Stability results for stochastic delayed recurrent neural networks with discrete and distributed delays

Size: px
Start display at page:

Download "Stability results for stochastic delayed recurrent neural networks with discrete and distributed delays"

Transcription

1 Sably resuls for sochasc delayed recurren neural neworks wh dscree and dsrbued delays Gulng Chen, Onno van Gaans, Sjoerd Verduyn Lunel Mahemacal Insue, Leden Unversy, P.O.Box 9512, 23 RA, Leden, The Neherlands Reor MI Absrac We resen new condons for asymoc sably and exonenal sably of a class of sochasc recurren neural neworks wh dscree and dsrbued me varyng delays. Our aroaches are based on he mehod usng fxed on heory and he mehod usng an arorae negral neualy, whch do no resor o any Lyaunov funcon. Our resuls neher reure he boundedness, monooncy and dfferenably of he acvaon funcons nor dfferenably of he me varyng delays. In arcular, a class of neural neworks whou sochasc erurbaons s also consdered by usng he wo aroaches. xamles are gven o llusrae our man resuls. Keywords: Fxed on heory, asymoc sably, exonenal sably, sochasc recurren neural neworks, rval soluon, varable delays, Burkholder-Davs-Gundy neualy, Doob s neualy. 1. Inroducon and man resuls Neural neworks have receved an ncreasng neres n varous areas [5, 8]. The sably of neural neworks [6, 16, 37, 38] s crcal for sgnal rocessng, esecally n mage rocessng and solvng some classes of omzaon roblems. For he sochasc effecs o he dynamcal behavors of neural neworks, Lao and Mao [14, 15] naed he sudy of sably and nsably of sochasc neural neworks. Due o he fne swchng seed of neurons and amlfers, me delays whch may lead o nsably and bad erformance n neural rocessng and sgnal ransmsson are commonly encounered n boh bologcal and arfcal neural neworks. In addon, neural neworks usually have a saal exen due o he resence of a mulude of arallel ahways wh a varey of axon szes and lenghs [31]. Thus here wll be a dsrbuon of conducon veloces along hese ahways and a dsrbuon of roagaon delays [39]. In hese crcumsances he sgnal roagaon s no nsananeous and may no be suably modeled wh dscree delays. Therefore, a more arorae way whch ncororaes connuously dsrbued delays n neural nework models has been used. Furher, due o random flucuaons and robablsc causes n he nework, noses do exs n a neural nework. Thus, s necessary and rewardng o sudy sochasc effecs o he sably roery of neural neworks. Many neresng arcles [9, 1, 11, 28, 32] have consdered some classes of he sochasc neural neworks. Hu e al.[9] and Wan and Sun [32] suded a class of sochasc neural neworks wh he delays consan and dscree. The acvaon funcons aearng n [9] are reured o be bounded. Lao and Mao [17] nvesgaed exonenal sably of sochasc delay nerval sysems va Razumkhn-ye heorems develoed n [23], several exonenal sably resuls were rovded. However, he resuls are no only dffcul o verfy bu also resrc o he case of he nerval marces à = B = C =. Sun and Cao [28] nvesgaed he h momen exonenal sably of sochasc dfferenal euaons wh dscree bounded delays by usng he mehod of varaon arameer, neualy echnue and sochasc analyss. Ths mehod was frsly used n [32], whch does no reure he boundedness, monooncy and dfferenably of he acvaon funcons. However, he sably crera n [28] reures ha he delay funcons are bounded, dfferenable and her dervaves are smulaneously reured o be no greaer han 1. Ths may mose a very src consran on model because me delays somemes vary dramacally wh me n real crcus (see [36]). 1

2 Huang e al. [1, 11] nvesgaed he exonenal sably of sochasc dfferenal euaons wh dscree mevaryng delays wh he hel of a Lyaunov funcon and Dn dervave. However, he use of her crera deends very much on he choce of osve numbers k j ec. and a osve dagonal marx M (see Theorem 3.3 n [1] and Theorem 3.3 n [11]). Recenly, Buron [2] has ulzed he fxed on mehod o nvesgae he sably for deermnsc sysems, Luo [19] and Aleby [1] have aled hs mehod o deal wh he sably roblems for sochasc delay dfferenal euaons, and aferwards, a grea number of classes of sochasc delay dfferenal euaons are dscussed by usng fxed on mehod, see, for examle, [5, 2, 21, 25, 26]. I urned ou ha fxed on mehod s a owerful echnue n dealng wh sably roblems for dfferenal euaons wh delays and sochasc dfferenal euaons wh delays, and can yeld he exsence, unueness and sably crera of he consdered sysem n one se by a fxed on argumen, whch s mossble when usng he oher mehods. Chen [4, 3] has aled he mehod by usng an arorae negral neualy o sudy exonenal sably of some classes of sochasc delay dfferenal euaons, and urns ou ha s a convenen way o dscuss exonenal sably of a sysem. The am of hs aer s o sudy a general class of sochasc neural neworks by usng fxed on mehod and he mehod by emloyng an arorae negral neualy. Indeed, we consder he followng class of sochasc neural neworks wh varyng dscree and dsrbued delays whch s descrbed by or dx() = dx () = c x () a j f j (x j ()) b j g j (x j ( τ())) l j r() h j (x j (s)) ds d σ j (, x j (), x j ( τ())) dω j (), (1) [ ] Cx() A f (x()) Bg(x( τ()))w h(x(s)) ds dσ(, x(), x( τ())) dω(), r() for, 2, 3,, n, where x()=(x 1 (), x 2 (), x n ()) T R n s he sae vecor assocaed wh he neurons; C= dag(c 1, c 2, c n )> where c > reresens he rae wh whch he h un wll rese s oenal o he resng sae n solaon when dsconneced from he nework and he exernal sochasc erurbaons; A=(a j ) n n, B=(b j ) n n and W= (l j ) n n reresen he connecon wegh marx, delayed connecon wegh marx and dsrbued delayed connecon wegh marx, resecvely; f j, g j, h j are acvaon funcons, f (x())=( f 1 (x()), f 2 (x()), f n (x())) T R n, g(x()) = (g 1 (x()), g 2 (x()), g n (x())) T R n, h(x()) = (h 1 (x()), h 2 (x()), h n (x())) T R n, whereτ() and r() denoe dscree me varyng delay and dsrbued me varyng delay, resecvely. Moreover, ω() = (ω 1 (),ω 2 (), ω n ()) T R n s an n-dmensonal Brownon moon defned on a comlee robably sace (Ω,F,P) wh naural comlee flraon{f } (.e.f = comleon ofσ{ω(s) : s }) andσ : R R n R n R n n, σ=(σ j ) n n s he dffuson coeffcen marx. Denoeϑ=nf { τ(), r()}. The nal condon for he sysem (1) s gven by x()=φ(), [ϑ, ], (2) whereφ()=(φ 1 (),φ 2 (), φ n ()) T C ( [ϑ, ], L F (Ω;R n ) ) wh he norm defned as φ = su φ (s), ϑ s where denoes execaon wh resec o he robably measure P. To oban our man resuls, we suose he followng condons are sasfed: (A1) he delays τ(), r() are connuous funcons such ha τ() and r() as ; 2

3 (A2) f j (x), g j (x), and h j (x) sasfy Lschz condons. Tha s, for each, 2, 3, n, here exss consansα j, β j,γ j such ha for every x, y R n, f j (x) f j (y) α j x y, g j (x) g j (y) β j x y, h j (x) h j (y) γ j x y ; (A3) Assume ha f (), g(), h(),σ(,, ) ; (A4)σ(, x, y) sasfes a Lschz condon. Tha s, here are nonnegave consansµ andν such ha race [ (σ(, x, y) σ(, u, v)) T (σ(, x, y) σ(, u, v)) ] [ µ (x u ) 2 ν (y v ) 2]. I follows from [7, 22] ha under he hyohess (A2), (A3) and (A4), sysem (1) wh nal condon (2) has one unue global soluon whch s denoed by x(,,φ) or x(), and su s x(s,,φ) < for >. Clearly, sysem (1) adms he rval soluon x(,, ). Defnon 1.1. The rval soluon of sysem (1) s sad o be sable n h ( 2) momen f for arbrary gven ǫ>, here exss aδ> such ha φ <δyelds ha x(,,φ) <ǫ,. whereφ() C ( [ϑ, ], L F (Ω;R n ) ). In arcular, when =2, he rval soluon s sad o be mean suare sable. Defnon 1.2. The rval soluon of sysem (1) s sad o be asymocally sable n h ( 2) momen f s sable n h ( 2) momen and here exss a scalarσ>, such ha φ <σmles whereφ() C ( [ϑ, ], L F (Ω;R n ) ). lm x(,,φ) =. Defnon 1.3. The rval soluon of sysem (1) s sad o be h ( 2) momen exonenally sable f here exss a ar of consansλ, C> such ha x(,,φ) C φ e λ,, holds forφ() C ( [ϑ, ], L F (Ω;R n ) ). secally, when =2, we seak of exonenally sable n mean suare. Dfferen choces of norms are defned for sace of sochasc rocesses. The norms we choose should be such ha he sace under consderaon s comlee and he euaon yelds a conracon wh resec o he norm. For he sysem (1) wh nal condon (2), we consder he followng wo dfferen comlee saces whch are defned by usng wo yes of norms. DefneS φ he sace of allf -adaed rocessesϕ(,ω) : [ϑ, ) Ω R n such haϕ C ( [ϑ, ), L F (Ω;R n ) ). Moreover, we seϕ(,ω)=φ() for [ϑ, ] and n ϕ () as,, 2, n. If we defne he norm ϕ := su ϑ ϕ (), (3) hens φ s a comlee merc sace. Usng he conracve mang defned on he saces φ and alyng he conracon mang rncle, we oban our frs resul. Theorem 1.4. Suose ha he assumons (A1)-(A4) hold. If he followng condons are sasfed, () he dsrbued delay r() s bounded by a consan r; 3

4 () α 5 1 c a j α j ( r 5 1 c ) l j γ j / / 5 1 c 5 1 n 1 b j β j / ( c /2 µ /2 ν /2) < 1, whereµ=max{µ 1,µ 2, µ n },ν=max{ν 1,ν 2, ν n }, hen he rval soluon of (1) s h momen asymocally sable. Consder he case when boh he dscree delayτ() and he dsrbued delay r() are bounded by a consanτ. DefneC φ he sace of allf -adaed rocessesϕ(,ω) : [ τ, ) Ω R n such haϕ C ( [ τ, ), L F (Ω;R n ) ). Moreover, we seϕ(,ω)=φ() for [ τ, ] and for, n su ϕ (s). The norm onc φ s defned as ( ϕ = su su ϕ (s) ). (5) henc φ s a comlee merc sace. Usng he conracon defned on he sacec φ and alyng he conracon mang rncle, we oban our second resul. Theorem 1.5. Suose ha he assumons (A1)-(A4) hold. If he followng condons are sasfed, () he dscree delayτ() and he dsrbued delay r() are bounded by a consanτ; () / / α 5 1 e cτ c a j α j 5 1 e cτ c b j β j / 5 1 τ e cτ c l j γ j 5 1 K n e cτ c 1 /2 (2c) 1( µ /2 ν /2) < 1, (6) where c=mn{c 1, c 2, c n },µ=max{µ 1,µ 2, µ n },ν=max{ν 1,ν 2, ν n }, hen he rval soluon of (1) s h momen asymocally sable. Remark 1.6. In Theorem 1.5, we oban ha { [ lm su x(s,,φ) ]}=, ha s, for any funcon s x (s,,φ), we have ha lm x (,,φ) C[ τ,] =, whch mles lm x(,,φ) =. Remark 1.7. In some aers, see, for examle, [18, 19, 33, 34], he norm for he sace of sochasc rocess s defned as { ( )} 1/2 ϕ [,] = su ϕ(s,ω) 2. s [,] As n [19], n order o show P(S) S, we need o esmae su s [,] I 5 (s) 2, where I 5 (s)= s e s z h(u) du [c(z)x(z)e(z)x(z δ(z))] dω(z). However, I 5 (s) s no a local marngale (see Secon 8 for s roof). Hence, Burkholder-Davs-Gundy Ineualy can no be aled drecly. 4 (4)

