Global stability of Cohen-Grossberg neural network with both time-varying and continuous distributed delays

Size: px
Start display at page:

Download "Global stability of Cohen-Grossberg neural network with both time-varying and continuous distributed delays"

Transcription

1 Global stablty of Cohen-Grossberg neural network wth both tme-varyng and contnuous dstrbuted delays José J. Olvera Departamento de Matemátca e Aplcações and CMAT, Escola de Cêncas, Unversdade do Mnho, Campus de Gualtar, Braga, Portugal e-mal: olvera@math.umnho.pt Abstract In ths paper, a generalzed neural network of Cohen-Grossberg type wth both dscrete tme-varyng and dstrbuted unbounded delays s consdered. Based on M-matrx theory, suffcent condtons are establshed to ensure the exstence and global attractvty of an equlbrum pont. The global exponental stablty of the equlbrum s also addressed but for the model wth bounded dscrete tme-varyng delays. A comparson of results shows that these results generalze and mprove some earler publcatons. Keywords: Cohen-Grossberg neural network, unbounded delay, tme-varyng delay, dstrbuted delay, global asymptotc stablty, global exponental stablty Mathematcs Subect Classfcaton: 34K20, 92B20. 1 Introducton The Cohen-Grossberg neural network models, frst proposed and studed by Cohen and Grossberg 4], have been the subect of an actve research due to ther large applcaton n varous engneerng and scentfc areas such as neural-bology, populaton bology, and computng technology. The neural network n 4] can be descrbed by the followng system of ordnary dfferental equatons ẋ t) = a x t)) b x t)) c f x t)) + I, = 1,..., n. 1.1) In order to be more realstc, dfferental equatons descrbng neural networks should ncorporate tme delays to take nto account physcal and bologcal phenomena. It s known that tme delays may lead to oscllaton, dvergence, or nstablty to a system 14]. For over two decades, several generalzatons of model 1.1) wth constant, dscrete tmevaryng, or contnuous dstrbuted delays have been proposed and studed see 2], 3], 18], 19], 20], and references theren). Besdes the cted works, there s an extensve lterature dealng wth stablty of Cohen-Grossberg models wth delays. Recently, neural network models wth both dscrete tme-varyng and contnuous dstrbuted delays have been consdered 17], 21], and the followng Cohen-Grossberg neural network model ẋ t) = a x t)) b x t)) c g x t)) d f x t τ t))) q 1 k s)v x t + s))ds + I ], t 0, 1.2)

2 was studed n 13], 16]. In ths paper we consder a generalzaton of 1.2), see 2.1) below, and shall study the exstence of an equlbrum pont and ts global attractvty and global exponental stablty. We emphasze that, contrary to the usual approach n the lterature, nstead of Lyapunov functonal technques, we apply some qute dfferent technques, presented n 5], 6], 15]) to study the global stablty of an equlbrum pont of the model. Our method allows us to deal wth more general models. Now we set some defntons and notatons. We denote by BC = BC, 0]; R n ) the space of bounded and contnuous functons, φ :, 0] R n, equpped wth the norm φ = sup s 0 φs), where s the maxmum norm n R n,.e. x = max{ x : = 1,..., n} for x = x 1,..., x n ) R n. For a R n, we also use a to denote the constant functon ϕs) = a n BC. A vector c = c 1,..., c n ) R n s sad to be postve f c > 0 for = 1,..., n and n ths case we wrte c > 0. For a real matrx A = a ] n n we denote by A the absolute-value matrx gven by A = a ] n n. For an open set D BC and f : 0, + ) D R n contnuous, consder the functonal dfferental equaton FDE) gven n a general settng by ẋt) = ft, x t ), t 0, 1.3) where, as usual, x t denotes the functon x t :, 0] R n defned by x t s) = xt + s) for s 0. As we are nterested n neural network models, we always consder solutons of 1.3) wth ntal bounded condtons. It s well-known that the Banach space BC s not an admssble phase space for 1.3) n the sense of 10], consequently the standard exstence, unqueness, contnuous dependence type results are not avalable. Instead of BC, we consder the admssble Banach space { } UC g = φ C, 0]; R n φs) φs) ) : sup <, s unformly contnuous on, 0], s 0 gs) gs) equpped wth the norm φ g satsfyng: = sup s 0 φs), where g :, 0] 1, ) s a functon gs) g1) g s a non-ncreasng contnuous functon and g0) = 1; g2) gs + u) lm = 1 unformly on, 0]; u 0 gs) g3) gs) + as s. As BC UC g, then BC s a subspace of UC g, and we denote by BC g the space BC wth the norm g. As UC g s an admssble Banach space, we consder the FDE 1.3), n the phase space UC g, for a convenent functon g and under enough smooth propertes of f, and we have exstence and unqueness of soluton for the ntal value problem see 11]). We denote by xt, 0, ϕ) the soluton of 1.3) wth ntal condton x 0 = ϕ, ϕ UC g. In secton 2, we establsh suffcent condtons for the exstence and global attractvty of an equlbrum pont of a general Cohen-Grossberg neural network model. An equlbrum pont x of a FDE s sad to be globally attractve f any soluton xt) wth bounded ntal condton satsfes xt) x as t +. 2

3 In secton 3, we set suffcent condtons for the equlbrum pont to be globally exponentally stable. An equlbrum pont x of a neural network model s sad to be globally exponentally stable f there are postve constants ε, M such that xt, 0, ϕ) x Me εt ϕ x t 0, ϕ BC. As mentoned n 7], the defnton of global exponental stablty of an equlbrum x s the usual n the lterature on neural networks wth unbounded delays, but s does not even mply the stablty of x n the phase space UC g,.e. relatve to the norm g. To prove the boundedness solutons of the dfferental systems studed n ths paper, we use the followng result establshed n 7]. Lemma ] Consder equaton 1.3) n UC g, and suppose that f transforms closed bounded sets of 0, + ) D nto bounded sets of R n. If the hypothess H) for all t 0 and ϕ BC such that ϕs) < ϕ0) for s, 0), then ϕ 0)f t, ϕ) < 0 for some {1,..., n} such that ϕ0) = ϕ 0) holds, then any soluton xt) of 1.3) wth ntal condton ϕ BC s defned and bounded on 0, + ), and verfes xt) ϕ for all t 0. To prove the exstence and unqueness of an equlbrum pont, we make use of arguments n 15] and the followng lemma: Lemma ] If H : R n R n s a contnuous and nectve functon such that lm Hx) = +, x + then Hx) s a homeomorphsm of R n onto tself. 2 Global attractvty Consder the generalzed Cohen-Grossberg model wth both dscrete tme-varyng and contnuous dstrbuted nfnte delays gven by ẋ t) = a x t)) b x t)) + h p) x t τ p) t)))+ +f p) ))] g p) x t + s))dη p) s), t 0, 2.1) = 1,..., n, where a : R 0, ), b, h p), f p), gp) : R R, τ p) : 0, ) 0, ) are contnuous functons, and η p) :, 0] R are non-decreasng bounded functons, normalzed so that η p) 0) ηp) ) = 1, for all, {1,..., n}, p {1,..., P }. In 7], a functon g :, 0] 1, + ) was defned by ) gs) = 1 on r 1, 0]; ) g r n ) = n, n N; ) g s contnuous and pecewse lnear lnear on ntervals r n+1, r n ]), 3

4 where r n + s a sutable sequence of postve numbers, n such a way that condtons g1), g2), g3) hold and gs)dη p) < +,, = 1,..., n, p = 1,..., P. See more detals n Lemma 4.1 n 7]. Thus, we may consder the dfferental system 2.1) n the phase space UC g. As mentoned above, we only consder solutons of 2.1) wth bounded ntal data,.e., x 0 = ϕ, ϕ BC. 2.2) In the sequel, for 2.1) the followng hypotheses wll be consdered: A1) for each {1,..., n}, there s β > 0 such that b u) b v))/u v) β, u, v R, u v; A2) h p), f p), gp), respectvely, for, = 1,..., n, p = 1,..., P ; σ p) : R R are Lpschtz functons wth Lpschtz constants γ p), µp) A3) t τ p) t) as t, for, = 1,..., n, p = 1,..., P. Now, we defne the square real matrces,, and B = dagβ 1,..., β n ), L = l ] and N = B L, 2.3) where β 1,..., β n are as n A1) and l = P γp) + µ p) σp). We recall here the defnton of a non-sngular M-matrx. For further propertes of M- matrces, we refer the reader to Chapter 5 of 8]. Defnton. If D = d ] s a square matrx wth non-postve off-dagonal entres,.e., d 0 for all, we say that D s a non-sngular M-matrx f all the egenvalues of D have postve real part, or, equvalently, f all the prncpal mnors of D are postve. Theorem 2.1. Assume A1)-A3). For N defned n 2.3), f N s a non-sngular M- matrx, then there s a unque equlbrum pont x = x 1,..., x n) of 2.1) whch s globally attractve. Proof. Snce N s a non-sngular M-matrx, then see 8]) there s d = d 1,..., d n ) > 0 such that Nd > 0,.e., β > d 1 l d, = 1,..., n. 2.4) The change z t) = d 1 x t) transforms 2.1) nto ż t) = ā z t)) b z t)) + h t, z t )], t 0, = 1,..., n, 2.5) 4

