Hindawi Publishing Corporation Discrete Dynamics in Nature and Society Volume 2011, Article ID 958381, 22 pages doi:10.1155/2011/958381 Research Article Lyapunov Function and Exponential Trichotomy on Time Scales Ji Zhang College of Mathematics, Jilin University, Changchun 130012, China Correspondence should be addressed to Ji Zhang, zhangjiccf@gmail.com Received 20 February 2011; Revised 7 May 2011; Accepted 6 July 2011 Academic Editor: Jinde Cao Copyright q 2011 Ji Zhang. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. We study the relation between Lyapunov function and exponential trichotomy for the linear equation xΔ Atx on time scales. Furthermore, as an application of these results, we give the roughness of exponential trichotomy on time scales. 1. Introduction Exponential trichotomy is important for center manifolds theorems and bifurcation theorems. When people analyze the asymptotic behavior of dynamical systems, exponential trichotomy is a powerful tool. The conception of trichotomy was first introduced by Sacker and Sell 1. They described SS-trichotomy for linear differential systems by linear skew-product flows. Furthermore, Elaydi and Hájek 2, 3 gave the notions of exponential trichotomy for differential systems and for nonlinear differential systems, respectively. These notions are stronger notions than SS-trichotomy. In 1991, Papaschinopoulos 4 discussed the exponential trichotomy for linear difference equations. And in 1999 Hong and his partners 5, 6 studied the relationship between exponential trichotomy and the ergodic solutions of linear differential and difference equations with ergodic perturbations. Recently, Barreira and Valls 7, 8 gave the conception of nonuniform exponential trichotomy. From their papers, we can see that the exponential trichotomy studied before is just a special case of the nonuniform exponential trichotomy. For more information about exponential trichotomy we refer the reader to papers 9–14. Many phenomena in nature cannot be entirely described by discrete system, or by continuous system, such as insect population model, the large population in the summer, the number increases, a continuous function can be shown. And in the winter the insects freeze 2 Discrete Dynamics in Nature and Society to death or all sleep, their number reduces to zero, until the eggs hatch in the next spring, the number increases again, this process is a jump process. Therefore, this population model is a discontinuous jump function, it cannot be expressed by a single differential equation, or by a single difference equations. Is it possible to use a unified framework to represent the above population model? In 1988, Hilger 15 first introduced the theory of time scales. From then on, there are numerous works on this area see 16–23. Time scales provide a method to unify and generalize theories of continuous and discrete dynamical systems. In this paper, motivated by 8, we study the exponential trichotomy on time scales. We firstly introduce λ, μ-Lyapunov function on time scales. Then we study the relationship between exponential trichotomy and λ, μ-Lyapunov function on time scales. We obtain that the linear equation xΔ Atx admits exponential trichotomy, if it has two λ, μLyapunov functions with some property; conversely, the linear equation has two λ, μLyapunov functions, if it admits strict exponential trichotomy. At last, by using these results we investigate the roughness of exponential trichotomy on time scales. Above all, our paper gives a way to unify the analysis of continuous and discrete exponential trichotomy. This paper is organized as follows. In Section 2, we review some useful notions and basic properties on time scales. Our main results will be stated and proved in Section 3. Finally, in Section 4, we study the roughness of exponential trichotomy on time scales. 