On the Relationship Between Packet Size and Router Performance for Heavy-Tailed Traffic 1

Size: px
Start display at page:

Download "On the Relationship Between Packet Size and Router Performance for Heavy-Tailed Traffic 1"

Transcription

1 On the Relationshi Between Packet Size and Router Performance for Heavy-Tailed Traffic 1 Imad Antonios antoniosi1@southernct.edu CS Deartment MO117 Southern Connecticut State University 501 Crescent St. New Haven, CT Lester Lisky lester@engr.uconn.edu CSE Deartment University of Connecticut 371 Fairfield Road, Unit 2155 Storrs, CT Abstract The roblem of characterizing the relationshi between acket size and network delay has received little attention in the field. Research in that area has been limited to either simulation studies or emirical observations that are detached from analytic traffic modeling. From a queueing viewoint, it is simle to show that these three variables are inter-related, which necessitates a more careful study. We resent a traffic model of a router fed by ON/OFF-tye sources with heavy-tailed burst sizes. The traffic model considered is consistent with the evidence that Web traffic is heavytailed. The analysis cases that are considered establish a quantitative characterization of the comlex relationshi among acket ayload and header sizes, traffic burstiness, and router queueing delay. 1 Introduction From the advent of acket-switched networks, it has been recognized that the choice of acket size bears some significance with resect to erformance metrics such as user resonse time, and acket delay. The literature includes many simulation studies that attemt to determine the otimal MTU size in an Ethernet or a wireless network (e.g. [Lett98]), while other studies have aroached the roblem more indirectly by looking at how acket fragmentation across a wide area network affects erformance (e.g. [Cárc91, Roma95]). Although queueing delay contributes to a significant ortion of end-to-end delay on the Internet, there have been no analytic studies that quantify the relationshi between acket size and erformance. Considering the size difference between an Ethernet acket and an ATM cell, where the first is roughly 28 times larger than the second, it seems sensible to tout 1 This work was artially suorted by the Deutsche Telekom for a roblem suggested by Prof. Eike Jessen of The Technische Universität München.

2 where otimality lies when it comes to acket size, and more generally how acket size affects router delay. In this aer, we resent a queueing model that allows us to exlore the relationshi between acket size and router delay. The queue at the router is fed by a set of ON/OFF-tye sources transmitting data that comes in bursts that are heavy-tailed in size. The traffic generated by the model exhibits the long-range deendence roerty of actual Web traffic, which is generally attributed to heavy-tailed transmissions [Crov96]. The remainder of the aer is structured as follows. In the next section, we lay out our modeling assumtions and discuss how the roblem can be framed as a queueing roblem. In Section 3, we resent our traffic model, and set u closedform solutions for acket delay in two secial cases. In Section 4, we resent some calculations of acket delay in terms of acket ayload and header size, as well as another arameter describing the degree of traffic burstiness. Section 5 rovides a conclusion. 2 Packet Size and Queueing Theory Since we are only interested in erformance from a queueing modeling oint of view, we emloy the terms acket and cell interchangeably to refer to an atomic information entity that is transmitted through a routing device modeled as a queueing system. At the host level, an alication submits a burst of data for transmission, which is then divided into ackets by the lowerlevel network rotocol levels. A host cycles between a eriod of time during which it transmits a burst (termed the ON eriod), and one where it s idle (the OFF eriod). This behavior is that of an ON/OFF rocess, which is the basis of the N-Burst model described later. We will make the assumtion that ackets are exonentially distributed in size. This modeling assumtion is reasonable since the erformance roerties of a queueing system with exonential service time are only marginally different from one with deterministic service time. Inherent in the design of a acket is a header, which is tyically a fixed number of bytes that carries network control

3 information. The size of the header is related to the comlexity of the associated network rotocol. The remaining art of the acket is the ayload, which contains the data intended for transmission by the alication. The size of the acket is thus the sum of the size of the header and the ayload. It can be easily inferred that if the header-to-ayload size ratio is too large, then the network becomes inefficient because it sends a significant amount of time rocessing header data. On the other hand, as the acket size gets larger, resonse times of interactive network-based alications might suffer as data transmission is delayed until a acket is filled with data. Although these intuitions are correct, they do not aint a comlete icture of how acket size influences erformance, and certainly do not allow for a quantitative analysis. We revert to queueing theory to reframe the roblem of acket size and erformance. Let us first consider an arrival stream where acket inter-arrival time is exonential, with rate λ. Assume that ackets are of size n. If this stream is fed into a router with exonential holding time, then we have the equivalent of an M/M/1 queue, whose associated steady-state system time, T n is described by: tsn T n =, (1) 1 ρ where t is the mean service time for a acket of size n, and s ρ = t λ n s is the router utilization. Now n consider the same model, excet in this case the ackets are half the size, n 2 acket arrivals is still a Poisson rocess but the arrival rate on a er-acket basis doubles ( λ 2λ ). If we assume that the service time of acket is roortional to its size, then the service time is halved ( ts t 2 n s n ), ensuring that the router utilization remains the same ( ρ = 2 λ t s 2 = λ t n s ). The n result of this change is that the system time, T n, decreases roortionally to the factor increase in acket size since the utilization rate remains the same. As acket size aroaches 0, this urorts that system time also aroaches 0. This conclusion suddenly changes if we attach a fixed-size