5 Usng an arorae negral neualy, we oban suffcen condons for exonenal sably of (1) wh nal condon (2), whch s our hrd resul. Theorem 1.8. Suose ha he assumons (A1)-(A4) hold. If he followng condons are sasfed, () he dscree delayτ() and dsrbued delay r() are bounded by a consanτ; () / / 5 1 c a j α j 5 1 c b j β j (7) ( τ ) / 5 1 c l j γ j 5 1 n c /2 (µ /2 ν /2 )<1, where c=mn{c 1, c 2, c n },µ=max{µ 1,µ 2, µ n },ν=max{ν 1,ν 2, ν n }, hen he rval soluon of (1) s exonenally sable n h momen, Remark 1.9. The sably crera we rovded n our man resuls are only n erms of he sysem arameers c, a j, b j, l j, ec. Hence, hese crera can usually be verfed easly n alcaons. Remark 1.1. Many arcles, see, for examle, [27, 28] have suded he case and secal case of sochasc neural nework (1). However, he delays should sasfy he followng condon: (H) he dscree delay τ() s dfferenable funcon and he dsrbued delay r() s non-negave and bounded, ha s, here exs consansτ M,ζ, r M such ha τ() τ M, τ () ζ, r() r M, (8) In Sun and Cao [28], seems ha he consran condon (H) on he dscree delays can be relaxed asτ() s bounded. As an examle, we consder a wo-dmensonal sochascally erurbed Hofeld neural nework wh mevaryng delays, dx()=[ Cx()A f (x()) Bg(x τ ())] dσ(, x(), x τ ()) dω(), where f (x)= 1 5 arcan x, g(x)= 1 5 anh x= 1 5 (ex e x )/(e x e x ),τ()= 1 2 sn 1 2, ( ) ( ) ( ) C=, A= and B= In hs examle, he dscree delay s bounded bu no dfferenable. Consder he case when here are no sochasc effecs on he sysem (1), whch hen comes down o he followng neural nework descrbed by dx () = c x () a j f j (x j ()) b j g j (x j ( τ())) d j h j (x j (s)) ds,, 2, 3,, n, (9) d or dx() d r() = Cx() A f (x()) Bg(x τ())d h(x(s)) ds, r() where x( )=(x 1 ( ), x 2 ( ),, x n ( )) T s he neuron sae vecor of he ransformed sysem (9). The nal condon for he sysem (9) s x()=φ(), [ϑ, ], (1) whereφ s a connuous funcon wh he norm defned by φ =su ϑ s n φ (s). Assume ha (A1) (A3) are sasfed, hen (9) and (1) adm a rval soluon x=. Denoe by x(; s;φ)= (x 1 (; s,φ 1 ),, x n (; s,φ n )) T R n he soluon of (9) wh nal condon (1). 5

6 Defnon For he sysem (9) wh nal condon (1), we have ha () he rval soluon of (9) s sad o be sable f for anyǫ>, here exssδ> such ha for any nal condonφ(s) C([ϑ, ],R n ) sasfyng φ <δ, we have for he corresondng soluon ha x(, s,φ) <ǫfor ; () he rval soluon of (9) s sad o be asymocally sable f s sable and for any nal condonφ(s) C([ϑ, ],R n ) we have for he corresondng soluon ha lm x(, s,φ) =; () he rval soluon of (9) s sad o be globally exonenally sable f here exs scalars k> andα> such ha for any nal condonφ(s) C([ϑ, ],R n ), we have for he corresondng soluon ha x(, s,φ) αe k φ for. DefneH φ =H 1φ H 2φ H nφ, whereh φ s he sace conssng of connuous funconsϕ () : R R such haϕ (θ)=φ(θ) forϑ θ andϕ () as,, 2 n. For anyϕ()=(ϕ 1 (),ϕ 2 (),,ϕ n ()) H φ and η()=(η 1 (),η 2 (),,η n ()) H φ, f we defne he merc as d(ϕ,η)=su ϑ n ϕ () η (), henh φ becomes a comlee merc sace. Usng he conracon mang defned on he sace H φ and alyng he conracon mang rncle, we oban our fourh resul. Theorem Suose ha he assumons (A1)-(A3) hold. If he followng condons are sasfed, () he dsrbued delay r() s bounded by a consan r; () α 1 c max a jα j,2,,n 1 c hen he rval soluon of (9) s asymocally sable. max b jβ j,2,,n r c max d jγ j <1, (11),2,,n By esablshng an arorae negral neualy, we oban suffcen condons for exonenal sably of (9), whch s our ffh resul. Theorem Suose ha he assumons (A1)-(A3) hold. If he followng condons are sasfed, () he dscree delayτ() and he dsrbued delay r() are bounded by a consanτ; () 1 c max,2,,n a jα j 1 c max,2,,n b jβ j 1 c hen he rval soluon of (9) wh nal condon (1) s exonenally sable. τ max,2,,n d jγ j <1, c=mn{c 1, c 2, c n }, (12) Remark Several exonenal sably resuls [13, 29, 3] were rovded for he sysem (9), by consrucng an arorae Lyaunov funconal and emloyng lnear marx neualy (LMI) mehod. However, he delays n hose resuls should sasfy he followng condon (H). From our man resuls, we need no know he arcular form relaed o he delays, we only need o know ha he delays are bounded. Furhermore, Theorem 1.12 s an exenson and mrovemen of he resul n La and Zhang [12]. As an examle, we consder a cellular neural nework wh me varyng delays where C= dx() d ( = Cx()Ag(x()) Bg(x τ()), ) (.2, A= ) (.1.1, B=.2.1 ).

7 The acvaon funcon s descrbed by g(x)= x1 x 1 2. The me-varyng delay sτ()= 1 5 cos. I s clear ha he dscree delay s bounded bu no dfferenable. Hence, he resuls n [13, 29, 3] are no alcable. The res of hs aer s organzed as follows. In Secon 2, we resen a roof of Theorem 1.4. The roof of Theorem 1.5 s resened n Secon 3 and he roof of Theorem 1.8 s gven n Secon 4. we resen he roofs of Theorem 1.12 and Theorem 1.13 n Secon 5 and Secon 6, resecvely. Some examles are gven o llusrae our man resuls n Secon 7 and an aendx s gven n Secon Proof of Theorem 1.4 In hs secon, we rove Theorem 1.4. We sar wh some rearaons. Lemma 2.1. ([32]) Ifω()=(ω 1,ω 2,,ω n ) T s a n-dmensonal Brownan moon defned on a comlee robably sace (Ω, F, P), hen we have he followng formula ( ) f (s) dω (s) f j (s) dω j (s) = f (s) f j (s) d ω,ω s, where ω,ω s =δ j s are he cross-varaons,δ j s correlaon coeffcen, 1, j n. If we mully boh sdes of (1) by e c and negrae from o, we oban x () = e c x () for,, 2, 3,, n. e c ( s) s e c ( s) l j s r(s) a j f j (x j (s)) ds h j (x j (u)) du ds e c ( s) e c ( s) b j g j (x j (s τ(s))) ds (13) σ j (s, x j (s), x j (s τ(s))) dω j (s). Lemma 2.2. Defne an oeraor by (Qϕ)()=φ() for [ϑ, ], and for,, 2, 3,, n, (Qϕ )() = e c ϕ () e c ( s) s e c ( s) l j s r(s) a j f j (ϕ j (s)) ds h j (ϕ j (u)) du ds e c ( s) e c ( s) b j g j (ϕ j (s τ(s))) ds (14) σ j (s,ϕ j (s),ϕ j (s τ(s))) dω j (s). Suose ha he assumon (A1)-(A4) holds. If condon (4) holds, hen Q :S φ S φ and Q s a conracon mang. Proof. Denoe (Qϕ )() := J 1 () J 2 () J 3 () J 4 () J 5 (), where J 1 () J 3 () = J 4 () = J 5 () = = e c ϕ (), J 2 ()= e c ( s) e c ( s) e c ( s) e c ( s) b j g j (ϕ j (s τ(s))) ds, l j s s r(s) h j (ϕ j (u)) du ds, a j f j (ϕ j (s)) ds, σ j (s,ϕ j (s),ϕ j (s τ(s))) dω j (s). 7

8 Se1. From he defnon of Banach saces φ, we have ha n ϕ () <, for all,ϕ S φ. Se2. We rove he connuy n h momen of Q on [, ). Le x S φ, 1, r be suffcenly small and ake he lm r. We have 1 J 2 ( 1 r) J 2 ( 1 ) = ( e c ( 1 r s) e ) n c ( 1 s) a j f j (x j (s)) ds 1 r e c ( 1 r s) a j f j (x j (s)) ds as r. Smlarly, we have ha J 3 ( 1 r) J 3 ( 1 ) as r, In he followng, we check he connuy of J 5 (). J 5 ( 1 r) J 5 ( 1 ) = n r e c ( 1 r s) J 4 ( 1 r) J 4 ( 1 ) as r. ( e c ( 1 r s) e ) n c ( 1 s) σ j (s, x j (s), x j (s τ(s))) dω j (s) 1 r e c ( 1 r s) 1 1 (2n) 1 (2n) 1 = (2n) 1 [1 r 1 σ j (s, x j (s), x j (s τ(s))) dω j (s) ( e c ( 1 r s) e c ( 1 s) ) σ j (s, x j (s), x j (s τ(s))) dω j (s) σ j (s, x j (s), x j (s τ(s))) dω j (s) e c ( 1 r s) e c ( 1 s) σ j (s, x j (s), x j (s τ(s))) dω j (s) 1 e c ( 1 r s) σ j (s, x j (s), x j (s τ(s))) dω j (s) 1 e c ( 1 r s) e c ( 1 s) 2 σ 2 j (s, x j(s), x j (s τ(s))) ds 1 r Thus, Q s ndeed connuous n h momen on [, ). ] /2 e 2c ( 1 r s) σ 2 j (s, x j(s), x j (s τ(s))) ds as r. /2 Se3. We rove ha Q(S φ ) S φ. Qϕ () = 5 J j () J j (). (15) 8

9 Now, we esmae he erms on he rgh sdes of he above neualy. J 2 () = e c ( s) a j f j (ϕ j (s)) ds e c ( s) e c ( s) a j f j (ϕ j (s)) ds [ ] / e c( s) ds e c ( s) a j f j (ϕ j (s)) ds c / e c ( s) a j α j ϕ j (s) ds / c / a j α j e c( s) ϕ j (s) ds. (16) Snceϕ() S φ, we have ha lm n ϕ () =. Thus for anyǫ>, here exss T 1 > such ha T 1 mles n ϕ () <ǫ, combnng wh (16), we oban ha / T1 J 2 () c / a j α j e c( s) ϕ j (s) ds / c / a j α j e c( s) ϕ j (s) ds T 1 / / < c e c (e c T 1 1) a j α j su ϕ j (s) ǫ c a j α j Hence, from he fac ha c > (, 2,, n), we oban ha n J 2 () as. Wh he smlar comuaon as (16), we oban ha / J 3 () c / b j β j e c( s) ϕ j (s τ(s))) ds / J 4 () c / l j γ j e c( s) s ϕ j (u) du s r(s) ds ( r / s c ) / l j γ j e c ( s) ϕ j (u) du ds. (17) s T 1 s r(s) Usng Lemma 2.1, we oban ha J 5 () = e c ( s) σ j (s,ϕ j (s),ϕ j (s τ(s))) dω j (s) [ ] 2 n 1 e c( s) σ j (s,ϕ j (s),ϕ j (s τ(s))) dω j (s) /2 [ ] /2 = n 1 e 2c( s) σ 2 j (s,ϕ j(s),ϕ j (s τ(s))) ds 9 (18)

