Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function

By the method of combining the matrix-valued Lyapunov functional and comparison theorem, connected Lyapunov stability and practical stability of large scale delay system are studied deeply. A series of new sufficient conditions are proposed. These results are not only of theoretical but also of prac...

Повний опис

Збережено в:
Бібліографічні деталі
Опубліковано в: :Прикладная механика
Дата:2013
Автори: Sun, J.F., Wang, X.L.
Формат: Стаття
Мова:Англійська
Опубліковано: Інститут механіки ім. С.П. Тимошенка НАН України 2013
Онлайн доступ:https://nasplib.isofts.kiev.ua/handle/123456789/87804
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function / J.F. Sun, X.L. Wang // Прикладная механика. — 2013. — Т. 49, № 5. — С. 139-144. — Бібліогр.: 6 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine
_version_ 1859707548509143040
author Sun, J.F.
Wang, X.L.
author_facet Sun, J.F.
Wang, X.L.
citation_txt Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function / J.F. Sun, X.L. Wang // Прикладная механика. — 2013. — Т. 49, № 5. — С. 139-144. — Бібліогр.: 6 назв. — англ.
collection DSpace DC
container_title Прикладная механика
description By the method of combining the matrix-valued Lyapunov functional and comparison theorem, connected Lyapunov stability and practical stability of large scale delay system are studied deeply. A series of new sufficient conditions are proposed. These results are not only of theoretical but also of practical value. Методом об‘єднання матрично-значних функціоналів Ляпунова і теореми порівняння досліджено зв‘язну стійкість за Ляпуновим і практичну стійкість великих систем з запізненням. Запропоновано ряд нових достатніх умов. Результати мають не лише теоретичний сенс, але також практичне значення.
first_indexed 2025-12-01T04:00:02Z
format Article
fulltext 2013 ПРИКЛАДНАЯ МЕХАНИКА Том 49, № 5 ISSN0032–8243. Прикл. механика, 2013, 49, № 5 139 J . F . S u n , X . L . W a n g CONNECTED STABILITY ANALYSIS OF DELAY SYSTEMS VIA THE MATRIX-VALUED LYAPUNOV FUNCTION Harbin Institute of Technology (Weihai), Shangdong province 2, Wenhua Xi Road Weihai, China; e-mail: sunif_70@126.com Abstract. By the method of combining the matrix-valued Lyapunov functional and comparison theorem, connected Lyapunov stability and practical stability of large scale de- lay system are studied deeply. A series of new sufficient conditions are proposed. These results are not only of theoretical but also of practical value. Key words: delay system, connected Lyapunov stability, practical stability. 1. Designations. Let 0 0([ , 0], ), [ , ), 0nC C R J t t     . For any C  the norm 0 sup ( ) s s        is used. For nx R , max rx x , 1, 2,...,r n . If 0([ , ), )nx C t R   , then tx C is de- termined as ( ) ( )tx s x t s  , 0s   . We designate  :H nC C H    , where 0H  or H   . 2. Description of the system and decomposition. Consider the large scale system modeled by functional differential equation 0 0( , ), ,H t t n dx f t x x C dt    (1) where : H n nf J C R  . Provided that the vector-function f maps the bounded sets into the bounded sets, for each 0t J and 0 H nC  there exists a unique solution 00( , )( )tx t x t determined on some interval 0[ , ]t t  , 0  , and if 1H H is such that 0 0 1( , ( )x t t H  , then    . The system (1) is decomposed into m interconnected subsystems 1( ) ( , ) ( , , , ), i i m i t i t t dx t f t x g t x x dt    (2) where  1, 2,...,mi I m  , ( , )i i i H n i nf J C R  , ( , )inH i ng C J C R  and i i n n . We assume that the functions ( , )i tg t x depend on m m -matrix of interconnections [ ]ij t tE e , 1 1 2 2( , ) ( , , , , ),i i im m i t i t t t t t tg t x g t e x e x e x  (3) 140 when mi I , where the