4 header to every acket, which can result in a very inefficient network; the network now sends a significant amount of time transmitting header information. Researchers have for long recognized that the effective utilization of the network is an imortant measure of its success, and have sought to determine a acket size that does not incur significant overhead. Now what haens when we instead grou ackets into larger ackets while maintaining the Poisson roerty? Without any header data, and alying the same analysis to (1), the conclusion here is that larger ackets yield worse erformance. The intuition of this finding is that delay statistics of the larger acket only account for the time of the last sub-acket that is rocessed. The tradeoff in this case is that as we increase the size of the acket, keeing the useful data to be transmitted constant, the cost associated with transmitting header data goes down. 3 The Traffic Model The above discussion offers a baseline analysis of the tradeoffs in acket and header sizes and router erformance, but its modeling assumtions are too simlistic to cature the behavior of realworld traffic. We therefore base our formulation on a queueing model that rovides for the generation of heavy-tailed traffic, the N-Burst model [Schw00b]. The arrival rocess of this model allows for the secification of sources that transmit heavy-tailed bursts and is reresented as an MMPP, or more generally a semi-markov (SM) rocess. The erformance characteristics of the modeled router can be obtained by solving the corresonding SM/M/1 queue using techniques develoed in [Neut81], which is described in detail in [Schw00b]. 3.1 Truncated Power-Tail Distribution At the heart of our traffic modeling is a Matrix Exonential reresentation of a ower-tail (PT) distribution. This is motivated by the discovery that if the ON time (or burst size) distribution is heavy-tailed then the long-range deendence of the measured traffic can be accounted for, as shown

5 in [Crov96, Fior98, Schw00b]). There is a good reason to believe that such functions are involved in network traffic. It has been shown in numerous laces that the file size distribution in many facilities, including those transmitted over the Internet, tend to have PT behavior [Fior98, Crov96]. In our modeling, we make use of a family of heavy-tailed distributions, called Truncated Power- Tail distributions (for a full treatment of TPT see [Grei99]), which are hase distributions [Lis92, Neut81], and thus can be used in analytic models of Markovian-like systems. The TPT distribution is governed by several arameters, one of which is the truncation arameter T that determines the range of the tail. In Figure 1, we lot on a logarithmic scale the reliability function of a TPT with different values of T. As the value of T increases, it can be seen that the exonential dro-off of the function occurs for larger values of the burst length. Figure 1: Reliability function of the TPT for different values of T on a logarithmic scale. 3.2 Overview of the N-Burst Model The N-Burst model [Schw99a-b, Schw00a-b] serves as a owerful descritor of the behavior of data traffic sources, and rovides the means for carrying out detailed erformance analysis using a linear algebraic aroach to queueing theory. The model mimics the behavior of N statistically identical

6 and indeendent hosts that intermittently ut data onto a telecommunications line leading to a router. Each of the hosts is an ON/OFF rocess reresented as an MMPP with a ower-tail ON time and an exonential OFF time. Each source transmits ackets at a rate of κ, contributing to the total acket generation rate of λ, which is simly N κ. During an ON eriod, the source transmits ackets at a rate of λ, the eak rate. The burst arameter, b, is defined to be the fraction of the time that a source is OFF, namely OFF ( ON + OFF). This arameter is instrumental in characterizing the degree of burstiness of a traffic source, in that as b aroaches 0 traffic resembles a Poisson stream, whereas when b aroaches the other limit of 1 traffic is resented for transmission in a very short eriod of time and the source is idle the majority of time (a bulk arrival). Keeing κ constant, as b goes from 0 to 1, we can observe that the acket arrival rate during an ON eriod grows unbounded. One of the most significant results of this work is the discovery of the conditions that cause blowus in acket delay at the router. Performance blowu turned out to be a function of the router utilization and the burst arameter, while its magnitude is deendent on the range of the ower-tail function and the tail exonent. In the model, the router is reresented as an exonential server with an average service rate of ν. With the arrival rocess being an MMPP, the system becomes an SM/M/1 queue, which can be solved analytically. The utilization ratio of the router, ρ, is simly N κ / ν. For a rocess with one ower-tailed source, or the 1-Burst model, a blowu oint is located where b = 1 ρ or equivalently where λ = ν. This indicates that it is sufficient for the eak rate of the source to exceed the router rate in order for a blowu to occur, irresective of the average rate at which a source is submitting ackets for transmission.

7 If there are 2 or more sources, the situation becomes more comlicated. An N-Burst rocess gives rise to N blowu oints, each with increasing delay severity as b increases. The value of λ for each blowu is given by ν ( N i) κ λ ( i N) : =. i Since κ = ( 1 b) λ and ρ = Nκ / ν, the b values of the blowu oints can be given as ν Nκ 1 ρ b( i N) : = =. (2) ν ( N i) κ 1 ( N i) ρ / N An examle of this behavior is given in Figure 2, which shows the mean acket delay for a 3-Burst rocess with different truncation arameters. It can be seen that as T grows the blowu in delay grows in severity (Note that the y-axis is on a logarithmic scale). Figure 2: Mean Packet Delay for a 3-Burst rocess. T = 1 corresonds to exonential ON times. The T = 9 curve shows 2 of the three blowu oints: b ( i 3) = {0.5,0.6,0.75}.

8 3.3 Parameter Sace Reformulation The N-Burst model considers the acket as the most rimitive modeling unit, which further needs to be reformulated in order to deal with a acket at a finer level of detail. We now reresent a acket as having the following two comonents: h = Packet header in bytes = Mean acket ayload in bytes This refinement necessitates the introduction of the random variable Y, which denotes the size of a burst in bytes. E(Y ) is then the exected number of bytes in a burst, or in other terms the mean file size submitted for transmission by the host. Two other indeendent arameters are introduced to deal with the distinction between useful work and overhead as follows: ν = Router service rate in bytes er second λ = Maximal (eak) source transmission rate in bytes er second Let L be the random variable denoting the number of ackets in a burst. Then the mean number of ackets in a burst can now be exressed as follows: n = E( L) = E( Y ). The rate at which ackets can be rocessed by the router is: ν ν = + h The arameters N, OFF, λ, and λ can be understood using their earlier interretation in the N- Burst model. The burst arameter as well as the router utilization can be exressed in terms of the new arameters as follows: b = 1 E( Y ) λb λ