5 where, for each t 0, = 1,..., n, φ UC g, and u R, we have h t, φ) = d 1 h p) ā u) = a d u), b u) = d 1 b d u). ) d φ τ p) t)) + f p) g p) d φ s))dη p) )) s), Note that b u) b v))/u v) β for u, v R, u v,.e., condton A1) s satsfed by the functons b, {1,..., n}. Defne the contnuous functon H : R n R n, Hx) = b x ) + d 1 ) ) n h p) d x ) + f p) g p) d x ) =1, x = x 1,, x n ) R n. Then x R n s an equlbrum of the neural network 2.5) f and only f Hx ) = 0. Consequently, to prove the exstence and unqueness of an equlbrum pont, t s suffcent to prove that H s a homeomorphsm. Argung as n the proof of Lemma 2.4 n 15], we conclude that H s nectve and lm x Hx) =, thus, from Lemma 1.2, we obtan that H s a homeomorphsm, hence there s a unque equlbrum pont x R n of 2.5). Translatng the equlbrum to the orgn through the change y t) = z t) x, 2.5) becomes where f = f 1,..., f n ) : 0, ) UC g R n s defned by ẏt) = ft, y t ), t 0, 2.6) f t, φ) = ā φ 0) + x ) b φ 0) + x ) + h t, φ + x )]. 2.7) It s easy to see that f transforms closed bounded sets of 0, + ) UC g nto bounded sets of R n. Let t 0 and ϕ BC satsfyng ϕs) < ϕ0), for s, 0), and let {1,..., n} be such that ϕ0) = ϕ 0). Suppose that ϕ 0) > 0 the stuaton s analogous for ϕ 0) < 0). Then ϕ 0) = sup ϕs) and s 0 b ϕ 0) + x ) + h t, ϕ + x ) = b ϕ 0) + x ) b x ) +d 1 +f p) β ϕ 0) d 1 β ϕ 0) d 1 ) h p) d ϕ τ p) t)) + d x ) h p) d x ) ) g p) d ϕ s) + d x )dη p) s) γ p) d ϕ τ p) t)) p) + µ l d sup ϕs) = β d 1 s 0 ] f p) gp) d x )) σp) d ϕ s) dη p) s) ) ) l d sup ϕs) > 0, s 0 5

6 hence f satsfes hypothess H). Consequently, for yt) = y t)) n =1 a soluton of 2.6) wth ntal condton y 0 BC, from Lemma 1.1 we conclude that yt) s defned and bounded on R and verfes Set and yt) y 0, t ) v = lm nf y t), u = lm sup y t), = 1,..., n, t t v = max{v }, u = max{u }. Note that u, v R and v u. It s suffcent to prove that max{u, v} = 0. Assume e.g. that v u, so that max{u, v} = u. The stuaton u v s analogous). Let {1,..., n} such that u = u. Argung as n the proof of Theorem 3.2 n 7], we can conclude that there s a postve real sequence t k ) k N such that t k, y t k ) u and f t k, y tk ) 0, as k. 2.9) For convenence, we prove t here. Case 1. Assume that y t) s eventually monotone. In ths case, lm t y t) = u and, for T large, ether ẏ t) 0 for t T or ẏ t) 0 for t T. Assume e.g. that ẏ t) 0 for all t > T large the stuaton ẏ t) 0 s analogous). Then ẏt) = f t, y t ) 0 for t T large, hence lm sup f t, y t ) := c 0. t If c < 0, then there s t 0 > 0 such that f t, y t ) < c/2 for t t 0, mplyng that t y t) = y t 0 ) + f s, y s )ds y t 0 ) + c t 0 2 t t 0), then y t) as t, whch s not possble because of 2.8). Thus c = 0, whch proves 2.9). Case 2. Assume that y t) s not eventually monotone. In ths case there s a sequence t k ) k N such that t k, ẏ t k ) = 0 and y t k ) u, as k. Then f t k, y tk ) = 0 for all k N, and 2.9) holds. For the sake of contradcton, assume that u > 0. Fx ε > 0 and let T = T ε) > 0 be such that yt) < u ɛ := u + ε, for t T, and T dη p) s) < ε y 0,, {1,..., n}, p {1,..., P }. Snce t τ p) t) as t and y t k ) u > 0, then there s k 0 N such that, for k k 0, t k τ p) t k) > T, t k > 2T, and y t k ) > 0. Hence, from the hypotheses and 2.8) we 6

7 conclude that, for k > k 0, b y t k ) + x ) + h t k, y tk + x ) = = b y t k ) + x ) b x ) + d 1 +f p) β y t k ) d 1 β y t k ) d 1 β y t k ) d 1 β y t k ) d 1 ) g p) d y t k + s) + d x )dη p) s) ) h p) d y t k τ p) t k)) + d x ) h p) d x ) γ p) d y t k τ p) t k)) + µ p) Lettng k and ε 0, we get lm nf k ] f p) gp) d x )) σp) d y t k + s) dη p) s) ) T )] γ p) d u ε + µ p) σp) d y t k + s) dη p) s) + y t k + s) dη p) s) T )] γ p) d u ε + µ p) σp) d ε + u ε dη p) s) T ) γ p) d u ε + µ p) σp) d u 2ε β y t k ) d 1 b y t k ) + x ) + h t k, y tk + x )] β d 1 d l u 2ε. l d )u > 0. On the other hand, snce y t) s bounded and ā s postve and contnuous, there s K > 0 such that ā y t) + x ) > K for all t 0. Together wth the above nequalty, ths mples that f t k, y tk ) 0 as k, whch s a contradcton. Thus u = 0, hence v = 0 as well, and ths ends the proof of the theorem. Example 2.1. Consder the Cohen-Grossberg neural network model wth nfnte dscrete tme-varyng delays see 12]) ẋ t) = a x t)) b x t)) c g x t)) d f x t τ t))), = 1,..., n, 2.10) where c, d R and a : R 0, ), b : R R and τ : 0, ) 0, ) are contnuous functons, g, f : R R are Lpschtz functons, and t τ t) as t, for, = 1,..., n. Note that system 2.10) ncludes known models of cellular neural networks as partcular cases. System 2.10) has the form 2.1) f P = 2, h 1) u) = c g u), h 2) u) = d f u), f 1) u) = f 2) u) = 0, τ 1) t) = 0 and τ 2) t) = τ t), u R, t 0,, = 1,..., n. If 7

8 g, f : R R have Lpschtz constants G, F respectvely, then h 1), h2) are also Lpschtz functons, wth Lpschtz constants l 1) = c G, l 2) = d F respectvely, for all, {1,..., n}. Theorem 2.1 appled to system 2.10) gves the followng result: Corollary 2.2. Assume that A1) holds, τ : 0, ) 0, ) satsfy t τ t) as t, and that g, f : R R are Lpschtz functons wth Lpschtz constants G, F, for all, = 1,..., n. If N := B L, where B = dagβ 1,..., β n ), L = l 1) + l 2) ], 2.11) wth β as n A1) and l 1) = c G, l 2) = d F, s a non-sngular M-matrx, then there s a unque equlbrum pont of 2.10), whch s globally attractve. Remark 2.1. For system 2.10), the global attractvty of an equlbrum pont was already obtaned by T. Huang et al. 12], provded that: the functons f, g are Lpschtzan, the matrx N as above s a non-sngular M-matrx, b u) are dfferentable functons such that b u) β > 0 clearly a stronger hypothess than A1)) and the followng addtonal hypothess s satsfed: for each {1,..., n}, there exst a, a > 0 such that 0 < a a u) a, u R. Thus, the above Corollary 2.2 sgnfcantly mproves the crteron n 12]. We also remark that the global exponental stablty of the equlbrum pont of 2.10), wth bounded delays τ, was studed n 6]. In the followng example, a bdrectonal assocatve memory BAM) neural network model s treated as a partcular case of the generalzed Cohen-Grossberg model 2.1). Example 2.2. Consder the BAM neural networks model presented n 15] see also 1]) ẋ t) = a x t)) b x t)) + m ẏ t) = d y t)) c y t)) + =1 f p) g p) x t ω p) t))) y t τ p) f p) y t ρ p) t))) t))), = 1,..., n, g p) x t ζ p) t))), = 1,..., m, 2.12) for t 0 and n, m, P N, where a, d : R 0, ), b, c, g p), f p), g p), f p) : R R are contnuous functons and ω p), ρ p), τ p), ζp) : 0, ) 0, ) are contnuous functons such that t ω p) t), t ρ p) t), t τ p) t), and t ζp) t) as t, = 1,..., n, = 1,..., m and p = 1,..., P. Contrarly to what happens n 15], we do not assume that the tme-dependent delays ω p) t), ρ p) t), τ p) t), ζp) t) are bounded. Clearly, 2.12) s a partcular case of 2.1). Next result s a drect consequence of Theorem 2.1, and extends the stablty crteron presented n 15] to the stuaton wth unbounded dscrete tme-varyng delays. 8