2. Preliminaries on Time Scales In order to make our paper independent, some preliminary definitions and theories on time scales are listed below. Definition 2.1. Let T be a time scale which is an arbitrary nonempty closed subset of the real numbers. The forward jump operator is defined by σt : inf{s ∈ T : s > t}, 2.1 while the backward jump operator is defined by ρt : sup{s ∈ T : s < t}, 2.2 for every t ∈ T. If σt t, then t is called right-dense. And if ρt t, then t is called leftdense. Let μt : σt − t be the graininess function. For example, the set of real numbers R is a time scale with σt t and μt 0 for t ∈ T R and the set of integers Z is a time scale with σt t 1 and μt 1 for t ∈ T Z. Definition 2.2. A function f : T → Rn is called rd-continuous if it is continuous at right-dense points in T and left-sided limits exist at left dense points in T. We denote the set of rd-continuous functions by Crd T, Rn . Discrete Dynamics in Nature and Society 3 Definition 2.3. We say that a function f : T → Rn is differentiable at t ∈ T, if for any ε > 0, there exist T-neighborhood U of t and f Δ t ∈ Rn such that for any s ∈ U one has fσt − fs − f Δ tσt − s ≤ εσt − s, 2.3 and f Δ t is called the derivative of f at t. Relate to the differential properties, if f and g is differentiable at t, then we have the following equalities: fσt ft μtf Δ t, fg Δ t f Δ tgt fσtg Δ t. 2.4 2.5 Definition 2.4. We say that a function p : T → R is regressive if 1 μtpt / 0 for all t ∈ T. Furthermore, if 1μtpt > 0, p is called positive regressive. An n×n matrix valued function At on time scale T is called regressive if Id μtAt is invertible for all t ∈ T. Let a, b be regressive. Define a⊕bt atbtμtatbt and at −at/1 μtat for all t ∈ T. Then the regressive set is a Abelian group and it is not hard to verify that the following properties hold: 1 a a 0; 2 a a; 3 a b a − b/1 μtb; 4 a ⊕ b a ⊕ b, where a, b are regressive. In order to make our paper intelligible, the exponential function on time scales which we will use in our paper is defined as following. For the more general conception of exponential function on time scales, please refer to 17. Definition 2.5. Let p be positive regressive. We define the exponential function by t ep t, s exp s 1 ln 1 μτpτ Δτ , μτ 2.6 for all s, t ∈ T. Here the integral is always understood in Lebesgue’s sense and if f ∈ Crd T, Rn , for any t ∈ T one has σt fsΔs μtft. 2.7 t From the definition of exponential function we have that if a, b are positive regressive, then 1 e0 t, s 1 and ea t, t 1; 4 Discrete Dynamics in Nature and Society 2 ea t, s ea s, t; 3 ea t, sea s, r ea t, r; 4 ea t, seb t, s ea⊕b t, s; Δ 5 ea c, tΔ t −atea c, σt, where ea c, tt stands for the delta derivative of ea c, t with respect to t. Lemma 2.6 L’Hôpital’s Rule. Suppose that f and g are differentiable on T. Then if gt satisfies gt > 0, g Δ > 0 and limt → ∞ gt ∞, then the existence of limt → ∞ f Δ t/g Δ t implies that limt → ∞ ft/gt exists and limt → ∞ ft/gt limt → ∞ f Δ t/g Δ t. 3. Exponential Trichotomy and Lyapunov Function on Time Scales From now on, we always suppose that T is a two-sides infinite time scale and the graininess function μt is bounded, which means that there exists a M > 0 such that 0 ≤ μt ≤ M. We consider the following linear equation xΔ Atx, 3.1 where At is an n × n matrix valued function on time scale T, satisfying that At is rdcontinuous and regressive. To make it simple, in this paper we always require that At is regressive. If At is not regressive, from 20 we can know that the linear evolution operator T t, τ associated to 3.1 is not invertible and exists only for τ ≤ t which will make our paper complicated. Here, T t, τ is called the linear evolution operator if T t, τ satisfies the following conditions: 1 T τ, τ Id; 2 T t, τT τ, s T t, s; 3 the mapping t, τ → T t, τx is continuous for any fixed x ∈ Rn . Equation 3.1 is said to admit an exponential trichotomy on time scale T if there exist projections P t, Qt, Rt : Rn → Rn for each t ∈ T such that P t Qt Rt Id, T t, τP τ P tT t, τ, T t, τQτ QtT t, τ, T t, τRτ RtT t, τ 3.2 for t, τ ∈ T and there are some constants a > b ≥ 0 and D > 1 such that for t ≥ τ, t, τ ∈ T T t, τP τ ≤ Dea τ, t, T t, τRτ ≤ Deb t, τ, 3.3 T t, τRτ ≤ Deb τ, t. 3.4 and for t ≤ τ, t, τ ∈ T T t, τQτ ≤ Dea t, τ, Discrete Dynamics in Nature and Society 5 We say that 3.1 admits a strong exponential trichotomy on time scale T if there exist P t, Qt, Rt : Rn → Rn for each t ∈ T and constants a > b ≥ 0, D > 1 satisfying 3.2, 3.3, 3.4, as well as a constant c with c ≥ a and c > 1 such that T t, τP τ ≤ Dec τ, t, t ≤ τ, t, τ ∈ T, T t, τQτ ≤ Dec t, τ, t ≥ τ, t, τ ∈ T. 3.5 Obviously, if for any t ∈ T we have μt ≡ 0 or μt ≡ 1, then the notion of exponential trichotomy on time scale T becomes the exponential trichotomy for differential equation or difference equation, respectively. Next, we give the notion of Lyapunov function. Consider a function V : T × Rn → Rn . If the following two conditions are satisfied H1 for each τ ∈ T set Vτ : V τ, · and Cs Vτ : {0} ∪ Vτ−1 −∞, 0, Cu Vτ : {0} ∪ Vτ−1 0, ∞. 