Asymptotical Stability of Fractional Order Systems with Time Delay Via an Integral Inequality
Bin‐Bin He,Hua‐Cheng Zhou,YangQuan Chen,Chunhai Kou
DOI: https://doi.org/10.1049/iet-cta.2017.1144
IF: 2.67
2018-01-01
IET Control Theory and Applications
Abstract:IET Control Theory & ApplicationsVolume 12, Issue 12 p. 1748-1754 Research ArticleFree Access Asymptotical stability of fractional order systems with time delay via an integral inequality Bin-Bin He, Bin-Bin He College of Information Science and Technology, Donghua University, Shanghai, 201620 People's Republic of ChinaSearch for more papers by this authorHua-Cheng Zhou, Hua-Cheng Zhou School of Mathematics and Statistics, Central South University, Changsha, 410075 People's Republic of ChinaSearch for more papers by this authorYangQuan Chen, Corresponding Author YangQuan Chen ychen53@ucmerced.edu orcid.org/0000-0002-7422-5988 Mechatronics, Embedded Systems and Automation Lab, University of California, Merced, 95343 CA, USASearch for more papers by this authorChun-Hai Kou, Chun-Hai Kou Department of Applied Mathematics, Donghua University, Shanghai, 201620 People's Republic of ChinaSearch for more papers by this author Bin-Bin He, Bin-Bin He College of Information Science and Technology, Donghua University, Shanghai, 201620 People's Republic of ChinaSearch for more papers by this authorHua-Cheng Zhou, Hua-Cheng Zhou School of Mathematics and Statistics, Central South University, Changsha, 410075 People's Republic of ChinaSearch for more papers by this authorYangQuan Chen, Corresponding Author YangQuan Chen ychen53@ucmerced.edu orcid.org/0000-0002-7422-5988 Mechatronics, Embedded Systems and Automation Lab, University of California, Merced, 95343 CA, USASearch for more papers by this authorChun-Hai Kou, Chun-Hai Kou Department of Applied Mathematics, Donghua University, Shanghai, 201620 People's Republic of ChinaSearch for more papers by this author First published: 01 August 2018 https://doi.org/10.1049/iet-cta.2017.1144Citations: 31AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract In this study, the asymptotical stability for several classes of fractional order differential systems with time delay is investigated. The authors first present an integral inequality by which the Halanay inequality is extended to fractional order case. Based on the generalised Halanay inequality, the authors establish several asymptotical stability conditions under which the fractional order systems with time delay are asymptotically stable. It is worth to note that these stability conditions are easy to check without resorting to the solution expression of the systems. 1 Introduction Fractional calculus, more precisely, arbitrary order calculus, appears almost at the same time with the classical calculus. However, it did not attract a great attention due to the lack of application background and the complexity of computation, until the last few decades when researchers found that there are many anomalies that cannot be explained by classical calculus while fractional calculus is a useful tool to describe some abnormal diffusion process because of its memory and hereditary properties, then it is widely used in many fields such as blood flow phenomena, electrochemical processes, viscoelastic materials and so on [1–3]. In 1991, Oustaloup proposes a CRONE control, and in 1999, Podlubny [4] introduces the controller, fractional calculus has been used in control theory, and the researchers found fractional order controllers provide superior performances in both theory and practice, and many results are obtained (see [5–8]). Asymptotical stability analysis of fractional order systems, one of the fundamental issues in control theory, is to find some stability conditions under which systems are asymptotically stable. For linear fractional order systems, the first stability result is the well-known Matignon theorem [9]. This theorem enables us to determine the stability of linear fractional order systems through the location in the complex plane of the dynamic matrix eigenvalues of the state. Recently, many mathematical tools have been applied to analyse the asymptotical stability in fractional order linear systems. These include linear matrix inequality (LMI) approach [10], the Laplace transform method [11], the Lyapunov method [12] and the Riesz basis approach and the semigroup method [13]. In [14], it introduces the characteristic equation of the system to give some simple sufficient asymptotical stability conditions for interval linear fractional order neutral system with time delay. Laplace transform method is developed in [11], where a n -dimensional linear fractional order differential equation with multiple time delays is addressed and several sufficient conditions of globally