9 Corollary 2.3. Suppose that b and c satsfy A1) wth constants β and γ respectvely, a u) > 0 and d u) > 0 for all u R, f p), g p), f p), gp) are Lpschtz functons wth Lpschtz constants θ p), ξ p), θ p), ξp) respectvely, and ω p), ρ p), τ p), ζp) are contnuous functons such that { } t mn ω p),,p t), ρ p) t), τ p) t), ζp) t) as t. where Defne B G d N := G F C F d n+m) n+m) B = dagβ 1,..., β n ), C = dagγ 1,..., γ m ) G d = dag ξ p) 1,..., ξ n p), F d = dag θ p) 1,..., G = ξ p) m n, F = θ p), n m. θ m p) If N s a non-sngular M-matrx, then there s a unque equlbrum pont of 2.12), whch s globally attractve. 3 Exponental stablty In ths secton, we address the global exponental stablty of the equlbrum pont of 2.1). Here system 2.1) s consdered n the phase space UC g, wth gs) = e αs, s, 0], φs),c for a convenent α > 0 defned below, equpped wth the norm φ g,c = sup, s 0 gs) where,c s the norm n R n defned by y,c = y 1,..., y n ),c = max{c y } wth c = c 1,..., c n ) > 0. The general famly of functonal dfferental systems n the phase space UC g, ẋ t) = ρ t, x t )b x t)) + f t, x t )], = 1,..., n, t 0, 3.1) where ρ : 0, + ) UC g 0, + ), b : R R, and f : 0, + ) UC g R are contnuous functons, was consdered n 7]. For system 3.1), the followng set of hypotheses was mposed: E1) for each {1,..., n}, there s β > 0 such that b u) b v))/u v) β, u, v R, u v; E2) for each {1,..., n}, there s l > 0 such that f t, ϕ) f t, ψ) l ϕ ψ g,c ϕ, ψ UC g ;, 9

10 E3) ρ := nf{ρ t, ϕ) : t 0, ϕ BC g, = 1,..., n} > 0; E4) for each {1,..., n}, β > c l. In 7], the followng result was establshed. Theorem ] Assume E1)-E4). If x = x 1,..., x n) R n s an equlbrum pont of 3.1), then there are postve constants ε, M such that xt, 0, ϕ) x,c Me εt ϕ x,c, t 0, ϕ BC g. In order to apply the exponental stablty crteron presented n the prevous theorem to model 2.1), we now assume that the delayed functons τ p) p) s) are bounded,.e., 0 τ t) τ for some τ > 0, and there exsts ξ > 0 such that all the normalzed non-decreasng and bounded functons η p) satsfy e ξs dη p) s) <. 3.2) Theorem 3.2. Consder 2.1), where a : R 0, + ), b : R R, and τ p) 0, + ) are contnuous, h p), f p), and gp) γ p), µp) η p), and σp) respectvely, and η p) ) = 1,, = 1,..., n, p = 1,..., P. 0) ηp) Assume n addton that: ) E1) s satsfed; ) a := nf{a x) : x R} > 0 for = 1,..., n; : 0, + ) are Lpschtz functons wth Lpschz constants are non-decreasng, bounded and normalzed,.e. ) there exsts a constant ξ > 0 such that η p) satsfy 3.2) for all, = 1,..., n, p = 1,..., P ; v) there exsts τ > 0 such that 0 τ p) t) τ for t 0,, = 1,..., n, p = 1,..., P. If the matrx N defned n 2.3) s a non-sngular M-matrx, then there s a unque equlbrum pont of 2.1), whch s globally exponentally stable. Proof. Snce N s a non-sngular M-matrx, there s d = d 1,..., d n ) > 0 such that 2.4) holds see 8]), hence there s δ > 0 such that β > d 1 l 1 + δ)d, = 1,..., n. 3.3) As n the proof of Theorem 4.3 n 7], from 3.2) we can conclude that there s ς 0, ξ) such that e ςs dη p) s) < 1 + δ,, = 1,..., n, p = 1,..., P. 3.4) { } log1 + δ) Let α := mn ς, and consder system 2.1) n the phase space UC g, where τ gs) = e αs, s 0, equpped wth the norm φ g,c wth c = d 1 1,..., d 1 n ). From Theorem 10

11 2.1, we know that there exsts a unque equlbrum pont, x = x 1,..., x n) R n. System 2.1) has the form 3.1) wth f t, φ) = h p) ) )) φ τ p) t)) + f p) g p) φ s))dη p) s). For ϕ, φ UC g and t 0, snce h p), f p) non-decreasng, we have, gp) are Lpschtz functons and η p) are f t, φ) f t, ϕ) = = +f p) h p) φ τ p) t))) hp) ϕ τ p) t))) ) )) g p) φ s))dη p) s) f p) g p) ϕ s))dη p) s) γ p) γ p) d γ p) Ths means that φ τ p) d 1 t)) ϕ τ p) φ ϕ ) τ p) t)) e ατ p) ατ e t) t)) + µ p) p) t) + µ p) σp) ] ) g p) φ s)) g p) ϕ s)) dη p) s) p) 0 eατ t) d + µ p) σp) d e αs dη p) ) φ s) ϕ g,c e αs d 1 d φ ϕ )s) e αs ) γ p) 1 + δ)d + µ p) σp) d 1 + δ) φ ϕ g,c l 1 + δ)d ) φ ϕ g,c. f t, φ) f t, ϕ) l φ ϕ g,c, = 1,..., n, wth l := n l 1 + δ)d, and from 3.3) we have β > c l, {1,..., n}. Now, the concluson follows from Theorem 3.1. dη p) s) ) Remark 3.1 Snce the non-autonomous terms h p) were not ncluded n the famly of neural network models studed n 7], the prevous result mproves Theorem 4.3 n 7]. Example 3.1. If we take P = 2, h 1) u) = c g u), f 1) 1) 2) u) = 0, τ t) = 0, τ t) = τ t), h 2) u) = d f u), f 2) u) = q u + I, g 2) u) = v u), wth c, d, q, I R, and η 2) s) = s k v)dv, s, 0],, = 1,..., n, 11

12 where k : 0, + ) 0, + ) are contnuous functons, the model 2.1) becomes the followng Cohen-Grossberg neural networks model ẋ t) = a x t)) b x t)) c g x t)) d f x t τ t))) q for = 1,..., n, wth the connecton matrces k s)v x t + s))ds + I ], t 0, 3.5) C = c ] n n, D = d ] n n, and Q = q ] n n. Applyng Theorem 3.2 to model 3.5), we have the followng result. Corollary 3.3. Consder 3.5), where a : R 0, + ), b : R R, and τ : 0, + ) 0, + ) are contnuous, g, f, and v are Lpschtz functons wth Lpschz constants G, F, and V respectvely, and k are nonnegatve contnuous functons such that + 0 k t)dt = 1,, = 1,..., n. Assume n addton that: ) E1) s satsfed; ) a := nf{a x) : x R} > 0 for = 1,..., n; ) there exsts a constant ξ > 0 such that + 0 k t)e ξt dt < +,, = 1,..., n; v) there exsts τ > 0 such that 0 τ p) t) τ for t 0,, = 1,..., n, p = 1,..., P. If the matrx N = B C G D F Q V where B = dagβ 1,..., β n ), G = dagg 1,..., G n ), F = dagf 1,..., F n ), and V = dagv 1,..., V n ), s a non-sngular M-matrx, then there s a unque equlbrum pont of 3.5), whch s globally exponentally stable. Remark 3.2 Snce model 3.5) s only a partcular case of model 2.1), our Theorem 3.2 s more general than the man result n 16]. Example 3.2. Consder the followng model: ẋt) = 2 + cos xt)) 6xt) + g 1 xt)) + g 1 yt)) ] +g 1 xt τt))) + k s)yt + s)ds ẏt) = 2 + sn yt)) 4.5yt) + g 2 yt)) + g 2 xt τt))) ] +g 2 yt τt))) + k s)xt + s)ds, 3.6) 12