10 [ n 1 e ( ] /2 2c ( s) µ j ϕ 2 j (s)ν jϕ 2 j (s τ(s))) ds ( ) /2 ( ) /2 n 1 2 /2 1 e 2c( s) µ j ϕ 2 j (s) ds e 2c( s) ν j ϕ 2 j (s τ(s)) ds ( ) /2 1 n 1 2 /2 1 e 2c( s) ds e 2c( s) µ /2 j ϕ j (s) ds ( ) /2 1 n 1 2 /2 1 e 2c( s) ds e 2c( s) ν /2 j ϕ j (s τ(s)) ds n 1 c 1 /2 µ/2 e 2c( s) ϕ j (s) dsν/2 e 2c( s) ϕ j (s τ(s)) ds n 1 c 1 /2 µ/2 e c( s) ϕ j (s) dsν/2 e c( s) ϕ j (s τ(s)) ds. Snce n ϕ (), τ() and r() as, for eachǫ>, here exss T 2 > such ha T 2 mles n ϕ ( τ(s)) <ǫ and n ϕ ( r()) <ǫ. From (17), we oban ha J 3 () / T2 c / b j β j e c( s) ϕ j (s τ(s))) ds / c / b j β j e c( s) ϕ j (s τ(s))) ds T 2 ( ) / / 1 T2 < e c e c s ds b j β j su ϕ j (s τ(s))) ǫ c ϑ s T 2 c b j β j / and J 4 () < ( r c ) / l j γ j ( r c ) / l j γ j ( re c r c ) / l j γ j / T2 s e c ( s) s r(s) / s e c ( s) T 2 s r(s) / su ϑ u T 2 ϕ j (u) du ds ϕ j (u) du ds T2 ϕ j (u) e cs ds ( ǫr r c c ) / l j γ j /. Furher, from (18), we oban J 5 () 1

11 n 1 n 1 c 1 /2 c 1 /2 n 1 < n 1 c 1 /2 n 1 µ/2 e c( s) ϕ j (s) dsν/2 e c( s) ϕ j (s τ(s)) ds T2 T2 µ/2 e c( s) ϕ j (s) dsν/2 e c( s) ϕ j (s τ(s)) ds µ/2 e c( s) ϕ j (s) dsν/2 e c( s) ϕ j (s τ(s)) ds T 2 T 2 T2 µ/2 su ϕ j (s) ν/2 su ϕ j (s) e s T 2 ϑ s T 2 c( s) ds ( ǫ(µ /2 ν /2 ) ). c 1 /2 c 1 /2 Hence, le, from he fac ha c > (, 2,, n), we oban ha c J 3 (), J 4 (), and J 5 (). Thus, combnng wh (15), we oban ha n Qϕ () as n ϕ (). Therefore, Q :S φ S φ. Se4. We rove ha Q s a conracon mang. For anyϕ,ψ S φ, from (16)-(18), we oban su s ϑ Qϕ (s) Qψ (s) 4 1 su s ϑ s ( ) e c (s u) a j f j (x j (u)) f j (y j (u)) du 4 1 su s ϑ s ) e c (s u) b j (g j (x j (u τ(u))) g j (y j (u τ(u))) du 4 1 su s ϑ s s ( ) e c (s u) l j h j (ϕ j (v)) h j (ψ j (v)) dv du s r(s) 4 1 su s ϑ s ( ) e c (s u) σ j (s, x j (s), x j (u τ(u))) σ j (s, y j (s), y j (s τ(u))) dω j (u) / s 4 1 su c / a j α j e c(s u) ϕ j (u) ψ(u) du s ϑ / s 4 1 su c / b j β j e c(s u) ϕ j (u τ(u))) ψ j (u τ(u))) du s ϑ ( τ / s u 4 1 su s ϑ c ) / l j γ j e c (s u) ϕ j (v) ψ j (v) dv du u r(u) 4 1 n 1 s su c 1 /2 s ϑ µ/2 e c(s u) ϕ j (u) ψ j (u) du 11

12 s ν /2 e c(s u) ϕ j (u τ(u)) ψ j (u τ(u)) du / / 4 1 c a j α j c b j β j ( / r ( c ) l j γ j n 1 c /2 µ /2 ν /2) su s ϑ ϕ j (s) ψ j (s) =α su s ϑ ϕ j (s) ψ j (s). From (4), we oban ha Q :S φ S φ s a conracon mang. We are now ready o rove Theorem 1.4. Proof. From Lemma 2.2, by he conracon mang rncle, we oban ha Q has a unue fxed on x(), whch s a soluon of (1) wh x()=φ() as [ϑ, ] and n x () as. Now, we rove ha he rval soluon of (1) s h momen sable. Leǫ> be gven and chooseδ> (δ<ǫ) such ha 5 1 δ<(1 α)ǫ. If x()=(x 1 (), x 2 (), x n ()) T s a soluon of (1) wh he nal condon (2) sasfyng n φ () <δ, hen x()=(qx)() defned n (14). We clam ha n x () <ǫ for all. Noce ha n x () <ǫ for [ϑ, ], we suose ha here exss > such ha n x ( ) =ǫand n x () <ǫforϑ <, hen follows from (4), we oban ha / x ( ) 5 1 e c x () 5 1 c / a j α j e c ( s) x j (s) ds / 5 1 c / b j β j e c( s) x j (s τ(s))) ds ( r / s 5 1 c ) / l j γ j e c ( s) x j (u) du ds s r(s) 5 1 n 1 c 1 /2 µ/2 e c ( s) x j (s) ds ν /2 e c( s) x j (s τ(s)) ds / / 5 1 c a j α j 5 1 c b j β j ( r 5 1 c ) l j γ j < (1 α)ǫαǫ=ǫ. / 5 1 n 1 ( c /2 µ /2 ν /2) ǫ 5 1 δ whch s a conradcon. Therefore, he rval soluon of (1) s asymocally sable n h momen. Corollary 2.3. Suose ha he assumons (A1)-(A4) hold. If he followng condons are sasfed, () he dsrbued delay r() s bounded by a consan r; 12

13 () 5 c 2 a 2 j α2 j 5 c 2 b 2 j β2 j ( r 5 c ) 2 l j γ j 2n c 1 (µν) α, where c,µ,ν are defned as n Theorem 1.4, hen he rval soluon of (1) s asymocally sable n mean suare Consder he sochasc neural neworks whou dsrbued delays dx ()= c x () a j f j (x j ()) b j g j (x j ( τ())) d σ j (, x j (), x j ( τ())) dω j () (19) for, 2, 3,, n. Corollary 2.4. Suose ha he assumons (A1)-(A4) hold. The rval soluon of (19) s asymocally sable n h momen f he followng neualy holds, 4 1 c a j α j / 4 1 whereµ,ν are defned as n Theorem 1.4. c b j β j / 4 1 n 1 Remark 2.5. Noe ha he dscree delayτ() n Corollary 2.4 can be unbounded. ( c /2 µ /2 ν /2) α, (2) 3. Proof of Theorem 1.5 In hs secon, we rove Theorem 1.5. We sar wh some rearaons. Lemma 3.1. Defne an oeraor by (Pϕ)()=φ() for [ τ, ], and for, (Pϕ)() s defned as (14), f here s α (, 1) such ha (6) holds, hen P :C φ C φ s a conracon mang. Proof. From he roof of Theorem 1.4, we oban ha P s connuous n h momen on [, ). Now, we rove ha P(C φ ) C φ. [ su Qϕ (s) ]= su 5 5 [ ] J j (s) 5 1 su J j (s). We esmae he erms on he rgh-hand sde of he above neualy. Le c=mn{c 1, c 2, c 3,, c n }, su J 2 (s) s = su e c (s u) a j f j (ϕ j (u)) du s c / su e c(s u) a j α j ϕ j (u) du / s c / a j α j su e c(s u) ϕ j (u) du / n { [ s ]} c / a j α j e c(s u) ϕ j (u) du 13 su

14 / { c / a j α j / e cτ c / a j α j / ( e cτ c a j α j su [ [ s ( e c(s u) su e c( u) ( su u τ v u u τ v u ϕ j (v) ) ϕ j (v) ) ] du ]} du ) su x j (s). (21) Snce n su ϕ j (s) as, hen for anyǫ>, here exss such ha T 1 mles [ su ϕ j (s) ]<ǫ. Then, combnng wh (21), we oban ha / J 2 (s) < ecτ c a j α j ǫ. su Hence, we oban ha n su J 2 (s) as. Smlarly, we oban ha / s su J 3 (s) c / b j β j su e c(s u) ϕ j (u τ(u)) du / { [ s ]} c / b j β j su e c(s u) ϕ j (u τ(u)) du / { [ s ]} c / b j β j su e c(s u) su ϕ j (v) du u τ v u / [ ] e cτ c / b j β j e c( u) su ϕ j (v) du u τ v u / [ ] e cτ c b j β j su ϕ j (s). (22) Snce n [ su ϕ j (s) ] as, ha s, for anyǫ >, here exss such ha T 2 mles n [ su ϕ j (s) ] <ǫ, so combnng wh (22) we have su Hence, n su J 3 (s) as. su J 4 (s) c / l j γ j c / l j γ j / J 3 (s) < ecτ c b j β j ǫ. / / 14 su n { su s e c(s u) u u r(u) [ s u e c(s u) u r(u) ϕ j (v) dv ϕ j (v) dv du ]} du

15 / { τ c / l j γ j / τ e cτ c / l j γ j / n [ τ e cτ c l j β j su [ [ s e c(s u) su u τ v u e c( u) su u τ v u ] su ϕ j (s). ]} ϕ j (v) du ] ϕ j (v) du Hence, we have ha [ n su J 4 (s) ] as. Leµ=max{µ 1,µ 2,,µ n },ν=max{ν 1,ν 2,,ν n }, [ s J 5 (s) n 1 su e c(s u) σ j (u,ϕ j (u),ϕ j (u τ(u))) dω j (u) ] su n 1 n 1 { n 1 e cτ su su [ su τ r { [ su su s τ r s [ e c(r u) σ j (u,ϕ j (u),ϕ j (u τ(u))) dω j (u) ]} s e c(r u) σ j (u,ϕ j (u),ϕ j (u τ(u))) dω j (u) ]} e c( u) σ j (u,ϕ j (u),ϕ j (u τ(u))) dω j (u) ( s /2 K n 1 e cτ su e 2c( u) σ 2 j (u,ϕ j(u),ϕ j (u τ(u))) du) ( s K n 1 e cτ 2 /2 1 su e 2c( u)( /2 µ j ϕ 2 j du) (u)) ( s K n 1 e cτ 2 /2 1 su e 2c(T u)( /2 ν j ϕ 2 j du) (u τ(u))) ( s ) /2 1 K n 1 e cτ 2 /2 1 su e 2c( u) du ( s s )]} e 2c( u) µ /2 j ϕ j (u) du e 2c(T u) ν /2 j ϕ j (u τ(u)) du [ ] K n e cτ c 1 /2 (µ /2 ν /2 ) e 2c(T u) su ϕ j (v) du u τ v u [ ] K n e cτ c 1 /2 (2c) 1 (µ /2 ν /2 ) su ϕ j (s). (23) Snce n [ su ϕ j (s) ] as, ha s, for anyǫ>, here exss T 3 > such ha T 3 mles n [ su ϕ j (s) ] <ǫ, so combnng wh (23) we have su J 5 (s) < K n e cτ c 1 /2 (2c) 1 (µ /2 ν /2 )ǫ. Hence, [ n su J 5 (s) ] as. Thus, P(C φ ) C φ. ] 15

16 Fnally, we rove ha Q s a conracon mang. For anyϕ,ψ C φ, from (21)-(23), we oban ha su su Qϕ (s) Qψ (s) 4 1 su s ( ) su e c (s u) a j f j (ϕ j (u)) f j (ψ j (u)) du 4 1 su s ) su e c (s u) b j (g j (ϕ j (u τ(u))) g j (ψ j (u τ(u))) du 4 1 su s s ( ) su e c (s u) l j h j (ϕ j (v)) h j (ψ j (v)) dv du s r(s) 4 1 su s su e c (s u) [σ j (u,ϕ j (u),ϕ j (u τ(u))) σ j (u,ψ j (u),ψ j (u τ(u)))] dω j (u) / / / 4 1 ecτ c a j α j e cτ c b j β j τ e cτ c l j γ j K n e cτ c 1 /2 (2c) 1( µ /2 ν /2) [ ] su su ϕ j (s) ψ j (s) =α su [ ] su ϕ j (s) ψ j (s). From (6), we oban ha Q :C φ C φ s a conracon mang. We are now ready o rove Theorem 1.5 Proof. From Lemma 3.1, by he conracon mang rncle, we oban ha P has a unue fxed on x(), whch s a soluon of (1) wh x()=φ() as [ τ, ] and n [ su x (s) ] as. We rove ha he rval soluon of (1) s h momen sable. Leǫ > be gven and chooseδ> (δ<ǫ) sasfyng 5 1 e c δ<(1 α)ǫ. (24) If x()=(x 1 (), x 2 (), x n ()) T s a soluon of (1) wh he nal condon sasfyng φ <δ, hen x()=(px)() defned n (14). We clam ha x <ǫfor all. Noce ha φ() <ǫfor [ τ, ], we suose ha here exss > such ha n [ su τ s x (s) ] =ǫand n [ su x (s) ] <ǫfor τ <, hen follows from (4) and (24), we oban ha [ ] su x (s) τ s [ su J j (s) τ s ] / 5 1 e c δ5 1 ecτ c a j α j e cτ c b j β j / τ e cτ c l j γ j K n e cτ c 1 /2 (2c) 1( µ /2 ν /2) ǫ < (1 α)ǫαǫ=ǫ. whch s a conradcon. Therefore, he rval soluon of (1) s asymocally sable n h momen. 16 /