asymptotically stable are exploited. For non-linear fractional order systems, the Lyapunov method is a classical approach to deal with the stability problem. The key advantage of applying the Lyapunov method in obtaining the stability criteria for a given system is that it does not need to solve the system to obtain the explicit solution expression and the need is just the system structure. However, due to the memory effect and the weakly singular kernels of the fractional order derivative, the fractional Lyapunov method which is different from the classical Lyapunov method was not developed until 2009 in [12] and 2010 in [15], and the applicability of the fractional Lyapunov method was not available until 2014 in [16]. The Mittag–Leffler stability concept is introduced in [12] and this kind of stability implies the asymptotical stability. In [16], a simple but useful fractional order differential inequality is established, which proves the quadratic function, , to be a good Lyapunov candidate function. Very recently, a class of convex and positive definite function as Lyapunov candidate function is developed in [17], which makes the fractional Lyapunov theorem more applicable. In [18], an indirect approach to obtain the stability for non-linear fractional order systems is established by using the relation of the stability between the fractional order systems and the corresponding integer order systems. For the stability of fractional order system with time delay (FSTD), Mittag–Leffler stability is addressed in [19]. In [20, 21], frequency domain method is used to obtain the asymptotical stability results. Razumikhin method was generalised to fractional order systems with delay in [22]. However, to the best of our knowledge, there are few easily verifiable asymptotical stability criteria for FSTD. In this paper, we are concerned with the asymptotical stability for FSTD via a new integral inequality. Based on this inequality, we generalise the Halanay inequality proposed by Aristide Halanay in [23]. With these inequalities and the result of [16], we establish some stability criteria that are easy to check. The paper is organised as follows. Section 2 presented some basic concepts and lemmas about fractional calculus. A new integral inequality and a generalised Halanay inequality are introduced in Section 3. Section 4 is devoted to the asymptotical stability for several kinds of fractional order systems with time delay. Section 5 is a conclusion about the paper. 2 Preliminary In this section, some basic definitions and preliminaries are given which are useful throughout this paper. Definition 1.The Caputo's fractional derivative of order for a function is defined by [4, p. 79] with , where is the m -order derivative of , and is the Gamma function. In particular, when , we have Definition 2.The one-parameter Mittag–Leffler function and two-parameter Mittag–Leffler function are defined by [4, p. 16] respectively, where . Let be a constant, we suppose that . For any , we denote by an element of defined by , . Consider the FSTD (1)where is the state vector, is a non-linear functional vector that satisfies , is the continuous vector valued history function and , where denotes the Euclidean norm for vectors. Definition 3.For system (1), the trivial solution is said to be [24, p. 4] stable, if for any given there exists a such that implies for all . asymptotically stable, if it is stable and if in addition there exists a such that if , then as . unstable, if it is not stable. Lemma 1.Let be a continuous and differentiable function [16], then, for any , If , it holds that for and , Lemma 2.The Cauchy problem for fractional order time-delay system of order (2)is equivalent to the following integral equation: (3)where and is a continuous function. Proof.Taking the Laplace transform on both side of (2) gives (4)where are, respectively, the Laplace transform of functions . Therefore (5)Taking the inverse Laplace transform on both side of (5) yields where denotes the convolution operator, that is, for For integral equation (3), it equals to (6)Taking Laplace transform on both side of (6) and then taking the inverse Laplace transform, we can transfer (3) to equation (2). This ends the proof. □ 3 Integral inequality In this section, we establish an integral inequality which can be used to deal with the asymptotical stability for fractional order systems with time delay. The proof is based on ' ' method. Theorem 1.Let be bounded on and continuous on . Suppose that are two continuous functions which satisfy , , and . is a constant and the following inequality holds: (7)If , then, . Proof.This proof is divided into two steps. Step 1: We show that is bounded for all . Indeed, since for any , then , by (7), we obtain (8)which, together