13 where g 1 u) = 1 2 u + 1 u 1 ), g 2u) = 1 2 u u 1 ), τt) = 3 cos t + 1, and kt) = e t. The model 3.6) satsfes all hypotheses of Theorem 3.2 wth a 1 = a 2 = 1, β 1 = 6, β 2 = 4.5, τ = 4, and g 1, g 2 are Lpschtz functons wth Lpschtz constant 1. Thus, ) ) B =, L =, N = and t easy to see that N s a non-sngular M-matrx. From Theorem 3.2, we know that model 3.6) has a unque equlbrum pont whch s globally exponental stable. Snce there s not a R such that g 1 u) = ag 2 u), for all u R, the man result n 16] cannot be appled to prove the stablty of model 3.6). We used the Mathematca software to plot a numercal smulaton of the behavor of the soluton xt), yt)) of model 3.6) wth ntal condton ϕs) = 3 cos s, 2 sn s), s ) 0, for 4 1 τt) 4 and kt) 0. Note that, n ths smple case, the matrx N = s also a non-sngular M-matrx. Once agan, we cannot apply the result n 16]. x t 0.3 ) t 0.1 Fgure 1: Behavor of the frst neuron xt) n system 3.6) wth τt) 4, kt) 0, and ntal condton ϕs) = 3 cos s, 2 sn s) 13

14 y t t Fgure 2: Behavor of the second neuron yt) n system 3.6) wth τt) 4, kt) 0, and ntal condton ϕs) = 3 cos s, 2 sn s) Acknowledgments. Ths research was supported by FCT Portugal), through the research center CMAT. The author thanks Professor Teresa Fara, for helpful dscussons. The author also thanks the referee for valuable comments. References 1] J. Cao and Q. Song, Stablty n Cohen-Grossberg-type bdrectonal assocatve memory neural networks wth tme-varyng delays, Nonlnearty ) ] Y. Chen, Global asymptotc stablty of delayed Cohen-Grossberg neural networks, IEEE Trans. Crcuts Syst ) ] T. Chen and L. Rong, Robust global exponental stablty of Cohen-Grossberg neural networks wth tme delays, IEEE Trans. Neural Netw ) ] M. Cohen and S. Grossberg, Absolute stablty of global pattern formaton and parallel memory storage by compettve neural networks, IEEE Trans. Systems Man Cybernet ) ] T. Fara and J.J. Olvera, Local and global stablty for Lotka-Volterra systems wth dstrbuted delays and nstantaneous negatve feedbacks. J. Dfferental Equatons ) ] T. Fara and J.J. Olvera, Boundedness and global exponental stablty for delayed dfferental equatons wth applcatons, Appl. Math. Comput ) ] T. Fara and J.J. Olvera, General crtera for asymptotc and exponental stablty of neural network models wth unbounded delays, preprnt 8] M. Fedler, Specal Matrz and Ther Applcatons n Numercal Mathematcs, Martnus Nhoff Publ. Kluwer), Dordrecht, ] M. Fort and A. Tes, New condton for global stablty of neural networks wth applcaton to lnear, quadratc programmng problems, IEEE Trans. Crcuts Syst )

15 10] J. Haddock and W. Hornor, Precompactness and convergence n norm of postve orbts n a certan fadng memory space, Funkcal. Ekvac ) ] J.K. Hale, J. Kato, Phase Space for Retarded Equatons wth Infnte Delay, Funkcala Ekvaco, ) ] T. Huang, A. Chan, Y. Huang and J. Cao, Stablty of Cohen-Grossberg neural networks wth tme-varyng delays, Neural Networks, ) ] Y. J, X. Lou and B. Cu, Global output convergence of Cohen-Grossberg neural networks wth both tme-varyng and dstrbuted delays, Chaos Soltons & Fractals, ) ] C.M. Marcus and R.M. Westervelt, Stablty of analogy neural netwoeks wth delay, Physcs Revews A ) ] J.J. Olvera, Global asymptotc stablty for neural network models wth dstrbuted delays, Math. Comput. Modellng ) ] Q. Song and J. Cao, Stablty analyss of Cohen-Grossberg neural network wth both tme-varyng and contnuously dstrbuted delays, J. Comput. Appl. Math ) ] Q.K. Song and Z.J. Zhao, Global dsspatvty of neural networks wth both varable and unbounded delays, Chaos Soltons & Fractals, ) ] L. Wang, Stablty of Cohen-Grossberg neural networks wth dstrbuted delays, Appl. Math. Comput ) ] L. Wang and X. Zou, Harmless delays n Cohen-Grossberg neural networks, Physca D ) ] L. Wu, B.T. Cu and X.Y. Lou, Global exponental stablty of Cohen-Grossberg neural networks wth dstrbuted delays, Math. Comput. Modellng ) ] J.Y. Zhang, Y. Suda and T. Iwasa, Absolutely exponental stablty of a class of neural networks wth unbounded delay, Neural Networks )

Recurrence. 1 Definitions and main statements

Recurrence. 1 Definitions and main statements Recurrence 1 Defntons and man statements Let X n, n = 0, 1, 2,... be a MC wth the state space S = (1, 2,...), transton probabltes p j = P {X n+1 = j X n = }, and the transton matrx P = (p j ),j S def.

More information

8.5 UNITARY AND HERMITIAN MATRICES. The conjugate transpose of a complex matrix A, denoted by A*, is given by

8.5 UNITARY AND HERMITIAN MATRICES. The conjugate transpose of a complex matrix A, denoted by A*, is given by 6 CHAPTER 8 COMPLEX VECTOR SPACES 5. Fnd the kernel of the lnear transformaton gven n Exercse 5. In Exercses 55 and 56, fnd the mage of v, for the ndcated composton, where and are gven by the followng

More information

1 Example 1: Axis-aligned rectangles

1 Example 1: Axis-aligned rectangles COS 511: Theoretcal Machne Learnng Lecturer: Rob Schapre Lecture # 6 Scrbe: Aaron Schld February 21, 2013 Last class, we dscussed an analogue for Occam s Razor for nfnte hypothess spaces that, n conjuncton

More information

BERNSTEIN POLYNOMIALS

BERNSTEIN POLYNOMIALS On-Lne Geometrc Modelng Notes BERNSTEIN POLYNOMIALS Kenneth I. Joy Vsualzaton and Graphcs Research Group Department of Computer Scence Unversty of Calforna, Davs Overvew Polynomals are ncredbly useful

More information

On Lockett pairs and Lockett conjecture for π-soluble Fitting classes

On Lockett pairs and Lockett conjecture for π-soluble Fitting classes On Lockett pars and Lockett conjecture for π-soluble Fttng classes Lujn Zhu Department of Mathematcs, Yangzhou Unversty, Yangzhou 225002, P.R. Chna E-mal: ljzhu@yzu.edu.cn Nanyng Yang School of Mathematcs

More information

Product-Form Stationary Distributions for Deficiency Zero Chemical Reaction Networks

Product-Form Stationary Distributions for Deficiency Zero Chemical Reaction Networks Bulletn of Mathematcal Bology (21 DOI 1.17/s11538-1-9517-4 ORIGINAL ARTICLE Product-Form Statonary Dstrbutons for Defcency Zero Chemcal Reacton Networks Davd F. Anderson, Gheorghe Cracun, Thomas G. Kurtz

More information

Luby s Alg. for Maximal Independent Sets using Pairwise Independence

Luby s Alg. for Maximal Independent Sets using Pairwise Independence Lecture Notes for Randomzed Algorthms Luby s Alg. for Maxmal Independent Sets usng Parwse Independence Last Updated by Erc Vgoda on February, 006 8. Maxmal Independent Sets For a graph G = (V, E), an ndependent

More information

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ).

benefit is 2, paid if the policyholder dies within the year, and probability of death within the year is ). REVIEW OF RISK MANAGEMENT CONCEPTS LOSS DISTRIBUTIONS AND INSURANCE Loss and nsurance: When someone s subject to the rsk of ncurrng a fnancal loss, the loss s generally modeled usng a random varable or

More information

Support Vector Machines

Support Vector Machines Support Vector Machnes Max Wellng Department of Computer Scence Unversty of Toronto 10 Kng s College Road Toronto, M5S 3G5 Canada wellng@cs.toronto.edu Abstract Ths s a note to explan support vector machnes.