3.6 let rs and ru be, respectively, the maximal dimensions of linear subspaces inside Cs Vτ and Cu Vτ , then we have rs ru n; H2 for every t ≥ τ, t, τ ∈ T and x ∈ Rn , we have T τ, tCs Vt ⊂ Cs Vτ and T t, τCu Vτ ⊂ Cu Vt , then we say that the function V is a Lyapunov function. Let Hτs : T τ, rCs Vr ⊂ Cs Vτ , r∈T Hτu : T τ, rCu Vr ⊂ Cu Vτ . 3.7 r∈T By the notion of Lyapunov function, we can see that there are subspaces Dτs V and Dτu V such that Dτs V ⊕ Dτu V Rn . The function V is called a λ, μ-Lyapunov function if V is a Lyapunov function and for t ≥ τ, t, τ ∈ T, V satisfies the following conditions: L1 |V t, x| ≤ Nx, L2 for x ∈ Dτs V , V 2 t, T t, τx ≤ eλ t, τV 2 τ, x, L3 for x ∈ Dτu V , V 2 t, T t, τx ≥ eμ t, τV 2 τ, x, L4 for x ∈ Dτs V Dτu V , |V t, x| ≥ x/N, where λ < μ and N is a constant with N > 1. Furthermore, let St for each t ∈ T be a symmetric invertible n × n matrix. Set Ht, x : Stx, x. If V t, x : − sign Ht, x |Ht, x| is a Lyapunov function λ, μLyapunov function, then V t, x is called a quadric Lyapunov function λ, μ-quadric Lyapunov function. Let St : supx∈Rn |Ht, x|/x2 . Now, we state and prove our main results. 6 Discrete Dynamics in Nature and Society Theorem 3.1. Suppose that 3.1 has two λ, μ-quadratic Lyapunov functions, V t, x with r1 , r2 and Wt, x with l1 , l2 satisfying r1 < r2 < 0 < l1 < l2 . If the symmetric invertible n × n matrixes St and T t for V t, x and Wt, x, respectively, satisfy supt∈T St < ∞ and supt∈T T t < ∞, then 3.1 has an exponential trichotomy. In order to prove Theorem 3.1, we need some lemmas. Lemma 3.2. The subspaces Dτs V , Dτu V , Dτs W, and Dτu W according to Lyapunov functions V and W have the following relations Dτs V ⊂ Dτs W, Dτu W ⊂ Dτu V . 3.8 Proof. By the knowledge of mathematical analysis, we can easily get that ln1 lx/x is a decreasing function for any l ∈ R. Thus, when 0 ≤ x ≤ M, we have ln1 lM/M ≤ ln1 lx/x ≤ l. Now, we prove this lemma by contradiction. Suppose that there is x ∈ Dτs V \ Dτs W. Then there are y ∈ Dτs W and z ∈ Dτu W \ {0} such that x y z. Since r1 < l1 < l2 , then by Lemma 2.6 we get 1 1 T t, τx ln|V t, T t, τx| ≥ lim ln t → ∞ t t → ∞ t N lim ≥ lim 1 t → ∞ t ln T t, τz − T t, τy V t, T t, τz 1 ln − N V t, T t, τy t → ∞ t N el2 t, τV τ, z 1 ≥ lim ln − N el1 t, τ V τ, y t → ∞ t N ≥ lim 3.9 1 ln el1 t, τ t → ∞ t 1 t 1 ln 1 μsl1 Δs lim t → ∞ 2t τ μs ≥ lim ≥ lim t → ∞ t − τ ln1 l1 M ln1 l1 M > 0. 2t M 2M But for x ∈ Dτs V , one has 1 1 r1 lim ln|V t, T t, τx| ≤ lim ln < 0. er1 t, τ|V τ, x| ≤ t → ∞ t t → ∞ t 2 3.10 It is a contradiction. So we have Dτs V ⊂ Dτs W. By the concept of Lyapunov function, we get that for t ≤ τ, V τ, x ≤ eλ τ, tV t, T t, τx, V τ, x ≥ eμ τ, tV t, T t, τx, x ∈ Dτs , x ∈ Dτu , 3.11 Discrete Dynamics in Nature and Society 7 where V V or W, λ r1 or l1 and μ r2 or l2 . Using above inequalities and the similar way by which we get Dτs V ⊂ Dτs W, we can directly obtain that Dτu W ⊂ Dτu V . Here we omit the detailed proofing. Thus, the proof is completed. Obviously, from Lemma 3.2 we have Dτu V Dτs W Dτu W {0}, Dτu V ∩ Dτs W Dτs V {0} and Dτu W Dτs V {0}. Furthermore, dim Dτu V Dτs W dim Dτu V dim Dτs W − dimDτu V Dτs W 3.12 n − dim Dτs V n − dim Dτu W − dimDτu V Dτs W. Thus, dim Dτu V Dτs W dim Dτs V dim Dτu W ≥ 2n − dimDτu V Dτs W ≥ n. 3.13 So we obtain Dτu V Dτs W ⊕ Dτs V ⊕ Dτu W Rn . 