Stability analysis of a max-min fair Rate Control Protocol (RCP) in a small buffer regime

Size: px
Start display at page:

Download "Stability analysis of a max-min fair Rate Control Protocol (RCP) in a small buffer regime"

Transcription

1 Stabiity anaysis of a max-min fair Rate Contro Protoco RCP) in a sma buffer regime Thomas Voice and Gaurav Raina Cambridge Consutants and IIT Madras Abstract In this note we anayse various stabiity properties of a max-min fair Rate Contro Protoco RCP) operating with sma buffers. We first tacke the issue of stabiity for networks with arbitrary topoogies. We prove that the max-min fair RCP fuid mode is gobay stabe in the absence of propagation deays, and aso derive a set of conditions for oca stabiity when arbitrary heterogeneous propagation deays are present. The network deay stabiity resut assumes that, at equiibrium, there is ony one botteneck ink aong each route. Lasty, in the simper setting of a singe ink, singe deay mode, we investigate the impact of the oss of oca stabiity via a Hopf bifurcation. keywords rate contro, max-min fairness, sma buffers, stabiity, bifurcation. 1 Introduction The Rate Contro Protoco RCP) [4, 7, 9] takes a radicay different approach for managing fow and congestion contro as compared to the existing Transmission Contro Protoco TCP) congestion avoidance framework. The TCP framework has, embedded in it, an impicit mechanism for detecting congestion within the network. Loss of a packet, which may be caused by the overfow of a buffer, is expected to provide the necessary feedback information. In sharp contrast, RCP aims to achieve fast fow competion times by communicating expicit rate feedback between routers and end-stations [8]. RCP is cosey reated to the expicit Contro Protoco XCP) [16], which aso proposes the use of expicit rate feedback. In fact, there is growing interest in designing scaabe agorithms for rate and congestion contro that use expicit feedback from routers; for exampe, see the recenty proposed JetMax [33] and references therein. RCP, XCP and the JetMax agorithms are intended to converge to a max-min fair resource aocation [4, 7, 16, 33]; see [22, 23, 24, 29] for a sampe of the iterature exporing various issues reated to different notions of fairness in a networking context. The performance of congestion contro agorithms is often couped with the choice of certain parameters in routers in the network. For exampe, such parameters may correspond to different choices of Active Queue Management AQM) 1

2 schemes [6, 14, 19, 21], or the size of buffers in routers [12, 26, 27]. Researchers have begun questioning the design rues for sizing buffers in routers in the Internet [3] and some recent iterature [10, 26, 32] suggests that buffers shoud indeed be much smaer as compared to the current design principes. In the ight of the separate deveopments made towards the understanding of RCP [4, 7, 9] and buffer sizing [10, 26, 32] it is natura to investigate the two, i.e., RCP and sma buffers, together. Such an investigation is aso motivated by the observation that previous buffer sizing studies have focused primariy on the impact of smaer buffers with TCP. In this note, the focus of our anaysis wi be on the imiting regime of a max-min fair RCP fuid mode operating over a communication network with sma buffers. Previous contro theoretic anaysis of RCP has focused on a singe botteneck, where the queue is modeed as a saturated integrator [4]. In our work, we assume that the size of the buffers is so sma that it is unnecessary to mode the queue expicity. Rather, with such sma buffers, at the time scae of operation of the congestion controers, it is the distribution of queue size that pays the important roe [26, 32]. It is worth stating that we do not in any way suggest exacty how sma the size of buffers ought to be. The mode we anayse is simper than the one anaysed by [4] in that we do not expicity mode the queue, but it is more invoved in the sense that our framework represents a heterogeneous network of arbitrary topoogy. In our fuid mode for RCP, apart from the mode for the queue, a other parameters are exacty the same as specified in the origina RCP agorithm [7]. The styes of anaysis we empoy have a been popuar in the study of congestion contro: for exampe, goba stabiity without propagation deays [28, 31], oca stabiity with deays [12, 14, 17, 19, 21, 28, 30], and aso an anaysis of the dynamic system when stabiity may not be guaranteed [13, 18, 25]. We now outine the essence of our theoretica contribution. First, we prove that the fuid mode of RCP is gobay stabe in the absence of propagation deays. Goba stabiity is a desirabe property, and it is reassuring that the fuid mode is gobay stabe, even if it is without propagation time deays. The next natura step is to investigate the stabiity properties of the system with time deays. To that end, we derive a set of conditions for oca stabiity when arbitrary heterogeneous propagation deays are present. The network deay stabiity resut depends on the assumption that, at equiibrium, there is ony one botteneck ink aong each route. Furthermore, in the simper setting of a singe ink, singe deay mode, we are abe to provide a sighty more in-depth study of the RCP agorithm. Using bifurcation anaysis, we investigate the impact of the oss of oca stabiity in a specia case where we show that the RCP agorithm woud aways give rise to a super-critica Hopf bifurcation. An overview of this note foows. In Section 2, we anayse some stabiity properties of RCP over networks of arbitrary topoogies. In Section 3, we investigate oca instabiity of RCP in a singe ink, singe deay mode, via a Hopf bifurcation anaysis. Finay, in Section 4 we summarise our contributions and discuss various avenues for further research. 2

3 2 RCP over arbitrary topoogy networks In this section, our objective is to anayse some stabiity properties of RCP in a sma buffer regime, over networks with arbitrary topoogies. We first show goba stabiity for the RCP fuid mode in the absence of propagation deays, and then derive a set of conditions for oca stabiity when arbitrary heterogeneous propagation deays are present. We begin with the mode description. 2.1 Mode description At the eve of theoretica abstraction, our set-up for a communication network foows the commony adopted framework [28]. We suppose that the communication network consists of an interconnection of a set of routes, S, with a set of inks, J. Each route r S represents a user of the network. Associated with each route is a set of inks which represents the path aong which that user transmits information through the network. Further, a route r has associated with it a fow rate x r t) 0, which represents a dynamic fuid approximation to the rate at which the user is sending packets aong route r, at time t. The fow rate for each r S is determined by the inks r, via the use of expicit rate feedback. Each ink J has associated with it a fow rate R t), which represents the maximum fow rate aowed for routes which pass through. Each packet which is sent through the network carries, in its header, an expicit rate feedback variabe. This variabe is initiay set to the maximum desired fow rate for r. As the packet passes through each r, if the feedback variabe is greater than R t), then ink sets it equa to R t). When the packet reaches its destination, an acknowedgement packet ack), containing the fina vaue of the expicit rate feedback variabe is returned to the origin of r, and the fow rate x r t) is updated accordingy. For each route r and ink r, we et τ r denote the propagation deay from the origin of r to, i.e. the amount of time it takes for a packet to trave from the origin to ink aong route r. Let τ r denote the propagation deay from to the origin of r, i.e. the amount of time it takes for the expicit rate feedback information from ink to reach the user who is transmitting aong route r. In RCP, a packet must reach its destination before an acknowedgement packet ack) is returned to its source. Furthermore, as we are interested in a network with sma buffers, we may safey assume that queueing deays are a negigibe component of the tota end-to-end deay. Thus for a r, τ r + τ r = τ r, the round-trip time for route r. We have the foowing mode for the end-system behaviour of RCP: for each route r S x r t) = min r R t τ r ). 1) Now, for each ink J, R t) is updated depending upon the tota aggregate fow through ink at time t. Athough the update rue is discrete, we can mode it via a fuid approximation with the foowing differentia equation [4]: for each 3

4 ink J α Ṙ t) = R t) C y t)) β ) + q t) d C d 2 C, 2) R t) where α, β are positive constants, C is the capacity at ink, d is the average round-trip time of the fows passing through ink, y t) = r: r x r t τ r ), 3) and q t) is the queue size at time t. Here we use the notation a = b) + c to mean that a = 0 if b < 0 and c 0, otherwise a = b. In the origina RCP mode [7], α = α and β = β for each J for some α and β. We have aowed these constants to vary between inks in order to study their effect on stabiity. Mode for the queue. The basis of our investigation is to study a regime where the buffers are so sma that it is inappropriate to mode the queue as a saturated integrator. This assumption simpy expresses the idea that with sma enough buffers, we may use the approximation q t) = p y t)), 4) where p ) is a continuousy differentiabe function representing the mean queue ength of ink. This is consistent with the observation in [26, 32] that in a sma buffer regime, it is the distribution of the queue size that pays the prominent roe in the dynamics of the congestion contro framework. At the eve of interest in this note, we do not propose any expicit functiona form for the mean queue ength. Different functiona forms may be suitabe candidates, but our primary focus is to investigate some stabiity properties of the dynamica system defined by 1-4). We observe that in some recent anaysis of sma buffer networks it has been shown that, under TCP congestion contro, uness fow access speeds are imited, the arriva and departure of fows to a sma buffer ink can degrade performance [20]. This study has dwet on the dynamics of TCP users arriving and departing. However, under RCP the access speed of each new fow that passes through a ink is bounded by R t). Reation to some previous work in ATM networks. We note that the RCP agorithm, basicay that given by 2), shares simiarities with some agorithms proposed in the past in the context of Asynchronous Transfer Mode ATM) networks: see [1, 2, 11, 15] for some agorithms and anaysis. The most noticeabe difference is that in prior work β = 0, and hence they do not consider the impact of feedback based on queue size. The current state of the art in the anaysis of max-min fair expicit congestion contro agorithms, with feedback deays, is [4], which proposes two forms of feedback: rate mismatch and queue size. Thus far there appears to be no definitive concusion about the excusion of feedback based on queue size and so the anaysis of system 1-4) where β > 0 4