More information

A Novel Methodology of Working Capital Management for Large. Public Constructions by Using Fuzzy S-curve Regression

A Novel Methodology of Working Capital Management for Large. Public Constructions by Using Fuzzy S-curve Regression Novel Methodology of Workng Captal Management for Large Publc Constructons by Usng Fuzzy S-curve Regresson Cheng-Wu Chen, Morrs H. L. Wang and Tng-Ya Hseh Department of Cvl Engneerng, Natonal Central Unversty,

More information

PERRON FROBENIUS THEOREM

PERRON FROBENIUS THEOREM PERRON FROBENIUS THEOREM R. CLARK ROBINSON Defnton. A n n matrx M wth real entres m, s called a stochastc matrx provded () all the entres m satsfy 0 m, () each of the columns sum to one, m = for all, ()

More information

Stability, observer design and control of networks using Lyapunov methods

Stability, observer design and control of networks using Lyapunov methods Stablty, observer desgn and control of networks usng Lyapunov methods von Lars Naujok Dssertaton zur Erlangung des Grades enes Doktors der Naturwssenschaften - Dr. rer. nat. - Vorgelegt m Fachberech 3

More information

v a 1 b 1 i, a 2 b 2 i,..., a n b n i.

v a 1 b 1 i, a 2 b 2 i,..., a n b n i. SECTION 8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS 455 8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS All the vector spaces we have studed thus far n the text are real vector spaces snce the scalars are

More information

REGULAR MULTILINEAR OPERATORS ON C(K) SPACES

REGULAR MULTILINEAR OPERATORS ON C(K) SPACES REGULAR MULTILINEAR OPERATORS ON C(K) SPACES FERNANDO BOMBAL AND IGNACIO VILLANUEVA Abstract. The purpose of ths paper s to characterze the class of regular contnuous multlnear operators on a product of

More information

+ + + - - This circuit than can be reduced to a planar circuit

+ + + - - This circuit than can be reduced to a planar circuit MeshCurrent Method The meshcurrent s analog of the nodeoltage method. We sole for a new set of arables, mesh currents, that automatcally satsfy KCLs. As such, meshcurrent method reduces crcut soluton to

More information

NON-CONSTANT SUM RED-AND-BLACK GAMES WITH BET-DEPENDENT WIN PROBABILITY FUNCTION LAURA PONTIGGIA, University of the Sciences in Philadelphia

NON-CONSTANT SUM RED-AND-BLACK GAMES WITH BET-DEPENDENT WIN PROBABILITY FUNCTION LAURA PONTIGGIA, University of the Sciences in Philadelphia To appear n Journal o Appled Probablty June 2007 O-COSTAT SUM RED-AD-BLACK GAMES WITH BET-DEPEDET WI PROBABILITY FUCTIO LAURA POTIGGIA, Unversty o the Scences n Phladelpha Abstract In ths paper we nvestgate

More information

n + d + q = 24 and.05n +.1d +.25q = 2 { n + d + q = 24 (3) n + 2d + 5q = 40 (2)

n + d + q = 24 and.05n +.1d +.25q = 2 { n + d + q = 24 (3) n + 2d + 5q = 40 (2) MATH 16T Exam 1 : Part I (In-Class) Solutons 1. (0 pts) A pggy bank contans 4 cons, all of whch are nckels (5 ), dmes (10 ) or quarters (5 ). The pggy bank also contans a con of each denomnaton. The total

More information

We are now ready to answer the question: What are the possible cardinalities for finite fields?

We are now ready to answer the question: What are the possible cardinalities for finite fields? Chapter 3 Fnte felds We have seen, n the prevous chapters, some examples of fnte felds. For example, the resdue class rng Z/pZ (when p s a prme) forms a feld wth p elements whch may be dentfed wth the

More information

Embedding lattices in the Kleene degrees

Embedding lattices in the Kleene degrees F U N D A M E N T A MATHEMATICAE 62 (999) Embeddng lattces n the Kleene degrees by Hsato M u r a k (Nagoya) Abstract. Under ZFC+CH, we prove that some lattces whose cardnaltes do not exceed ℵ can be embedded

More information

Production. 2. Y is closed A set is closed if it contains its boundary. We need this for the solution existence in the profit maximization problem.

Production. 2. Y is closed A set is closed if it contains its boundary. We need this for the solution existence in the profit maximization problem. Producer Theory Producton ASSUMPTION 2.1 Propertes of the Producton Set The producton set Y satsfes the followng propertes 1. Y s non-empty If Y s empty, we have nothng to talk about 2. Y s closed A set

More information

An Alternative Way to Measure Private Equity Performance

An Alternative Way to Measure Private Equity Performance An Alternatve Way to Measure Prvate Equty Performance Peter Todd Parlux Investment Technology LLC Summary Internal Rate of Return (IRR) s probably the most common way to measure the performance of prvate

More information

Ring structure of splines on triangulations

Ring structure of splines on triangulations www.oeaw.ac.at Rng structure of splnes on trangulatons N. Vllamzar RICAM-Report 2014-48 www.rcam.oeaw.ac.at RING STRUCTURE OF SPLINES ON TRIANGULATIONS NELLY VILLAMIZAR Introducton For a trangulated regon

More information

THE DISTRIBUTION OF LOAN PORTFOLIO VALUE * Oldrich Alfons Vasicek

THE DISTRIBUTION OF LOAN PORTFOLIO VALUE * Oldrich Alfons Vasicek HE DISRIBUION OF LOAN PORFOLIO VALUE * Oldrch Alfons Vascek he amount of captal necessary to support a portfolo of debt securtes depends on the probablty dstrbuton of the portfolo loss. Consder a portfolo

More information

Generalizing the degree sequence problem

Generalizing the degree sequence problem Mddlebury College March 2009 Arzona State Unversty Dscrete Mathematcs Semnar The degree sequence problem Problem: Gven an nteger sequence d = (d 1,...,d n ) determne f there exsts a graph G wth d as ts

More information

Logistic Regression. Lecture 4: More classifiers and classes. Logistic regression. Adaboost. Optimization. Multiple class classification

Logistic Regression. Lecture 4: More classifiers and classes. Logistic regression. Adaboost. Optimization. Multiple class classification Lecture 4: More classfers and classes C4B Machne Learnng Hlary 20 A. Zsserman Logstc regresson Loss functons revsted Adaboost Loss functons revsted Optmzaton Multple class classfcaton Logstc Regresson

More information

Imperial College London

Imperial College London F. Fang 1, C.C. Pan 1, I.M. Navon 2, M.D. Pggott 1, G.J. Gorman 1, P.A. Allson 1 and A.J.H. Goddard 1 1 Appled Modellng and Computaton Group Department of Earth Scence and Engneerng Imperal College London,

More information

What is Candidate Sampling

What is Candidate Sampling What s Canddate Samplng Say we have a multclass or mult label problem where each tranng example ( x, T ) conssts of a context x a small (mult)set of target classes T out of a large unverse L of possble

More information

On the Approximation Error of Mean-Field Models

On the Approximation Error of Mean-Field Models On the Approxmaton Error of Mean-Feld Models Le Yng School of Electrcal, Computer and Energy Engneerng Arzona State Unversty Tempe, AZ 85287 le.yng.2@asu.edu ABSTRACT Mean-feld models have been used to

More information

OPTIMAL INVESTMENT POLICIES FOR THE HORSE RACE MODEL. Thomas S. Ferguson and C. Zachary Gilstein UCLA and Bell Communications May 1985, revised 2004

OPTIMAL INVESTMENT POLICIES FOR THE HORSE RACE MODEL. Thomas S. Ferguson and C. Zachary Gilstein UCLA and Bell Communications May 1985, revised 2004 OPTIMAL INVESTMENT POLICIES FOR THE HORSE RACE MODEL Thomas S. Ferguson and C. Zachary Glsten UCLA and Bell Communcatons May 985, revsed 2004 Abstract. Optmal nvestment polces for maxmzng the expected

More information

Loop Parallelization

Loop Parallelization - - Loop Parallelzaton C-52 Complaton steps: nested loops operatng on arrays, sequentell executon of teraton space DECLARE B[..,..+] FOR I :=.. FOR J :=.. I B[I,J] := B[I-,J]+B[I-,J-] ED FOR ED FOR analyze

More information

Solution: Let i = 10% and d = 5%. By definition, the respective forces of interest on funds A and B are. i 1 + it. S A (t) = d (1 dt) 2 1. = d 1 dt.