9 ρ λ λb E( Y )( + h) = [0 ν ν = ] The above assumes that the burst arameter and the intra-burst acket rate are indeendent of h, which means that each source sends ayload data at the maximal rate λ 0]. One of the imortant characteristics of this reformulation is that the new model reduces to the N-Burst model for h = 0, which rovides a very useful mechanism to validate our calculations relative to earlier results. We can observe that when h is zero, the router utilization becomes indeendent of acket size. For other values of h, it becomes useful to further refine our notion of router utilization. First, let then we can define: δ = h, [ ρ( δ ) = ρ(0)(1 + δ ), where ρ(0) = λ b E( Y ) ν. (3) We interret ρ(0) as being the useful work utilization, which is the fraction of time sent by the router rocessing ayload information, while ρ(δ ) is the fraction of time the router is busy. It can be seen that if the header-to-ayload size ratio increases, that is, if ackets are made smaller or the header made bigger, ρ(δ ) will be bigger for the same ayload. Since ρ(δ ) should not exceed 1, δ must never be allowed to exceed 1 ρ(0) 1in our calculations. It is not immediately clear how this affects router erformance metrics over the range of values of the burst arameter as this requires intensive calculations. However, for values of b near the endoints, we can draw qualitative conclusions about the model behavior based on our understanding of the base model. When the burst arameter b is near zero, the system behaves like an M M / 1 queue as the λ / ν OFF eriods vanish and the overall inter-acket arrival rocess becomes governed by the intraburst acket arrival rocess, which is Poisson. Since the merging of several Poisson streams yields a Poisson stream, the model behavior is indeendent of the number of sources (see [Lis01] for

10 details). For the other endoint, namely when b is close to one, the system behaves like an / [ ] M ON λ M ν b /1 queue, which is a bulk arrival queue with Poisson inter-burst arrivals, where the size of each burst is described by the ON-time distribution. This model is once again indeendent of the number of sources. Our distribution of interest for the ON times is a TPT, which as shown in Figure 2 causes blowu oints for sufficiently large values of T. As the blowu oints are a function of the router utilization, we next investigate how δ affects erformance in the corresonding blowu regions. 3.4 Shift in Blowu Points In the 1-Burst rocess, blowu occurs when the acket arrival rate during an ON eriod exceeds the rate at which the router can rocess them. In terms of the reformulated arameters, a blowu occurs when λ ν + h or λ ν 1 + δ. This can be rewritten in a form that is more at for interretation, namely ν λ 1 + δ. The numerator describes the rate at which the router could rocess ackets of size in the absence of overhead. Therefore, if it is necessary to stay below the blowu oint in the resence of acket overhead, the intra-burst acket rate must be reduced by the factor 1 (1 + δ ). Of course, every acket must have a header of finite size, so an alternate comarison can be made between two different acket sizes with the same h. Suose we wish to comare erformance between a system with acket size (system 1), and one with 2 (system 2). The latter sends twice as many ackets as the former, but in the same amount of time. This leads to the two blowu oints occurring at:

11 ν λ 1 = and 1 + δ λ ν = 1+ 2δ 2. So, halving the acket size will cause a reduction in the accetable maximal data flow rate by the factor ( 1+ δ ) (1 + 2δ ) TPT30, α = 1.4, θ = 0.5, T = 30, ρ(0) = 0.25, h = 40, E(Y) = = 48 = 100 = 500 = 1000 = mcd Burst arameter b Figure 3: Mean cell delay calculations for various ayload sizes. The blowu oints can also be exressed in terms of the burst arameter. From (2), we have b ( 1 1) = 1 ρ ( δ ) = (1 + δ ). Figure 3 shows how the blowu oints shift with δ, which takes on the values, { 0.833, 0.4, 0.08, 0.04, }. This corresonds to blowu oints in the range b < The figure also reveals the comlexity in characterizing the relationshi between introducing overhead and the degree of traffic burstiness. The value of δ for which the mean Cell Delay (mcd) is a minimum is a very comlicated function of b. For instance, if we consider the lots with the lowest and highest values of, we observe that when b equals 0, the mcd for the first lot exhibits the best erformance, and when b is one, erformance is reversed. In

12 a similar fashion, one can find the blowu oints for the N-Burst rocess by using (2), relacing ρ with ρ ( 0)(1 + δ ) as follows: 1 ρ(0)(1 + δ ) b( i N) = for 1 i N. 1 ( N i) ρ(0)(1 + δ ) / N In this form, it is clear that the blowu oints deend on and h only in their ratio, δ. 3.5 Otimality for b = 0 At b = 0, the mean acket delay for the base model satisfies (1), yielding: 1 1 mcd ( b = 0) = = ν λ ν + h λ b E( Y ). This equation is indeendent of N, as long as λ b is interreted as the total burst rate from all sources. In order to find for what value of is this exression a minimum, we hold h constant and solve from which we obtain: mpd = 0, 2 ν = λ E( Y )(1 + δ ). b We know that ν = λ E( Y ) / ρ(0), which means that otimality occurs when ρ (0)(1 δ ) 2 = 1, which can also be exressed as: b 1 ρot ( δ ) =. 1 + δ + ot In other words, as δ is increased from 0, the router becomes busier according to (3), but the mean cell delay decreases until = 1 ρ(0) 1, after which oint it begins to increase. It s imortant δ ot to recognize that δ ot is only otimal for b = 0.

13 3.6 Behavior at b = 1 The analysis of acket delay in the case where b = 1is similar to that of the revious section excet that it is more comlex. As b aroaches 1, a host transmits ackets at a faster rate, becoming a bulk arrival rocess in the limit. The governing equation for mcd given a bulk arrival rocess can be exressed as follows (see [Coo81] for details): 1 ν mcd( b = 1) = D, where 1 ρ E( L( L 1)) D =. ) + 2 E( L With L being the random variable denoting the number of ackets generated in one burst, which is roortional to the ON-time distribution, we can derive the acket size distribution from the MMPP of our model. It is imortant to note that D above grows linearly with the variance of the burst size, which for TPT distributions can be unboundedly large, even if the router utilization is small (see [Anto04] for details). 4 Analysis Results We aim to cature the relationshi between the relevant arameters of our traffic model and delay at the router. From the N-Burst model, we know that router erformance is deendent on the burst arameter b and the ON-time distribution. Having augmented the N-Burst model to define a acket in terms of a ayload and a header, we are interested in revealing the erformance interlay between these last arameters and the ones from the N-Burst model. We first look at mcd as a function of ayload and acket size for a 1-Burst rocess with exonential ON time, which is shown in Figure 4. The mean number of bytes in a burst, E( Y ), is , and the burst arameter is set to 0.5. It can be observed that as the header-to-ayload ratio increases (for a given header size), mcd grows