elements ([ ,0],[0,1])ij te C   depend in general case on the delay ( ) ( ), [ ,0].ij ij i t te x e t x t        We designate by tE the fundamental matrix of interactions with the elements 1 ij te  if jx is contained in ( , )i tg t x ; 0 ij te  if jx is not contained in ( , )i tg t x . For 0tE  we get from system (2) the independent subsystems of functional differen- tial equations of smaller dimensions 0 0 ( ) ( , ), , i i i i i t t dx t f t x x C dt    (4) where in ix R and i i i H n i nf J C R   . Moreover, we assume ( ,0) 0if t  and for subsys- tems (2) ( ,0) ( ,0) 0i if t g t  for all t J and mi I , i.e. the state 1 ... 0mx x x    is the unique equilibrium state of system (1) and subsystems (2). For subsystem (2), whose functions ( , ), 1, 2,...,i tg t x i m , depend on the matrix of in- teractions tE , the problem on practical stability of motion is reduced to the establishment of conditions under which the solution 00( , ) ( )tx t x t of system (2) possesses certain qualitative properties for given estimates of initial and subsequent deviations on the infinite interval. 3. Matrix-valued functional. For system (2) we construct the matrix-valued functional ( , ) [ ( , )], ,ij mU t t i j I    (5) with the elements satisfying the following conditions. 1H . The elements ( , ) ( , )i i i H mi ii nt C J C R    , 1 i im n  , ( ,0) 0,ii t  are locally Lipschitz in i ; 2H .The elements ( , , ) ( , ),i jI I i i m mH Hi j ij n nt C J C C R      are locally Lipschitz in i and j for all ( ) mi j I  . By means of the real vector , 0mR   , we construct the functional ( , , ) ( , ) ,TV t U t     (6) which is continuous and definite on the set H nJ C by conditions 1 2H H . The upper de- rivative of functional (6) along solutions of system (2) is determined by the formula ( , , ) ( , ) ,TD V t D U t      (7) where ( , )D U t     0 1 lim sup ( , ( , )) ( , ) .tU t x t U t          Note that ( , )D U t  is com- puted element-wise. 4. Definitions of connected stability of system (2). Taking into account the results of paper [3] we shall cite the definitions of stability no- tion incorporated in this paper. Definition 1. The equilibrium state 0x  of system (1) is called a) connectedly stable if for every 0  and 0 0t  there exists 0( , ),t   such that 0( , )( )x t t  whenever 0[ , ]nC t t   for all t tE E ; b) uniformly connectedly stable if in definition (a) the value  does not depend on 0t ; 141 c) asymptotically connectedly stable if it is connectedly stable and for any 0 0t  there exists 0  such that 0( , ( ) 0,x t t  as ,t  whenever ,nC  for all t tE E ; d) uniformly asymptotically connectedly stable if it is uniformly connectedly stable and there exists some 0  and for every 0  there exists 0  such that 0( , ( ) ,x t t  whenever 0 0[ , ]nC t t   for all t tE E . 5. Conditions of connected stability of system (2). Using matrix-valued functional (5) and its derivative (7) and applying the theorems of comparison principle for functional-differential equations (see [1]) we shall set out a series of sufficient conditions for connected stability of the equilibrium state 0x  of system (1). Theorem 1. Let system of functional-differential equations (1) be such that 1) there exists the matrix-valued functional ( , ) ( , )H m m nU t C J C R   , ( ,0) 0U t  for all t J and ( , )U t  is locally Lipschitz in  for every t J ; 2) there exist m m constant matrices 1( )A  and 1( )B  , real vector mR  , 0  and comparison functions 1 2( (0) ), ( ),i i i i mu u i I   , of Hahn class K so that 1 1 1( (0) ) ( ) ( (0) )Tu A u   , 1 ( , ) m i j ij i j u t     2 1 2( ) ( ) ( )Tu B u   for all t J and H nC  ; 3) there exists the comparison function ( , )W C J R R  such that ( , , ) ( , ( , , ))D V t W t V t     (8) for all ( , ) H nt J C   and all matrices of interaction t tE E . Then the certain type of sta- bility of zero solution to the comparison equation 0 0( , ), ( ) 0 du W t u u t u dt    (9) and the restrictions on the matrices 1 1( ), ( )A B  imply the corresponding type of connected stability of the equilibrium state of system (1) with decomposition (2). Proof. Provided that the matrices 1( )A  and 1( )B  are positive definite, functional (6) is positive definite and decreasing. Further, we apply Theorem 4.4.3 from [1] and determine certain type of connected stability of system (1). Corollary 1. Let 1) conditions (1) and (2) of Theorem 1 be satisfied; 2) the matrix 1( )A  be positive definite, the matrix 1( ) 0B   and the comparison func- tion ( , ( , , )) 0W t V t    . Then the equilibrium state 0x  of system (1) with decomposition (2) is connectedly stable. Corollary 2. Let 1) conditions (1) and (2) of Theorem 1 be satisfied; 2) the matrices 1( )A  and 1( )B  be positive definite and the comparison function ( , ( , , )) 0W t V t    . Then the equilibrium state 0x  of system (1) with decomposition (2) is uniformly connectedly stable. Corollary 3. Let 1) conditions (1) and (2) of Theorem 1 be satisfied; 2) the matrices 1( )A  and 1( )B  be positive definite; 3) the zero solution of comparison equation (9) be uniformly asymptotically stable. 142 Then the equilibrium state 0x  of system (1) with decomposition (2) is uniformly as- ymptotically connectedly stable. Theorem 2. Let system of functional differential equations (1) be such that 1) conditions (1) and (2) of Theorem 1 are satisfied; 2) there exist a constant m m matrix 1( ), , 0mC R    and functions 3 ( )i i tu x , 3iu is of class K for all mi I , such that 3 1 3( , , ) ( ) ( ) ( )T t tD V t u x C u x    for any ( , ) H nt J C   and any matrices of interactions t tE E , where 3 ( )T tu x = 31 3( ( ),..., ( ))l m t m tu x u x ; 3) the matrices 1( )A  and 1( )B  are positive definite and the matrix 1( )C  is negative definite. Then the equilibrium state 0x  of system (1) with decomposition (2) is uniformly as- ymptotically connectedly stable. Theorem 3. Let in system of equations (1) the vector function ( , )f t  be bounded in  and 1) conditions (1) and (2) of Theorem 1 are satisfied; 2) there exist a constant m m matrix 2 ( ), , 0mC R    and functions 4 ( )i i tu x of class K for all mi I such that 4 2 4( , , ) ( ) ( ) ( )T t tD V t u x C u x    for all ( , ) H nt J C   and any matrices of interconnections t tE E ; 3) the matrices 1( )A  and 1( )B  are positive definite and the matrix 2 ( )C  is negative definite. Then the equilibrium state 0x  of system (1) with decomposition (2) is uniformly as- ymptotically connectedly stable. 6. Matrix-valued function on space product. For system (4) we construct the matrix-valued function ( , , ) [ ( , , )], , 1, 2,...,t ij tU t x x v t x x i j m  , (10) with the elements satisfying the following conditions. 3H . The elements ( , ), ( , 0, 0) 0i i H ii iinv C J C C R v t    are locally Lipschitz in ix ; 4H . The elements ( , ),ji i j HH ij n nv C J C C C C R     ( , 0, 0, 0) 0ijv t  are locally Lipschitz in ,i jx x for all ( ) mi j I  . By means of the real vector , 0mR   , we construct the function ( , , , ) ( , , ) ,T t tV t x x U t x x   (11) which is definite on the space product nR C and locally Lipschitz in x , providing condi- tions of assumptions 3H and 4H are satisfied. Further we define ( , , , ) ( , , ) ,T t tD V t x x D U t x x    (12) where ( , , ) lim sup ( ,tD U t x x U t    1( , ), ( )) ( , , )] :t t h tx f t x x U t x x      0  . (13) Note that when formula (12) is properly applied, ( , , )tD U t x x is computed element-wise. 7. Conditions of connected practical stability of system (2). In