3.14 Then from the conditions in Theorem 3.1 we have the following results. When t ≥ τ, for x ∈ Dτs V , we get T t, τx ≤ N|V t, T t, τx| ≤ N er1 t, τ|V τ, x| ≤ N 2 er1 t, τx, and for x ∈ Dτu V 3.15 Dτs W, we get T t, τx ≤ N|V t, T t, τx| ≤ N el1 t, τ|V τ, x| ≤ N 2 el1 t, τx. 3.16 When t ≤ τ, for x ∈ Dτu W, we get T t, τx ≤ N|V t, T t, τx| ≤ N el2 t, τ|V τ, x| ≤ N 2 el2 t, τx, and for x ∈ Dτu V 3.17 Dτs W, we get T t, τx ≤ N|V t, T t, τx| ≤ N er2 t, τ|V τ, x| ≤ N 2 er2 t, τx. 3.18 Let PV τ : Rn −→ Dτs V , QV τ : Rn −→ Dτu V , PW τ : Rn −→ Dτs W, QW τ : Rn −→ Dτu W 3.19 8 Discrete Dynamics in Nature and Society be projections. Set P τ : PV τ, Qτ : QW τ and Rτ : PW τ QV τ. From Lemma 3.2, we have QV τPW τ PW τQV τ 0. Thus, by simply deducing, we get that Rτ is also a projection. Set L1 τ Id − P τ. Next, we will give the boundedness of P τ, Qτ, and Rτ. By the conditions of Theorem 3.1, we suppose that there is a constant L > 0 such that supt∈T St ≤ L and supt∈T T t ≤ L. We have the following lemma. Lemma 3.3. P tx ≤ √ 2N 2 x, Qtx ≤ √ 2N 2 x, √ Rtx ≤ 2 2N 2 x. 3.20 Proof. Firstly, one has V t, P tx StP tx, P tx ≥ P tx2 , N2 Qtx2 . V t, L1 tx −StL1 tx, L1 tx ≥ N2 3.21 Since 2 2 N2 1 1 N2 P − L Stx Stx − tx tx 1 2 2 2 2 N N 1 1 N2 2 2 tx tx P L Stx2 1 2 N2 N2 − P tx, Stx L1 tx, Stx 3.22 ≤ StP tx, P tx − StL1 tx, L1 tx N2 Stx2 − P tx, Stx L1 tx, Stx 2 N2 Stx2 , 2 then we obtain N2 N2 Stx P tx ≤ P tx − Stx 2 2 N2 N2 ≤ √ Stx Stx 2 2 √ √ ≤ 2N 2 Stx ≤ 2N 2 Lx. 3.23 Discrete Dynamics in Nature and Society 9 Similarly, we get Qtx ≤ √ 2N 2 Lx, √ Rtx ≤ 2 2N 2 Lx. 3.24 Proof of Theorem 3.1. Firstly, we need to show that there are constants r1 , r2 < 0 and l1 , l2 > 0 such that when t ≥ τ, one has er1 t, τ ≤ er1 t, τ, el1 t, τ ≤ el1 t, τ, 3.25 3.26 and when t ≤ τ, one has el2 t, τ ≤ el2 t, τ, er2 t, τ ≤ er2 t, τ. 3.27 For inequality 3.25, by the definition of exponential function, we only need to find a constant r1 < 0 such that 2 1 μtr1 ≤ 1 μtr1 . 3.28 That means r1 ≤ 2r1 μtr12 . 3.29 Thus, if r1 /2 ≤ r1 < 0, 3.25 always holds. Similarly, we get that if l1 ≥ l1 /2, 0 < l2 < 1 Ml2 − 1/M and − 1 Mr2 − 1/M < r2 < 0, then 3.26, 3.27 always hold, respectively. Here, we notice that since r2 is positive regressive, then 1 Mr2 is significative. By 3.15–3.18 and Lemma 3.3, we get that when t ≥ τ, T t, τP τx ≤ T t, τ|Dτs V P τx √ 2N 2 L er1 t, τx ≤ 2N 2 Ler1 t, τx, T t, τRτx ≤ T t, τ|Dτs W Dτu V Rτx ≤ √ √ √ ≤ 2 2N 2 L el1 t, τx ≤ 2 2N 2 Lel1 t, τx, 3.30 10 Discrete Dynamics in Nature and Society and when t ≤ τ, T t, τQτx ≤ T t, τ|Dτu W Qτx √ 2N 2 L el2 t, τx ≤ 2N 2 Lel2 t, τx, T t, τRτx ≤ T t, τ|Dτs W Dτu V Rτx ≤ √ 3.31 √ √ ≤ 2 2N 2 L er2 t, τx ≤ 2 2N 2 Ler2 t, τx. √ Set D : 2 2N 2 L, a : min{|r1 |, l2 }, and b : max{|r2 |, l1 }. Obviously, we have a > b and the above inequalities imply that for t ≥ τ, t, τ ∈ T one has T t, τP τ ≤ Dea τ, t, T t, τRτ ≤ Deb t, τ, 3.32 T t, τRτ ≤ Deb τ, t. 3.33 and for t ≤ τ, t, τ ∈ T one has T t, τQτ ≤ Dea t, τ, Thus, from the definition of exponential trichotomy on time scales, we get that 3.1 has an exponential trichotomy on time scale T. This completes the proof of Theorem 3.1. Next, we give an approximately converse result of Theorem 3.1. Theorem 3.4. If 3.1 admits a strict exponential trichotomy, then for 3.1 there exist two λ, μLyapunov functions, V t, x with r1 , r2 and W t, x with l1 , l2 satisfying r1 < r2 < 0 < l1 < l2 . Proof. Suppose that 3.1 admits a strict exponential trichotomy with projections P t, Qt, Rt, and constants c ≥ a > b ≥ 0 satisfying 3.2–3.5. Set ∞ St : T σv, tP t∗ T σv, tP tec1 ⊕b σv, tΔv t − t −∞ 3.34 ∗ T σv, tL1 t T σv, tL1 tec2 ⊕b σv, tΔv, where L1 t Qt Rt and b < c2 < c1 < a. Let Ht, x : Stx, x. Discrete Dynamics in Nature and Society 11 When t ≥ τ, t, τ ∈ T, we get Ht, T t, τx ∞ t − ≤ T σv, τP τx2 ec1 ⊕b σv, tΔv t −∞ ∞ T σv, τL1 τx2 ec2 ⊕b σv, tΔv T σv, τP τx2 ec1 ⊕b σv, τΔvec1 ⊕b τ, t 3.35 τ − τ −∞ T σv, τL1 τx2 ec2 ⊕b σv, τΔvec2 ⊕b τ, t ≤ ec2 ⊕b τ, tHτ, x. Set V t, x : −signHt, x |Ht, x|. Then if V τ, x > 0, that is, x ∈ Cu Vτ , we have Hτ, x < 0. By 3.35 we obtain that Ht, T t, τ < 0. Then we get V t, T t, τx > 0, that is, T t, τx ∈ Cu Vt . That means T t, τCu Vτ ⊂ Cu Vt . Similarly, we have T τ, tCs Vt ⊂ Cs Vτ . Thus, V t, x is a Lyapunov function. By 3.3, 3.4 and the characters of exponential function on time scales, one has |Ht, x| ≤ ∞ t ≤ ≤ −∞ T σv, tL1 tx2 ec2 ⊕b σv, tΔv D2 ea t, σvea t, σvec1 ⊕b σv, tΔvx2 t −∞ t −∞ ∞ t t ∞ t T σv, tP tx2 ec1 ⊕b σv, tΔv D2 ea σv, tea σv, tec2 ⊕b σv, tΔvx2 D2 eb t, σveb t, σvec2 ⊕b σv, tΔvx2 D2 eab t, σvΔvx2 t −∞ D Δ 2 −eab t, vv t −∞ −∞ D2 ea⊕a⊕c2 ⊕b σv, tΔvx2 D2 ec2 b σv, tΔvx2 ∞ t t D2 ab Δ − ea⊕a⊕c2 ⊕b t, v v Δvx Δvx2 D a ⊕ a ⊕ c2 ⊕ b −∞ 2 t −ebc2 t, vΔ v Δvx2 . b c2 2 3.36 12 Discrete Dynamics in Nature and Society Since μt ≤ M, then 1 μtb 1 Mb 1 ≤ , ab a−b a−b 1 μta ⊕ a ⊕ c2 ⊕ b −1 1 1 ≤ M ≤ M, 3.37 a ⊕ a ⊕ c2 ⊕ b a ⊕ a ⊕ c2 ⊕ b a ⊕ a ⊕ c2 ⊕ b 2a c2 b 1 μtc2 1 Mc2 −1 ≤ . b c2 c2 − b c2 − b Thus, 1 D2 1 Mb 2 2 1 Mc2 D M D |Ht, x| ≤ x2 . a−b 2a c2 b c2 − b 3.38 Set K : D2 1 Mb/a − b D2 1/2a c2 b M D2 1 Mc2 /c2 − b. Let Hτs V : P τRn and Hτu V : L1 τRn RτRn ⊕ QτRn . For x ∈ Hτs V , one has Ht, x ∞ T σv, tP tx2 ec1 ⊕b σv, tΔv t ∞ 3.39 2 T σv, tx ec1 ⊕b σv, tΔv. t Obviously, one has T t, τ ≤ T t, τP τ T t, τQτ T t, τRτ. 3.40 Then by 3.3–3.5, we obtain that T t, τ ≤ Dea τ, t Deb t, τ Dec t, τ for t ≥ τ, 3.41 T t, τ ≤ Dea t, τ Deb τ, t Dec τ, t for t ≤ τ. 3.42 Discrete Dynamics in Nature and Society 13 Thus, |Ht, x| ≥ ∞ T t, σv2 t ≥ x2 ec1 ⊕b σv, tΔv ∞ x2 Dea t, σv Deb σv, t Dec σv, t2 x2 ∞ ec1 ⊕b σv, tec⊕c t, σvΔv ≥ 9D2 t Δ x2 ∞ − ec⊕cc1 ⊕b t, v v Δv. c ⊕ c c1 ⊕ b 9D2 t ec1 ⊕b σv, tΔv t 3.43 Since c ⊕ c c1 ⊕ b 2c μsc2 − c1 b μsc1 b/1 μsc1 b μsb ≤ 2c − c1 − b Mc2 − Mc1 b, then we get |Ht, x| ≥ x2 . 9D2 2c − c1 − b Mc2 − Mc1 b 3.44 Similarly, for x ∈ Hτu V , we get Ht, x − t −∞ T σv, tx2 ec2 ⊕b σv, tΔv, x2 . |Ht, x| ≥ 9D2 c2 b 2c M2c2 c 2bc c2 M2 bc c2 c2 bc2 M3 bc2 3.45 Let L : max{c2 b 2c M2c2 c 2bc c2 M2 bc c2 c2 bc2 M3 bc2 , 2c − c1 − b Mc2 − Mc1 b}. Then when x ∈ Hτs V Hτu V , we get |Ht, x| ≥ x2 /9D2 L . Let N 2 : max{9D2 L , K}. We get |Ht, x| ≤ K ≤ N 2 x2 for any t ∈ T, x ∈ Rn , and when x ∈ Hτs V Hτu V , we get |Ht, x| ≥ x2 /N 2 . Since c2 < c1 , then we have c2 ⊕ b < c1 ⊕ b and there are constants r1 and r2 with c2 ⊕ b ≤ r2 < r1 ≤ c1 ⊕ b. Thus, when x ∈ Hτs V , for any t ≥ τ we get Ht, T t, τx ∞ T σv, τP τx2 ec1 ⊕b σv, tΔv ≤ ec1 ⊕b τ, tHτ, x, 3.46 t that is, V 2 t, T t, τx ≤ ec1 ⊕b τ, tV 2 τ, x ≤ er1 τ, tV 2 τ, x, 3.47 and when x ∈ Hτu V , for any t ≥ τ we get Ht, T t, τx − t −∞ T σv, τL1 τx2 ec2 ⊕b σv, tΔv ≤ ec2 ⊕b τ, tHτ, x, 3.48 14 Discrete Dynamics in Nature and Society that is, V 2 t, T t, τx ≥ ec2 ⊕b τ, tV 2 τ, x ≥ er2 τ, tV 2 τ, x. 3.49 Therefore, V t, x is a r1 , r2 -Lyapunov function. Set ∞ T t : T σv, tL2 t∗ T σv, tL2 ted1 ⊕b t, σvΔv t − t −∞ 3.50 ∗ T σv, tQt T σv, tQted2 ⊕b t, σvΔv, P t Rt and b < d1 < d2 < a. Let Gt, x : T tx, x and Wt, x : where L2 t − sign Gt, x |Gt, x|. Then Gt, x ∞ t − T σv, tL2 tx2 ed1 ⊕b t, σvΔv 3.51 t 2 −∞ T σv, tQtx ed2 ⊕b t, σvΔv. Let Hτs W : L2 τRn and Hτu W : QτRn . Similar to the consideration for V t, x, we can see that Gt, x is also a Lyapunov function satisfying Gt, T t, τx ≤ ed2 ⊕b t, τGτ, x, 1 1 Mb 1 Ma M |Gt, x| ≤ D2 x2 , a−b 2a d1 b d1 − b |Gt, x| ≥ x2 , 9D2 L for x ∈ Hτs W 3.52 Hτu W, where L max{d1 b 2c M2d1 c 2bc c2 M2 bc d1 c2 bc2 M3 bc2 , 2c − d2 − b Mc2 − Md2 b}. Since d1 < d2 , then we have d1 ⊕ b < d2 ⊕ b and there are constants l1 and l2 with d1 ⊕ b ≤ l1 < l2 ≤ d2 ⊕ b. Thus, for any t ≥ τ, we obtain that when x ∈ Hτs W, Gt, T t, τx ∞ T σv, τL2 τx2 ed1 ⊕b t, σvΔv ≤ ed1 ⊕b t, τGτ, x, 3.53 t that is, W 2 t, T t, τx ≤ ed1 ⊕b t, τW 2 τ, x ≤ el1 t, τW 2 τ, x, and when x ∈ Hτu W, 3.54 Discrete Dynamics in Nature and Society Gt, T t, τx − t −∞ T σv, τL1 τx2 ed2 ⊕b t, σvΔv ≤ ed2 ⊕b t, τGτ, x, 15 3.55 that is, W 2 t, T t, τx ≥ ed2 ⊕b t, τW 2 τ, x ≥ el2 t, τV 2 τ, x. 3.56 Thus, Wt, x is a l1 , l2 -Lyapunov function. According to all the discussions above we get that when 3.1 has a strict exponential trichotomy, there exists two λ, μ-Lyapunov functions, V t, x with r1 , r2 and Wt, x with l1 , l2 satisfying r1 < r2 < 0 < l1 < l2 . This completes the proof. 4. Roughness of Exponential Trichotomy on Time Scales The roughness of exponential dichotomy on time scales had been studied by Zhang and his cooperators in their paper 24 in 2010. In this section, we go further to study the roughness of exponential trichotomy on time scales, using the results which we get from Section 3. We have known that the notion of exponential trichotomy plays a central role when we study center manifolds, so it is important to understand how exponential trichotomy vary under perturbations. Here, we discuss the following linear perturbed equation xΔ Atx Btx, 4.1 where Bt is an n × n matrix valued function on time scale T