14 until it blows u as indicated by the sikes. Aside from this blowu, the surface in the figure does not reveal any region where mcd is not accetable. When the ON-time distribution is a TPT, erformance looks radically different as shown in Figure 5. All other arameters values are the same as in Figure 4. There are three henomena that are imortant to extract from this figure. First, we see that for small header sizes when ayload size increases, mcd grows. This behavior is consistent with our analysis from Section 2 that as acket size increases, so does system time. However, for a realistic range of ayload sizes such as the ones in the figure, mcd is still in an accetable range. The second region of interest is defined by the blowu in mcd as the header size increases over the range of ayload sizes. This blowu corresonds to the one defined for the N-Burst model, excet there it was exressed in terms of the burst arameter. In Figure 5, we see that for a fixed value of b, the location of the blowu oint occurs earlier for small ayloads than large ayloads. In other words, for small ayload sizes, the range of values of header size that yield accetable erformance is smaller than it is for larger ayload sizes. This is exlained by (3), which effectively shifts the blowu oint as the router utilization increases, keeing the burst arameter constant. As such, we are able to define a region of accetable service with resect to mean cell delay at the router. There is a third blowu equivalent to the one seen with the exonential ON time, and that is due to high header-to-ayload size ratio, which is indicated by the sikes in the figure. We last look at mcd as a function of b and ayload size for a fixed header size of 5000, as shown in Figure 6. As before, an increase in mcd as ayload grows can be observed. A region of accetable service can also be exressed here, but in this case it is in terms of allowable burst arameter values over a range of ayload sizes for a fixed header. As ayload size increases, the blowu in mcd occurs for larger values of b. The increase in ayload size reduces δ, which lowers the router utilization, effectively shifting the value of b at which a blowu occurs.

15 mcd calculations for 1 Burst: T = 1, θ = 0.5, ρ(0) = 0.25, b = mcd x ayload header x 10 4 Figure 4: Mean Cell Delay as a function of ayload and header size for a 1-Burst rocess with exonential ON time. mcd calculations for 1 Burst: T = 30, θ = 0.5, ρ(0) = 0.25, b = mcd x ayload header x 10 4 Figure 5: Mean Cell Delay as a function of ayload and header size for a 1-Burst rocess with ower-tailed ON time.

16 mcd calculations: TPT30, θ = 0.5, ρ(0) = 0.5, h = mcd x ayload burst arameter Figure 6: Mean Cell Delay as a function of ayload size and the burst arameter for a 1-Burst rocess with ower-tailed ON-time and fixed header size. 5 Conclusion In this aer, we have resented an analytic model with a rich arameter set based on the N-Burst model. Our formulation rovides us with the means to study the intricate relationshi among ayload, header size and mean cell delay at the router in the resence of heavy-tailed traffic. For secial cases of the burst arameter, namely when the arrival rocess reduces to either a Poisson or a bulk arrival rocess, we rovide closed-form solutions. When the ON time distribution is exonential, we show that delay is tame as long as the header-to-ayload size ratio is within a realistic range. When burst durations are ower-tailed, however, we are able to characterize of region of accetable service, which is defined by an allowable range of header sizes for each ayload size, beyond which delay blows u according to the results of the N-Burst model.

17 References [Anto04] [Cáce91] [Coo81] [Crov96] [Fior98] [Grei99] [Lis01] [Lis92] [Lett98] [Neut81] [Roma95] [Schw99a] Antonios, I., Semi-Markov Processes as Models for Telecommunications Systems with Heavy-Tailed Traffic, PhD Dissertation, in rogress, University of Connecticut, Cáceres, R., Efficiency of Asynchronous Transfer Mode Networks in Transorting Wide Area Network Data, TR International Comuter Science Institute, Berkeley, CA, July Cooer, R.B., Introduction to Queueing Theory, 2 nd ed. New York: North Holland, Crovella, M., and Bestavros, A., Self-Similarity in World Wide Web Traffic: Evidence and Possible Causes, Proceedings of SIGMETRICS Conference on Measurement and Modeling of Comuter Systems, Fiorini, P., Modeling Telecommunication Systems with Self-Similar Data Traffic, PhD Dissertation, Deartment of Comuter Science, University of Connecticut, May Greiner, M., Jobmann, M., and Lisky, L., The Imortance of Power-Tail Distributions for Telecommunications Traffic Models, Oerations Research, vol. 41, no. 6, March Lisky, L., Schwefel, H.-P., Greiner, M., and Jobmann, M., Comarison of The Analytic N-Burst Model with Other Aroximations to Self-similar Telecommunications Traffic, IEEE NCA_01 Sym. on Network Comuting, Cambridge, MA. February Lisky, L., QUEUEING THEORY: a Linear Algebraic Aroach, MacMillan Publishing Comany, New York, Lettieri, P., and Srivastava, M.B., Adative Frame Length Control for Imroving Wireless Link Throughut, Range, and Energy Efficiency, IEEE Infocom'98, San Francisco, , March Neuts, M., MATRIX-GEOMETRIC SOLUTIONS IN STOCHASTIC MODELS, John Hokins University Press, London, Romanov, A., and Floyd, S., Dynamics of TCP Traffic over ATM Networks, IEEE Journal on Selected Areas in Communications, May, Schwefel, H.-P., and Lisky, L., Performance Results For Models of Traffic in Telecommunication Systems, Based on Multile ON-OFF Sources With Self-Similar Behavior, Teletraffic Engineering in a Cometitive world, Vol 3A, P.Key and D. Smith, Eds., , Elsevier Science B.V., 1999.