view of the results from [1, 4] we shall formulate the following definitions. Definition 2. System (2) is called 143 a) connectedly practically stable, if given estimates of ( , ), 0A A   , the condition 0 nC  implies 0 0( , )( )x t t A  for all 0t t and all t tE E ; b) connectedly asymptotically practically stable, if conditions of definition (a) are satis- fied and 0 0lim ( , )( ) 0 t x t t   . The other definitions of connected practical stability can be formulated in terms of Definition 2. Theorem 4. Let system of functional differentional equations (1) be such that 1) there exists a matrix-valued function ( , )H m m nU C J C C R    , ( ,0,0) 0U t  for all t J and ( , , )tU t x x is locally Lipschitz in x for ( , , ) ( ) ( )tt x x J S A C A   ; 2) there exist a real vector , 0R   , constant m m matrices ( )A  and ( )B  and a comparison function 1 2( ), ( ( ) ), 1, 2,...,i i tu x u x i m  , 1 2,i iu u K , such that 1 1( ) ( ) ( )Tu x A u x  , 1 ( , , ) m i j ij t i j v t x x       2 2( ) ( ) ( )T t tu x B u x  for all ( , , ) ( ) ( )tt x x J S A C A   ; 3) there exists a comparison function ( , )W C J R R  such that ( , , , )tD V t x x   ( , ( , , , ))tW t V t x x  for all ( , , ) ( ) ( )tt x x J S A C A   and all matrices of interactions t tE E ; 4) the matrices A and B are positive definite and ( ) ( ) ( ) ( )M mB a A b    where ( )m A is the minimal and ( )M B is the maximal eigenvalues of the matrices A and B re- spectively and ,a b are of class K . Then the certain type of practical stability of zero solution to the equation 0 0( , ), ( ) 0 du W t u u t u dt    (14) implies the certain type of connected practical stability of system (2). Proof. Note first that under conditions (1) and (2) of Theorem 4 for the function ( , , )tV t x x determined by (11) the estimate ( ) ( ) ( , , ) ( ) ( ( ) )m t M tA b x V t x x B a x    (15) is true. This follows from the fact that for function 1 2, , 1, 2,...,i iu u K i m  , there exist functions ( ( ) )ta x  and ( )b x of class K such that 1 1( ) ( ) ( )Tb x u x u x and 2 2( ( ) ) ( ( ) ) ( ( ) ).T t t ta x u x u x    Further we have from condition (3) of Theorem 4 for the function 0 00 0( ) ( , ( , )( ), ( , ))t tm t V t x t x t x t x ( ) ( , ( ))D m t W t m t  which together with the condition 00 0( , , )tV t x x u yield the estimate 0 00 0 0 0( , ( , )( ), ( , )) ( , , ),t t tV t x t x t x t x r t t u 0t t (16) according to the comparison principle (see[1] Theorem 4.1.1). Let the zero solution of equa- tion (14) be practically stable. Given ( ( ) ( ), ( ) ( ))M mB a A b A   , we have 0 0( , , ) ( ) ( )mu t t u A b A , (17) provided that 0 ( ) ( )Mu B a  . (18) 144 Let 0x  and 0 ( )tx   . (19) We shall demonstrate that 00( , )( )tx t x t A for all 0 .t t Assume that this is not true and that there exists 1 0t t such that for the solution 00( , ) ( )tx t x t with initial condition (19) the correlations 00 1( , ) ( )tx t x t A and 00( , ) ( )tx t x t A hold for 0 1t t t  . Estimate (15) yields 0 1 01 0 1 0( , ( , )( ), ( , )) ( ) ( )t t t mV t x t x t x t x A b A (20) Let 0 0 00 0 0 0 0( , ( , )( ), ( , ))t t tu V t x t x t x t x . Then for all 0 1t t t  , estimate (16) is valid, where 0 0( , , )r t t u is the maximal solution of equation (14). Since 0 2( ) T Mu B u  0 02( ( ) ) ( ( ) )t tx u x   ( ) ( ),M B a  we find by the comparison principle and inequalities (15). 1 0 1 0( ) ( ) ( ) ( ) ( )T m mA b A A u x u x   0 1 01 0 1 0( , ( , )( ), ( , ))t t tV t x t x t x t x 1 0 0( , , ) ( ) ( )mr t t u A b A  . (21) The obtained contradiction shows that 1t J and therefore system (2) is connectedly practi- cally stable. Р Е ЗЮМ Е . Методом об‘єднання матрично-значних функціоналів Ляпунова і теореми порів- няння досліджено зв‘язну стійкість за Ляпуновим і практичну стійкість великих систем з запізнен- ням. Запропоновано ряд нових достатніх умов. Результати мають не лише теоретичний сенс, але також практичне значення. 