5 is sti important. We aso highight the anaysis of [5], which deveops robust controers to dea with a regime where, due to queue dynamics, the end systems have to cope with varying deays. In contrast, the focus of our anaysis is on controing the distribution of the queue size; a regime aso anaysed in [26]. A key focus of this note is a contro theoretic anaysis of the system described by 1-4), with heterogeneous feedback deays. This aspect has not previousy been anayzed, and so our anaysis presents a step forward. 2.2 Goba stabiity without propagation deays In this subsection, we show that the fuid mode of RCP is gobay stabe when the effects of propagation deays are not taken into account. This provides a check on the goba behaviour of the agorithm and compements our ater oca stabiity anaysis where we incorporate deays into the mode. We note that goba anaysis of fuid modes for congestion contro without deays foowed by a oca anaysis of the same mode with deays is a commony adopted practice in fow contro theory [28]. With RCP, one way of thinking about this is to note that the deay-free mode becomes more accurate when, for a J, the α and β terms in 2) are sma, and so the agorithm evoves on a sower timescae than propagation deays. We first deduce that with sufficienty sma α RCP is gobay stabe, and then use oca anaysis to deveop insight into when, for arger vaues of α, oca instabiities may occur. To mode RCP without the effects of deays we use 1-4) setting τ r = 0, for a r S, but eave a other parameters fixed. Ideay we woud ike to show that our deay-free mode of RCP, at equiibrium, is gobay stabe. Unfortunatey, as the foowing exampe demonstrates, we cannot expect the equiibrium points of 1-4) to be unique, or even isoated. Consider a network consisting of two identica inks,, j and one route r = {, j}. Then, since p ) is stricty increasing, there exists a unique y such that α C y) β p y) d C d 2 C = 0. By inspection, there are equiibrium points at R = y, y ) and R = y, y) for a y y. Aternativey consider the situation where and j are not competey identica, and C j > C ; then, there exists a unique y such that α j C j y ) β jp j y ) d j C j d 2 j C = 0. j Now y > y, so if x r y then Ṙj wi be stricty positive; however, if x r > y then Ṙ wi be stricty negative. Thus, for this exampe, no equiibrium point exists. Hence there may be an entire manifod of equiibrium vectors for R, or there may be no equiibrium point at a; but, the same is not true for the equiibrium vector of fow rates x. 5

6 Theorem 2.1. Suppose that xt) evoves according to 1-4) with τ r = 0 for a r S. Then there exists a unique vector x such that xt) x, as t. Proof. We prove this resut recursivey for the more genera system where, for a J, we repace 2) with Ṙ t) = R t) u t) + α C y t)) β ) + p y t)) d C d 2 C, 5) R t) where u t) 0 as t. Let N be the number of r S such that r. Set y equa to the unique vaue such that Ṙt) = 0 when y t) = y, and set R = min J y /N. We sha show that, for a r S with R = y /N for some r, x r R as t. This aows us to remove such an x r t) from the system, by repacing it with R pus a vanishing term which we incorporate into the u t), for each r. Since this wi aways remove at east one r from S, this is sufficient to prove our resut. It remains to show that x r R for r S with R = y /N, for some r. Now, for any ɛ > 0 there exists a T such that u t) < ɛ α d C, for a J, t > T. From 5) we see that, if, for t > T, for any J, R t) < y /N ) 2ɛ, then Ṙt) > δ for some δ > 0. Thus, for some T, for a t > T, J, R t) y /N ) 2ɛ. Let us assume that ɛ is sma enough so that, for any j J, with y j /N j ) > R, ) y j R > max N N ɛ. j J Let r S be such that R = y /N for some r. Let r be the set of r with R = y /N. Suppose that for some t > T, min R t) > R + 2ɛ max N j. r j J If x r t) = R j t) and j r then x r t) > R + 2ɛ max j J N j. Otherwise x r t) = R j t) with j r, but since t > T, x r > R + 2ɛ max j J N j sti hods. For any r, x r t) is sufficienty high so that y t) must be greater than N R + 2ɛ. Thus, Ṙ t) < δ for some δ > 0. Therefore, for some T, for a t > T, Hence for a t > T, min R t) R + 2ɛ max N j. r j J xr t) R 2ɛ max j J N j. Since ɛ was arbitrary, the resut foows. Note that the vector x wi be cose to that of a max-min fair aocation of fow rates, because of its construction. 6

7 2.3 Loca stabiity with propagation deays In this subsection, we derive conditions for the oca stabiity of 1-4) when propagation deays are present. Our resut depends on the assumption that there is ony one botteneck ink aong each route; that is, for each r S there is ony one r such that x r t) = R t) at equiibrium. For each r S, J, we et x r be the equiibrium vaue of x r t) and R be the maximum of x s for a s S such that s. For each J, we et y be the unique vaue such that Ṙt) = 0 whenever y t) = y. We can assume, without oss of generaity, that for a J y = r: r x r. If this does not hod, then when the system is cose to equiibrium, R t) wi increase for a t. If R t) is initiay much bigger than x r for a r S, then it wi have no effect on a ocay stabe system and can be ignored. The proof of oca stabiity proceeds as a recursive argument. From our assumptions, for each r S there is a unique j r such that x r = R j. Thus for a r S, r, whenever the system is cose to equiibrium, either x r t) = R t τ r ), or ese x r < R and R t τ r ) has no effect on x r t). Starting with the J such that R is minima, we recursivey isoate each J and mode the effect of the rest of the system on R t) as a perturbation. For each J, the stabiity of the isoated systems for j J with R j < R impies that this perturbation is vanishing. We initiay restrict our attention to the singe ink case, J = {}, and we consider the foowing generaisation of 2) ) ) + Ṙ t) = R t) f y t) N R + u t), 6) R t) where y t) is defined as in 3), f ) is an increasing differentiabe function with f 0) = 0 and u t) 0 as t. When the recursive argument is compete, the vector ut) represents the behaviour of R j t) for j J such that R j < R. We examine oca stabiity for the singe ink case 6) by considering what happens if R t) moves away from R. We show that under the conditions provided, as R t) gets cose to moving the furthest away from R, R t) must turn back towards R. As R t) can ony reach smaer and smaer distances away from R, convergence is ensured. Theorem 2.2. Consider the case where the network consists of ony a singe ink, J =, with fow rates foowing 6). If f 0)R τ r < 1, r S then for a ɛ, there exists a and u such that, if R t) R a for a t 0 and u t) < u for a t, then R t) R ɛ for a t and R t) R as t. 7

8 Proof. Let τ be max r S τ r, the maximum round-trip time. Suppose for some t, γ 1, a and u, R t) R = γa, R t ) R a and u t ) < u a for a t [t 2τ, t]. Then for a t [t τ, t], Ṙt ) R + a ) ) u a + f 0) S a + oa ). Thus for a r S So R t τ r ) R t) τ r R f 0) S a + τ r R + a ) a u + oa ). 7) Ṙ t) R t) = f 0) S R y t) ) u t) + oa ) = f 0) S R R t) ) u t) + f 0) r S = f 0) S γa + δ, ) R t) R t τ r ) + oa ) where δ f 0)R τ r f 0) S a + S τ R + a ) a u + oa ) r S = f 0) 2 S 2 R d a + S τ R + a ) a u + oa ). Since f 0) S R d < 1, there exists a vaue of a < ɛ, u and γ < 1 such that, if a a then δ is guaranteed to be ess than γf 0) S a. In which case Ṙt) must have the same sign as R R t). We can take u = u a, with a as given above, and if R t) R a for a t 0 and u t) < u for a t, then R t) R ɛ for a t. Furthermore, for a a < a, there exists a time T such that, for a t > T, u t) < u a. From the above anaysis, we know that after T, if R t) is beyond a of R then it wi converge to R at a rate of γ. Thus there is some T > T such that R t) R a for a t > T. Since a was arbitrary, R t) R as required. We now proceed with our argument by recursivey appying the above Theorem 2.2 to each ink in the genera network case. Theorem 2.3. For system 1-4) the equiibrium point R is ocay asymptoticay stabe provided that, for each J, α + β p y ) ) d C d 2 C R τ r < 1. 8) r: r,r =x r Proof. Let δ be the minimum of R R j /2 for j J. If for a t and a J, R t) R < δ then each J wi evove according to ) ) Ṙ t) = R t) f t) y eq + u t), 9) y eq 8