18 [Schw99b] [Schw00a] [Schw00b] Schwefel, H.-P., Fiorini, P., and Lisky, L., Analytic Model of Performance in Telecommunication Systems, Based on ON-OFF Traffic Sources With Self-Similar Behavior, 7th International Conference on Telecommunication Systems, Nashville, TN, March 18-21, Schwefel, H.-P., Imact of Aggregated, Self-Similar ON/OFF Traffic on Delay in Stationary Queueing Models (extended version), Performance Evaluation, Schwefel, H.-P., Performance Analysis of Intermediate Systems Serving Aggregated ON/OFF Traffic with Long-Range Deendent Proerties, PhD Dissertation, Technische Universität München, 2000.

MATHEMATICAL MODELLING OF THE WIRELESS COMMUNICATION NETWORK

MATHEMATICAL MODELLING OF THE WIRELESS COMMUNICATION NETWORK Comuter Modelling and ew Technologies, 5, Vol.9, o., 3-39 Transort and Telecommunication Institute, Lomonosov, LV-9, Riga, Latvia MATHEMATICAL MODELLIG OF THE WIRELESS COMMUICATIO ETWORK M. KOPEETSK Deartment

More information

An Analysis of TCP over Random Access Satellite Links

An Analysis of TCP over Random Access Satellite Links An Analysis of over Random Access Satellite Links Chunmei Liu and Eytan Modiano Massachusetts Institute of Technology Cambridge, MA 0239 Email: mayliu, modiano@mit.edu Abstract This aer analyzes the erformance

More information

Shadow Computing: An Energy-Aware Fault Tolerant Computing Model

Shadow Computing: An Energy-Aware Fault Tolerant Computing Model Shadow Comuting: An Energy-Aware Fault Tolerant Comuting Model Bryan Mills, Taieb Znati, Rami Melhem Deartment of Comuter Science University of Pittsburgh (bmills, znati, melhem)@cs.itt.edu Index Terms

More information

Combining Logistic Regression with Kriging for Mapping the Risk of Occurrence of Unexploded Ordnance (UXO)

Combining Logistic Regression with Kriging for Mapping the Risk of Occurrence of Unexploded Ordnance (UXO) Combining Logistic Regression with Kriging for Maing the Risk of Occurrence of Unexloded Ordnance (UXO) H. Saito (), P. Goovaerts (), S. A. McKenna (2) Environmental and Water Resources Engineering, Deartment

More information

A Simple Throughput Model for TCP Veno

A Simple Throughput Model for TCP Veno A Simle Throughut Model for TCP Veno Bin Zhou, Cheng Peng Fu, Dah-Ming Chiu, Chiew Tong Lau, and Lek Heng Ngoh School of Comuter Engineering, Nanyang Technological University, Singaore 639798 Email: {zhou00,

More information

4. Score normalization technical details We now discuss the technical details of the score normalization method.

4. Score normalization technical details We now discuss the technical details of the score normalization method. SMT SCORING SYSTEM This document describes the scoring system for the Stanford Math Tournament We begin by giving an overview of the changes to scoring and a non-technical descrition of the scoring rules

More information

arxiv:cond-mat/ v2 25 Sep 2002

arxiv:cond-mat/ v2 25 Sep 2002 Energy fluctuations at the multicritical oint in two-dimensional sin glasses arxiv:cond-mat/0207694 v2 25 Se 2002 1. Introduction Hidetoshi Nishimori, Cyril Falvo and Yukiyasu Ozeki Deartment of Physics,

More information

MODELING THE RELIABILITY OF C4ISR SYSTEMS HARDWARE/SOFTWARE COMPONENTS USING AN IMPROVED MARKOV MODEL

MODELING THE RELIABILITY OF C4ISR SYSTEMS HARDWARE/SOFTWARE COMPONENTS USING AN IMPROVED MARKOV MODEL Technical Sciences and Alied Mathematics MODELING THE RELIABILITY OF CISR SYSTEMS HARDWARE/SOFTWARE COMPONENTS USING AN IMPROVED MARKOV MODEL Cezar VASILESCU Regional Deartment of Defense Resources Management

More information

dn i where we have used the Gibbs equation for the Gibbs energy and the definition of chemical potential

dn i where we have used the Gibbs equation for the Gibbs energy and the definition of chemical potential Chem 467 Sulement to Lectures 33 Phase Equilibrium Chemical Potential Revisited We introduced the chemical otential as the conjugate variable to amount. Briefly reviewing, the total Gibbs energy of a system

More information

Age of Information: Whittle Index for Scheduling Stochastic Arrivals

Age of Information: Whittle Index for Scheduling Stochastic Arrivals Age of Information: Whittle Index for Scheduling Stochastic Arrivals Yu-Pin Hsu Deartment of Communication Engineering National Taiei University yuinhsu@mail.ntu.edu.tw arxiv:80.03422v2 [math.oc] 7 Ar

More information

On the Role of Finite Queues in Cooperative Cognitive Radio Networks with Energy Harvesting

On the Role of Finite Queues in Cooperative Cognitive Radio Networks with Energy Harvesting On the Role of Finite Queues in Cooerative Cognitive Radio Networks with Energy Harvesting Mohamed A. Abd-Elmagid, Tamer Elatt, and Karim G. Seddik Wireless Intelligent Networks Center (WINC), Nile University,

More information

Developing A Deterioration Probabilistic Model for Rail Wear

Developing A Deterioration Probabilistic Model for Rail Wear International Journal of Traffic and Transortation Engineering 2012, 1(2): 13-18 DOI: 10.5923/j.ijtte.20120102.02 Develoing A Deterioration Probabilistic Model for Rail Wear Jabbar-Ali Zakeri *, Shahrbanoo

More information

John Weatherwax. Analysis of Parallel Depth First Search Algorithms

John Weatherwax. Analysis of Parallel Depth First Search Algorithms Sulementary Discussions and Solutions to Selected Problems in: Introduction to Parallel Comuting by Viin Kumar, Ananth Grama, Anshul Guta, & George Karyis John Weatherwax Chater 8 Analysis of Parallel