17 4. Proof of Theorem 1.8 In hs secon, we rove Theorem 1.8. We sar wh a lemma resenng an negral neualy. Lemma 4.1. Consderγ>, osve consansλ 1,λ 2,λ 3 and a funcon y : [ τ, ) [, ). Ifλ 1 λ 2 τλ 3 < c and he followng neualy holds, y e c λ 1 e c( s) y(s) dsλ 2 e c( s) y(s τ(s)) dsλ 3 e c( s) s y(u) du ds, s r(s) y() (25) y e c, [ τ, ], ( ) hen we have y() y e γ ( τ), whereγs a osve roo of he algebrac euaon 1 c γ λ1 e γτ λ 2 eγτ 1 γ λ 3 = 1. ( ) Proof. Le F(γ)= 1 c γ λ1 e γτ λ 2 eγτ 1 γ λ 3 1. We have F()F(c )<, ha s, here exss a osve consan γ (, c) such ha F(γ)=. For anyǫ>, le To rove he lemma, we clam ha (25) mles C ǫ =ǫ y. y() C ǫ e γ, τ. (26) I s easly shown ha (26) holds for [ τ, ]. Assume ha here exss 1 > such ha Combnng wh (25), we have y()<c ǫ e γ, [ τ, 1 ), y( 1 )=C ǫe γ 1. (27) y(1 ) y 1 e c 1 λ1 e c( 1 s) 1 y(s) dsλ 2 e c( 1 s) 1 s y(s τ(s)) dsλ 3 e c( 1 s) 1 < y e c 1 Cǫ λ 1 e c( 1 s) e γs 1 ds C ǫ λ 2 e c( 1 s) e γ(s τ(s)) 1 s ds C ǫ λ 3 e c( 1 s) = [ y C ( )] ǫ λ 1 e γτ λ 2 eγτ 1 C λ 3 e c 1 ǫ c γ γ c γ From he defnon of C ǫ, we have y C ǫ c γ ( ( ) λ 1 e γτ λ 2 eγτ 1 λ 3 e γ 1. γ λ 1 e γτ λ 2 eγτ 1 λ 3 )=y C ǫ <. γ s r(s) y(u) du ds s r(s) e γu du ds Then, ogeher wh he defnon ofγ, we oban ha y( 1 )< C ǫe γ 1, whch conradcs (27), so (26) holds. Asǫ> s arbrarly small, n vew of (26), follows ha y() y e γ, for τ. Proof. For he reresenaon (13), usng (16)-(18), we oban ha / x () 5 1 e c φ () 5 1 c / a j α j e c( s) x j (s) ds / 5 1 c / b j β j e c( s) x j (s τ(s))) ds ( τ ) / / s 5 1 c l j γ j e c( s) x j (u) du ds s r(s) 5 1 n c 1 /2 µ/2 e c( s) x j (s) dsν/2 e c( s) x j (s τ(s)) ds. 17

18 Hence, by usng Lemma 4.1 and (7), we oban ha he rval soluon of (1) s exonenally sable n h momen. Corollary 4.2. Suose ha he assumons (A1)-(A4) hold. If he followng condons are sasfed, () he dscree delayτ() and dsrbued delay r() are bounded by a consanτ; () 5c 2 a 2 j α2 j 5c 2 where c,µ,ν are defned as n Theorem 1.4, b 2 j β2 j 5c 2 τ 2 hen he rval soluon of (1) s exonenally sable n mean suare, l 2 j γ2 j 2n2 c 1 (µν)<1, Corollary 4.3. Le 2. Suose ha he assumons (A1)-(A4) hold. If he followng condons are sasfed, () he dscree delayτ() and dsrbued delay r() are bounded by a consanτ; () / / 4 1 c a j α j 4 1 c b j β j 4 1 n c /2 (µ /2 ν /2 )<1, where c,µ,ν are defned as n Theorem 1.4, hen he rval soluon of (19) s exonenally sable n h momen. 5. Proof of Theorem 1.12 In hs secon, we rove Theorem We sar wh some rearaons. Mully boh sdes of (9) by e c and negrae from o, we oban ha for, x () = e c x () e c ( s) a j g j (x j (s)) ds e c ( s) b j g j (x j (s τ(s))) ds e c ( s) d j s s r(s) g j (x j (u)) du ds,, 2, 3,, n. (28) Lemma 5.1. Defne an oeraor by (Px)(θ) = φ(θ), for ϑ θ, and for, Px () = e c x () e c ( s) a j g j (x j (s)) ds e c ( s) b j g j (x j (s τ(s))) ds e c ( s) d j s s r(s) g j (x j (u)) du ds := I 1 () I 2 () I 3 () I 4 (). (29) If here exssα (, 1) such ha (11) holds, hen P :H φ H φ and P s a conracon mang. Proof. Frs, we rove ha PH φ H φ. In vew of (29), we have ha, for fxed me 1, s easy o check ha lm r [(Px )( 1 r) (Px )( 1 )]=. Thus, P s connuous on [, ). Noe ha (Px )(θ)=φ(θ) forϑ θ, we oban ha P s ndeed connuous on [ϑ, ). 18

19 Nex, we rove ha lm (Px )()= for x () H φ. Snce x () H φ, we have ha lm x ()=. Then for anyǫ>, here exss T > such ha s T mles x (s) <ǫ. Choose T= max,2,,n {T }, combnng wh condon (A2), I 2 () = e c ( s) a j f j (x j (s)) ds T e c ( s) a j k j x j (s) ds e c ( s) a j α j x j (s) ds e c a j α j su x j (s) s T T a j α j su x j (s) s T T e c ( s) dsǫ T a j α j e c s ds ǫ c T e c ( s) ds a j α j. (3) From he fac ha c > (, 2,, n) and esmae (3), we have ha I 2 () as. Snce x () and τ() as, for eachǫ >, here exss T > such ha s T mles x (s τ(s)) <ǫ for, 2,, n. Choose T = max,2, n {T }, we oban I 3 () = e c ( s) b j g j (x j (s τ(s))) ds T e c ( s) b j β j x j (s τ(s)) ds e c ( s) b j k j x j (s τ(s)) ds T e c b j β j su ϑ s T x j (s) T e c s ds ǫ c b j β j. (31) From he fac ha c > (, 2,, n) and esmae (31), we have ha I 3 () as. Snce x () and r() as, for eachǫ >, here exss T > such ha s T mles x (s r(s)) <ǫ for, 2,, n. Choose T = max,2, n {T }, we oban s I 4 () = e c ( s) d j h j (x j (u)) du ds s r(s) T s e c ( s) d j γ j x j (u) du dsǫr e c ( s) d j γ j ds s r(s) T T r d j γ j su x j (u) e c( s) ds ǫr d j γ j. (32) ϑ u T c From he fac ha c > (, 2,, n) and esmae (32), we have ha I 4 () as. From he above esmae, we conclude ha lm (Px )()= for x () H φ. Therefore, P :H φ H φ. Now, we rove ha P s a conracon mang. For any x(), y() H φ, from (3) and (32), we oban ha (Px )() (Py )() max a jα j,2,,n max b jβ j,2,,n e c ( s) e c ( s) x j (s) y j (s) ds x j (s τ(s)) y j (s τ(s)) ds 19

20 { 1 c = α su ϑ s max d jγ j,2,,n max,2,,n a jα j 1 c x j (s) y j (s). From (11), we oban ha P s a conracon mang. We are now ready o rove Theorem s e c ( s) s r(s) max,2,,n b jβ j r c x j (u) y j (u) du ds } max d jγ j su,2,,n ϑ s x j (s) y j (s) Proof. Le P be defned as n Lemma 5.1, by conracon mang rncle, P has a unue fxed on x H φ wh x(θ)=φ(θ) onϑ θ and x() as. To oban asymocally sable, we need o rove ha he rval eulbrum x = of (9) s sable. For any ǫ >, chooseσ> andσ<ǫ sasfyng he condonσǫα<ǫ. If x(, s,φ)=(x 1 (, s,φ), x 2 (, s,φ),, x n (, s,φ)) s he soluon of (9) wh he nal condon φ <σ, he we clam ha x(, s,φ) <ǫ for all. Indeed, we suose ha here exss > such ha x ( ; s,φ) =ǫ, and x (; s,φ) <ǫ for <. (33) From (11) and (28), we oban x ( ; s,φ) e c x () e c ( s) s e c ( s) d j 1 < σǫ max a jα j,2,,n c s r(s) 1 c a j f j (x j (s)) ds e c ( s) h j (x j (u)) du ds max b jβ j,2,,n r whch conradcs (33). Therefore, x(, s,φ) <ǫ for all. Ths comlees he roof. c b j g j (x j (s τ(s))) ds max d jγ j σǫα<ǫ.,2,,n Le d j for, 2, n,, 2, n, he sysem s reduced o dx () = c x () a j f j (x j ()) b j g j (x j ( τ())), (34) d whch s he descron of cellular neural nework wh me-varyng delays. Followng he resul of heorem 1.12, we have he followng corollary. Corollary 5.2. Suose ha he assumons (A1)-(A3) hold. If he followng condon s sasfed, 1 max a 1 jα j max b jβ j <1, (35),2,,n,2,,n hen he rval soluon of (34) s asymocally sable. c Remark 5.3. Noe ha he delay n Corollary 5.2 can be unbounded. La and Zhang [12] suded he asymoc sably (34) as well. However, he addonal condon 1 max a j k j 1 b j k j c < 1 (36) n,2, n c s needed n Theorem 4.1 of [12]. I s clearly ha Corollary 5.2 s an mrovemen of he resul n [12]. c 2

21 6. Proof of Theorem 1.13 Proof. From he reresenon (28), we oban ha x () e c x () max { b jk j },2, n max { d jk j },2, n e c( s) e c( s) s e c( s) s r(s) max { a jk j },2, n x j (s τ(s)) ds x j (s) ds x j (u) du ds. Combnng wh Lemma 4.1, we oban ha he rval soluon of (9) wh nal condon (1) s exonenally sable. For he cellular neural nework (34), we have he followng resul. Corollary 6.1. Suose ha he assumons (A1)-(A3) hold. If he followng condons are sasfed, () he dscree delayτ() and dsrbued delay r() are bounded by a consanτ; () max a jk j,2,,n max b jk j <c, c=mn{c 1, c 2, c 1 },,2,,n hen he rval soluon of (34) wh nal condon (1) s exonenally sable. 7. xamles xamle 7.1. Consder he followng wo-dmensonal cellular neural nework where C= ( ) c1 = c 2 ( 3 3 dx() d = Cx()Ag(x()) Bg(x τ()), ) ( ) ( a11 a, A= 12 = a 21 a 22 6/7 3/7 1/7 1/7 ) ( ) ( b11 b, B= 12 6/7 2/7 = b 21 b 22 3/7 1/7 The acvaon funcon s descrbed by g (x)= x1 x 1 2, for, 2. The me-varyng delayτ() s connuous and τ() τ, whereτs a consan. I s clear haα =β = 1 for, 2. We check he condon (35) n Corollary 5.2, 2 1 c max,2 a jα j 2 1 max c b jβ j 1 ( 6, ) = < 1. Hence, by Corollary 5.2, he rval eulbrum x = of hs cellular neural nework s asymocally sable. However, he condon (36) becomes 1 2 max a j α j 1 2 b j β j,2 c c = > 1. 2 ). Hence, Theorem 4.1 of [12] s no alcable. 21