9 where u t) represents the effect of R j t) for a R j < R, and y eq t) = x r t τ r ), r: r,x r=r with y eq equa to the equiibrium vaue of y eq t). For each J, we have f w) = α w + y d C C ) + β p y + w) d 2 C. By definition, 9) is an exampe of 6) for the reduced network formed by and a routes r S with r and x r = R. Furthermore, 8) is precisey the condition of Theorem 2.2 for this reduced system. Suppose R t) R < δ for a J, for a t. By differentiabiity of p ), for any u we can find an ɛ such that if Rj t) R j < ɛ for a t and for a j such that R j < R, then u t) < u for a t. This aows us to prove oca stabiity recursivey. We begin with the J such that R is maxima. We appy Theorem 2.2 to find conditions under which R t) converges to R and R t) R < δ for a t. This gives us a and u, where the initia conditions of R t) shoud be within a of R and u t) shoud tend to zero and aways be bounded by u. We can find an ɛ < δ so that if Rj t) R j < ɛ for a t and Rj t) converges to R j for a j, then u t) < u for a t and u t) converges to 0. Hence we set R t) [ R a, R + a ] for t 0 as our initia condition for, remove from the network and repeat this process for δ = ɛ. Since we remove one ink each time, eventuay we wi find suitabe initia conditions for the entire network. Note that each ink does not necessariy need to keep track of which fows are under its contro in order to meet 8). The condition hods if, for a J, α + β p y ) ) y d d 2 d p C < 1, where d p is the average per packet round-trip time. Here we say that the roundtrip time of a packet is that of the fow which has transmitted it. Thus d p = 1 x r τ r. y r: r This suggests that perhaps it shoud be d p, and not d, in the RCP controer 2). If for each J we et γ be such that p y ) = γ p y )/y, then p y ) = γ α d β C y y < γ α d β, thus 8) is ensured if, for a J, α = d d p 1 + γ ). 10) 9

10 REMARK 1. This parameter choice scheme is attractive, because it is fairy decentraized and ony requires oca information. Each ink ony needs to determine or estimate the average round-trip times for packets passing through that ink. In current specifications of RCP, sources actuay measure their own round-trip times and advertise them to inks in packet headers. REMARK 2. The network deay stabiity resut depends on the weak assumption that, at equiibrium, each route has ony one botteneck ink. That is, at equiibrium, for each route there is a unique ink with minima per-fow bandwidth avaiabe. Note that a network with mutipe botteneck inks may be transformed into one with ony a singe botteneck ink, for exampe, by making a sma change in the parameters β, J. Preiminary numerica resuts suggest that Theorem 2.3 may hod in genera; but finding an anaytic resut to confirm this, or a counter-exampe to disprove it, remains an open probem. From a practica networking context, there is growing interest in deveoping agorithms that may be used in conjunction with an expicit congestion contro architecture [9, 16, 33]. However, it is imperative to have theoretica guidance to steer a proposed set of agorithms towards a successfu practica impementation. This note represents an advance towards that goa. In the next section we investigate the impact of the oss of oca stabiity in RCP. 3 Loca bifurcation anaysis In the choice of parameters for any congestion contro proposa, a key aim is to ensure that they ead to a stabe equiibrium. Most proposas for congestion contro, for exampe see [28], ead to the anaysis of noninear time deayed dynamica systems. For such noninear systems, typicay sufficient conditions for oca stabiity guide parameter choices. Foowing oca stabiity, a oca bifurcation theoretic anaysis can make us comfortabe in running the system cose to the edge of stabiity. We first recapituate a resut about the oss of oca stabiity in a noninear retarded functiona differentia equation, and then use it to anayse the RCP fuid mode. Foowing the anaysis in [25] we outine a oca Hopf bifurcation resut for the foowing noninear deay equation ut) = η ξ x ut τ) ± ξ xy ut)ut τ) ), 11) where η, τ, ξ x, ξ xy > 0. The parameter η has been intentionay introduced to just tip the above equation over the edge of oca) stabiity. This exogenous, non-dimensiona, parameter wi act as our bifurcation parameter. We now state the foowing Theorem about equation 11): 10

11 Theorem 3.1. A necessary and sufficient condition for oca stabiity is ηξ x τ < π/2, and treating η as the bifurcation parameter, the first oca Hopf bifurcation occurs with period 4τ at η = η c, where η c ξ x τ = π/2. Further, as the Hopf condition is just vioated, the equation wi aways undergo a super-critica Hopf bifurcation where the ampitude of the stabe bifurcating soutions wi be proportiona to ξ x 20πη ηc ). 3π 2 ξ xy 3.1 RCP: singe ink, singe deay mode In our bifurcation anaysis, for the sake of simpicity, we sha eave the term that represents feedback based on queue size out of the mode by taking β = 0. Consider the foowing singe ink, singe deay, RCP fuid mode α ) Ṙt) = ηrt) Cτ C yt)), 12) where yt) = s Rt τ) is the aggregate oad at the ink and η is the nondimensiona bifurcation parameter. Define ut) = Rt) R, and take a Tayor expansion of 12) to obtain ut) = η α τ α ut τ) η ut)ut τ). τr Theorem 3.1 eads us to the foowing concusions about equation 12). A necessary and sufficient condition for oca stabiity is ηα < π/2, and treating η as the bifurcation parameter, the first oca Hopf bifurcation occurs with period 4τ at η = η c, where η c α = π/2. If the Hopf condition is just vioated, the equation wi aways undergo a super-critica Hopf bifurcation where the ampitude of the stabe bifurcating soutions wi be proportiona to 20πη ηc ) R. 3π 2 We emphasise that equation 12) cannot produce a sub-critica Hopf bifurcation; but, in the bifurcation anaysis, we have omitted any additiona noninear effects that woud arise if β > 0 in the RCP mode. In the next section, we summarise our contributions in this note and outine some avenues for further research. 11

12 4 Contributions and avenues for further research As RCP aims for max-min fairness, it is appropriate to first anayze a mode that embodies the origina formuation [7], as we did in this note. We first proved that the RCP fuid mode is gobay stabe in the absence of propagation deays. Then, we derived a set of conditions for oca stabiity when arbitrary heterogeneous propagation deays are present. The network deay stabiity resut depends on the weak assumption that, at equiibrium, there is ony one botteneck ink aong each route. An interesting avenue for research woud be to show if this resut may hod in genera, or to provide a counter-argument to disprove it. The stabiity conditions have the attractive property of being fairy decentraized, and so RCP is potentiay suitabe for depoyment in very arge networks. Observe that the sma buffer regime has enabed us to tacke the question of stabiity for arge networks with arbitrary topoogies. Lasty, in a singe ink, singe deay mode, we investigate the impact of the oss of oca stabiity in a specia case β = 0) where we show that the RCP agorithm woud aways give rise to an innocuous-ooking super-critica Hopf bifurcation. We now outine some natura avenues for research. Choice of parameters. We ony considered the imiting case of very sma buffers. The choice of β impacts the rate at which the queue is drained. Sma vaues of β drain the queue sowy; so with sma β and for arge enough buffers it is appropriate to mode the queue as a saturated integrator. With arge vaues of β, the queue may drain fast enough to necessitate the deveopment of finer queueing theoretic modes. Such queuing modes woud certainy contribute to the noninearity in the RCP dynamica system, and hence affect the stabiity of the time deayed network. Further, the resuts from oca bifurcation theory coud aso be subte. In our bifurcation theoretic anaysis we set β = 0, which removes any noninearity that may arise from feedback based on queue size. In the above modeing regimes, it woud be important to understand the impact of the RCP parameters α and β) on stabiity, instabiity and rate of convergence. Fairness and stabiity. A parameterized famiy of α-fair rate aocations was introduced in [23]. The parameter α ies in the range 0, ), and the cases α 0, α = 1 and α correspond respectivey to an aocation which achieves maximum throughput, is proportionay fair or is max-min fair. In this note we ony considered a max-min fair rate aocation; as originay specified [7]. An important direction for research is to incorporate different notions of fairness into the RCP framework, and anayse some stabiity properties of such networks. Expicit congestion contro. A key question is the scae of the step-change in rate that is necessary, at a resource, to accommodate new fows. Another reevant concern is how resources may be shared between TCP users and those using a rate controed architecture. 12