More information

Analysis of M/M/n/K Queue with Multiple Priorities

Analysis of M/M/n/K Queue with Multiple Priorities Analysis of M/M/n/K Queue with Multile Priorities Coyright, Sanjay K. Bose For a P-riority system, class P of highest riority Indeendent, Poisson arrival rocesses for each class with i as average arrival

More information

Estimation of the large covariance matrix with two-step monotone missing data

Estimation of the large covariance matrix with two-step monotone missing data Estimation of the large covariance matrix with two-ste monotone missing data Masashi Hyodo, Nobumichi Shutoh 2, Takashi Seo, and Tatjana Pavlenko 3 Deartment of Mathematical Information Science, Tokyo

More information

Analysis of execution time for parallel algorithm to dertmine if it is worth the effort to code and debug in parallel

Analysis of execution time for parallel algorithm to dertmine if it is worth the effort to code and debug in parallel Performance Analysis Introduction Analysis of execution time for arallel algorithm to dertmine if it is worth the effort to code and debug in arallel Understanding barriers to high erformance and redict

More information

Understanding and Using Availability

Understanding and Using Availability Understanding and Using Availability Jorge Luis Romeu, Ph.D. ASQ CQE/CRE, & Senior Member C. Stat Fellow, Royal Statistical Society Past Director, Region II (NY & PA) Director: Juarez Lincoln Marti Int

More information

Optimal Random Access and Random Spectrum Sensing for an Energy Harvesting Cognitive Radio with and without Primary Feedback Leveraging

Optimal Random Access and Random Spectrum Sensing for an Energy Harvesting Cognitive Radio with and without Primary Feedback Leveraging 1 Otimal Random Access and Random Sectrum Sensing for an Energy Harvesting Cognitive Radio with and without Primary Feedback Leveraging Ahmed El Shafie, Member, IEEE, arxiv:1401.0340v3 [cs.it] 27 Ar 2014

More information

Modeling and Estimation of Full-Chip Leakage Current Considering Within-Die Correlation

Modeling and Estimation of Full-Chip Leakage Current Considering Within-Die Correlation 6.3 Modeling and Estimation of Full-Chi Leaage Current Considering Within-Die Correlation Khaled R. eloue, Navid Azizi, Farid N. Najm Deartment of ECE, University of Toronto,Toronto, Ontario, Canada {haled,nazizi,najm}@eecg.utoronto.ca

More information

Analysis of Multi-Hop Emergency Message Propagation in Vehicular Ad Hoc Networks

Analysis of Multi-Hop Emergency Message Propagation in Vehicular Ad Hoc Networks Analysis of Multi-Ho Emergency Message Proagation in Vehicular Ad Hoc Networks ABSTRACT Vehicular Ad Hoc Networks (VANETs) are attracting the attention of researchers, industry, and governments for their

More information

Lower Confidence Bound for Process-Yield Index S pk with Autocorrelated Process Data

Lower Confidence Bound for Process-Yield Index S pk with Autocorrelated Process Data Quality Technology & Quantitative Management Vol. 1, No.,. 51-65, 15 QTQM IAQM 15 Lower onfidence Bound for Process-Yield Index with Autocorrelated Process Data Fu-Kwun Wang * and Yeneneh Tamirat Deartment

More information

A Qualitative Event-based Approach to Multiple Fault Diagnosis in Continuous Systems using Structural Model Decomposition

A Qualitative Event-based Approach to Multiple Fault Diagnosis in Continuous Systems using Structural Model Decomposition A Qualitative Event-based Aroach to Multile Fault Diagnosis in Continuous Systems using Structural Model Decomosition Matthew J. Daigle a,,, Anibal Bregon b,, Xenofon Koutsoukos c, Gautam Biswas c, Belarmino

More information

A PEAK FACTOR FOR PREDICTING NON-GAUSSIAN PEAK RESULTANT RESPONSE OF WIND-EXCITED TALL BUILDINGS

A PEAK FACTOR FOR PREDICTING NON-GAUSSIAN PEAK RESULTANT RESPONSE OF WIND-EXCITED TALL BUILDINGS The Seventh Asia-Pacific Conference on Wind Engineering, November 8-1, 009, Taiei, Taiwan A PEAK FACTOR FOR PREDICTING NON-GAUSSIAN PEAK RESULTANT RESPONSE OF WIND-EXCITED TALL BUILDINGS M.F. Huang 1,

More information

Statistical Multiplexing Gain of Link Scheduling Algorithms in QoS Networks (Short Version)

Statistical Multiplexing Gain of Link Scheduling Algorithms in QoS Networks (Short Version) 1 Statistical Multilexing Gain of Link Scheduling Algorithms in QoS Networks (Short Version) Technical Reort: University of Virginia, CS-99-23, July 1999 Robert Boorstyn Almut Burchard Jörg Liebeherr y

More information

Understanding and Using Availability

Understanding and Using Availability Understanding and Using Availability Jorge Luis Romeu, Ph.D. ASQ CQE/CRE, & Senior Member Email: romeu@cortland.edu htt://myrofile.cos.com/romeu ASQ/RD Webinar Series Noviembre 5, J. L. Romeu - Consultant

More information

NUMERICAL AND THEORETICAL INVESTIGATIONS ON DETONATION- INERT CONFINEMENT INTERACTIONS

NUMERICAL AND THEORETICAL INVESTIGATIONS ON DETONATION- INERT CONFINEMENT INTERACTIONS NUMERICAL AND THEORETICAL INVESTIGATIONS ON DETONATION- INERT CONFINEMENT INTERACTIONS Tariq D. Aslam and John B. Bdzil Los Alamos National Laboratory Los Alamos, NM 87545 hone: 1-55-667-1367, fax: 1-55-667-6372