22 xamle 7.2. Consder he wo-dmensonal sochasc recurren neural nework wh me-varyng delays ( )( ) ( )( ) 6 x1 () anh(x1 ()) dx() = d d 5 x 2 () anh(x 2 ()) ( )( ) ( ) anh(x1 ( τ 1 ())) 1 2 r() d anh(x 2 ( τ 2 ())) anh(x 1(s)) ds.2 anh(x r() 2(s)) ds) d σ(, x(), x( τ())) dw(), (37) whereτ(), r() are connuous funcons such ha τ() as and r() 1,σ : R R 2 R 2 R 2 R 2 sasfes race [ σ T (, x, y)σ(, x, y) ].3(x 2 1 x2 2 y2 1 y2 2 ). We suose =2, and akeµ =ν =.3 for, 2, by smle comuaon, we haveα =.2, for, 2, c=mn{c 1, c 2 }=5,µ=ν=.3. From Corollary 2.3, we have ha 5 2 c 2 2 a 2 j α2 j 5 2 c 2 2 b 2 j α2 j 2 ( τ 2 5 c ) 2 l 2 j α2 j c 1 (µν)<.256<1. Then he rval soluon of (37) s mean suare asymocally sable. If τ() s bounded, from Corollary 4.2, we oban ha c 2 a 2 j α2 j 5c 2 b 2 j α2 j 5c 2 τ 2 l 2 j α2 j 2 4c 1 (µν)<.298. Hence, he rval soluon of (37) s mean suare exonenally sable. xamle 7.3. Consder a wo-dmensonal sochascally erurbed HNN wh me-varyng delays, dx()=[ Cx()A f (x()) Bg(x τ ())] dσ(, x(), x τ ()) dω(), (38) where f (x)= 1 5 arcan x, g(x)= 1 5 anh x= 1 5 (ex e x )/(e x e x ),τ()= 1 2 sn 1 2, C= ( ) (, A= ) and B= ( In hs examle, le =3, akeα j =.2,β j =.2,, 2,σ : R R 2 R 2 R 2 R 2 sasfes race [ σ T (, x, y)σ(, x, y) ].1(x 2 1 x2 2 y2 1 y2 2 ). Noe ha he exonenal sably of (38) has been suded n Sun and Cao [28] by emloyng he mehod of varaon arameer, neualy echnue and sochasc analyss. Now, we check he condon n Corollary 4.3, / / c (1/) a j α j 4 1 c (1/) b j β j c /2 (µ /2 ν /2 )<.18<1. From Corollary 4.3, he rval soluon of (38) s exonenally sable. ). 22

23 8. Aendx In hs secon, we frs show ha I 5 (s) n [19] s no a local marngale and hen we resen some examles abou Banach saces. Defnon 8.1. A real valuedf -adaed rocess M={M() : } s a marngale f M() < for all and [M() F s ]= M(s), a.s. for all s<<. Lemma 8.2. For connuous funconσ(), e c( s) σ(s) dω(s) s no a marngale. Proof. In fac, for u, [ ] e c( s) σ(s) dω(s) F u [ u = = u [ e c( s) σ(s) dω(s) F u ] e c( s) σ(s) dω(s) u u e c( s) σ(s) dω(s) F u ] e c(u s) σ(s) dω(s). (39) Lemma 8.3. ([24]) If M() s a local marngale and for every, su s [,] M(s) <, hen M() s a marngale. Lemma 8.4. For connuous funconσ(), e c( s) σ(s) dω(s) s no a local marngale. Proof. We suose ha e c( s) σ(s) dω(s) s a local marngale. For every, we have ha s su e c(s u) σ(u) dω(u) s = su e cs e cu σ(u) dω(u) s [,] s [,] s su e cu σ(u) dω(u) s [,] ( ) 1/2 ( 1/2 K 1 e 2cu σ 2 (u) du K 1 e 2cu σ 2 (u) du) <. From Lemma 8.3, we oban ha M s a marngale. However, from Lemma 8.2, we know ha e c( s) σ(s) dω(s) s no a marngale, whch s a conradcon. A normed lnear sace s a merc sace wh resec o he merc d derved from s norm, where d(x, y)= x y. Defnon 8.5. A Banach sace s a normed lnear sace ha s comlee merc sace wh resec o he merc derved from s norm. Here are some examles of Banach saces. xamle 8.6. The sace C([a, b]) of connuous, real-valued (or comlex-valued) funcons on [a, b] wh he sunormed s a Banach sace. More generally, we have he followng examles. () If X s a Banach sace, he sace C([a, b]; X) of connuous, X-valued funcons on [a, b] eued wh he su-norm s a Banach sace. () If X s a Banach sace, he sace BC([a, b]; X) :={ϕ ϕ C([a, b]; X), ϕ < } of bounded connuous, X- valued funcons on [a, b] eued wh he su-norm s a Banach sace. () If X s a Banach sace, he sace{ϕ ϕ C([a, b]; X), lm ϕ()=} and he sace { } ϕ ϕ C([a, b]; X), ϕ = su ϕ(s) s bounded and lmϕ()= s [a,b] are Banach saces wh resec o he su-norm. Clearly, he sace { } C ([a, b]; L (Ω,R n )) := ϕ ϕ C([a, b]; L (Ω,R n )), lm ϕ() = s a Banach saces wh resec o he norm defned by ϕ := su s [ ϕ(s) ]. 23

24 The followng lemma resens a Banach sace ha s used n hs aer. Lemma 8.7. Suose haf s comlee, ha s, conans all null ses. Then he sace s a closed subsace of C ([a, b]; L (Ω,R n )). D :={ϕ C ([a, b]; L (Ω,R n )),ϕ() sf measurable for all } Proof. Leϕ(),ψ() D, henϕ() andψ() aref -measurable for all, soϕ()ψ() andαϕ() (α C) aref - measurable for all. Suose ha he seuenceϕ 1 (),ϕ 2 (), ϕ n () D,ϕ() C ([a, b]; L (Ω,R n )) andϕ n () ϕ() as n for all, we clam haϕ() sf -measurable. In fac, snceϕ n () ϕ() as n, hen [ su ϕn (s) ϕ(s) ] as n. s Ω So, for every, we oban ha ϕ n (s) ϕ(s) as n, whch mles ha here exss a subseuence (ϕ nk ()) k such haϕ nk () ϕ() a.e. onω. On he oher hand,f s comlee. Hence, we oban haϕ() sf -measurable, whch mles ha D s a closed subsace of he sace C ([a, b]; L (Ω,R n )). References [1] J.A.D. Aleby, Fxed ons, sably and harmless sochasc erurbaons, Prern, 28. [2] T.A. Buron, Sably by fxed on heory for funconal dfferenal euaons. Dover Publcaon, New York, 26. [3] H.B. Chen, Inegral neualy and exonenal sably for neural sochasc aral dfferenal euaons wh delays, Journal of neuales and alcaons 29, Ar. ID , 15 ages. [4] H.B. Chen, Imulsve-negral neualy and exonenal sably for sochasc aral dfferenal euaons wh delays, Sascs and Probably Leers 8 (21), [5] A. Djoud, R. Khems, Fxed ons echnues and sably for neural nonlnear dfferenal euaons wh unbounded delays, Georgan Mahemacal Journal 13(26), No.1, [6] M. For, A. Tes, New condons for global sably of neural neworks wh alcaon o lnear and uadrac rogrammng roblems, I Trans. Crc. Sys. I, 42 (1995), [7] A. Fredman, Sochasc dfferenal euaons and alcaons, Academc ress, New York, [8] S. Haykn Neural neworks, Prence-Hall, nglewood Clffs, [9] J. Hu, S. Zhong, L. Lang, xonenal sably analyss of sochasc delayed cellular neural nework, Chaos Solon Frac. 27 (26), [1] C.X. Huang, Y.G. He, H.N. Wang, Mean suare exonenal sably of sochasc recurren neural neworks wh me-varyng delays, Comuers and Mahemacs wh Alcaons 56 (28), [11] C.X. Huang, Y.G. He, L.H. Huang, W.J. Zhu, h momen sably analyss of sochasc recurren neural neworks wh me-varyng delays, Informaon Scence 178 (28), [12] X.H. La, Y.T. Zhang, Fxed on and asymoc analyss of cellular neural neworks, Journal of Aled Mahemacs 212 (212), Arcle ID , 12 ages do:1.1155/212/ [13] T. L, Q. Luo, C.Y. Sun, B.Y. Zhang, xonenal sably of recurren neural neworks wh me-varyng dscree and dsrbued delays, Nonlnear Analyss: Real World Alcaons 1 (29), [14] X. Lao, X. Mao, xonenal sably and nsably of sochasc neural neworks, Sochasc Analyss and Alcaon 14 (2) (1996), [15] X. Lao, X. Mao, Sably of sochasc neural neworks, Neural, Parallel and Scenfc Comuaons 4 (2) (1996), [16] X.X. Lao, J. Wang, Algebrac crera for global exonenal sably of cellular neural neworks wh mulle me delays, I TARANS. Crc. Sya. I. 5 (23), [17] X. Lao, X. Mao. xonenal sably of sochasc delay nerval sysems, Sys. Conrol. Le., 2, 4: [18] D.Z. Lu, G.Y. Yang and W. Zhang, The sably of neural sochasc delay dfferenal euaons wh osson jums by fxed ons, Journal of Comuaonal and Aled Mahemacs 235(211), [19] J.W. Luo, Fxed ons and sably of neural sochasc delay dfferenal euaons. J. Mah. Anal. Al., 334(27), [2] J.W. Luo and T. Tanguch, Fxed ons and sably of sochasc neural aral dfferenal euaons wh nfne delays, Sochasc Analyss and Alcaons 27 (29), [21] J.W. Luo, Fxed ons and exonenal sably of mld soluons of sochasc aral dfferenal euaons wh delays, J.Mah.Anal.Al. 342 (28), [22] S..A. Mohammed, Sochasc funconal dfferenal euaons, Longman Scenfc and Techncal, [23] X.R. Mao, Razumkhn-ye heorems on exonenal sably of sochasc funconal dfferenal euaons, Sochas. Proc. Al., 1996, 65: [24] P.. Proer, Sochasc Inegraon and Dfferenal uaons, 2nd don, 24. [25] R. Sakhvel, J. Luo, Asymoc sably of nonlnear mulsve sochasc dfferenal euaons, Sas. Probab. Le. 79 (29),

25 [26] R. Sakhvel, J. Luo, Asymoc sably of mulsve sochasc aral dfferenal euaons wh nfne delays. J. Mah. Anal. Al. 356 (29), 1-6. [27] R. Sakhvel, R. Samdura, S.M. Anhon, Asymoc sably of sochasc delayed recurren neural neworks wh mulsve effecs, J Om Theory Al 147 (21), [28] Y.H. Sun, J.D. Cao, h momen exonenal sably of sochasc recurren neural neworks wh me-varyng delays, Nonlnear Analyss: Real World Alcaons 8 (27) [29] J.K. Tan, S.M. Zhong, Y. Wang, Imroved exonenal sably crera for neural neworks wh me-varyng delays, Neurocomung 97 (212), [3] J.K. Tan, S.M. Zhong, New delay-deenden exonenal sably crera for neural neworks wh dscree and dsrbued me-varyng delays, Neurocomung 74 (211), [31] B.D. Vres, J.C. Prncle, The gamma model- a new neural model for emoral rocessng, Neural Neworks, 5 (1992), [32] L. Wan, J. Sun, Mean suare exonenal sably of sochasc delayed Hofeld neural neworks. Phys. Le. A 343 (25), [33] M. Wu, N.J. Huang and C.W. Zhao, Fxed ons and sably n neural sochasc dfferenal euaons wh varable delays, Fxed Pon Theory and Alcaons, Volume 28, Arcle ID 47352, 11 ages. do: /28/ [34] M. Wu, N.J. Huang and C.W. Zhao, Sably of half-lnear neural sochasc dfferenal euaons wh delays, Bull.Aus.Mah.Soc. 8 (29), [35] J.J. Yu, K.J. Zhang, S.M. Fe, T. L, Smlfed exonenal sably analyss for recurren neural neworks wh dscree and dsrbued me-varyng delays, Aled Mahemacs and Comuaon 25 (28), [36] Z. Yuan, L. Yuan, L. Huang, Dynamcs of erodc Cohen-Grossberg neural neworks wh varyng delays, Neurocomung 7 (26), [37] Z.G. Zeng, D.S. Wang, Global sably of a general class of dscree-me recurren neural neworks, Neural Comuaon, 18 (26), [38] Z.G. Zeng, J. Wang, Comlee sably of cellular neural neworks wh me-varyng delays, I Trans. Crc. Sys. I., 53 (26), [39] H.Y. Zhao, Global asymoc sably of Hofeld neural nework nvolvng dsrbued delays, Neural Neworks 17 (24),