13 Acknowedgements The authors are gratefu to Frank Key for comments on earier drafts. Gaurav woud aso ike to thank Tony White for some suggestions on the exposition of the text. The usua caveat appies. References [1] S.P. Abraham and A. Kumar. Max-min fair rate contro of ABR connections with nonzero MCRs. Proceedings of IEEE Gobecom, [2] S.P. Abraham and A. Kumar. A stochastic approximation approach for max-min fair adaptive rate contro of ABR sessions with MCRs. Proceedings of IEEE Infocom, [3] G. Appenzeer, I. Kesassy, and N. McKeown. Sizing router buffers. Computer Communication Review, vo. 34, no. 4, pp , [4] H. Baakrishnan, N. Dukkipati, N. McKeown, and C.J. Tomin. Stabiity anaysis of expicit congestion contro protocos. IEEE Communications Letters, vo. 11, no. 10, pp , [5] F. Banchini, R. Cigno, and R. Tempo. Robust rate contro for integrated services packet networks. IEEE Transactions on Networking, vo. 10, no. 5, pp , [6] S. Deb and R. Srikant. Rate-based versus queue-based modes of congestion contro. Proceedings of ACM Sigmetrics, [7] N. Dukkipati, M. Kobayashi, R. Zhang-Shen, and N. McKeown. Processor sharing fows in the Internet. Thirteenth Internationa Workshop on Quaity of Service, [8] N. Dukkipati and N. McKeown. Why fow-competion time is the right metric for congestion contro. Computer Communication Review, vo. 36, no. 1, pp , [9] N. Dukkipati, N. McKeown, and A.G. Fraser. RCP-AC: congestion contro to make fows compete quicky in any environment. Proceedings of IEEE Infocom, [10] M. Enachescu, Y. Ganjai, A. Goe, N. McKeown, and T. Roughgarden. Part III: routers with very sma buffers. Computer Communication Review, vo. 35, no. 3, pp , [11] C. Futon, S.Q. Li, and C.S. Lim. An ABR feedback contro scheme with tracking. Proceedings of IEEE Infocom,

14 [12] H. Han, C.V. Hoot, Y. Chait, and V. Misra. TCP networks stabiized by buffer-based AQMs. Proceedings of IEEE Infocom, [13] H. Han, C.V. Hoot, D. Towsey, and Y. Chait. Synchronization of TCP fows in networks with sma droptai buffers. Proceedings of IEEE Conference on Decision and Contro, [14] C.V. Hoot, V. Misra, D. Towsey, and W. Gong. Anaysis and design of controers for AQM routers supporting TCP fows. IEEE Transactions on Automatic Contro, vo. 47, no. 6, pp , [15] S. Kayanaraman, R. Jain, S. Fahmy, R. Goya, and B. Vandaore. The ERICA switch agorithm for ABR traffic management in ATM networks. IEEE Transactions on Networking, vo. 8, no. 1, pp , [16] D. Katabi, M. Handey, and C. Rohrs. Congestion contro for high bandwidth-deay product networks. Computer Communication Review, vo. 32, no. 4, pp , [17] S.S. Kunniyur and R. Srikant. Stabe, scaabe, fair congestion contro and AQM schemes that achieve high utiization in the Internet. IEEE Transactions on Automatic Contro, vo. 48, no. 11, pp , [18] R.J. La. Instabiity of a tandem network and its propagation under RED. IEEE Transactions on Automatic Contro, vo. 49, no. 6, pp , [19] A. Lakshmikantha, C. Beck, and R. Srikant. Robustness of rea and virtua queue-based active queue management schemes. IEEE Transactions on Networking, vo. 13, no. 1, pp , [20] A. Lakshmikantha, R. Srikant, and C. Beck. Impact of fie arrivas and departures on buffer sizing in core routers. Proceedings of IEEE Infocom, [21] S. Liu, T. Basar, and R. Srikant. Exponentia-RED: a stabiizing AQM scheme for ow- and high-speed TCP protocos. IEEE Transactions on Networking, vo. 13, no. 5, pp , [22] P. Marbach. Priority service and max-min fairness. IEEE Transactions on Networking, vo. 11, no. 5, pp , [23] J. Mo and J. Warand. Fair end-to-end window-based congestion contro. IEEE Transactions on Networking, vo. 8, no. 5, pp , [24] B. Radunovic and J.-Y. Le Boudec. Rate performance objectives of mutihop wireess networks. IEEE Transactions on Mobie Computing, vo. 3, no. 4, pp ,

15 [25] G. Raina. Loca bifurcation anaysis of some dua congestion contro agorithms. IEEE Transactions on Automatic Contro, vo. 50, no. 8, pp , [26] G. Raina, D. Towsey, and D. Wischik. Part II: contro theory for buffer sizing. Computer Communication Review, vo. 35, no. 3, pp , [27] G. Raina and D. Wischik. Buffer sizes for arge mutipexers: TCP queueing theory and instabiity anaysis. Proceedings of Next Generation Internet Networks, [28] R. Srikant. The Mathematics of Internet Congestion Contro. Birkhäuser, [29] L. Tan, X. Zhang, L. Andrew, and M. Zukerman. Price-based max-min fair rate aocation in wireess muti-hop networks. IEEE Communications Letters, vo. 10, no.1, pp , [30] G. Vinnicombe. On the stabiity of networks operating TCP-ike congestion contro. Proceedings of Internationa Federation of Automatic Contro, [31] T. Voice. A goba stabiity resut for prima-dua congestion contro agorithms with routing. Computer Communication Review, vo. 34, no. 3, pp , [32] D. Wischik and N. McKeown. Part I: buffer sizes for core routers. Computer Communication Review, vo. 35, no. 3, pp , [33] Y. Zhang, D. Leonard, and D. Loguinov. Jetmax: scaabe max-min congestion contro for high-speed heterogeneous networks. Computer Networks, vo. 52, no. 6, pp ,

Equilibrium of Heterogeneous Congestion Control Protocols

Equilibrium of Heterogeneous Congestion Control Protocols Equiibrium of Heterogeneous Congestion Contro Protocos Ao Tang Jiantao Wang Steven H. Low EAS Division, Caifornia Institute of Technoogy Mung Chiang EE Department, Princeton University Abstract When heterogeneous

More information

Recursive Constructions of Parallel FIFO and LIFO Queues with Switched Delay Lines

Recursive Constructions of Parallel FIFO and LIFO Queues with Switched Delay Lines Recursive Constructions of Parae FIFO and LIFO Queues with Switched Deay Lines Po-Kai Huang, Cheng-Shang Chang, Feow, IEEE, Jay Cheng, Member, IEEE, and Duan-Shin Lee, Senior Member, IEEE Abstract One

More information

<C 2 2. λ 2 l. λ 1 l 1 < C 1

<C 2 2. λ 2 l. λ 1 l 1 < C 1 Teecommunication Network Contro and Management (EE E694) Prof. A. A. Lazar Notes for the ecture of 7/Feb/95 by Huayan Wang (this document was ast LaT E X-ed on May 9,995) Queueing Primer for Muticass Optima

More information

STABILITY OF A PARAMETRICALLY EXCITED DAMPED INVERTED PENDULUM 1. INTRODUCTION

STABILITY OF A PARAMETRICALLY EXCITED DAMPED INVERTED PENDULUM 1. INTRODUCTION Journa of Sound and Vibration (996) 98(5), 643 65 STABILITY OF A PARAMETRICALLY EXCITED DAMPED INVERTED PENDULUM G. ERDOS AND T. SINGH Department of Mechanica and Aerospace Engineering, SUNY at Buffao,

More information

Power Control and Transmission Scheduling for Network Utility Maximization in Wireless Networks

Power Control and Transmission Scheduling for Network Utility Maximization in Wireless Networks ower Contro and Transmission Scheduing for Network Utiity Maximization in Wireess Networks Min Cao, Vivek Raghunathan, Stephen Hany, Vinod Sharma and. R. Kumar Abstract We consider a joint power contro

More information

A Brief Introduction to Markov Chains and Hidden Markov Models

A Brief Introduction to Markov Chains and Hidden Markov Models A Brief Introduction to Markov Chains and Hidden Markov Modes Aen B MacKenzie Notes for December 1, 3, &8, 2015 Discrete-Time Markov Chains You may reca that when we first introduced random processes,

More information

Lecture 6: Moderately Large Deflection Theory of Beams

Lecture 6: Moderately Large Deflection Theory of Beams Structura Mechanics 2.8 Lecture 6 Semester Yr Lecture 6: Moderatey Large Defection Theory of Beams 6.1 Genera Formuation Compare to the cassica theory of beams with infinitesima deformation, the moderatey

More information

Multilayer Kerceptron

Multilayer Kerceptron Mutiayer Kerceptron Zotán Szabó, András Lőrincz Department of Information Systems, Facuty of Informatics Eötvös Loránd University Pázmány Péter sétány 1/C H-1117, Budapest, Hungary e-mai: szzoi@csetehu,

More information

Unconditional security of differential phase shift quantum key distribution

Unconditional security of differential phase shift quantum key distribution Unconditiona security of differentia phase shift quantum key distribution Kai Wen, Yoshihisa Yamamoto Ginzton Lab and Dept of Eectrica Engineering Stanford University Basic idea of DPS-QKD Protoco. Aice