More information

State Estimation with ARMarkov Models

State Estimation with ARMarkov Models Deartment of Mechanical and Aerosace Engineering Technical Reort No. 3046, October 1998. Princeton University, Princeton, NJ. State Estimation with ARMarkov Models Ryoung K. Lim 1 Columbia University,

More information

Power Aware Wireless File Downloading: A Constrained Restless Bandit Approach

Power Aware Wireless File Downloading: A Constrained Restless Bandit Approach PROC. WIOP 204 Power Aware Wireless File Downloading: A Constrained Restless Bandit Aroach Xiaohan Wei and Michael J. Neely, Senior Member, IEEE Abstract his aer treats ower-aware throughut maximization

More information

Signaled Queueing. Laura Brink, Robert Shorten, Jia Yuan Yu ABSTRACT. Categories and Subject Descriptors. General Terms. Keywords

Signaled Queueing. Laura Brink, Robert Shorten, Jia Yuan Yu ABSTRACT. Categories and Subject Descriptors. General Terms. Keywords Signaled Queueing Laura Brink, Robert Shorten, Jia Yuan Yu ABSTRACT Burstiness in queues where customers arrive indeendently leads to rush eriods when wait times are long. We roose a simle signaling scheme

More information

An Improved Generalized Estimation Procedure of Current Population Mean in Two-Occasion Successive Sampling

An Improved Generalized Estimation Procedure of Current Population Mean in Two-Occasion Successive Sampling Journal of Modern Alied Statistical Methods Volume 15 Issue Article 14 11-1-016 An Imroved Generalized Estimation Procedure of Current Poulation Mean in Two-Occasion Successive Samling G. N. Singh Indian

More information

Spatial Outage Capacity of Poisson Bipolar Networks

Spatial Outage Capacity of Poisson Bipolar Networks Satial Outage Caacity of Poisson Biolar Networks Sanket S. Kalamkar and Martin Haenggi Deartment of Electrical Engineering, University of Notre Dame, Notre Dame, IN 46556, USA E-mail: skalamka@nd.edu,

More information

COMMUNICATION BETWEEN SHAREHOLDERS 1

COMMUNICATION BETWEEN SHAREHOLDERS 1 COMMUNICATION BTWN SHARHOLDRS 1 A B. O A : A D Lemma B.1. U to µ Z r 2 σ2 Z + σ2 X 2r ω 2 an additive constant that does not deend on a or θ, the agents ayoffs can be written as: 2r rθa ω2 + θ µ Y rcov

More information

Approximating min-max k-clustering

Approximating min-max k-clustering Aroximating min-max k-clustering Asaf Levin July 24, 2007 Abstract We consider the roblems of set artitioning into k clusters with minimum total cost and minimum of the maximum cost of a cluster. The cost

More information

Characteristics of Beam-Based Flexure Modules

Characteristics of Beam-Based Flexure Modules Shorya Awtar e-mail: shorya@mit.edu Alexander H. Slocum e-mail: slocum@mit.edu Precision Engineering Research Grou, Massachusetts Institute of Technology, Cambridge, MA 039 Edi Sevincer Omega Advanced

More information

Plotting the Wilson distribution

Plotting the Wilson distribution , Survey of English Usage, University College London Setember 018 1 1. Introduction We have discussed the Wilson score interval at length elsewhere (Wallis 013a, b). Given an observed Binomial roortion

More information

arxiv: v2 [q-bio.pe] 5 Jan 2018

arxiv: v2 [q-bio.pe] 5 Jan 2018 Phenotyic switching of oulations of cells in a stochastic environment arxiv:76.7789v [-bio.pe] 5 Jan 8 Peter G. Hufton,, Yen Ting Lin,,, and Tobias Galla, Theoretical Physics, School of Physics and stronomy,

More information

Distributed Rule-Based Inference in the Presence of Redundant Information

Distributed Rule-Based Inference in the Presence of Redundant Information istribution Statement : roved for ublic release; distribution is unlimited. istributed Rule-ased Inference in the Presence of Redundant Information June 8, 004 William J. Farrell III Lockheed Martin dvanced

More information

An Analysis of Reliable Classifiers through ROC Isometrics

An Analysis of Reliable Classifiers through ROC Isometrics An Analysis of Reliable Classifiers through ROC Isometrics Stijn Vanderlooy s.vanderlooy@cs.unimaas.nl Ida G. Srinkhuizen-Kuyer kuyer@cs.unimaas.nl Evgueni N. Smirnov smirnov@cs.unimaas.nl MICC-IKAT, Universiteit

More information

Scaling Multiple Point Statistics for Non-Stationary Geostatistical Modeling

Scaling Multiple Point Statistics for Non-Stationary Geostatistical Modeling Scaling Multile Point Statistics or Non-Stationary Geostatistical Modeling Julián M. Ortiz, Steven Lyster and Clayton V. Deutsch Centre or Comutational Geostatistics Deartment o Civil & Environmental Engineering

More information

Pulse Propagation in Optical Fibers using the Moment Method

Pulse Propagation in Optical Fibers using the Moment Method Pulse Proagation in Otical Fibers using the Moment Method Bruno Miguel Viçoso Gonçalves das Mercês, Instituto Suerior Técnico Abstract The scoe of this aer is to use the semianalytic technique of the Moment

More information

A Bound on the Error of Cross Validation Using the Approximation and Estimation Rates, with Consequences for the Training-Test Split

A Bound on the Error of Cross Validation Using the Approximation and Estimation Rates, with Consequences for the Training-Test Split A Bound on the Error of Cross Validation Using the Aroximation and Estimation Rates, with Consequences for the Training-Test Slit Michael Kearns AT&T Bell Laboratories Murray Hill, NJ 7974 mkearns@research.att.com

More information

System Reliability Estimation and Confidence Regions from Subsystem and Full System Tests

System Reliability Estimation and Confidence Regions from Subsystem and Full System Tests 009 American Control Conference Hyatt Regency Riverfront, St. Louis, MO, USA June 0-, 009 FrB4. System Reliability Estimation and Confidence Regions from Subsystem and Full System Tests James C. Sall Abstract

