23) |
Hence
lim inft→+∞ln(V[0,1]wx(⋅,t)+V[0,1]wt(⋅,t))t≥c2ln1+√52>0. |
Therefore, the system (3.1) is chaotic in the sense of Definition 1.1.
In this section, we mainly consider the oscillation problem governed by 1D wave equation with a fixed end. We will show that only the effect of self-regulation effect at one of the two ends can cause the onset of chaos. But if the fixed end is replaced by a free end, the system never has chaos. As we know, both a fixed end and a free end are called the energy-conserving boundary condition. Even though either a fixed end or a free end has the same effect to the energy of the system, the systems may show completely different dynamics: one with chaos, the other without chaos. Thus the relationship between chaos and energy of the system is much more complicated.
Theorem 4.1. Consider an initial-boundary problem governed by 1D wave equation with a fixed end as follows:
{ztt(x,t)−c2zxx(x,t)=0, x∈(0,1), t>0,z(0,t)=0, t>0,zx(1,t)=h(zt(1,t)), t>0,z0(x)=z(x,0),z1(x)=zt(x,0),0≤x≤1. | (4.1) |
There exists a piecewise linear function h so that the system is chaotic in the sense of Definition 1.1.
Proof. Without loss of generality, we take c=1. Put
hθ(x)={−2x+2(θ2−3), x∈[2(θ2−3),−2),(1−θ2)x, x∈[−2,0],(θ−1)x, x∈(0,1],−2x+(θ+1), x∈(1,θ+1],−x, else, | (4.2) |
where θ∈(0,1) is a fixed parameter. It is clear that hθ∈C0(R) and
ϕθ:x↦12(hθ(x)−x) |
is strictly monotonically decreasing and ϕθ(R)=R. Then ϕ−1θ exists. Let
Qθ(y)=12(hθ(y)+y), |
then
ψθ(x)=x+ϕ−1θ(x)=Qθ∘ϕ−1θ(x). |
Let u=zx+zt2 and u=zx−zt2 be the Riemann invariants of system 4.1. For some initial date (u0,v0), suppose (u,v) is solved for t≤2. Then for t=2n+τ with τ∈[0,2) and n∈N, we have
u(x,t)=ψnθ(u(x,τ)),v(x,t)=ψnθ(v(x,τ)), x∈[0,1]. | (4.3) |
Put
I1=[−θ−1,12θ−1], I2=[12θ−1,0], I3=[0,θ2]. |
They are monotone intervals of ψθ, respectively. A simple calculation shows that
ψθ(−θ−1)=0, ψθ(12θ−1)=12θ, ψθ(0)=0, ψθ(θ2)=θ2−2. |
Let θ∈(0,1) be sufficiently small such that 12θ≥θ2 and θ2−2≤−θ−1, which implies
I3⊆ψθ(I1)∩ψθ(I2), I1∪I2⊆ψθ(I3). |
Let n∈N. Put
xn=VI1ψnθ, yn=VI2ψnθ, zn=VI3ψnθ. |
Lemma 3.2 shows that
xn+1≥Vψθ(I1)ψnθ≥VI3ψnθ=zn, yn+1≥zn, zn+2≥xn+1+yn+1, |
which implies
zn+2≥2zn. |
Inductively, we obtain
z2n+2≥2nz2>0, z2n+1≥2nz1>0. |
Therefore,
lim infn→+∞lnznn≥12ln2>0. | (4.4) |
A simple calculation shows that
ψθ(x)=0, x∈(−∞,−θ−1]∪[2(3−θ2),+∞]. |
By (4.3), |u|+|v| is always uniformly bounded for any initial date. Analogous to the proof of Theorem 3.1, there exists a large class of initial data such that
lim inft→+∞ln(V[0,1]zx(⋅,t)+V[0,1]zt(⋅,t))t≥14ln2>0. |
Therefore, the proof completes.
We first consider a problem about perturbations at boundaries, described by the following model:
{wtt(x,t)−c2wxx(x,t)=0, x∈(0,1), t>0,wx(0,t)=−ηwt(0,t), η≠c−1, t>0,wx(1,t)=−c−1wt(1,t)+εsin(wt(1,t)), t>0,w0(x)=w(x,0),w1(x)=wt(x,0),0≤x≤1. | (5.1) |
If ε=0, for any initial data (w0,w1) and η≠c−1, by using wave propagation method we can obtain
∀t>2,x∈[0,1], wx(x,t)=wt(x,t)=0. |
That shows system (5.1) is of global asymptotical stability.
Assume ε≠0 and |ε| is sufficiently small. Put
hε(y)=−c−1y+εsin(y) |
It is evident that i)y↦hε(y)+c−1y is piecewise monotone, ii)y↦hε(y)−c−1y strictly monotonically decreasing, iii) there are infinite solutions to the equation hε(y)+c−1y=0 in R∗. That shows the hypotheses of Theorem 3.1 hold. Therefore, system (5.1) is chaotic in the sense of Definition 1.1.
The above analysis shows that this kind of system is no longer stable under small perturbations at one of the boundaries. That is to say this kind of system is not of structural stability.
As we have stated in the previous sections, if the feedback control law is even chaos never happen. In particular, consider the system described by the following model:
{ztt(x,t)−zxx(x,t)=0, x∈(0,1), t>0,zx(0,t)=−ηzt(0,t), η≠1, or z(0,t)=0, t>0,zz(1,t)=cos(zt(1,t))−1, t>0,z0(x)=z(x,0),z1(x)=zt(x,0),0≤x≤1. | (5.2) |
The feedback control law at the right end is periodic and even. Therefore there is no chaos in the system (5.2). Moreover, there should exists a compact attractor in the system 5.2.