More information

CS229 Lecture notes. Andrew Ng

CS229 Lecture notes. Andrew Ng CS229 Lecture notes Andrew Ng Part IX The EM agorithm In the previous set of notes, we taked about the EM agorithm as appied to fitting a mixture of Gaussians. In this set of notes, we give a broader view

More information

Stochastic Complement Analysis of Multi-Server Threshold Queues. with Hysteresis. Abstract

Stochastic Complement Analysis of Multi-Server Threshold Queues. with Hysteresis. Abstract Stochastic Compement Anaysis of Muti-Server Threshod Queues with Hysteresis John C.S. Lui The Dept. of Computer Science & Engineering The Chinese University of Hong Kong Leana Goubchik Dept. of Computer

More information

Tracking Control of Multiple Mobile Robots

Tracking Control of Multiple Mobile Robots Proceedings of the 2001 IEEE Internationa Conference on Robotics & Automation Seou, Korea May 21-26, 2001 Tracking Contro of Mutipe Mobie Robots A Case Study of Inter-Robot Coision-Free Probem Jurachart

More information

Throughput Optimal Scheduling for Wireless Downlinks with Reconfiguration Delay

Throughput Optimal Scheduling for Wireless Downlinks with Reconfiguration Delay Throughput Optima Scheduing for Wireess Downinks with Reconfiguration Deay Vineeth Baa Sukumaran vineethbs@gmai.com Department of Avionics Indian Institute of Space Science and Technoogy. Abstract We consider

More information

Asynchronous Control for Coupled Markov Decision Systems

Asynchronous Control for Coupled Markov Decision Systems INFORMATION THEORY WORKSHOP (ITW) 22 Asynchronous Contro for Couped Marov Decision Systems Michae J. Neey University of Southern Caifornia Abstract This paper considers optima contro for a coection of

More information

Scalable Spectrum Allocation for Large Networks Based on Sparse Optimization

Scalable Spectrum Allocation for Large Networks Based on Sparse Optimization Scaabe Spectrum ocation for Large Networks ased on Sparse Optimization innan Zhuang Modem R&D Lab Samsung Semiconductor, Inc. San Diego, C Dongning Guo, Ermin Wei, and Michae L. Honig Department of Eectrica

More information

Integrating Factor Methods as Exponential Integrators

Integrating Factor Methods as Exponential Integrators Integrating Factor Methods as Exponentia Integrators Borisav V. Minchev Department of Mathematica Science, NTNU, 7491 Trondheim, Norway Borko.Minchev@ii.uib.no Abstract. Recenty a ot of effort has been

More information

Stabilized MAX-MIN Flow Control Using PID and PII 2 Controllers

Stabilized MAX-MIN Flow Control Using PID and PII 2 Controllers Stabiized MAX-MIN Fow Contro Using PID and PII Controers Jeong-woo Cho and Song Chong Department o Eectrica Engineering and Computer Science Korea Advanced Institute o Science and Technoogy Daejeon 5-71,

More information

T.C. Banwell, S. Galli. {bct, Telcordia Technologies, Inc., 445 South Street, Morristown, NJ 07960, USA

T.C. Banwell, S. Galli. {bct, Telcordia Technologies, Inc., 445 South Street, Morristown, NJ 07960, USA ON THE SYMMETRY OF THE POWER INE CHANNE T.C. Banwe, S. Gai {bct, sgai}@research.tecordia.com Tecordia Technoogies, Inc., 445 South Street, Morristown, NJ 07960, USA Abstract The indoor power ine network

More information

TCP and Network Coding: Equilibrium and Dynamic Properties

TCP and Network Coding: Equilibrium and Dynamic Properties TCP and Network Coding: Equiibrium and Dynamic Properties Hamet Medina Ruiz, Miche Kieffer, Beatrice Pesquet-Popescu To cite this version: Hamet Medina Ruiz, Miche Kieffer, Beatrice Pesquet-Popescu. TCP

More information

Introduction to Simulation - Lecture 13. Convergence of Multistep Methods. Jacob White. Thanks to Deepak Ramaswamy, Michal Rewienski, and Karen Veroy

Introduction to Simulation - Lecture 13. Convergence of Multistep Methods. Jacob White. Thanks to Deepak Ramaswamy, Michal Rewienski, and Karen Veroy Introduction to Simuation - Lecture 13 Convergence of Mutistep Methods Jacob White Thans to Deepa Ramaswamy, Micha Rewiensi, and Karen Veroy Outine Sma Timestep issues for Mutistep Methods Loca truncation

More information

Duality Model of TCP/AQM

Duality Model of TCP/AQM Optimization and Contro of Network Duaity Mode of TCP/AQM Lijun Chen 02/08/2016 Agenda Utiity maimization and dua decompoition An introduction to TCP congetion contro A genera dynamica mode of TCP/AQM

More information

The basic equation for the production of turbulent kinetic energy in clouds is. dz + g w

The basic equation for the production of turbulent kinetic energy in clouds is. dz + g w Turbuence in couds The basic equation for the production of turbuent kinetic energy in couds is de TKE dt = u 0 w 0 du v 0 w 0 dv + g w q 0 q 0 e The first two terms on the RHS are associated with shear

More information

MARKOV CHAINS AND MARKOV DECISION THEORY. Contents

MARKOV CHAINS AND MARKOV DECISION THEORY. Contents MARKOV CHAINS AND MARKOV DECISION THEORY ARINDRIMA DATTA Abstract. In this paper, we begin with a forma introduction to probabiity and expain the concept of random variabes and stochastic processes. After

More information

arxiv:quant-ph/ v3 6 Jan 1995

arxiv:quant-ph/ v3 6 Jan 1995 arxiv:quant-ph/9501001v3 6 Jan 1995 Critique of proposed imit to space time measurement, based on Wigner s cocks and mirrors L. Diósi and B. Lukács KFKI Research Institute for Partice and Nucear Physics

More information

Uniprocessor Feasibility of Sporadic Tasks with Constrained Deadlines is Strongly conp-complete

Uniprocessor Feasibility of Sporadic Tasks with Constrained Deadlines is Strongly conp-complete Uniprocessor Feasibiity of Sporadic Tasks with Constrained Deadines is Strongy conp-compete Pontus Ekberg and Wang Yi Uppsaa University, Sweden Emai: {pontus.ekberg yi}@it.uu.se Abstract Deciding the feasibiity

More information

Delay Analysis of Maximum Weight Scheduling in Wireless Ad Hoc Networks

Delay Analysis of Maximum Weight Scheduling in Wireless Ad Hoc Networks 1 Deay Anaysis o Maximum Weight Scheduing in Wireess Ad Hoc Networks ong Bao e, Krishna Jagannathan, and Eytan Modiano Abstract This paper studies deay properties o the weknown maximum weight scheduing

More information

Introduction to Simulation - Lecture 14. Multistep Methods II. Jacob White. Thanks to Deepak Ramaswamy, Michal Rewienski, and Karen Veroy

Introduction to Simulation - Lecture 14. Multistep Methods II. Jacob White. Thanks to Deepak Ramaswamy, Michal Rewienski, and Karen Veroy Introduction to Simuation - Lecture 14 Mutistep Methods II Jacob White Thans to Deepa Ramaswamy, Micha Rewiensi, and Karen Veroy Outine Sma Timestep issues for Mutistep Methods Reminder about LTE minimization

More information

A Survey on Delay-Aware Resource Control. for Wireless Systems Large Deviation Theory, Stochastic Lyapunov Drift and Distributed Stochastic Learning

A Survey on Delay-Aware Resource Control. for Wireless Systems Large Deviation Theory, Stochastic Lyapunov Drift and Distributed Stochastic Learning A Survey on Deay-Aware Resource Contro 1 for Wireess Systems Large Deviation Theory, Stochastic Lyapunov Drift and Distributed Stochastic Learning arxiv:1110.4535v1 [cs.pf] 20 Oct 2011 Ying Cui Vincent

More information

Coupling of LWR and phase transition models at boundary

Coupling of LWR and phase transition models at boundary Couping of LW and phase transition modes at boundary Mauro Garaveo Dipartimento di Matematica e Appicazioni, Università di Miano Bicocca, via. Cozzi 53, 20125 Miano Itay. Benedetto Piccoi Department of

More information

Symbolic models for nonlinear control systems using approximate bisimulation

Symbolic models for nonlinear control systems using approximate bisimulation Symboic modes for noninear contro systems using approximate bisimuation Giordano Poa, Antoine Girard and Pauo Tabuada Abstract Contro systems are usuay modeed by differentia equations describing how physica

More information

Methods for Ordinary Differential Equations. Jacob White

Methods for Ordinary Differential Equations. Jacob White Introduction to Simuation - Lecture 12 for Ordinary Differentia Equations Jacob White Thanks to Deepak Ramaswamy, Jaime Peraire, Micha Rewienski, and Karen Veroy Outine Initia Vaue probem exampes Signa

More information

12.2. Maxima and Minima. Introduction. Prerequisites. Learning Outcomes

12.2. Maxima and Minima. Introduction. Prerequisites. Learning Outcomes Maima and Minima 1. Introduction In this Section we anayse curves in the oca neighbourhood of a stationary point and, from this anaysis, deduce necessary conditions satisfied by oca maima and oca minima.

More information

Stochastic Automata Networks (SAN) - Modelling. and Evaluation. Paulo Fernandes 1. Brigitte Plateau 2. May 29, 1997

Stochastic Automata Networks (SAN) - Modelling. and Evaluation. Paulo Fernandes 1. Brigitte Plateau 2. May 29, 1997 Stochastic utomata etworks (S) - Modeing and Evauation Pauo Fernandes rigitte Pateau 2 May 29, 997 Institut ationa Poytechnique de Grenobe { IPG Ecoe ationae Superieure d'informatique et de Mathematiques

More information

Exploring the Throughput Boundaries of Randomized Schedulers in Wireless Networks

Exploring the Throughput Boundaries of Randomized Schedulers in Wireless Networks Exporing the Throughput Boundaries of Randomized Scheduers in Wireess Networks Bin Li and Atia Eryimaz Abstract Randomization is a powerfu and pervasive strategy for deveoping efficient and practica transmission

More information

Joint Congestion Control and Routing Optimization: An Efficient Second-Order Distributed Approach

Joint Congestion Control and Routing Optimization: An Efficient Second-Order Distributed Approach IEEE/ACM TRANSACTIONS ON NETWORKING Joint Congestion Contro and Routing Optimization: An Efficient Second-Order Distributed Approach Jia Liu, Member, IEEE, Ness B. Shroff, Feow, IEEE, Cathy H. Xia, and

More information

A Novel Learning Method for Elman Neural Network Using Local Search

A Novel Learning Method for Elman Neural Network Using Local Search Neura Information Processing Letters and Reviews Vo. 11, No. 8, August 2007 LETTER A Nove Learning Method for Eman Neura Networ Using Loca Search Facuty of Engineering, Toyama University, Gofuu 3190 Toyama

More information

King Fahd University of Petroleum & Minerals

King Fahd University of Petroleum & Minerals King Fahd University of Petroeum & Mineras DEPARTMENT OF MATHEMATICAL SCIENCES Technica Report Series TR 369 December 6 Genera decay of soutions of a viscoeastic equation Saim A. Messaoudi DHAHRAN 3161

More information

THE REACHABILITY CONES OF ESSENTIALLY NONNEGATIVE MATRICES

THE REACHABILITY CONES OF ESSENTIALLY NONNEGATIVE MATRICES THE REACHABILITY CONES OF ESSENTIALLY NONNEGATIVE MATRICES by Michae Neumann Department of Mathematics, University of Connecticut, Storrs, CT 06269 3009 and Ronad J. Stern Department of Mathematics, Concordia

More information

Coalitions in Routing Games: A Worst-Case Perspective

Coalitions in Routing Games: A Worst-Case Perspective Coaitions in Routing Games: A Worst-Case Perspective Gideon Bocq and Arie Orda, Feow, IEEE arxiv:30.3487v3 [cs.ni] 27 Mar 206 Abstract We investigate a routing game that aows for the creation of coaitions,

More information

IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 15, NO. 2, FEBRUARY

IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 15, NO. 2, FEBRUARY IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 5, NO. 2, FEBRUARY 206 857 Optima Energy and Data Routing in Networks With Energy Cooperation Berk Gurakan, Student Member, IEEE, OmurOze,Member, IEEE,

More information

Distributed average consensus: Beyond the realm of linearity

Distributed average consensus: Beyond the realm of linearity Distributed average consensus: Beyond the ream of inearity Usman A. Khan, Soummya Kar, and José M. F. Moura Department of Eectrica and Computer Engineering Carnegie Meon University 5 Forbes Ave, Pittsburgh,

More information

Schedulability Analysis of Deferrable Scheduling Algorithms for Maintaining Real-Time Data Freshness

Schedulability Analysis of Deferrable Scheduling Algorithms for Maintaining Real-Time Data Freshness 1 Scheduabiity Anaysis of Deferrabe Scheduing Agorithms for Maintaining Rea- Data Freshness Song Han, Deji Chen, Ming Xiong, Kam-yiu Lam, Aoysius K. Mok, Krithi Ramamritham UT Austin, Emerson Process Management,

More information

Lecture Note 3: Stationary Iterative Methods

Lecture Note 3: Stationary Iterative Methods MATH 5330: Computationa Methods of Linear Agebra Lecture Note 3: Stationary Iterative Methods Xianyi Zeng Department of Mathematica Sciences, UTEP Stationary Iterative Methods The Gaussian eimination (or

More information

SydU STAT3014 (2015) Second semester Dr. J. Chan 18

SydU STAT3014 (2015) Second semester Dr. J. Chan 18 STAT3014/3914 Appied Stat.-Samping C-Stratified rand. sampe Stratified Random Samping.1 Introduction Description The popuation of size N is divided into mutuay excusive and exhaustive subpopuations caed

More information

Atomic Hierarchical Routing Games in Communication Networks

Atomic Hierarchical Routing Games in Communication Networks Atomic Hierarchica Routing Games in Communication etworks Vijay Kambe, Eitan Atman, Rachid E-Azouzi Vinod Sharma Dept. of Industria Engineering Management, IIT - Kharagpur, West Benga, India Maestro group,

More information

A unified framework for design and analysis of networked and quantized control systems

A unified framework for design and analysis of networked and quantized control systems 1 A unified framework for design and anaysis of networked and quantized contro systems Dragan ešić and Danie Liberzon Abstract We generaize and unify a range of recent resuts in quantized contro systems

More information

Formulas for Angular-Momentum Barrier Factors Version II

Formulas for Angular-Momentum Barrier Factors Version II BNL PREPRINT BNL-QGS-06-101 brfactor1.tex Formuas for Anguar-Momentum Barrier Factors Version II S. U. Chung Physics Department, Brookhaven Nationa Laboratory, Upton, NY 11973 March 19, 2015 abstract A

More information

A Statistical Framework for Real-time Event Detection in Power Systems

A Statistical Framework for Real-time Event Detection in Power Systems 1 A Statistica Framework for Rea-time Event Detection in Power Systems Noan Uhrich, Tim Christman, Phiip Swisher, and Xichen Jiang Abstract A quickest change detection (QCD) agorithm is appied to the probem

More information

arxiv:hep-ph/ v1 15 Jan 2001

arxiv:hep-ph/ v1 15 Jan 2001 BOSE-EINSTEIN CORRELATIONS IN CASCADE PROCESSES AND NON-EXTENSIVE STATISTICS O.V.UTYUZH AND G.WILK The Andrzej So tan Institute for Nucear Studies; Hoża 69; 00-689 Warsaw, Poand E-mai: utyuzh@fuw.edu.p

More information

First-Order Corrections to Gutzwiller s Trace Formula for Systems with Discrete Symmetries

First-Order Corrections to Gutzwiller s Trace Formula for Systems with Discrete Symmetries c 26 Noninear Phenomena in Compex Systems First-Order Corrections to Gutzwier s Trace Formua for Systems with Discrete Symmetries Hoger Cartarius, Jörg Main, and Günter Wunner Institut für Theoretische

More information

A Model for Stochastic Hybrid Systems with Application to Communication Networks

A Model for Stochastic Hybrid Systems with Application to Communication Networks A Mode for Stochastic Hybrid Systems with Appication to Communication Networks João P. Hespanha Dept. of Eectrica and Computer Eng., Univ. of Caifornia, Santa Barbara, CA 9216-956 hespanha@ece.ucsb.edu

More information

Module 22: Simple Harmonic Oscillation and Torque

Module 22: Simple Harmonic Oscillation and Torque Modue : Simpe Harmonic Osciation and Torque.1 Introduction We have aready used Newton s Second Law or Conservation of Energy to anayze systems ike the boc-spring system that osciate. We sha now use torque

More information

More Scattering: the Partial Wave Expansion

More Scattering: the Partial Wave Expansion More Scattering: the Partia Wave Expansion Michae Fower /7/8 Pane Waves and Partia Waves We are considering the soution to Schrödinger s equation for scattering of an incoming pane wave in the z-direction

More information

Analysis of Emerson s Multiple Model Interpolation Estimation Algorithms: The MIMO Case

Analysis of Emerson s Multiple Model Interpolation Estimation Algorithms: The MIMO Case Technica Report PC-04-00 Anaysis of Emerson s Mutipe Mode Interpoation Estimation Agorithms: The MIMO Case João P. Hespanha Dae E. Seborg University of Caifornia, Santa Barbara February 0, 004 Anaysis