A New Generalized Gronwall-Bellman Type Inequality

A New Generalized Gronwall-Bellman Type Inequality 22 Inernaonal Conference on Image, Vson and Comung (ICIVC 22) IPCSIT vol. 5 (22) (22) IACSIT Press, Sngaore DOI:.7763/IPCSIT.22.V5.46 A New Generalzed Gronwall-Bellman Tye Ineualy Qnghua Feng School of

More information

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas)

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas) Lecure 8: The Lalace Transform (See Secons 88- and 47 n Boas) Recall ha our bg-cure goal s he analyss of he dfferenal equaon, ax bx cx F, where we emloy varous exansons for he drvng funcon F deendng on

More information

EP2200 Queuing theory and teletraffic systems. 3rd lecture Markov chains Birth-death process - Poisson process. Viktoria Fodor KTH EES

EP2200 Queuing theory and teletraffic systems. 3rd lecture Markov chains Birth-death process - Poisson process. Viktoria Fodor KTH EES EP Queung heory and eleraffc sysems 3rd lecure Marov chans Brh-deah rocess - Posson rocess Vora Fodor KTH EES Oulne for oday Marov rocesses Connuous-me Marov-chans Grah and marx reresenaon Transen and

More information

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation Global Journal of Pure and Appled Mahemacs. ISSN 973-768 Volume 4, Number 6 (8), pp. 89-87 Research Inda Publcaons hp://www.rpublcaon.com Exsence and Unqueness Resuls for Random Impulsve Inegro-Dfferenal

More information

FI 3103 Quantum Physics

FI 3103 Quantum Physics /9/4 FI 33 Quanum Physcs Aleander A. Iskandar Physcs of Magnesm and Phooncs Research Grou Insu Teknolog Bandung Basc Conces n Quanum Physcs Probably and Eecaon Value Hesenberg Uncerany Prncle Wave Funcon

More information

PHYS 705: Classical Mechanics. Canonical Transformation

PHYS 705: Classical Mechanics. Canonical Transformation PHYS 705: Classcal Mechancs Canoncal Transformaon Canoncal Varables and Hamlonan Formalsm As we have seen, n he Hamlonan Formulaon of Mechancs,, are ndeenden varables n hase sace on eual foong The Hamlon

More information

On One Analytic Method of. Constructing Program Controls

On One Analytic Method of. Constructing Program Controls Appled Mahemacal Scences, Vol. 9, 05, no. 8, 409-407 HIKARI Ld, www.m-hkar.com hp://dx.do.org/0.988/ams.05.54349 On One Analyc Mehod of Consrucng Program Conrols A. N. Kvko, S. V. Chsyakov and Yu. E. Balyna

More information

Relative controllability of nonlinear systems with delays in control

Relative controllability of nonlinear systems with delays in control Relave conrollably o nonlnear sysems wh delays n conrol Jerzy Klamka Insue o Conrol Engneerng, Slesan Techncal Unversy, 44- Glwce, Poland. phone/ax : 48 32 37227, {jklamka}@a.polsl.glwce.pl Keywor: Conrollably.

More information

Advanced time-series analysis (University of Lund, Economic History Department)

Advanced time-series analysis (University of Lund, Economic History Department) Advanced me-seres analss (Unvers of Lund, Economc Hsor Dearmen) 3 Jan-3 Februar and 6-3 March Lecure 4 Economerc echnues for saonar seres : Unvarae sochasc models wh Box- Jenns mehodolog, smle forecasng

More information

CONSISTENT ESTIMATION OF THE NUMBER OF DYNAMIC FACTORS IN A LARGE N AND T PANEL. Detailed Appendix

CONSISTENT ESTIMATION OF THE NUMBER OF DYNAMIC FACTORS IN A LARGE N AND T PANEL. Detailed Appendix COSISE ESIMAIO OF HE UMBER OF DYAMIC FACORS I A LARGE AD PAEL Dealed Aendx July 005 hs verson: May 9, 006 Dane Amengual Dearmen of Economcs, Prnceon Unversy and Mar W Wason* Woodrow Wlson School and Dearmen

More information

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model BGC1: Survval and even hsory analyss Oslo, March-May 212 Monday May 7h and Tuesday May 8h The addve regresson model Ørnulf Borgan Deparmen of Mahemacs Unversy of Oslo Oulne of program: Recapulaon Counng

More information

Part II CONTINUOUS TIME STOCHASTIC PROCESSES

Part II CONTINUOUS TIME STOCHASTIC PROCESSES Par II CONTINUOUS TIME STOCHASTIC PROCESSES 4 Chaper 4 For an advanced analyss of he properes of he Wener process, see: Revus D and Yor M: Connuous marngales and Brownan Moon Karazas I and Shreve S E:

More information

Comparison of Differences between Power Means 1

Comparison of Differences between Power Means 1 In. Journal of Mah. Analyss, Vol. 7, 203, no., 5-55 Comparson of Dfferences beween Power Means Chang-An Tan, Guanghua Sh and Fe Zuo College of Mahemacs and Informaon Scence Henan Normal Unversy, 453007,

More information

Let s treat the problem of the response of a system to an applied external force. Again,

Let s treat the problem of the response of a system to an applied external force. Again, Page 33 QUANTUM LNEAR RESPONSE FUNCTON Le s rea he problem of he response of a sysem o an appled exernal force. Agan, H() H f () A H + V () Exernal agen acng on nernal varable Hamlonan for equlbrum sysem

More information

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim Korean J. Mah. 19 (2011), No. 3, pp. 263 272 GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS Youngwoo Ahn and Kae Km Absrac. In he paper [1], an explc correspondence beween ceran

More information

Track Properities of Normal Chain

Track Properities of Normal Chain In. J. Conemp. Mah. Scences, Vol. 8, 213, no. 4, 163-171 HIKARI Ld, www.m-har.com rac Propes of Normal Chan L Chen School of Mahemacs and Sascs, Zhengzhou Normal Unversy Zhengzhou Cy, Hennan Provnce, 4544,

More information

Inverse Joint Moments of Multivariate. Random Variables

Inverse Joint Moments of Multivariate. Random Variables In J Conem Mah Scences Vol 7 0 no 46 45-5 Inverse Jon Momens of Mulvarae Rom Varables M A Hussan Dearmen of Mahemacal Sascs Insue of Sascal Sudes Research ISSR Caro Unversy Egy Curren address: Kng Saud

More information

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Journal of Appled Mahemacs and Compuaonal Mechancs 3, (), 45-5 HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Sansław Kukla, Urszula Sedlecka Insue of Mahemacs,

More information

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS R&RATA # Vol.) 8, March FURTHER AALYSIS OF COFIDECE ITERVALS FOR LARGE CLIET/SERVER COMPUTER ETWORKS Vyacheslav Abramov School of Mahemacal Scences, Monash Unversy, Buldng 8, Level 4, Clayon Campus, Wellngon

More information

MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES. Institute for Mathematical Research, Universiti Putra Malaysia, UPM Serdang, Selangor, Malaysia

MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES. Institute for Mathematical Research, Universiti Putra Malaysia, UPM Serdang, Selangor, Malaysia Malaysan Journal of Mahemacal Scences 9(2): 277-300 (2015) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES Journal homeage: h://ensemumedumy/journal A Mehod for Deermnng -Adc Orders of Facorals 1* Rafka Zulkal,

More information

NONLOCAL BOUNDARY VALUE PROBLEM FOR SECOND ORDER ANTI-PERIODIC NONLINEAR IMPULSIVE q k INTEGRODIFFERENCE EQUATION

NONLOCAL BOUNDARY VALUE PROBLEM FOR SECOND ORDER ANTI-PERIODIC NONLINEAR IMPULSIVE q k INTEGRODIFFERENCE EQUATION Euroean Journal of ahemac an Comuer Scence Vol No 7 ISSN 59-995 NONLOCAL BOUNDARY VALUE PROBLE FOR SECOND ORDER ANTI-PERIODIC NONLINEAR IPULSIVE - INTEGRODIFFERENCE EQUATION Hao Wang Yuhang Zhang ngyang

More information

Method of upper lower solutions for nonlinear system of fractional differential equations and applications

Method of upper lower solutions for nonlinear system of fractional differential equations and applications Malaya Journal of Maemak, Vol. 6, No. 3, 467-472, 218 hps://do.org/1.26637/mjm63/1 Mehod of upper lower soluons for nonlnear sysem of fraconal dfferenal equaons and applcaons D.B. Dhagude1 *, N.B. Jadhav2

More information

NPTEL Project. Econometric Modelling. Module23: Granger Causality Test. Lecture35: Granger Causality Test. Vinod Gupta School of Management

NPTEL Project. Econometric Modelling. Module23: Granger Causality Test. Lecture35: Granger Causality Test. Vinod Gupta School of Management P age NPTEL Proec Economerc Modellng Vnod Gua School of Managemen Module23: Granger Causaly Tes Lecure35: Granger Causaly Tes Rudra P. Pradhan Vnod Gua School of Managemen Indan Insue of Technology Kharagur,

More information

SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β

SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β SARAJEVO JOURNAL OF MATHEMATICS Vol.3 (15) (2007), 137 143 SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β M. A. K. BAIG AND RAYEES AHMAD DAR Absrac. In hs paper, we propose

More information

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy Arcle Inernaonal Journal of Modern Mahemacal Scences, 4, (): - Inernaonal Journal of Modern Mahemacal Scences Journal homepage: www.modernscenfcpress.com/journals/jmms.aspx ISSN: 66-86X Florda, USA Approxmae

More information

Pavel Azizurovich Rahman Ufa State Petroleum Technological University, Kosmonavtov St., 1, Ufa, Russian Federation

Pavel Azizurovich Rahman Ufa State Petroleum Technological University, Kosmonavtov St., 1, Ufa, Russian Federation VOL., NO. 5, MARCH 8 ISSN 89-668 ARN Journal of Engneerng and Aled Scences 6-8 Asan Research ublshng Nework ARN. All rghs reserved. www.arnjournals.com A CALCULATION METHOD FOR ESTIMATION OF THE MEAN TIME

More information

Solution in semi infinite diffusion couples (error function analysis)

Solution in semi infinite diffusion couples (error function analysis) Soluon n sem nfne dffuson couples (error funcon analyss) Le us consder now he sem nfne dffuson couple of wo blocks wh concenraon of and I means ha, n a A- bnary sysem, s bondng beween wo blocks made of

More information

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC CH.3. COMPATIBILITY EQUATIONS Connuum Mechancs Course (MMC) - ETSECCPB - UPC Overvew Compably Condons Compably Equaons of a Poenal Vecor Feld Compably Condons for Infnesmal Srans Inegraon of he Infnesmal

More information

CS286.2 Lecture 14: Quantum de Finetti Theorems II

CS286.2 Lecture 14: Quantum de Finetti Theorems II CS286.2 Lecure 14: Quanum de Fne Theorems II Scrbe: Mara Okounkova 1 Saemen of he heorem Recall he las saemen of he quanum de Fne heorem from he prevous lecure. Theorem 1 Quanum de Fne). Le ρ Dens C 2

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4 CS434a/54a: Paern Recognon Prof. Olga Veksler Lecure 4 Oulne Normal Random Varable Properes Dscrmnan funcons Why Normal Random Varables? Analycally racable Works well when observaon comes form a corruped

More information

Department of Economics University of Toronto

Department of Economics University of Toronto Deparmen of Economcs Unversy of Torono ECO408F M.A. Economercs Lecure Noes on Heeroskedascy Heeroskedascy o Ths lecure nvolves lookng a modfcaons we need o make o deal wh he regresson model when some of