In this section, we will give some numerical simulations to validate the theoretical results of this paper. We first consider the system as follow:
{ztt(x,t)−zxx(x,t)=0, x∈(0,1), t>0,z(0,t)=0, t>0,zx(1,t)=h(zt(1,t)), t>0,z0(x)=z(x,0),z1(x)=zt(x,0),0≤x≤1, | (6.1) |
where
h(x)={−2x−112, x∈[−112,−2),34x, x∈[−2,0],−12x, x∈(0,1],−2x+32, x∈(1,32],−x, else. | (6.2) |
Choose
w0=0, w1(x)=14sin4(2πx), x∈[0,1], |
as the initial data of system (6.1). Put
ψ∞(x)={13x−116, x∈(14,112],−7x, x∈(0,14],−13x, x∈(−34,0],13x+12, x∈[−32,−34],0, else. | (6.3) |
As we show in Section 4, the solution (z,zx,zt) can be represented by the iterations of this ψ∞ and initial data (z0,z1). Note that h and ψ∞ are piecewise linear, thus the feedback control at the boundaries is easy to implement and the solutions of system are much simpler to be solved and represented.
We present the graphics in some detail, for zx,zt for 98≤t≤100. Figure 1 shows that zx,zt are extremely oscillatory in every direction of space and time.
This work was supported by the National Natural Science Foundation of China under Grant 11671410 and the Natural Science Foundation of Guangdong Province, China (No.2022A1515012153).
The authors declare there is no conflicts of interest.
[1] |
Y.Huang, Growth rates of total variations of snapshots of the 1D linear wave equation with composite nonlinear boundary reflection, Int. J. Bifurcation Chaos, 13 (2003), 1183–1195. https://doi.org/10.1142/S0218127403007138 doi: 10.1142/S0218127403007138
![]() |
[2] | T. Y. Li, J. A.Yorke, Period Three Implies Chaos, Am. Math. Mon., 82 (1975), 985–992. https://doi.org/10.1007/978-0-387-21830-4 |
[3] |
M. Štefánková, Inheriting of chaos in uniformly convergent nonautonomous dynamical systems on the interval, Discrete Contin. Dyn. Syst., 36 (2016), 3435–3443. https://doi.org/10.3934/dcds.2016.36.3435 doi: 10.3934/dcds.2016.36.3435
![]() |
[4] |
J. S. Cánocas, T. Puu, M. R. Marín, Detecting chaos in a duopoly model via symbolic dynamics, Discrete Contin. Dyn. Syst.-B, 13 (2010), 269–278. https://doi.org/10.3934/dcdsb.2010.13.269 doi: 10.3934/dcdsb.2010.13.269
![]() |
[5] |
G. Chen, S. B. Hsu, J. Zhou, Chaotic vibrations of the one-dimensional wave equation due to a self-excitation boundary condition, Part I: controlled hysteresis, Trans. Amer. Math. Soc., 350 (1998), 4265–4311. https://doi.org/10.1090/S0002-9947-98-02022-4 doi: 10.1090/S0002-9947-98-02022-4
![]() |
[6] |
G. Chen, , S.B. Hsu, J. Zhou, , Chaotic vibrations of the one-dimensional wave equation due to a self-excitation boundary condition, Part II: Energy injection, period doubling and homoclinic orbits, Int. J. Bifurcation Chaos, 8 (1998), 423–445. https://doi.org/10.1142/S0218127498000280 doi: 10.1142/S0218127498000280
![]() |
[7] |
Y. Huang, J. Luo, Z. L. Zhou, Rapid fluctuations of snapshots of one-dimensional linear wave equations with a van der Pol nonlinear boundary conditions, Int. J. Bifurcation Chaos, 15 (2005), 567–580. https://doi.org/10.1142/S0218127405012223 doi: 10.1142/S0218127405012223
![]() |
[8] |
L. L. Li, Y. Huang, Growth of total varitions of snapshots of 1D linear wave equations with nonlinear right-end boundary conditions, J. Math. Anal. Appl., 361 (2010), 69–85. https://doi.org/10.1016/j.jmaa.2009.09.011 doi: 10.1016/j.jmaa.2009.09.011
![]() |
[9] |
J. Liu, Y. Huang, H. Sun, M. Xiao, Numerical methods for weak solution of wave equation with van der Pol type nonlinear boundary conditions, Numer. Meth. Part. D. E., 32 (2016), 373–398. https://doi.org/10.1002/num.21997 doi: 10.1002/num.21997
![]() |
[10] | P. C. Etter, Underwater Acoustic Modeling and Simulation, Spon Press, London, New Yoek, 2003. https://doi.org/10.4324/9780203417652 |
[11] | L. S. Block, W. A. Coppel, Dynamics in One Dimension, Lecture Notes in Mathematics, Springer-Verlag, NY, Heidelberg Berlin, 1992. https://doi.org/10.1007/BFb0084762 |
[12] |
G. Chen, T. Huang, Y. Huang, Chaotic behavior of interval maps and total variations of iterates, Int. J. Bifurcation Chaos, 14 (2004), 2161–2186. https://doi.org/10.1142/S0218127404010540 doi: 10.1142/S0218127404010540
![]() |
1. | Xu Zhang, Iram Hussan, Manyu Zhao, Qigui Yang, Li–Yorke Chaos in Partial Differential Equations with or Without Energy Injection, 2025, 35, 0218-1274, 10.1142/S0218127425500245 |