Solution: Let i = 10% and d = 5%. By definition, the respective forces of interest on funds A and B are. i 1 + it. S A (t) = d (1 dt) 2 1. = d 1 dt. Chapter 9 Revew problems 9.1 Interest rate measurement Example 9.1. Fund A accumulates at a smple nterest rate of 10%. Fund B accumulates at a smple dscount rate of 5%. Fnd the pont n tme at whch the forces

More information

When Network Effect Meets Congestion Effect: Leveraging Social Services for Wireless Services

When Network Effect Meets Congestion Effect: Leveraging Social Services for Wireless Services When Network Effect Meets Congeston Effect: Leveragng Socal Servces for Wreless Servces aowen Gong School of Electrcal, Computer and Energy Engeerng Arzona State Unversty Tempe, AZ 8587, USA xgong9@asuedu

More information

PSYCHOLOGICAL RESEARCH (PYC 304-C) Lecture 12

PSYCHOLOGICAL RESEARCH (PYC 304-C) Lecture 12 14 The Ch-squared dstrbuton PSYCHOLOGICAL RESEARCH (PYC 304-C) Lecture 1 If a normal varable X, havng mean µ and varance σ, s standardsed, the new varable Z has a mean 0 and varance 1. When ths standardsed

More information

where the coordinates are related to those in the old frame as follows.

where the coordinates are related to those in the old frame as follows. Chapter 2 - Cartesan Vectors and Tensors: Ther Algebra Defnton of a vector Examples of vectors Scalar multplcaton Addton of vectors coplanar vectors Unt vectors A bass of non-coplanar vectors Scalar product

More information

COLLOQUIUM MATHEMATICUM

COLLOQUIUM MATHEMATICUM COLLOQUIUM MATHEMATICUM VOL. 74 997 NO. TRANSFERENCE THEORY ON HARDY AND SOBOLEV SPACES BY MARIA J. CARRO AND JAVIER SORIA (BARCELONA) We show that the transference method of Cofman and Wess can be extended

More information

THE METHOD OF LEAST SQUARES THE METHOD OF LEAST SQUARES

THE METHOD OF LEAST SQUARES THE METHOD OF LEAST SQUARES The goal: to measure (determne) an unknown quantty x (the value of a RV X) Realsaton: n results: y 1, y 2,..., y j,..., y n, (the measured values of Y 1, Y 2,..., Y j,..., Y n ) every result s encumbered

More information

Dynamics of heterogeneous peer-to-peer networks

Dynamics of heterogeneous peer-to-peer networks Dynamcs of heterogeneous peer-to-peer networks Fernando Pagann, Andrés Ferragut and Martín Zubeldía Unversdad ORT Uruguay Abstract Dynamc models of populaton n peer-to-peer fle sharng systems have focused

More information

DEFINING %COMPLETE IN MICROSOFT PROJECT

DEFINING %COMPLETE IN MICROSOFT PROJECT CelersSystems DEFINING %COMPLETE IN MICROSOFT PROJECT PREPARED BY James E Aksel, PMP, PMI-SP, MVP For Addtonal Informaton about Earned Value Management Systems and reportng, please contact: CelersSystems,

More information

Causal, Explanatory Forecasting. Analysis. Regression Analysis. Simple Linear Regression. Which is Independent? Forecasting

Causal, Explanatory Forecasting. Analysis. Regression Analysis. Simple Linear Regression. Which is Independent? Forecasting Causal, Explanatory Forecastng Assumes cause-and-effect relatonshp between system nputs and ts output Forecastng wth Regresson Analyss Rchard S. Barr Inputs System Cause + Effect Relatonshp The job of

More information

Fisher Markets and Convex Programs

Fisher Markets and Convex Programs Fsher Markets and Convex Programs Nkhl R. Devanur 1 Introducton Convex programmng dualty s usually stated n ts most general form, wth convex objectve functons and convex constrants. (The book by Boyd and

More information

Bandwdth Packng E. G. Coman, Jr. and A. L. Stolyar Bell Labs, Lucent Technologes Murray Hll, NJ 07974 fegc,stolyarg@research.bell-labs.com Abstract We model a server that allocates varyng amounts of bandwdth

More information

Module 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Module 2 LOSSLESS IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur Module LOSSLESS IMAGE COMPRESSION SYSTEMS Lesson 3 Lossless Compresson: Huffman Codng Instructonal Objectves At the end of ths lesson, the students should be able to:. Defne and measure source entropy..

More information

SPEE Recommended Evaluation Practice #6 Definition of Decline Curve Parameters Background:

SPEE Recommended Evaluation Practice #6 Definition of Decline Curve Parameters Background: SPEE Recommended Evaluaton Practce #6 efnton of eclne Curve Parameters Background: The producton hstores of ol and gas wells can be analyzed to estmate reserves and future ol and gas producton rates and

More information

A Lyapunov Optimization Approach to Repeated Stochastic Games

A Lyapunov Optimization Approach to Repeated Stochastic Games PROC. ALLERTON CONFERENCE ON COMMUNICATION, CONTROL, AND COMPUTING, OCT. 2013 1 A Lyapunov Optmzaton Approach to Repeated Stochastc Games Mchael J. Neely Unversty of Southern Calforna http://www-bcf.usc.edu/

More information

How Sets of Coherent Probabilities May Serve as Models for Degrees of Incoherence

How Sets of Coherent Probabilities May Serve as Models for Degrees of Incoherence 1 st Internatonal Symposum on Imprecse Probabltes and Ther Applcatons, Ghent, Belgum, 29 June 2 July 1999 How Sets of Coherent Probabltes May Serve as Models for Degrees of Incoherence Mar J. Schervsh

More information

IDENTIFICATION AND CORRECTION OF A COMMON ERROR IN GENERAL ANNUITY CALCULATIONS

IDENTIFICATION AND CORRECTION OF A COMMON ERROR IN GENERAL ANNUITY CALCULATIONS IDENTIFICATION AND CORRECTION OF A COMMON ERROR IN GENERAL ANNUITY CALCULATIONS Chrs Deeley* Last revsed: September 22, 200 * Chrs Deeley s a Senor Lecturer n the School of Accountng, Charles Sturt Unversty,

More information

FINANCIAL MATHEMATICS. A Practical Guide for Actuaries. and other Business Professionals

FINANCIAL MATHEMATICS. A Practical Guide for Actuaries. and other Business Professionals FINANCIAL MATHEMATICS A Practcal Gude for Actuares and other Busness Professonals Second Edton CHRIS RUCKMAN, FSA, MAAA JOE FRANCIS, FSA, MAAA, CFA Study Notes Prepared by Kevn Shand, FSA, FCIA Assstant

More information

The descriptive complexity of the family of Banach spaces with the π-property

The descriptive complexity of the family of Banach spaces with the π-property Arab. J. Math. (2015) 4:35 39 DOI 10.1007/s40065-014-0116-3 Araban Journal of Mathematcs Ghadeer Ghawadrah The descrptve complexty of the famly of Banach spaces wth the π-property Receved: 25 March 2014

More information

Forecasting the Direction and Strength of Stock Market Movement

Forecasting the Direction and Strength of Stock Market Movement Forecastng the Drecton and Strength of Stock Market Movement Jngwe Chen Mng Chen Nan Ye cjngwe@stanford.edu mchen5@stanford.edu nanye@stanford.edu Abstract - Stock market s one of the most complcated systems

More information

A Probabilistic Theory of Coherence

A Probabilistic Theory of Coherence A Probablstc Theory of Coherence BRANDEN FITELSON. The Coherence Measure C Let E be a set of n propostons E,..., E n. We seek a probablstc measure C(E) of the degree of coherence of E. Intutvely, we want

More information

1. Fundamentals of probability theory 2. Emergence of communication traffic 3. Stochastic & Markovian Processes (SP & MP)

1. Fundamentals of probability theory 2. Emergence of communication traffic 3. Stochastic & Markovian Processes (SP & MP) 6.3 / -- Communcaton Networks II (Görg) SS20 -- www.comnets.un-bremen.de Communcaton Networks II Contents. Fundamentals of probablty theory 2. Emergence of communcaton traffc 3. Stochastc & Markovan Processes

More information

Combinatorial Agency of Threshold Functions

Combinatorial Agency of Threshold Functions Combnatoral Agency of Threshold Functons Shal Jan Computer Scence Department Yale Unversty New Haven, CT 06520 shal.jan@yale.edu Davd C. Parkes School of Engneerng and Appled Scences Harvard Unversty Cambrdge,

More information

The literature on many-server approximations provides significant simplifications toward the optimal capacity

The literature on many-server approximations provides significant simplifications toward the optimal capacity Publshed onlne ahead of prnt November 13, 2009 Copyrght: INFORMS holds copyrght to ths Artcles n Advance verson, whch s made avalable to nsttutonal subscrbers. The fle may not be posted on any other webste,

More information

NMT EE 589 & UNM ME 482/582 ROBOT ENGINEERING. Dr. Stephen Bruder NMT EE 589 & UNM ME 482/582

NMT EE 589 & UNM ME 482/582 ROBOT ENGINEERING. Dr. Stephen Bruder NMT EE 589 & UNM ME 482/582 NMT EE 589 & UNM ME 482/582 ROBOT ENGINEERING Dr. Stephen Bruder NMT EE 589 & UNM ME 482/582 7. Root Dynamcs 7.2 Intro to Root Dynamcs We now look at the forces requred to cause moton of the root.e. dynamcs!!