More information

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005 Dynamc Team Decson Theory EECS 558 Proec Shruvandana Sharma and Davd Shuman December 0, 005 Oulne Inroducon o Team Decson Theory Decomposon of he Dynamc Team Decson Problem Equvalence of Sac and Dynamc

More information

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5 TPG460 Reservor Smulaon 08 page of 5 DISCRETIZATIO OF THE FOW EQUATIOS As we already have seen, fne dfference appromaons of he paral dervaves appearng n he flow equaons may be obaned from Taylor seres

More information

Variational method to the second-order impulsive partial differential equations with inconstant coefficients (I)

Variational method to the second-order impulsive partial differential equations with inconstant coefficients (I) Avalable onlne a www.scencedrec.com Proceda Engneerng 6 ( 5 4 Inernaonal Worksho on Aomoble, Power and Energy Engneerng Varaonal mehod o he second-order mlsve aral dfferenal eqaons wh nconsan coeffcens

More information

( ) () we define the interaction representation by the unitary transformation () = ()

( ) () we define the interaction representation by the unitary transformation () = () Hgher Order Perurbaon Theory Mchael Fowler 3/7/6 The neracon Represenaon Recall ha n he frs par of hs course sequence, we dscussed he chrödnger and Hesenberg represenaons of quanum mechancs here n he chrödnger

More information

On Local Existence and Blow-Up of Solutions for Nonlinear Wave Equations of Higher-Order Kirchhoff Type with Strong Dissipation

On Local Existence and Blow-Up of Solutions for Nonlinear Wave Equations of Higher-Order Kirchhoff Type with Strong Dissipation Inernaonal Journal of Modern Nonlnear Theory and Alcaon 7 6-5 h://wwwscrorg/journal/jna ISSN Onlne: 67-987 ISSN Prn: 67-979 On Local Exsence and Blow-U of Soluons for Nonlnear Wave Euaons of Hgher-Order

More information

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany Herarchcal Markov Normal Mxure models wh Applcaons o Fnancal Asse Reurns Appendx: Proofs of Theorems and Condonal Poseror Dsrbuons John Geweke a and Gann Amsano b a Deparmens of Economcs and Sascs, Unversy

More information

THE POLYNOMIAL TENSOR INTERPOLATION

THE POLYNOMIAL TENSOR INTERPOLATION Pease ce hs arce as: Grzegorz Berna, Ana Ceo, The oynoma ensor neroaon, Scenfc Research of he Insue of Mahemacs and Comuer Scence, 28, oume 7, Issue, ages 5-. The webse: h://www.amcm.cz./ Scenfc Research

More information

Delay-Range-Dependent Stability Analysis for Continuous Linear System with Interval Delay

Delay-Range-Dependent Stability Analysis for Continuous Linear System with Interval Delay Inernaonal Journal of Emergng Engneerng esearch an echnology Volume 3, Issue 8, Augus 05, PP 70-76 ISSN 349-4395 (Prn) & ISSN 349-4409 (Onlne) Delay-ange-Depenen Sably Analyss for Connuous Lnear Sysem

More information

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s Ordnary Dfferenal Equaons n Neuroscence wh Malab eamples. Am - Gan undersandng of how o se up and solve ODE s Am Undersand how o se up an solve a smple eample of he Hebb rule n D Our goal a end of class

More information

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair TECHNI Inernaonal Journal of Compung Scence Communcaon Technologes VOL.5 NO. July 22 (ISSN 974-3375 erformance nalyss for a Nework havng Sby edundan Un wh ang n epar Jendra Sngh 2 abns orwal 2 Deparmen

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 0 Canoncal Transformaons (Chaper 9) Wha We Dd Las Tme Hamlon s Prncple n he Hamlonan formalsm Dervaon was smple δi δ Addonal end-pon consrans pq H( q, p, ) d 0 δ q ( ) δq ( ) δ

More information

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading Onlne Supplemen for Dynamc Mul-Technology Producon-Invenory Problem wh Emssons Tradng by We Zhang Zhongsheng Hua Yu Xa and Baofeng Huo Proof of Lemma For any ( qr ) Θ s easy o verfy ha he lnear programmng

More information

On the numerical treatment ofthenonlinear partial differentialequation of fractional order

On the numerical treatment ofthenonlinear partial differentialequation of fractional order IOSR Journal of Mahemacs (IOSR-JM) e-iss: 2278-5728, p-iss: 239-765X. Volume 2, Issue 6 Ver. I (ov. - Dec.26), PP 28-37 www.osrjournals.org On he numercal reamen ofhenonlnear paral dfferenalequaon of fraconal

More information

ON THE WEAK LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS

ON THE WEAK LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS ON THE WEA LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS FENGBO HANG Absrac. We denfy all he weak sequenal lms of smooh maps n W (M N). In parcular, hs mples a necessary su cen opologcal

More information

arxiv: v2 [math.pr] 2 Nov 2015

arxiv: v2 [math.pr] 2 Nov 2015 Weak and srong momens of l r -norms of log-concave vecors arxv:1501.01649v2 [mah.pr] 2 Nov 2015 Rafa l Laa la and Mara Srzelecka revsed verson Absrac We show ha for 1 and r 1 he -h momen of he l r -norm

More information

Stability Analysis of Fuzzy Hopfield Neural Networks with Timevarying

Stability Analysis of Fuzzy Hopfield Neural Networks with Timevarying ISSN 746-7659 England UK Journal of Informaon and Compung Scence Vol. No. 8 pp.- Sably Analyss of Fuzzy Hopfeld Neural Neworks w mevaryng Delays Qfeng Xun Cagen Zou Scool of Informaon Engneerng Yanceng

More information

Notes on the stability of dynamic systems and the use of Eigen Values.

Notes on the stability of dynamic systems and the use of Eigen Values. Noes on he sabl of dnamc ssems and he use of Egen Values. Source: Macro II course noes, Dr. Davd Bessler s Tme Seres course noes, zarads (999) Ineremporal Macroeconomcs chaper 4 & Techncal ppend, and Hamlon

More information

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data Anne Chao Ncholas J Goell C seh lzabeh L ander K Ma Rober K Colwell and Aaron M llson 03 Rarefacon and erapolaon wh ll numbers: a framewor for samplng and esmaon n speces dversy sudes cology Monographs

More information

Endogeneity. Is the term given to the situation when one or more of the regressors in the model are correlated with the error term such that

Endogeneity. Is the term given to the situation when one or more of the regressors in the model are correlated with the error term such that s row Endogeney Is he erm gven o he suaon when one or more of he regressors n he model are correlaed wh he error erm such ha E( u 0 The 3 man causes of endogeney are: Measuremen error n he rgh hand sde

More information

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria IOSR Journal of Mahemacs (IOSR-JM e-issn: 78-578, p-issn: 9-765X. Volume 0, Issue 4 Ver. IV (Jul-Aug. 04, PP 40-44 Mulple SolonSoluons for a (+-dmensonalhroa-sasuma shallow waer wave equaon UsngPanlevé-Bӓclund

More information

Dynamic Poverty Measures

Dynamic Poverty Measures heorecal Economcs Leers 63-69 do:436/el34 Publshed Onlne November (h://wwwscrporg/journal/el) Dynamc Povery Measures Absrac Eugene Kouass Perre Mendy Dara Seck Kern O Kymn 3 Resource Economcs Wes Vrgna

More information

Convex Games in Banach Spaces

Convex Games in Banach Spaces Convex Games n Banach Saces Karhk Srdharan TTI-Chcago karhk@c.edu Ambuj Tewar TTI-Chcago ewar@c.edu Absrac We sudy he regre of an onlne learner layng a mul-round game n a Banach sace B agans an adversary

More information

SELFSIMILAR PROCESSES WITH STATIONARY INCREMENTS IN THE SECOND WIENER CHAOS

SELFSIMILAR PROCESSES WITH STATIONARY INCREMENTS IN THE SECOND WIENER CHAOS POBABILITY AD MATEMATICAL STATISTICS Vol., Fasc., pp. SELFSIMILA POCESSES WIT STATIOAY ICEMETS I TE SECOD WIEE CAOS BY M. M A E J I M A YOKOAMA AD C. A. T U D O LILLE Absrac. We sudy selfsmlar processes

More information

A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION

A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION S19 A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION by Xaojun YANG a,b, Yugu YANG a*, Carlo CATTANI c, and Mngzheng ZHU b a Sae Key Laboraory for Geomechancs and Deep Underground Engneerng, Chna Unversy

More information

A NOVEL NETWORK METHOD DESIGNING MULTIRATE FILTER BANKS AND WAVELETS

A NOVEL NETWORK METHOD DESIGNING MULTIRATE FILTER BANKS AND WAVELETS A NOVEL NEWORK MEHOD DESIGNING MULIRAE FILER BANKS AND WAVELES Yng an Deparmen of Elecronc Engneerng and Informaon Scence Unversy of Scence and echnology of Chna Hefe 37, P. R. Chna E-mal: yan@usc.edu.cn

More information

Comb Filters. Comb Filters

Comb Filters. Comb Filters The smple flers dscussed so far are characered eher by a sngle passband and/or a sngle sopband There are applcaons where flers wh mulple passbands and sopbands are requred Thecomb fler s an example of

More information

Should Exact Index Numbers have Standard Errors? Theory and Application to Asian Growth

Should Exact Index Numbers have Standard Errors? Theory and Application to Asian Growth Should Exac Index umbers have Sandard Errors? Theory and Applcaon o Asan Growh Rober C. Feensra Marshall B. Rensdorf ovember 003 Proof of Proposon APPEDIX () Frs, we wll derve he convenonal Sao-Vara prce

More information

Volatility Interpolation

Volatility Interpolation Volaly Inerpolaon Prelmnary Verson March 00 Jesper Andreasen and Bran Huge Danse Mares, Copenhagen wan.daddy@danseban.com brno@danseban.com Elecronc copy avalable a: hp://ssrn.com/absrac=69497 Inro Local

More information

@FMI c Kyung Moon Sa Co.

@FMI c Kyung Moon Sa Co. Annals of Fuzzy Mahemacs and Informacs Volume 8, No. 2, (Augus 2014), pp. 245 257 ISSN: 2093 9310 (prn verson) ISSN: 2287 6235 (elecronc verson) hp://www.afm.or.kr @FMI c Kyung Moon Sa Co. hp://www.kyungmoon.com

More information

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue. Lnear Algebra Lecure # Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons

More information

First-order piecewise-linear dynamic circuits

First-order piecewise-linear dynamic circuits Frs-order pecewse-lnear dynamc crcus. Fndng he soluon We wll sudy rs-order dynamc crcus composed o a nonlnear resse one-por, ermnaed eher by a lnear capacor or a lnear nducor (see Fg.. Nonlnear resse one-por

More information

FTCS Solution to the Heat Equation

FTCS Solution to the Heat Equation FTCS Soluon o he Hea Equaon ME 448/548 Noes Gerald Reckenwald Porland Sae Unversy Deparmen of Mechancal Engneerng gerry@pdxedu ME 448/548: FTCS Soluon o he Hea Equaon Overvew Use he forward fne d erence

More information

STABILITY OF NONLINEAR NEUTRAL DELAY DIFFERENTIAL EQUATIONS WITH VARIABLE DELAYS

STABILITY OF NONLINEAR NEUTRAL DELAY DIFFERENTIAL EQUATIONS WITH VARIABLE DELAYS Elecronic Journal of Differenial Equaions, Vol. 217 217, No. 118, pp. 1 14. ISSN: 172-6691. URL: hp://ejde.mah.xsae.edu or hp://ejde.mah.un.edu STABILITY OF NONLINEAR NEUTRAL DELAY DIFFERENTIAL EQUATIONS

More information

Sampling Procedure of the Sum of two Binary Markov Process Realizations

Sampling Procedure of the Sum of two Binary Markov Process Realizations Samplng Procedure of he Sum of wo Bnary Markov Process Realzaons YURY GORITSKIY Dep. of Mahemacal Modelng of Moscow Power Insue (Techncal Unversy), Moscow, RUSSIA, E-mal: gorsky@yandex.ru VLADIMIR KAZAKOV