with Bt ≤ δ. For this linear perturbed 4.1 we get the following theorem. Theorem 4.1. Suppose that 3.1 has a strict exponential trichotomy. If δ > 0 is sufficiently small, then the linear perturbed 4.1 also has an exponential trichotomy. Proof. Since 3.1 has a strict exponential trichotomy, then by Theorem 3.4 there are two λ, μLyapunov functions, V t, x with r1 , r2 and Wt, x with l1 , l2 satisfying r1 < r2 < 0 < l1 < l2 . Here, V t, x and Wt, x are defined in Theorem 3.4. Let xt be a solution of 3.1 satisfying xτ x. Differentiating on both sides of Ht, xt Stxt, xt, one has H Δ t, xt Stxt, xtΔ SΔ txt, xt Sσt xΔ t, xt Sσtxσt, xΔ t SΔ t SσtAt A∗ tSσt μtA∗ tSσtAt xt, xt . 4.2 16 Discrete Dynamics in Nature and Society Let T t, τ be the linear evolution operator related to 3.1. Then we get xt T t, τx. And we can see that P txt P tT t, τx T t, τP τx is also a solution of 3.1. By 4.2 we get H Δ t, P txt SΔ t SσtAt A∗ tSσt μtA∗ tSσtAt P txt, P txt . 4.3 At the same time, by 2.4, 2.7, and 3.46 one has μtH Δ t, P txt Hσt, P σtxσt − Ht, P txt ≤ Ht, P txtec1 ⊕b t, σt − Ht, P txt σt 4.4 1 ln 1 c1 ⊕ bμs Δs − 1 Ht, P txt exp μs t c1 ⊕ bμtHt, P txt. Thus, we obtain that H Δ t, P txt ≤ c1 ⊕ bHt, P txt. 4.5 Then by 4.3 and 4.5 we can see SΔ t SσtAt A∗ tSσt μtA∗ tSσtAt ≤ c1 ⊕ bSt. 4.6 Now, Let yt be a solution of equation xΔ Atx Btx satisfying yτ y. Since yσt yt μtyΔ t yt μtAt Btyt, one has Δ H Δ t, yt Styt, yt SΔ tyt, yt SσtyΔ t, yt Sσtyσt, yΔ t SΔ tyt, yt SσtAtyt, yt A∗ tSσtyt, yt μtA∗ tSσtAtyt, yt SσtBtyt, yt μtSσtBtyt, Atyt Sσtyt, Btyt μtSσtAtyt, Btyt μtSσtBtyt, Btyt 2 ≤ c1 ⊕ bH t, yt SσtBtyt 2 2 2μtAtSσtBtyt SσtBtyt 2 μtSσtBt2 yt . 4.7 Discrete Dynamics in Nature and Society 17 By the variation of constants formula, we have T t, s Id by 2.7 we obtain T σt, t I σt t s AvT v, sΔv. Thus, AvT v, tΔv 4.8 t Id μtAtT t, t Id μtAt. Therefore, by Definition 2.5 and 2.7, 3.41, we obtain Id μtAt ≤ Dea t, σt Deb σt, t Dec σt, t D 1 μta D 1 μtb D 1 μtc 4.9 ≤ 3D DMb c. Thus, one has μtAt ≤ Id μtAt Id ≤ 3D DMb c 1. 4.10 From Theorem 3.4, we know St sup Ht, x ≤ N2, xt2 yt ≤ N V t, yt . 4.11 x∈Rn Thus, H Δ t, yt ≤ c1 ⊕ bH t, yt 2N 2 δ 23D 1 DMb DMcN 2 δ 4.12 MN δ N H t, yt . 2 2 2 Let Mδ : 2N 2 δ 23D 1 DMb DMcN 2 δ MN 2 δ2 N 2 . It is not hard to see that Mδ → 0 when δ → 0. Set Eδ : c1 ⊕ b Mδ. Then when δ is sufficiently small, we have Eδ < 0 and 1 μtEδ 1 μt − c1 ⊕ b Mδ 1 μtc1 ⊕ b ! 1 μtMδ > 0. 1 μtc1 ⊕ b 4.13 So Eδ is positive regressive. Then for Ht, yt > 0, 4.12 can be written as H Δ t, y ≤ EδH t, yt . 4.14 18 Discrete Dynamics in Nature and Society Notice that Δ H t, yt eEδ t, τ H Δ t, yt eEδ σt, τ EδeEδ t, τH t, yt H Δ t, yt eEδ σt, τ EδeEδ t, σteEδ σt, τH t, yt H Δ t, yt eEδ σt, τ 4.15 1 eEδ σt, τH t, yt 1 Eδμt H Δ t, y − EδH t, yt eEδ σt, τ ≤ 0. Eδ Thus, we get H t, yt ≤ eEδ t, τH τ, yτ . 4.16 Then we can see that if yt ∈ Cs Vt , we have V t, yt < 0 and thus Ht, yt > 0. By 4.16 we have Hτ, yτ > 0 and thus V τ, yτ < 0, that is, yτ ∈ Cs Vτ . Let Ut, τ be the linear evolution operator associated to 4.1. Then yt Ut, τyτ. Thus, we get Uτ, tCs Vt ⊂ Cs Vτ . 4.17 Let L1 t Qt Rt. We can see that L1 txt is also a solution of 3.1. By 4.2 one has H Δ t, L1 txt SΔ t SσtAt A∗ tSσt μtA∗ tSσtAt L1 txt, L1 txt . 4.18 Similar to the proof of 4.5, we have H Δ t, L1 txt ≤ c2 ⊕ bHt, L1 txt. Then we get SΔ t SσtAt A∗ tSσt μtA∗ tSσtAt ≤ c2 ⊕ bSt. 4.19 Following the process for getting 4.12, one has H Δ t, yt ≤ c2 ⊕ bH t, yt MδH t, yt . 4.20 Let Fδ : c2 ⊕ b − Mδ. When δ is sufficiently small, we obtain Fδ < 0 and 1 μtFδ 1 μt − c2 ⊕ b − Mδ 1 μtc2 ⊕ b 1 − Mδ. 1 μta ⊕ b ! 4.21 Discrete Dynamics in Nature and Society 19 So Fδ is positive regressive. Then for Ht, yt < 0, 4.20 can be written as H Δ t, y ≤ FδH t, yt . 4.22 Following the similar process as 4.15, we get H t, yt ≤ eFδ t, τH τ, yτ . 4.23 Then if yτ ∈ Cu Vτ , we have V τ, yτ > 0 and thus Hτ, yτ < 0. By 4.23 we have Ht, yt < 0 and thus V t, yt > 0, that is, yt ∈ Cu Vt . So we get Ut, τCu Vτ ⊂ Cu Vt . 