with the boundedness of , gives Noting that we get (9)Since , it follows from (9) that (10)that is, is bounded on . Step 2: We show that . Denote . Since is non-negative and is bounded on just proved in step 1, is well-defined and is non-increasing with respect to t, which implies that exists. Moreover, . Thus, to show , it suffices to prove . Indeed, for any given , there exists such that for all . It follows from (7) and (10) that (11)By the definition of and (11), we get for (12)From and , we have (13)Taking the limitation on both side of (12) and noting that , by (13), we derive which, jointly with , yields . By the arbitrariness of , we get . Proof is completed. Several remarks about Theorem 1 are given in order. Remark 1.In general, the inequality (7) does not yield that tends to zero, exponentially. For example, consider the following integral equation: (14)where for , and We claim that the solution of (14) satisfies the inequality (7). Indeed, by Lemma 2, it is seen that the integral equation (14) is actually equivalent to the following fractional order differential equation with time delay: (15)The solution is explicitly found to be (16)which gives . On the other hand, since is monotone decreasing on and for , we have . Thus, by (14), satisfies (7) with , , , and . From the proof of Corollary 1 below, we get , which implies satisfies (7). However, by (16), is asymptotically convergent to zero and cannot decay exponentially as t goes to infinity. Remark 2.If and decay to zero exponentially, then may tend to zero exponentially. To illustrate this, we consider the following integral equation: (17)where for . Similar to Remark 1, the solution of (17) satisfies the inequality (7) with , . This system (17) is equivalent to the following first-order differential equation with time delay: (18)A simple computation shows the solution of (18) is explicitly given by for all . We can see that and decays to zero exponentially. Fig. 1 shows the evolutions of the solution of system (15) with and the solution of system (18). It is seen that the solution of system (18) decays quicker than the solution of system (15). Fig. 1Open in figure viewerPowerPoint Evolutions of the solutions of systems (15) and (18) By using Theorem 1, we arrive at the following corollary. Corollary 1.Let be bounded on and continuous on . Assume that for some positive constants , the following inequality holds: (19)where . Then, . Proof.By (19), we know that there exists a non-negative function satisfying (20)Taking the Laplace transform on both sides of (20) gives (21)where (22)are, respectively, the Laplace transform of the functions , and . Therefore, by (21), we have that (23)Taking the inverse Laplace transform on both sides of (23) yields (24)where denotes the convolution operator. Since , and are non-negative functions, it follows that (25)We can see that (25) is the form of (7) in Theorem 1 with and . It is obvious that , and . To show , it suffices to prove that . Indeed, it follows from [25, p. 50, formula 1.10.7] that which gives us (26)It follows from [25, p. 43, formula 1.8.28] that with , which yields On the other hand, from (26), since is non-decreasing with respect t, so does for . Thus (27)which, jointly with , implies (28)Thus, it follows from Theorem 1 that . □ Remark 3.The above corollary can be regarded as a generalisation of Halanay inequality [23, p. 378, Lemma], where and tends to zero, exponentially, as t goes to infinity.However, when , we generally cannot expect that decay exponentially because there is a memory effect in the equation due to the tail of time. A typical example can be seen in Remark 1, system (15) satisfies the conditions in Corollary 1, the solution of (15) satisfies but cannot decay exponentially. 4 Asymptotical stability In this section, we investigate the asymptotical stability for five classes of fractional order systems with time delay. Example 1.Consider the following fractional order differential system: (29)where , is the state, a, b are two constants and is a continuous function and satisfies for .In order to obtain the stability condition of (29), we denote and choose two constants . Finding Caputo's derivative of with respect to t along the solution of (29) yields (30)if (31)which is equivalents to and , that is, and . Since , LMI (31) reduces to . Therefore, if , then LMI (31) is feasible, and by Corollary 1, the solution of (29) is asymptotically stable.In system (29), taking , , the numerical simulations of this system with delays and , respectively, are plotted in Fig. 2, from which, we can see that system (29) with both and are asymptotically stable and the solution of the system (29) with time-varying delay decays faster than that of the system with constant delay. Taking and the rest of