More information

Extending Probabilistic Dynamic Epistemic Logic

Extending Probabilistic Dynamic Epistemic Logic Extendng Probablstc Dynamc Epstemc Logc Joshua Sack May 29, 2008 Probablty Space Defnton A probablty space s a tuple (S, A, µ), where 1 S s a set called the sample space. 2 A P(S) s a σ-algebra: a set

More information

On the Optimal Control of a Cascade of Hydro-Electric Power Stations

On the Optimal Control of a Cascade of Hydro-Electric Power Stations On the Optmal Control of a Cascade of Hydro-Electrc Power Statons M.C.M. Guedes a, A.F. Rbero a, G.V. Smrnov b and S. Vlela c a Department of Mathematcs, School of Scences, Unversty of Porto, Portugal;

More information

Can Auto Liability Insurance Purchases Signal Risk Attitude?

Can Auto Liability Insurance Purchases Signal Risk Attitude? Internatonal Journal of Busness and Economcs, 2011, Vol. 10, No. 2, 159-164 Can Auto Lablty Insurance Purchases Sgnal Rsk Atttude? Chu-Shu L Department of Internatonal Busness, Asa Unversty, Tawan Sheng-Chang

More information

8 Algorithm for Binary Searching in Trees

8 Algorithm for Binary Searching in Trees 8 Algorthm for Bnary Searchng n Trees In ths secton we present our algorthm for bnary searchng n trees. A crucal observaton employed by the algorthm s that ths problem can be effcently solved when the

More information

On Robust Network Planning

On Robust Network Planning On Robust Network Plannng Al Tzghadam School of Electrcal and Computer Engneerng Unversty of Toronto, Toronto, Canada Emal: al.tzghadam@utoronto.ca Alberto Leon-Garca School of Electrcal and Computer Engneerng

More information

On the Solution of Indefinite Systems Arising in Nonlinear Optimization

On the Solution of Indefinite Systems Arising in Nonlinear Optimization On the Soluton of Indefnte Systems Arsng n Nonlnear Optmzaton Slva Bonettn, Valera Ruggero and Federca Tnt Dpartmento d Matematca, Unverstà d Ferrara Abstract We consder the applcaton of the precondtoned

More information

An Evaluation of the Extended Logistic, Simple Logistic, and Gompertz Models for Forecasting Short Lifecycle Products and Services

An Evaluation of the Extended Logistic, Simple Logistic, and Gompertz Models for Forecasting Short Lifecycle Products and Services An Evaluaton of the Extended Logstc, Smple Logstc, and Gompertz Models for Forecastng Short Lfecycle Products and Servces Charles V. Trappey a,1, Hsn-yng Wu b a Professor (Management Scence), Natonal Chao

More information

CHAPTER 14 MORE ABOUT REGRESSION

CHAPTER 14 MORE ABOUT REGRESSION CHAPTER 14 MORE ABOUT REGRESSION We learned n Chapter 5 that often a straght lne descrbes the pattern of a relatonshp between two quanttatve varables. For nstance, n Example 5.1 we explored the relatonshp

More information

Equlbra Exst and Trade S effcent proportionally

Equlbra Exst and Trade S effcent proportionally On Compettve Nonlnear Prcng Andrea Attar Thomas Marott Franços Salané February 27, 2013 Abstract A buyer of a dvsble good faces several dentcal sellers. The buyer s preferences are her prvate nformaton,

More information

Period and Deadline Selection for Schedulability in Real-Time Systems

Period and Deadline Selection for Schedulability in Real-Time Systems Perod and Deadlne Selecton for Schedulablty n Real-Tme Systems Thdapat Chantem, Xaofeng Wang, M.D. Lemmon, and X. Sharon Hu Department of Computer Scence and Engneerng, Department of Electrcal Engneerng

More information

On fourth order simultaneously zero-finding method for multiple roots of complex polynomial equations 1

On fourth order simultaneously zero-finding method for multiple roots of complex polynomial equations 1 General Mathematcs Vol. 6, No. 3 (2008), 9 3 On fourth order smultaneously zero-fndng method for multple roots of complex polynomal euatons Nazr Ahmad Mr and Khald Ayub Abstract In ths paper, we present

More information

Joint Scheduling of Processing and Shuffle Phases in MapReduce Systems

Joint Scheduling of Processing and Shuffle Phases in MapReduce Systems Jont Schedulng of Processng and Shuffle Phases n MapReduce Systems Fangfe Chen, Mural Kodalam, T. V. Lakshman Department of Computer Scence and Engneerng, The Penn State Unversty Bell Laboratores, Alcatel-Lucent

More information

Lecture 2: Single Layer Perceptrons Kevin Swingler

Lecture 2: Single Layer Perceptrons Kevin Swingler Lecture 2: Sngle Layer Perceptrons Kevn Sngler kms@cs.str.ac.uk Recap: McCulloch-Ptts Neuron Ths vastly smplfed model of real neurons s also knon as a Threshold Logc Unt: W 2 A Y 3 n W n. A set of synapses

More information

Stochastic epidemic models revisited: Analysis of some continuous performance measures

Stochastic epidemic models revisited: Analysis of some continuous performance measures Stochastc epdemc models revsted: Analyss of some contnuous performance measures J.R. Artalejo Faculty of Mathematcs, Complutense Unversty of Madrd, 28040 Madrd, Span A. Economou Department of Mathematcs,

More information

Do Hidden Variables. Improve Quantum Mechanics?

Do Hidden Variables. Improve Quantum Mechanics? Radboud Unverstet Njmegen Do Hdden Varables Improve Quantum Mechancs? Bachelor Thess Author: Denns Hendrkx Begeleder: Prof. dr. Klaas Landsman Abstract Snce the dawn of quantum mechancs physcst have contemplated

More information

A hybrid global optimization algorithm based on parallel chaos optimization and outlook algorithm

A hybrid global optimization algorithm based on parallel chaos optimization and outlook algorithm Avalable onlne www.ocpr.com Journal of Chemcal and Pharmaceutcal Research, 2014, 6(7):1884-1889 Research Artcle ISSN : 0975-7384 CODEN(USA) : JCPRC5 A hybrd global optmzaton algorthm based on parallel

More information

Compiling for Parallelism & Locality. Dependence Testing in General. Algorithms for Solving the Dependence Problem. Dependence Testing

Compiling for Parallelism & Locality. Dependence Testing in General. Algorithms for Solving the Dependence Problem. Dependence Testing Complng for Parallelsm & Localty Dependence Testng n General Assgnments Deadlne for proect 4 extended to Dec 1 Last tme Data dependences and loops Today Fnsh data dependence analyss for loops General code

More information

A Secure Password-Authenticated Key Agreement Using Smart Cards

A Secure Password-Authenticated Key Agreement Using Smart Cards A Secure Password-Authentcated Key Agreement Usng Smart Cards Ka Chan 1, Wen-Chung Kuo 2 and Jn-Chou Cheng 3 1 Department of Computer and Informaton Scence, R.O.C. Mltary Academy, Kaohsung 83059, Tawan,

More information

Lecture 3: Force of Interest, Real Interest Rate, Annuity

Lecture 3: Force of Interest, Real Interest Rate, Annuity Lecture 3: Force of Interest, Real Interest Rate, Annuty Goals: Study contnuous compoundng and force of nterest Dscuss real nterest rate Learn annuty-mmedate, and ts present value Study annuty-due, and

More information

The Choice of Direct Dealing or Electronic Brokerage in Foreign Exchange Trading

The Choice of Direct Dealing or Electronic Brokerage in Foreign Exchange Trading The Choce of Drect Dealng or Electronc Brokerage n Foregn Exchange Tradng Mchael Melvn & Ln Wen Arzona State Unversty Introducton Electronc Brokerage n Foregn Exchange Start from a base of zero n 1992

More information

Cautiousness and Measuring An Investor s Tendency to Buy Options

Cautiousness and Measuring An Investor s Tendency to Buy Options Cautousness and Measurng An Investor s Tendency to Buy Optons James Huang October 18, 2005 Abstract As s well known, Arrow-Pratt measure of rsk averson explans a ratonal nvestor s behavor n stock markets

More information

"Research Note" APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES *

Research Note APPLICATION OF CHARGE SIMULATION METHOD TO ELECTRIC FIELD CALCULATION IN THE POWER CABLES * Iranan Journal of Scence & Technology, Transacton B, Engneerng, ol. 30, No. B6, 789-794 rnted n The Islamc Republc of Iran, 006 Shraz Unversty "Research Note" ALICATION OF CHARGE SIMULATION METHOD TO ELECTRIC

More information

CALL ADMISSION CONTROL IN WIRELESS MULTIMEDIA NETWORKS

CALL ADMISSION CONTROL IN WIRELESS MULTIMEDIA NETWORKS CALL ADMISSION CONTROL IN WIRELESS MULTIMEDIA NETWORKS Novella Bartoln 1, Imrch Chlamtac 2 1 Dpartmento d Informatca, Unverstà d Roma La Sapenza, Roma, Italy novella@ds.unroma1.t 2 Center for Advanced