4.24 By the notion of Lyapunov function and 4.17, 4.24 we can see that V t, x is a Lyapunov function of 4.1. Let Hτu V : Uτ, rCu Vr ⊂ Cu Vτ , r∈T Hτs V : 4.25 Uτ, rCs Vr ⊂ Cs Vτ . r∈T By the notion of Lyapunov function, we know that there exist two subspaces Dτu V and Dτs V such that Dτu V ⊂ Hτu V , Dτs V ⊂ Hτs V , and Dτu V ⊕ Dτs V Rn . Now, we prove that V t, x is a λ, μ-Lyapunov function. Since c1 > c2 , we have c1 ⊕b > c2 ⊕ b. By simple computation, we obtain c1 ⊕ b < c2 ⊕ b. Therefore, when δ is sufficiently small, we get Eδ < Fδ. Then there are constants r1 , r2 such that Eδ ≤ r1 < r2 ≤ Fδ. Thus, by inequality 4.16, for yτ ∈ Dτs one has V 2 t, yt ≤ eEδ t, τV 2 τ, yτ ≤ er1 t, τV 2 τ, yτ . 4.26 By inequality 4.23, for yτ ∈ Dτu , we obtain V 2 t, yt ≥ eFδ t, τV 2 τ, yτ ≥ er2 t, τV 2 τ, yτ . 4.27 Then by 4.26 and 4.27 we can see that V t, y is a r1 , r2 -Lyapunov function of 4.1. Next, we will prove that Wt, y is also a λ, μ-Lyapunov function of 4.1. Similar to the consideration of Ht, y, for Gt, y we have that if y ∈ Cs Wt , then Wt, yt < 0, that is, Gt, yt > 0. Thus, we get GΔ t, yt ≤ d1 ⊕ b MδG t, yt . 4.28 20 Discrete Dynamics in Nature and Society Let E δ : d1 ⊕ b Mδ. By the similar deducing as 4.16, one has G t, yt ≤ eE δ t, τG τ, yτ . 4.29 Thus, when Gt, yt > 0, that is, Wt, yt < 0, we have Gτ, yτ > 0, that is, Wτ, yτ < 0. That means Uτ, tCs Wt ⊂ Cs Wτ . 4.30 When y ∈ Cu Wt , we have Wt, yt > 0, that is, Gt, yt < 0. Then GΔ t, yt ≤ d2 ⊕ b − MδG t, yt . 4.31 Let F δ d2 ⊕ b − Mδ. By the similar deducing as 4.15, one has G t, yt ≤ eF δ t, τG τ, yτ . 4.32 Then when Gτ, yτ < 0, that is, Wτ, yτ > 0, we have Gt, yt < 0, that is, Wτ, yτ > 0. That means Ut, τCu Wτ ⊂ Cu Wt . 4.33 By 4.30 and 4.33, we know that Wt, y is a Lyapunov function of 4.1. Since d1 < d2 , we have d1 ⊕ b < d2 ⊕ b. Then when δ is sufficiently small, we get E δ < F δ. Therefore, there are constants l1 , l2 such that E δ ≤ l1 < l2 ≤ F δ. Let Hτu W : Uτ, rCu Wr ⊂ Cu Wτ , r∈T Hτs W : 4.34 Uτ, rCs Wr ⊂ Cs Wτ . r∈T By the notion of Lyapunov function, we can know that there exist Dτu W and Dτs W such that Dτu W ⊂ Hτu W, Dτs W ⊂ Hτs W, and Dτu W ⊕ Dτs W Rn . Furthermore, by 4.29 and 4.32, for any t ≥ τ, when yτ ∈ Dτs W, one has W 2 t, yt ≤ eE δ t, τW 2 τ, yτ ≤ el1 t, τW 2 τ, yτ , 4.35 and when yτ ∈ Dτu W, one has W 2 t, yt ≥ eF δ t, τW 2 τ, yτ ≥ el2 t, τW τ, yτ . 4.36 Thus, we can see that Wt, y is a l1 , l2 -Lyapunov function of 4.1. Therefore, we know that 4.1 has two λ, μ-quadratic Lyapunov functions, V t, x with r1 , r2 and Wt, x with Discrete Dynamics in Nature and Society 21 l1 , l2 satisfying r1 < r2 < 0 < l1 < l2 . Then by Theorem 3.1 we get that 4.1 has an exponential trichotomy. The proof is completed. Acknowledgment The author expresses her sincere thanks to Professors Yong Li and Zhenxin Liu for their encouragement and valuable suggestions. References 1 R. J. Sacker and G. R. Sell, “Existence of dichotomies and invariant splittings for linear differential systems. III,” Journal of Differential Equations, vol. 22, no. 2, pp. 497–522, 1976. 2 S. Elaydi and O. Hájek, “Exponential trichotomy of differential systems,” Journal of Mathematical Analysis and Applications, vol. 129, no. 2, pp. 362–374, 1988. 3 S. Elaydi and O. Hájek, “Exponential dichotomy and trichotomy of nonlinear differential equations,” Differential and Integral Equations, vol. 3, no. 6, pp. 1201–1224, 1990. 4 G. Papaschinopoulos, “On exponential trichotomy of linear difference equations,” Applicable Analysis, vol. 40, no. 2-3, pp. 89–109, 1991. 5 J. Hong, R. Obaya, and A. S. Gil, “Exponential trichotomy and a class of ergodic solutions of differential equations with ergodic perturbations,” Applied Mathematics Letters, vol. 12, no. 1, pp. 7–13, 1999. 6 J. Hong, R. Obaya, and A. Sanz, “Existence of a class of ergodic solutions implies exponential trichotomy,” Applied Mathematics Letters, vol. 12, no. 4, pp. 43–45, 1999. 7 L. Barreira and C. Valls, “Robustness of nonuniform exponential trichotomies in Banach spaces,” Journal of Mathematical Analysis and Applications, vol. 351, no. 1, pp. 373–381, 2009. 8 L. Barreira and C. Valls, “Lyapunov functions for trichotomies with growth rates,” Journal of Differential Equations, vol. 248, no. 1, pp. 151–183, 2010. 9 S.-N. Chow and Y. Yi, “Center manifold and stability for skew-product flows,” Journal of Dynamics and Differential Equations, vol. 6, no. 4, pp. 543–582, 1994. 