1. Lakshimikantham V., Leela S., Martynyuk A.A. Practical Stability of Nonlinear Systems. – Singapore: World Scientific, 1990. – 215 p. 2. Lakshmikantham V.,Leela S., Sivasundaram S. Lyapunov functions on product spaces and theory of delay differential equation // J. Math. Anal. and Appl. – 1991. – 154 – P. 391 – 402. 3. Martynyuk A.A., Sun Z.Q. A matrix-valued Lyapunov functional and stability of systems with delay // Dokl. Akad. Nauk. – 1998. – 359, N 2. –P.165 – 167. 4. Martynyuk A.A., Sun Z.Q. On Connected Practical Stability of Motion of Systems with Delay// Int. Appl. Mech. – 1999. – 35, N 1. – P. 87 – 92. 5. .Martynyuk A.A., Sun Z.Q. Stability analysis for nonlinear system with small parameter. – Beijin: Science Publishing House, 2006. – P. 212 – 214. 6. Martynyuk. A.A., Martynyuk-Chernienko Yu.A. Uncertain Dynamical Systems: Stability and Motion Con- trol. – Boca – Raton: CRC Press, 2012. – 237 p. From the Editorial Board: The article corresponds completely to submitted manuscript. Поступила 10.09.2012 Утверждена в печать 26.06.2013
id nasplib_isofts_kiev_ua-123456789-87804
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
issn 0032-8243
language English
last_indexed 2025-12-01T04:00:02Z
publishDate 2013
publisher Інститут механіки ім. С.П. Тимошенка НАН України
record_format dspace
spelling Sun, J.F.
Wang, X.L.
2015-10-25T18:43:15Z
2015-10-25T18:43:15Z
2013
Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function / J.F. Sun, X.L. Wang // Прикладная механика. — 2013. — Т. 49, № 5. — С. 139-144. — Бібліогр.: 6 назв. — англ.
0032-8243
https://nasplib.isofts.kiev.ua/handle/123456789/87804
By the method of combining the matrix-valued Lyapunov functional and comparison theorem, connected Lyapunov stability and practical stability of large scale delay system are studied deeply. A series of new sufficient conditions are proposed. These results are not only of theoretical but also of practical value.
Методом об‘єднання матрично-значних функціоналів Ляпунова і теореми порівняння досліджено зв‘язну стійкість за Ляпуновим і практичну стійкість великих систем з запізненням. Запропоновано ряд нових достатніх умов. Результати мають не лише теоретичний сенс, але також практичне значення.
en
Інститут механіки ім. С.П. Тимошенка НАН України
Прикладная механика
Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function
Анализ связной устойчивости систем с задержкой с помощью матрично-значной функции Ляпунова
Article
published earlier
spellingShingle Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function
Sun, J.F.
Wang, X.L.
title Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function
title_alt Анализ связной устойчивости систем с задержкой с помощью матрично-значной функции Ляпунова
title_full Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function
title_fullStr Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function
title_full_unstemmed Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function
title_short Connected Stability Analysis of Delay Systems via the Matrix-Valued Lyapunov Function
title_sort connected stability analysis of delay systems via the matrix-valued lyapunov function
url https://nasplib.isofts.kiev.ua/handle/123456789/87804
work_keys_str_mv AT sunjf connectedstabilityanalysisofdelaysystemsviathematrixvaluedlyapunovfunction
AT wangxl connectedstabilityanalysisofdelaysystemsviathematrixvaluedlyapunovfunction
AT sunjf analizsvâznoiustoičivostisistemszaderžkoispomoŝʹûmatričnoznačnoifunkciilâpunova
AT wangxl analizsvâznoiustoičivostisistemszaderžkoispomoŝʹûmatričnoznačnoifunkciilâpunova