parameters as before, the numerical simulations plotted in Fig. 3 show that system (29) with both and are still asymptotically stable. The solution of the system (29) with time-varying delay decays faster than that of the system with constant delay.The stability for system (29) is considered in [22], where the author used the Lyapunov-like theorem to prove the asymptotical stability result. Fig. 4 shows the regions of that make the system (29) to be asymptotically stable. Proposition 1.Suppose that , then given , system (29) with is asymptotically stable for all fast-varying delay . Example 2.Now we consider the following fractional order differential system: (32)where , is the state, are constant matrices, and is a continuous function satisfies for .Similar to Example 1, let Lyapunov function . Then if (33) Fig. 2Open in figure viewerPowerPoint System (29) with , , , and with delay and Fig. 3Open in figure viewerPowerPoint System (29) with , , , and with delay and . Fig. 4Open in figure viewerPowerPoint Regions of which make the system (29) to be asymptotically stable From Corollary 1, we have the following proposition. Proposition 2.Suppose that there exist two positive constants such that the LMI (33) is feasible. Then, for any given , system (32) with is asymptotically stable for all fast-varying delay . Remark 4.When the delay is independent of time, i.e. for some constant , the asymptotical stability of (32) is considered in [11], where the Laplace transform method is used. Note that when the delay is dependent of time, the asymptotical stability cannot follow from the Laplace transform method. Example 3.Consider the following fractional order differential system with distributed delay: (34)where , are the states, is a constant, and , are continuous functions. Proposition 3.Suppose that the function h satisfies with some and there exist three positive constants and with such that (35)and (36)Then, the system (34) with , is asymptotically stable. Proof.Let be a Lyapunov function. Finding Caputo's derivative of with respect to t along the solution of (34) gives (37)Noting the fact that , by Young's inequality, and (35)–(37), we get which, jointly with Corollary 1, implies that as . Thus, the system (34) is asymptotically stable. □ Corollary 2.Suppose that the functions f,g,h satisfy and with some and there exist three positive constants and with such that (38)and (39)Then, the system (34) with , is asymptotically stable. Proof.By proposition 3, it suffices to prove that (36) holds. By conditions (38), we have which implies that (36) holds. This ends the proof. □ Example 4.Consider the following fractional order non-linear system with distributed delay: (40)where , , , the initial function is . are two continuous vector-valued functions such that (40) has a global solution on . Proposition 4.Suppose that there exist two positive constants such that (41)Then the system (40) with is asymptotically stable. Proof.Let be the Lyapunov function. By Lemma 1, computing the Caputo's derivative of along the solution of (40) gives (42)Since , we have Thus, it follows from (42) that which, jointly with Corollary 1, implies that as . Thus, system (40) is asymptotically stable. □ Corollary 3.Suppose that the function g satisfies with some and there exist three positive constants , and with such that (43)Then the system (40) with is asymptotically stable. Proof.By Proposition 4, it suffices to show that (41) holds. Indeed, by Young's inequality (43) and the fact that , we have which proves that (41) holds. This ends the proof. □ Example 5.Consider the fractional order non-linear system with distributed delay governed by (44)where , , the initial function is . The time delay functions are continuous. Functions are continuous such that (44) has a global solution on . Proposition 5.For any , suppose that functions satisfy with some , , and there exist three positive constants , and with such that (45)where . Then, given , the system (44) with is asymptotically stable. Proof.The Proposition can be established by mimicking the proof of Corollary 3. We leave details to the interested reader. □ Remark 5.The key idea of Lyapunov method to get the stability is that we do not need to solve the differential systems but just use the structure of systems. Clearly, the conditions (36), (43) and (45) do not involve the solution of (34), (40) and (44), respectively, thus (36), (43) and (45) are easy to verify for application purpose. The results here could provide some insights into the qualitative analysis of fractional order systems with time delay. 