10 A. Kelley, “The stable, center-stable, center, center-unstable, unstable manifolds,” Journal of Differential Equations, vol. 3, pp. 546–570, 1967. 11 X. Liu, “Exponential trichotomy and homoclinic bifurcation with saddle-center equilibrium,” Applied Mathematics Letters, vol. 23, no. 4, pp. 409–416, 2010. 12 B. Sasu and A. L. Sasu, “Exponential trichotomy and p-admissibility for evolution families on the real line,” Mathematische Zeitschrift, vol. 253, no. 3, pp. 515–536, 2006. 13 A. L. Sasu and B. Sasu, “Exponential trichotomy for variational difference equations,” Journal of Difference Equations and Applications, vol. 15, no. 7, pp. 693–718, 2009. 14 D. Zhu, “Exponential trichotomy and heteroclinic bifurcations,” Nonlinear Analysis: Theory, Methods & Applications, vol. 28, no. 3, pp. 547–557, 1997. 15 S. Hilger, Ein Maßkettenkalk ül mit Anwendung auf Zentrumsmannigfaltigkeiten, Ph.D. thesis, University of Würzburg, Würzburg, Germany, 1988. 16 B. Aulbach and S. Hilger, “A unified approach to continuous and discrete dynamics, Differential equations: qualitative theory,” in ColloquiaMathematica Societatis J’anos Bolyai, vol. 48, North-Holland, Amsterdam, The Netherlands, 1991. 17 M. Bohner and A. Peterson, Dynamic Equations on Time Scales: An Introduction with Applications, Birkhäuser, Boston, Mass, USA, 2001. 18 M. Błaszak, B. Silindir, and B. M. Szablikowski, “The R-matrix approach to integrable systems on time scales,” Journal of Physics A, vol. 41, no. 38, Article ID 385203, 18 pages, 2008. 19 S. Hilger, “Analysis on measure chains—a unified approach to continuous and discrete calculus,” Results in Mathematics, vol. 18, no. 1-2, pp. 18–56, 1990. 20 C. Pötzsche, “Exponential dichotomies for dynamic equations on measure chains,” Nonlinear Analysis: Theory, Methods & Applications, vol. 47, no. 2, pp. 873–884, 2001. 21 C. Pötzsche, “Exponential dichotomies of linear dynamic equations on measure chains under slowly varying coefficients,” Journal of Mathematical Analysis and Applications, vol. 289, no. 1, pp. 317–335, 2004. 22 Discrete Dynamics in Nature and Society 22 Y. Xia, J. Cao, and M. Han, “A new analytical method for the linearization of dynamic equation on measure chains,” Journal of Differential Equations, vol. 235, no. 2, pp. 527–543, 2007. 23 M. Bohner and A. Peterson, Advances in Dynamic Equations on Time Scales, Birkhäuser, Boston, Mass, USA, 2003. 24 J. Zhang, M. Fan, and H. Zhu, “Existence and roughness of exponential dichotomies of linear dynamic equations on time scales,” Computers & Mathematics with Applications, vol. 59, no. 8, pp. 2658–2675, 2010. Advances in Operations Research Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Advances in Decision Sciences Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Mathematical Problems in Engineering Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Journal of Algebra Hindawi Publishing Corporation http://www.hindawi.com Probability and Statistics Volume 2014 The Scientific World Journal Hindawi Publishing Corporation http://www.hindawi.com Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 International Journal of Differential Equations Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Volume 2014 Submit your manuscripts at http://www.hindawi.com International Journal of Advances in Combinatorics Hindawi Publishing Corporation http://www.hindawi.com Mathematical Physics Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Journal of Complex Analysis Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 International Journal of Mathematics and Mathematical Sciences Journal of Hindawi Publishing Corporation http://www.hindawi.com Stochastic Analysis Abstract and Applied Analysis Hindawi Publishing Corporation http://www.hindawi.com Hindawi Publishing Corporation http://www.hindawi.com International Journal of Mathematics Volume 2014 Volume 2014 Discrete Dynamics in Nature and Society Volume 2014 Volume 2014 Journal of Journal of Discrete Mathematics Journal of Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Applied Mathematics Journal of Function Spaces Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Optimization Hindawi Publishing Corporation http://www.hindawi.com Volume 2014 Hindawi Publishing Corporation http://www.hindawi.com Volume 2014