More information

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System Communcaons n Sascs Theory and Mehods, 34: 475 484, 2005 Copyrgh Taylor & Francs, Inc. ISSN: 0361-0926 prn/1532-415x onlne DOI: 10.1081/STA-200047430 Survval Analyss and Relably A Noe on he Mean Resdual

More information

Lecture 2 M/G/1 queues. M/G/1-queue

Lecture 2 M/G/1 queues. M/G/1-queue Lecure M/G/ queues M/G/-queue Posson arrval process Arbrary servce me dsrbuon Sngle server To deermne he sae of he sysem a me, we mus now The number of cusomers n he sysems N() Tme ha he cusomer currenly

More information

Density Matrix Description of NMR BCMB/CHEM 8190

Density Matrix Description of NMR BCMB/CHEM 8190 Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon If we say wh one bass se, properes vary only because of changes n he coeffcens weghng each bass se funcon x = h< Ix > - hs s how we calculae

More information

Density Matrix Description of NMR BCMB/CHEM 8190

Density Matrix Description of NMR BCMB/CHEM 8190 Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon Alernae approach o second order specra: ask abou x magnezaon nsead of energes and ranson probables. If we say wh one bass se, properes vary

More information

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION INTERNATIONAL TRADE T. J. KEHOE UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 27 EXAMINATION Please answer wo of he hree quesons. You can consul class noes, workng papers, and arcles whle you are workng on he

More information

, t 1. Transitions - this one was easy, but in general the hardest part is choosing the which variables are state and control variables

, t 1. Transitions - this one was easy, but in general the hardest part is choosing the which variables are state and control variables Opmal Conrol Why Use I - verss calcls of varaons, opmal conrol More generaly More convenen wh consrans (e.g., can p consrans on he dervaves More nsghs no problem (a leas more apparen han hrogh calcls of

More information

Chapter 3: Signed-rank charts

Chapter 3: Signed-rank charts Chaer : gned-ran chars.. The hewhar-ye conrol char... Inroducon As menoned n Chaer, samles of fxed sze are aen a regular nervals and he long sasc s hen loed. The queson s: Whch qualy arameer should be

More information

On computing differential transform of nonlinear non-autonomous functions and its applications

On computing differential transform of nonlinear non-autonomous functions and its applications On compung dfferenal ransform of nonlnear non-auonomous funcons and s applcaons Essam. R. El-Zahar, and Abdelhalm Ebad Deparmen of Mahemacs, Faculy of Scences and Humanes, Prnce Saam Bn Abdulazz Unversy,

More information

Cubic Bezier Homotopy Function for Solving Exponential Equations

Cubic Bezier Homotopy Function for Solving Exponential Equations Penerb Journal of Advanced Research n Compung and Applcaons ISSN (onlne: 46-97 Vol. 4, No.. Pages -8, 6 omoopy Funcon for Solvng Eponenal Equaons S. S. Raml *,,. Mohamad Nor,a, N. S. Saharzan,b and M.

More information

2 Aggregate demand in partial equilibrium static framework

2 Aggregate demand in partial equilibrium static framework Unversy of Mnnesoa 8107 Macroeconomc Theory, Sprng 2009, Mn 1 Fabrzo Perr Lecure 1. Aggregaon 1 Inroducon Probably so far n he macro sequence you have deal drecly wh represenave consumers and represenave

More information

Epistemic Game Theory: Online Appendix

Epistemic Game Theory: Online Appendix Epsemc Game Theory: Onlne Appendx Edde Dekel Lucano Pomao Marcano Snscalch July 18, 2014 Prelmnares Fx a fne ype srucure T I, S, T, β I and a probably µ S T. Le T µ I, S, T µ, βµ I be a ype srucure ha

More information

P R = P 0. The system is shown on the next figure:

P R = P 0. The system is shown on the next figure: TPG460 Reservor Smulaon 08 page of INTRODUCTION TO RESERVOIR SIMULATION Analycal and numercal soluons of smple one-dmensonal, one-phase flow equaons As an nroducon o reservor smulaon, we wll revew he smples

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm H ( q, p, ) = q p L( q, q, ) H p = q H q = p H = L Equvalen o Lagrangan formalsm Smpler, bu

More information

2/20/2013. EE 101 Midterm 2 Review

2/20/2013. EE 101 Midterm 2 Review //3 EE Mderm eew //3 Volage-mplfer Model The npu ressance s he equalen ressance see when lookng no he npu ermnals of he amplfer. o s he oupu ressance. I causes he oupu olage o decrease as he load ressance

More information

F-Tests and Analysis of Variance (ANOVA) in the Simple Linear Regression Model. 1. Introduction

F-Tests and Analysis of Variance (ANOVA) in the Simple Linear Regression Model. 1. Introduction ECOOMICS 35* -- OTE 9 ECO 35* -- OTE 9 F-Tess and Analyss of Varance (AOVA n he Smple Lnear Regresson Model Inroducon The smple lnear regresson model s gven by he followng populaon regresson equaon, or

More information

How about the more general "linear" scalar functions of scalars (i.e., a 1st degree polynomial of the following form with a constant term )?

How about the more general linear scalar functions of scalars (i.e., a 1st degree polynomial of the following form with a constant term )? lmcd Lnear ransformaon of a vecor he deas presened here are que general hey go beyond he radonal mar-vecor ype seen n lnear algebra Furhermore, hey do no deal wh bass and are equally vald for any se of

More information

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes.

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes. umercal negraon of he dffuson equaon (I) Fne dfference mehod. Spaal screaon. Inernal nodes. R L V For hermal conducon le s dscree he spaal doman no small fne spans, =,,: Balance of parcles for an nernal

More information

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue. Mah E-b Lecure #0 Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons are

More information

Discrete time approximation of decoupled Forward-Backward SDE with jumps

Discrete time approximation of decoupled Forward-Backward SDE with jumps Dscree me approxmaon of decoupled Forward-Backward SD wh jumps Bruno Bouchard, Romuald le To ce hs verson: Bruno Bouchard, Romuald le Dscree me approxmaon of decoupled Forward-Backward SD wh jumps Sochasc

More information

3. OVERVIEW OF NUMERICAL METHODS

3. OVERVIEW OF NUMERICAL METHODS 3 OVERVIEW OF NUMERICAL METHODS 3 Inroducory remarks Ths chaper summarzes hose numercal echnques whose knowledge s ndspensable for he undersandng of he dfferen dscree elemen mehods: he Newon-Raphson-mehod,

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm Hqp (,,) = qp Lqq (,,) H p = q H q = p H L = Equvalen o Lagrangan formalsm Smpler, bu wce as

More information

TitleA random walk analogue of Levy's th. Studia scientiarum mathematicarum H Citation

TitleA random walk analogue of Levy's th. Studia scientiarum mathematicarum H Citation TleA random walk analogue of Levy's h Auhor(s) Fuja, Takahko Suda scenarum mahemacarum H Caon 3-33 Issue 008-06 Dae Type Journal Arcle Tex Verson auhor URL hp://hdlhandlene/10086/15876 Ths s an auhor's

More information

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study)

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study) Inernaonal Mahemacal Forum, Vol. 8, 3, no., 7 - HIKARI Ld, www.m-hkar.com hp://dx.do.org/.988/mf.3.3488 New M-Esmaor Objecve Funcon n Smulaneous Equaons Model (A Comparave Sudy) Ahmed H. Youssef Professor

More information

arxiv: v1 [math.pr] 6 Mar 2019

arxiv: v1 [math.pr] 6 Mar 2019 Local law and Tracy Wdom lm for sparse sochasc block models Jong Yun Hwang J Oon Lee Wooseok Yang arxv:1903.02179v1 mah.pr 6 Mar 2019 March 7, 2019 Absrac We consder he specral properes of sparse sochasc

More information

Tight results for Next Fit and Worst Fit with resource augmentation

Tight results for Next Fit and Worst Fit with resource augmentation Tgh resuls for Nex F and Wors F wh resource augmenaon Joan Boyar Leah Epsen Asaf Levn Asrac I s well known ha he wo smple algorhms for he classc n packng prolem, NF and WF oh have an approxmaon rao of

More information

Preamble-Assisted Channel Estimation in OFDM-based Wireless Systems

Preamble-Assisted Channel Estimation in OFDM-based Wireless Systems reamble-asssed Channel Esmaon n OFDM-based reless Sysems Cheong-Hwan Km, Dae-Seung Ban Yong-Hwan Lee School of Elecrcal Engneerng INMC Seoul Naonal Unversy Kwanak. O. Box 34, Seoul, 5-600 Korea e-mal:

More information

Application of ARIMA Model for River Discharges Analysis

Application of ARIMA Model for River Discharges Analysis Alcaon of ARIMA Model for Rver Dscharges Analyss Bhola NS Ghmre Journal of Neal Physcal Socey Volume 4, Issue 1, February 17 ISSN: 39-473X Edors: Dr. Go Chandra Kahle Dr. Devendra Adhkar Mr. Deeendra Parajul

More information

The Finite Element Method for the Analysis of Non-Linear and Dynamic Systems

The Finite Element Method for the Analysis of Non-Linear and Dynamic Systems Swss Federal Insue of Page 1 The Fne Elemen Mehod for he Analyss of Non-Lnear and Dynamc Sysems Prof. Dr. Mchael Havbro Faber Dr. Nebojsa Mojslovc Swss Federal Insue of ETH Zurch, Swzerland Mehod of Fne

More information

Advanced Macroeconomics II: Exchange economy

Advanced Macroeconomics II: Exchange economy Advanced Macroeconomcs II: Exchange economy Krzyszof Makarsk 1 Smple deermnsc dynamc model. 1.1 Inroducon Inroducon Smple deermnsc dynamc model. Defnons of equlbrum: Arrow-Debreu Sequenal Recursve Equvalence

More information

e-journal Reliability: Theory& Applications No 2 (Vol.2) Vyacheslav Abramov

e-journal Reliability: Theory& Applications No 2 (Vol.2) Vyacheslav Abramov June 7 e-ournal Relably: Theory& Applcaons No (Vol. CONFIDENCE INTERVALS ASSOCIATED WITH PERFORMANCE ANALYSIS OF SYMMETRIC LARGE CLOSED CLIENT/SERVER COMPUTER NETWORKS Absrac Vyacheslav Abramov School

More information

グラフィカルモデルによる推論 確率伝搬法 (2) Kenji Fukumizu The Institute of Statistical Mathematics 計算推論科学概論 II (2010 年度, 後期 )

グラフィカルモデルによる推論 確率伝搬法 (2) Kenji Fukumizu The Institute of Statistical Mathematics 計算推論科学概論 II (2010 年度, 後期 ) グラフィカルモデルによる推論 確率伝搬法 Kenj Fukuzu he Insue of Sascal Maheacs 計算推論科学概論 II 年度 後期 Inference on Hdden Markov Model Inference on Hdden Markov Model Revew: HMM odel : hdden sae fne Inference Coue... for any Naïve

More information

by Lauren DeDieu Advisor: George Chen

by Lauren DeDieu Advisor: George Chen b Laren DeDe Advsor: George Chen Are one of he mos powerfl mehods o nmercall solve me dependen paral dfferenal eqaons PDE wh some knd of snglar shock waves & blow-p problems. Fed nmber of mesh pons Moves

More information

Brownian Motion and Stochastic Calculus. Brownian Motion and Stochastic Calculus

Brownian Motion and Stochastic Calculus. Brownian Motion and Stochastic Calculus Browa Moo Sochasc Calculus Xogzh Che Uversy of Hawa a Maoa earme of Mahemacs Seember, 8 Absrac Ths oe s abou oob decomoso he bascs of Suare egrable margales Coes oob-meyer ecomoso Suare Iegrable Margales

More information

Department of Economics University of Warsaw Warsaw, Poland Długa Str. 44/50.

Department of Economics University of Warsaw Warsaw, Poland Długa Str. 44/50. MIGRATIOS OF HETEROGEEOUS POPULATIO OF DRIVERS ACROSS CLASSES OF A BOUS-MALUS SYSTEM BY WOJCIECH OTTO Dearmen of Economcs Unversy of Warsaw 00-24 Warsaw Poland Długa Sr. 44/50 woo@wne.uw.edu.l . ITRODUCTIO

More information

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS THE PREICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS INTROUCTION The wo dmensonal paral dfferenal equaons of second order can be used for he smulaon of compeve envronmen n busness The arcle presens he

More information