More information

NOISE-INDUCED STABILIZATION OF STOCHASTIC DIFFERENTIAL EQUATIONS

NOISE-INDUCED STABILIZATION OF STOCHASTIC DIFFERENTIAL EQUATIONS NOISE-INDUCED STABILIZATION OF STOCHASTIC DIFFERENTIAL EQUATIONS TONY ALLEN, EMILY GEBHARDT, AND ADAM KLUBALL 3 ADVISOR: DR. TIFFANY KOLBA 4 Abstract. The phenomenon of noise-induced stabiization occurs

More information

Centralized Coded Caching of Correlated Contents

Centralized Coded Caching of Correlated Contents Centraized Coded Caching of Correated Contents Qianqian Yang and Deniz Gündüz Information Processing and Communications Lab Department of Eectrica and Eectronic Engineering Imperia Coege London arxiv:1711.03798v1

More information

8 Digifl'.11 Cth:uits and devices

8 Digifl'.11 Cth:uits and devices 8 Digif'. Cth:uits and devices 8. Introduction In anaog eectronics, votage is a continuous variabe. This is usefu because most physica quantities we encounter are continuous: sound eves, ight intensity,

More information

Cryptanalysis of PKP: A New Approach

Cryptanalysis of PKP: A New Approach Cryptanaysis of PKP: A New Approach Éiane Jaumes and Antoine Joux DCSSI 18, rue du Dr. Zamenhoff F-92131 Issy-es-Mx Cedex France eiane.jaumes@wanadoo.fr Antoine.Joux@ens.fr Abstract. Quite recenty, in

More information

$, (2.1) n="# #. (2.2)

$, (2.1) n=# #. (2.2) Chapter. Eectrostatic II Notes: Most of the materia presented in this chapter is taken from Jackson, Chap.,, and 4, and Di Bartoo, Chap... Mathematica Considerations.. The Fourier series and the Fourier

More information

Some Measures for Asymmetry of Distributions

Some Measures for Asymmetry of Distributions Some Measures for Asymmetry of Distributions Georgi N. Boshnakov First version: 31 January 2006 Research Report No. 5, 2006, Probabiity and Statistics Group Schoo of Mathematics, The University of Manchester

More information

Physics 127c: Statistical Mechanics. Fermi Liquid Theory: Collective Modes. Boltzmann Equation. The quasiparticle energy including interactions

Physics 127c: Statistical Mechanics. Fermi Liquid Theory: Collective Modes. Boltzmann Equation. The quasiparticle energy including interactions Physics 27c: Statistica Mechanics Fermi Liquid Theory: Coective Modes Botzmann Equation The quasipartice energy incuding interactions ε p,σ = ε p + f(p, p ; σ, σ )δn p,σ, () p,σ with ε p ε F + v F (p p

More information

General Certificate of Education Advanced Level Examination June 2010

General Certificate of Education Advanced Level Examination June 2010 Genera Certificate of Education Advanced Leve Examination June 2010 Human Bioogy HBI6T/Q10/task Unit 6T A2 Investigative Skis Assignment Task Sheet The effect of using one or two eyes on the perception

More information

Arbitrary Throughput Versus Complexity Tradeoffs in Wireless Networks using Graph Partitioning

Arbitrary Throughput Versus Complexity Tradeoffs in Wireless Networks using Graph Partitioning University of Pennsyvania SchoaryCommons Departmenta Papers (ESE) Department of Eectrica & Systems Engineering November 2006 Arbitrary Throughput Versus Compexity Tradeoffs in Wireess Networks using Graph

More information

Accelerated Dual Descent for Constrained Convex Network Flow Optimization

Accelerated Dual Descent for Constrained Convex Network Flow Optimization 52nd IEEE Conference on Decision and Contro December 10-13, 2013. Forence, Itay Acceerated Dua Descent for Constrained Convex Networ Fow Optimization Michae Zargham, Aejandro Ribeiro, Ai Jadbabaie Abstract

More information

Smoothness equivalence properties of univariate subdivision schemes and their projection analogues

Smoothness equivalence properties of univariate subdivision schemes and their projection analogues Numerische Mathematik manuscript No. (wi be inserted by the editor) Smoothness equivaence properties of univariate subdivision schemes and their projection anaogues Phiipp Grohs TU Graz Institute of Geometry

More information

Steepest Descent Adaptation of Min-Max Fuzzy If-Then Rules 1

Steepest Descent Adaptation of Min-Max Fuzzy If-Then Rules 1 Steepest Descent Adaptation of Min-Max Fuzzy If-Then Rues 1 R.J. Marks II, S. Oh, P. Arabshahi Λ, T.P. Caude, J.J. Choi, B.G. Song Λ Λ Dept. of Eectrica Engineering Boeing Computer Services University

More information

Convergence Property of the Iri-Imai Algorithm for Some Smooth Convex Programming Problems

Convergence Property of the Iri-Imai Algorithm for Some Smooth Convex Programming Problems Convergence Property of the Iri-Imai Agorithm for Some Smooth Convex Programming Probems S. Zhang Communicated by Z.Q. Luo Assistant Professor, Department of Econometrics, University of Groningen, Groningen,

More information

Separation of Variables and a Spherical Shell with Surface Charge

Separation of Variables and a Spherical Shell with Surface Charge Separation of Variabes and a Spherica She with Surface Charge In cass we worked out the eectrostatic potentia due to a spherica she of radius R with a surface charge density σθ = σ cos θ. This cacuation

More information

Copyright information to be inserted by the Publishers. Unsplitting BGK-type Schemes for the Shallow. Water Equations KUN XU

Copyright information to be inserted by the Publishers. Unsplitting BGK-type Schemes for the Shallow. Water Equations KUN XU Copyright information to be inserted by the Pubishers Unspitting BGK-type Schemes for the Shaow Water Equations KUN XU Mathematics Department, Hong Kong University of Science and Technoogy, Cear Water

More information

Schedulability Analysis of Deferrable Scheduling Algorithms for Maintaining Real-Time Data Freshness

Schedulability Analysis of Deferrable Scheduling Algorithms for Maintaining Real-Time Data Freshness 1 Scheduabiity Anaysis of Deferrabe Scheduing Agorithms for Maintaining Rea-Time Data Freshness Song Han, Deji Chen, Ming Xiong, Kam-yiu Lam, Aoysius K. Mok, Krithi Ramamritham UT Austin, Emerson Process

More information

A Fundamental Storage-Communication Tradeoff in Distributed Computing with Straggling Nodes

A Fundamental Storage-Communication Tradeoff in Distributed Computing with Straggling Nodes A Fundamenta Storage-Communication Tradeoff in Distributed Computing with Stragging odes ifa Yan, Michèe Wigger LTCI, Téécom ParisTech 75013 Paris, France Emai: {qifa.yan, michee.wigger} @teecom-paristech.fr

More information

arxiv: v2 [cond-mat.stat-mech] 14 Nov 2008

arxiv: v2 [cond-mat.stat-mech] 14 Nov 2008 Random Booean Networks Barbara Drosse Institute of Condensed Matter Physics, Darmstadt University of Technoogy, Hochschustraße 6, 64289 Darmstadt, Germany (Dated: June 27) arxiv:76.335v2 [cond-mat.stat-mech]

More information

An Adaptive Opportunistic Routing Scheme for Wireless Ad-hoc Networks

An Adaptive Opportunistic Routing Scheme for Wireless Ad-hoc Networks An Adaptive Opportunistic Routing Scheme for Wireess Ad-hoc Networks A.A. Bhorkar, M. Naghshvar, T. Javidi, and B.D. Rao Department of Eectrica Engineering, University of Caifornia San Diego, CA, 9093

More information

Conditions for Saddle-Point Equilibria in Output-Feedback MPC with MHE

Conditions for Saddle-Point Equilibria in Output-Feedback MPC with MHE Conditions for Sadde-Point Equiibria in Output-Feedback MPC with MHE David A. Copp and João P. Hespanha Abstract A new method for soving output-feedback mode predictive contro (MPC) and moving horizon

More information

Random maps and attractors in random Boolean networks

Random maps and attractors in random Boolean networks LU TP 04-43 Rom maps attractors in rom Booean networks Björn Samuesson Car Troein Compex Systems Division, Department of Theoretica Physics Lund University, Sövegatan 4A, S-3 6 Lund, Sweden Dated: 005-05-07)

More information

Available online at ScienceDirect. IFAC PapersOnLine 50-1 (2017)

Available online at   ScienceDirect. IFAC PapersOnLine 50-1 (2017) Avaiabe onine at www.sciencedirect.com ScienceDirect IFAC PapersOnLine 50-1 (2017 3412 3417 Stabiization of discrete-time switched inear systems: Lyapunov-Metzer inequaities versus S-procedure characterizations

More information

Competitive Diffusion in Social Networks: Quality or Seeding?