More information

The Power of Slightly More than One Sample in Randomized Load Balancing

The Power of Slightly More than One Sample in Randomized Load Balancing The Power of Slghtly More than One Sample n Randomzed oad Balancng e Yng, R. Srkant and Xaohan Kang Abstract In many computng and networkng applcatons, arrvng tasks have to be routed to one of many servers,

More information

Mathematical Option Pricing

Mathematical Option Pricing Mark H.A.Davs Mathematcal Opton Prcng MSc Course n Mathematcs and Fnance Imperal College London 11 January 26 Department of Mathematcs Imperal College London South Kensngton Campus London SW7 2AZ Contents

More information

Brigid Mullany, Ph.D University of North Carolina, Charlotte

Brigid Mullany, Ph.D University of North Carolina, Charlotte Evaluaton And Comparson Of The Dfferent Standards Used To Defne The Postonal Accuracy And Repeatablty Of Numercally Controlled Machnng Center Axes Brgd Mullany, Ph.D Unversty of North Carolna, Charlotte

More information

A Penalty Method for American Options with Jump Diffusion Processes

A Penalty Method for American Options with Jump Diffusion Processes A Penalty Method for Amercan Optons wth Jump Dffuson Processes Y. d Hallun, P.A. Forsyth, and G. Labahn March 9, 23 Abstract The far prce for an Amercan opton where the underlyng asset follows a jump dffuson

More information

STATISTICAL DATA ANALYSIS IN EXCEL

STATISTICAL DATA ANALYSIS IN EXCEL Mcroarray Center STATISTICAL DATA ANALYSIS IN EXCEL Lecture 6 Some Advanced Topcs Dr. Petr Nazarov 14-01-013 petr.nazarov@crp-sante.lu Statstcal data analyss n Ecel. 6. Some advanced topcs Correcton for

More information

Abstract. 260 Business Intelligence Journal July IDENTIFICATION OF DEMAND THROUGH STATISTICAL DISTRIBUTION MODELING FOR IMPROVED DEMAND FORECASTING

Abstract. 260 Business Intelligence Journal July IDENTIFICATION OF DEMAND THROUGH STATISTICAL DISTRIBUTION MODELING FOR IMPROVED DEMAND FORECASTING 260 Busness Intellgence Journal July IDENTIFICATION OF DEMAND THROUGH STATISTICAL DISTRIBUTION MODELING FOR IMPROVED DEMAND FORECASTING Murphy Choy Mchelle L.F. Cheong School of Informaton Systems, Sngapore

More information

Duality for nonsmooth mathematical programming problems with equilibrium constraints

Duality for nonsmooth mathematical programming problems with equilibrium constraints uu et al. Journal of Inequaltes and Applcatons (2016) 2016:28 DOI 10.1186/s13660-016-0969-4 R E S E A R C H Open Access Dualty for nonsmooth mathematcal programmng problems wth equlbrum constrants Sy-Mng

More information

The Stock Market Game and the Kelly-Nash Equilibrium

The Stock Market Game and the Kelly-Nash Equilibrium The Stock Market Game and the Kelly-Nash Equlbrum Carlos Alós-Ferrer, Ana B. Ana Department of Economcs, Unversty of Venna. Hohenstaufengasse 9, A-1010 Venna, Austra. July 2003 Abstract We formulate the

More information

L10: Linear discriminants analysis

L10: Linear discriminants analysis L0: Lnear dscrmnants analyss Lnear dscrmnant analyss, two classes Lnear dscrmnant analyss, C classes LDA vs. PCA Lmtatons of LDA Varants of LDA Other dmensonalty reducton methods CSCE 666 Pattern Analyss

More information

J. Parallel Distrib. Comput.

J. Parallel Distrib. Comput. J. Parallel Dstrb. Comput. 71 (2011) 62 76 Contents lsts avalable at ScenceDrect J. Parallel Dstrb. Comput. journal homepage: www.elsever.com/locate/jpdc Optmzng server placement n dstrbuted systems n

More information

On the Interaction between Load Balancing and Speed Scaling

On the Interaction between Load Balancing and Speed Scaling On the Interacton between Load Balancng and Speed Scalng Ljun Chen, Na L and Steven H. Low Engneerng & Appled Scence Dvson, Calforna Insttute of Technology, USA Abstract Speed scalng has been wdely adopted

More information

Solving Factored MDPs with Continuous and Discrete Variables

Solving Factored MDPs with Continuous and Discrete Variables Solvng Factored MPs wth Contnuous and screte Varables Carlos Guestrn Berkeley Research Center Intel Corporaton Mlos Hauskrecht epartment of Computer Scence Unversty of Pttsburgh Branslav Kveton Intellgent

More information

How To Calculate The Accountng Perod Of Nequalty

How To Calculate The Accountng Perod Of Nequalty Inequalty and The Accountng Perod Quentn Wodon and Shlomo Ytzha World Ban and Hebrew Unversty September Abstract Income nequalty typcally declnes wth the length of tme taen nto account for measurement.

More information

Comparison of Control Strategies for Shunt Active Power Filter under Different Load Conditions

Comparison of Control Strategies for Shunt Active Power Filter under Different Load Conditions Comparson of Control Strateges for Shunt Actve Power Flter under Dfferent Load Condtons Sanjay C. Patel 1, Tushar A. Patel 2 Lecturer, Electrcal Department, Government Polytechnc, alsad, Gujarat, Inda

More information

Point cloud to point cloud rigid transformations. Minimizing Rigid Registration Errors

Point cloud to point cloud rigid transformations. Minimizing Rigid Registration Errors Pont cloud to pont cloud rgd transformatons Russell Taylor 600.445 1 600.445 Fall 000-014 Copyrght R. H. Taylor Mnmzng Rgd Regstraton Errors Typcally, gven a set of ponts {a } n one coordnate system and

More information

The Development of Web Log Mining Based on Improve-K-Means Clustering Analysis

The Development of Web Log Mining Based on Improve-K-Means Clustering Analysis The Development of Web Log Mnng Based on Improve-K-Means Clusterng Analyss TngZhong Wang * College of Informaton Technology, Luoyang Normal Unversty, Luoyang, 471022, Chna wangtngzhong2@sna.cn Abstract.

More information

Risk-based Fatigue Estimate of Deep Water Risers -- Course Project for EM388F: Fracture Mechanics, Spring 2008

Risk-based Fatigue Estimate of Deep Water Risers -- Course Project for EM388F: Fracture Mechanics, Spring 2008 Rsk-based Fatgue Estmate of Deep Water Rsers -- Course Project for EM388F: Fracture Mechancs, Sprng 2008 Chen Sh Department of Cvl, Archtectural, and Envronmental Engneerng The Unversty of Texas at Austn

More information

The OC Curve of Attribute Acceptance Plans

The OC Curve of Attribute Acceptance Plans The OC Curve of Attrbute Acceptance Plans The Operatng Characterstc (OC) curve descrbes the probablty of acceptng a lot as a functon of the lot s qualty. Fgure 1 shows a typcal OC Curve. 10 8 6 4 1 3 4

More information

Energies of Network Nastsemble

Energies of Network Nastsemble Supplementary materal: Assessng the relevance of node features for network structure Gnestra Bancon, 1 Paolo Pn,, 3 and Matteo Marsl 1 1 The Abdus Salam Internatonal Center for Theoretcal Physcs, Strada

More information

Estimation of Dispersion Parameters in GLMs with and without Random Effects

Estimation of Dispersion Parameters in GLMs with and without Random Effects Mathematcal Statstcs Stockholm Unversty Estmaton of Dsperson Parameters n GLMs wth and wthout Random Effects Meng Ruoyan Examensarbete 2004:5 Postal address: Mathematcal Statstcs Dept. of Mathematcs Stockholm

More information

1. Measuring association using correlation and regression

1. Measuring association using correlation and regression How to measure assocaton I: Correlaton. 1. Measurng assocaton usng correlaton and regresson We often would lke to know how one varable, such as a mother's weght, s related to another varable, such as a

More information

RESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL. Yaoqi FENG 1, Hanping QIU 1. China Academy of Space Technology (CAST) yaoqi.feng@yahoo.

RESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL. Yaoqi FENG 1, Hanping QIU 1. China Academy of Space Technology (CAST) yaoqi.feng@yahoo. ICSV4 Carns Australa 9- July, 007 RESEARCH ON DUAL-SHAKER SINE VIBRATION CONTROL Yaoq FENG, Hanpng QIU Dynamc Test Laboratory, BISEE Chna Academy of Space Technology (CAST) yaoq.feng@yahoo.com Abstract

More information

SIMPLE LINEAR CORRELATION

SIMPLE LINEAR CORRELATION SIMPLE LINEAR CORRELATION Smple lnear correlaton s a measure of the degree to whch two varables vary together, or a measure of the ntensty of the assocaton between two varables. Correlaton often s abused.

More information