5 Concluding remarks In this paper, we have obtained a new integral inequality, which is our novelty. Applying this integral inequality and Laplace transform, Halanay inequality is extended to Caputo fractional order case. Several examples validate that this generalised Halanay inequality can be easily applied to obtain the asymptotical stability conditions for fractional order systems with time delay, where the delay is bounded. If the delay is unbounded, a new technique to prove the integral inequality should be developed. Stability of fractional order systems with time-varying delay or time constant delay is a wide open and fertile area for future research. The presented integral inequality can be used to obtain the stability of Caputo fractional order systems with multiple time delays. 6 Acknowledgments This work is partially supported by the National Natural Science Foundation of China (no. 11526164) and the Fundamental Research Funds for the Central Universities (no. CUSF-DH-D-2017083). 7 References 1Mainardi F.: ' Fractional calculus and wave in linear viscoelasticity: an introduction to mathematical models' ( World Scientific, Singapore, 2010) 2Monje C.A. Chen Y.Q., and Vinagre B.M. et al.: ' Fractional-order systems and controls: fundamentals and applications' ( Springer-Verlag, London, 2010) 3Sun H.G. Chen W., and Chen Y.Q.: 'Variable-order fractional differential operators in anomalous diffusion modeling', Phys. A, 2009, 388, (21), pp. 4586– 4592 4Podlubny I.: ' Fractional differential equations' ( Academic Press, San Diego, 1999) 5Zhou X.: 'Integral-type approximate controllability of linear parabolic integro-differential equations', Syst. Control Lett., 2017, 105, pp. 44– 47 6Mur T., and Henríquez H.R.: 'Controllability of abstract systems of fractional order', Frac. Calc. Appl. Anal., 2015, 18, (6), pp. 1379– 1398 7Wei Y. Chen Y., and Cheng S. et al.: 'Completeness on the stability criterion of fractional order LTI systems', Frac. Calc. Appl. Anal., 2017, 20, (1), pp. 159– 172 8Wang Z. Huang X., and Shi G.: 'Analysis of nonlinear dynamics and chaos in a fractional order financial system with time delay', Comput. Math. Appl., 2011, 62, (3), pp. 1531– 1539 9Matignon D.: 'Stability results for fractional differential equations with applications to control processing', Computat. Eng. Syst., 1996, 2, pp. 963– 968 10Sabatier J. Moze M., and Farges C.: 'LMI stability conditions for fractional order systems', Comput. Math. Appl., 2010, 59, (5), pp. 1594– 1609 11Deng W.H. Li C.P., and Lü J.H.: 'Stability analysis of linear fractional differential system with multiple time delays', Nonlinear Dyn., 2007, 48, (4), pp. 409– 416 12Li Y. Chen Y.Q., and Podlubny I.: 'Mittag–Leffler stability of fractional order nonlinear dynamic systems', Automatica, 2009, 45, (8), pp. 1965– 1969 13Zhou H.C., and Guo B.Z.: 'Boundary feedback stabilization for an unstable time fractional reaction diffusion equation', SIAM J. Control Optim., 2018, 56, pp. 75– 101 14Li H. Zhong S.M., and Li H.B.: 'Stability analysis of linear fractional-order neutral systems with time delay', Sci. J. Circuits Syst., 2017, 6, (1), pp. 1– 5 15Li Y. Chen Y.Q., and Podlubny I.: 'Stability of fractional-order nonlinear dynamic systems: Lyapunov direct method and generalized Mittag–Leffler stability', Comput. Math. Appl., 2010, 59, (5), pp. 1810– 1821 16Aguila-Camacho N. Duarte-Mermoud M., and Gallegos J.A.: 'Lyapunov functions for fractional order systems', Commun. Nonlinear Sci. Numer. Simul., 2014, 19, (9), pp. 2951– 2957 17Chen W.S. Dai H., and Song Y.F. et al.: 'Convex Lyapunov functions for stability analysis of fractional order systems', IET Control Theory Appl., 2017, 11, (7), pp. 1070– 1074 18Chen Y. Wei Y., and Zhou X. et al.: 'Stability for nonlinear fractional order systems: an indirect approach', Nonlinear Dyn., 2017, 89, (2), pp. 1011– 1018 19Sadati S.J. Baleanu D., and Ranjbar A. et al.: 'Mittag–Leer stability theorem for fractional nonlinear systems with delay', Abstr. Appl. Anal., 2010, 2010, Article ID 108651, doi: 10.1155/2010/108651 20Bonnet C., and Partington J.R.: 'Analysis of fractional delay systems of retarded and neutral type', Automatica, 2002, 38, (7), pp. 1133– 1138 21Bonnet C., and Partington J.R.: 'Stabilization of some fractional delay systems of neutral type', Automatica, 2007, 43, (12), pp. 2047– 2053 22Stamova I.M.: 'On the Lyapunov theory for functional differential equations of fractional order', Proc. Am. Math. Soc., 2016, 144, pp. 1581– 1593 23Halanay A.: ' Differential equations: stability, oscillations, time lags' ( Academic Press, New York, 1966) 24Cesari L.: ' Asymptotic behavior and stability problems in ordinary differential equations' ( Academic Press, New York, 1963) 25Kilbas A.A. Srivastava H.M., and Trujillo J.J.: ' Theory and applications of fractional differential equations' ( Elsevier, Amsterdam, 2006) Citing Literature Volume12, Issue12August 2018Pages 1748-1754 FiguresReferencesRelatedInformation