Competitive Diffusion in Social Networks: Quality or Seeding? Competitive Diffusion in Socia Networks: Quaity or Seeding? Arastoo Fazei Amir Ajorou Ai Jadbabaie arxiv:1503.01220v1 [cs.gt] 4 Mar 2015 Abstract In this paper, we study a strategic mode of marketing and

More information

14 Separation of Variables Method

14 Separation of Variables Method 14 Separation of Variabes Method Consider, for exampe, the Dirichet probem u t = Du xx < x u(x, ) = f(x) < x < u(, t) = = u(, t) t > Let u(x, t) = T (t)φ(x); now substitute into the equation: dt

More information

Research Article Further Stability Criterion on Delayed Recurrent Neural Networks Based on Reciprocal Convex Technique

Research Article Further Stability Criterion on Delayed Recurrent Neural Networks Based on Reciprocal Convex Technique Mathematica Probems in Engineering Voume 2012, Artice ID 829037, 14 pages doi:10.1155/2012/829037 Research Artice Further Stabiity Criterion on Deayed Recurrent Neura Networks Based on Reciproca Convex

More information

Explicit overall risk minimization transductive bound

Explicit overall risk minimization transductive bound 1 Expicit overa risk minimization transductive bound Sergio Decherchi, Paoo Gastado, Sandro Ridea, Rodofo Zunino Dept. of Biophysica and Eectronic Engineering (DIBE), Genoa University Via Opera Pia 11a,

More information

MATH 172: MOTIVATION FOR FOURIER SERIES: SEPARATION OF VARIABLES

MATH 172: MOTIVATION FOR FOURIER SERIES: SEPARATION OF VARIABLES MATH 172: MOTIVATION FOR FOURIER SERIES: SEPARATION OF VARIABLES Separation of variabes is a method to sove certain PDEs which have a warped product structure. First, on R n, a inear PDE of order m is

More information

C. Fourier Sine Series Overview

C. Fourier Sine Series Overview 12 PHILIP D. LOEWEN C. Fourier Sine Series Overview Let some constant > be given. The symboic form of the FSS Eigenvaue probem combines an ordinary differentia equation (ODE) on the interva (, ) with a

More information

Reliability: Theory & Applications No.3, September 2006

Reliability: Theory & Applications No.3, September 2006 REDUNDANCY AND RENEWAL OF SERVERS IN OPENED QUEUING NETWORKS G. Sh. Tsitsiashvii M.A. Osipova Vadivosto, Russia 1 An opened queuing networ with a redundancy and a renewa of servers is considered. To cacuate

More information

Discrete Techniques. Chapter Introduction

Discrete Techniques. Chapter Introduction Chapter 3 Discrete Techniques 3. Introduction In the previous two chapters we introduced Fourier transforms of continuous functions of the periodic and non-periodic (finite energy) type, we as various

More information

Discrete Techniques. Chapter Introduction

Discrete Techniques. Chapter Introduction Chapter 3 Discrete Techniques 3. Introduction In the previous two chapters we introduced Fourier transforms of continuous functions of the periodic and non-periodic (finite energy) type, as we as various

More information

Effect of transport ratio on source term in determination of surface emission coefficient

Effect of transport ratio on source term in determination of surface emission coefficient Internationa Journa of heoretica & Appied Sciences, (): 74-78(9) ISSN : 975-78 Effect of transport ratio on source term in determination of surface emission coefficient Sanjeev Kumar and Apna Mishra epartment

More information

Linear Network Coding for Multiple Groupcast Sessions: An Interference Alignment Approach

Linear Network Coding for Multiple Groupcast Sessions: An Interference Alignment Approach Linear Network Coding for Mutipe Groupcast Sessions: An Interference Aignment Approach Abhik Kumar Das, Siddhartha Banerjee and Sriram Vishwanath Dept. of ECE, The University of Texas at Austin, TX, USA

More information

THINKING IN PYRAMIDS

THINKING IN PYRAMIDS ECS 178 Course Notes THINKING IN PYRAMIDS Kenneth I. Joy Institute for Data Anaysis and Visuaization Department of Computer Science University of Caifornia, Davis Overview It is frequenty usefu to think

More information

Analytical Mean-Field Approach to the Phase Diagram of Ultracold Bosons in Optical Superlattices

Analytical Mean-Field Approach to the Phase Diagram of Ultracold Bosons in Optical Superlattices Laser Physics, Vo. 5, No. 2, 25, pp. 36 365. Origina Text Copyright 25 by Astro, Ltd. Copyright 25 by MAIK Nauka /Interperiodica (Russia). PHYSICS OF COLD TRAPPED ATOMS Anaytica Mean-Fied Approach to the

More information

Best-effort networks: modeling and performance analysis via large networks asymptotics

Best-effort networks: modeling and performance analysis via large networks asymptotics IEEE INFOCOM 200 Best-effort networs: modeing and performance anaysis via arge networs asymptotics Guy Fayoe Arnaud de La Fortee Jean-Marc Lasgouttes Laurent Massouié James Roberts Abstract In this paper

More information

DISTRIBUTION OF TEMPERATURE IN A SPATIALLY ONE- DIMENSIONAL OBJECT AS A RESULT OF THE ACTIVE POINT SOURCE

DISTRIBUTION OF TEMPERATURE IN A SPATIALLY ONE- DIMENSIONAL OBJECT AS A RESULT OF THE ACTIVE POINT SOURCE DISTRIBUTION OF TEMPERATURE IN A SPATIALLY ONE- DIMENSIONAL OBJECT AS A RESULT OF THE ACTIVE POINT SOURCE Yury Iyushin and Anton Mokeev Saint-Petersburg Mining University, Vasiievsky Isand, 1 st ine, Saint-Petersburg,

More information

Appendix of the Paper The Role of No-Arbitrage on Forecasting: Lessons from a Parametric Term Structure Model

Appendix of the Paper The Role of No-Arbitrage on Forecasting: Lessons from a Parametric Term Structure Model Appendix of the Paper The Roe of No-Arbitrage on Forecasting: Lessons from a Parametric Term Structure Mode Caio Ameida cameida@fgv.br José Vicente jose.vaentim@bcb.gov.br June 008 1 Introduction In this

More information

4 Separation of Variables

4 Separation of Variables 4 Separation of Variabes In this chapter we describe a cassica technique for constructing forma soutions to inear boundary vaue probems. The soution of three cassica (paraboic, hyperboic and eiptic) PDE

More information

Multigrid Method for Elliptic Control Problems

Multigrid Method for Elliptic Control Problems J OHANNES KEPLER UNIVERSITÄT LINZ Netzwerk f ür Forschung, L ehre und Praxis Mutigrid Method for Eiptic Contro Probems MASTERARBEIT zur Erangung des akademischen Grades MASTER OF SCIENCE in der Studienrichtung

More information

c 2016 Georgios Rovatsos

c 2016 Georgios Rovatsos c 2016 Georgios Rovatsos QUICKEST CHANGE DETECTION WITH APPLICATIONS TO LINE OUTAGE DETECTION BY GEORGIOS ROVATSOS THESIS Submitted in partia fufiment of the requirements for the degree of Master of Science

More information

Statistical Learning Theory: A Primer

Statistical Learning Theory: A Primer Internationa Journa of Computer Vision 38(), 9 3, 2000 c 2000 uwer Academic Pubishers. Manufactured in The Netherands. Statistica Learning Theory: A Primer THEODOROS EVGENIOU, MASSIMILIANO PONTIL AND TOMASO

More information

Numerical solution of one dimensional contaminant transport equation with variable coefficient (temporal) by using Haar wavelet

Numerical solution of one dimensional contaminant transport equation with variable coefficient (temporal) by using Haar wavelet Goba Journa of Pure and Appied Mathematics. ISSN 973-1768 Voume 1, Number (16), pp. 183-19 Research India Pubications http://www.ripubication.com Numerica soution of one dimensiona contaminant transport

More information

General Certificate of Education Advanced Level Examination June 2010

General Certificate of Education Advanced Level Examination June 2010 Genera Certificate of Education Advanced Leve Examination June 2010 Human Bioogy HBI6T/P10/task Unit 6T A2 Investigative Skis Assignment Task Sheet The effect of temperature on the rate of photosynthesis

More information

17 Lecture 17: Recombination and Dark Matter Production

17 Lecture 17: Recombination and Dark Matter Production PYS 652: Astrophysics 88 17 Lecture 17: Recombination and Dark Matter Production New ideas pass through three periods: It can t be done. It probaby can be done, but it s not worth doing. I knew it was

More information

II. PROBLEM. A. Description. For the space of audio signals

II. PROBLEM. A. Description. For the space of audio signals CS229 - Fina Report Speech Recording based Language Recognition (Natura Language) Leopod Cambier - cambier; Matan Leibovich - matane; Cindy Orozco Bohorquez - orozcocc ABSTRACT We construct a rea time

More information