More information

Fisher Information in Flow Size Distribution Estimation

Fisher Information in Flow Size Distribution Estimation 1 Fisher Information in Flow Size Distribution Estimation Paul Tune, Member, IEEE, and Darryl Veitch, Fellow, IEEE Abstract The flow size distribution is a useful metric for traffic modeling and management

More information

An Investigation on the Numerical Ill-conditioning of Hybrid State Estimators

An Investigation on the Numerical Ill-conditioning of Hybrid State Estimators An Investigation on the Numerical Ill-conditioning of Hybrid State Estimators S. K. Mallik, Student Member, IEEE, S. Chakrabarti, Senior Member, IEEE, S. N. Singh, Senior Member, IEEE Deartment of Electrical

More information

Improved Capacity Bounds for the Binary Energy Harvesting Channel

Improved Capacity Bounds for the Binary Energy Harvesting Channel Imroved Caacity Bounds for the Binary Energy Harvesting Channel Kaya Tutuncuoglu 1, Omur Ozel 2, Aylin Yener 1, and Sennur Ulukus 2 1 Deartment of Electrical Engineering, The Pennsylvania State University,

More information

Modeling Residual-Geometric Flow Sampling

Modeling Residual-Geometric Flow Sampling Modeling Residual-Geometric Flow Samling Xiaoming Wang Amazon.com Seattle, WA 98101 USA Email: xmwang@gmail.com Xiaoyong Li Texas A&M University College Station, TX 77843 USA Email: xiaoyong@cse.tamu.edu

More information

Characterizing the Behavior of a Probabilistic CMOS Switch Through Analytical Models and Its Verification Through Simulations

Characterizing the Behavior of a Probabilistic CMOS Switch Through Analytical Models and Its Verification Through Simulations Characterizing the Behavior of a Probabilistic CMOS Switch Through Analytical Models and Its Verification Through Simulations PINAR KORKMAZ, BILGE E. S. AKGUL and KRISHNA V. PALEM Georgia Institute of

More information

Modeling Residual-Geometric Flow Sampling

Modeling Residual-Geometric Flow Sampling 1 Modeling Residual-Geometric Flow Samling Xiaoming Wang, Xiaoyong Li, and Dmitri Loguinov Abstract Traffic monitoring and estimation of flow arameters in high seed routers have recently become challenging

More information

Delay characterization of multi-hop transmission in a Poisson field of interference

Delay characterization of multi-hop transmission in a Poisson field of interference 1 Delay characterization of multi-ho transmission in a Poisson field of interference Kostas Stamatiou and Martin Haenggi, Senior Member, IEEE Abstract We evaluate the end-to-end delay of a multi-ho transmission

More information

Information collection on a graph

Information collection on a graph Information collection on a grah Ilya O. Ryzhov Warren Powell February 10, 2010 Abstract We derive a knowledge gradient olicy for an otimal learning roblem on a grah, in which we use sequential measurements

More information

Morten Frydenberg Section for Biostatistics Version :Friday, 05 September 2014

Morten Frydenberg Section for Biostatistics Version :Friday, 05 September 2014 Morten Frydenberg Section for Biostatistics Version :Friday, 05 Setember 204 All models are aroximations! The best model does not exist! Comlicated models needs a lot of data. lower your ambitions or get

More information

Pretest (Optional) Use as an additional pacing tool to guide instruction. August 21

Pretest (Optional) Use as an additional pacing tool to guide instruction. August 21 Trimester 1 Pretest (Otional) Use as an additional acing tool to guide instruction. August 21 Beyond the Basic Facts In Trimester 1, Grade 8 focus on multilication. Daily Unit 1: Rational vs. Irrational

More information

3312 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 65, NO. 5, MAY 2016

3312 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 65, NO. 5, MAY 2016 3312 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 65, NO. 5, MAY 2016 L -Problem-Based Transmission Rate Allocation With Packet Loss and Power Metrics Over Satellite Networks Igor Bisio, Senior Member,

More information

CHAPTER-II Control Charts for Fraction Nonconforming using m-of-m Runs Rules

CHAPTER-II Control Charts for Fraction Nonconforming using m-of-m Runs Rules CHAPTER-II Control Charts for Fraction Nonconforming using m-of-m Runs Rules. Introduction: The is widely used in industry to monitor the number of fraction nonconforming units. A nonconforming unit is

More information

1 Gambler s Ruin Problem

1 Gambler s Ruin Problem Coyright c 2017 by Karl Sigman 1 Gambler s Ruin Problem Let N 2 be an integer and let 1 i N 1. Consider a gambler who starts with an initial fortune of $i and then on each successive gamble either wins

More information

Proof: We follow thearoach develoed in [4]. We adot a useful but non-intuitive notion of time; a bin with z balls at time t receives its next ball at

Proof: We follow thearoach develoed in [4]. We adot a useful but non-intuitive notion of time; a bin with z balls at time t receives its next ball at A Scaling Result for Exlosive Processes M. Mitzenmacher Λ J. Sencer We consider the following balls and bins model, as described in [, 4]. Balls are sequentially thrown into bins so that the robability

More information

An Outdoor Recreation Use Model with Applications to Evaluating Survey Estimators

An Outdoor Recreation Use Model with Applications to Evaluating Survey Estimators United States Deartment of Agriculture Forest Service Southern Research Station An Outdoor Recreation Use Model with Alications to Evaluating Survey Estimators Stanley J. Zarnoch, Donald B.K. English,

More information

Notes on pressure coordinates Robert Lindsay Korty October 1, 2002

Notes on pressure coordinates Robert Lindsay Korty October 1, 2002 Notes on ressure coordinates Robert Lindsay Korty October 1, 2002 Obviously, it makes no difference whether the quasi-geostrohic equations are hrased in height coordinates (where x, y,, t are the indeendent

More information

pp physics, RWTH, WS 2003/04, T.Hebbeker

pp physics, RWTH, WS 2003/04, T.Hebbeker 1. PP TH 03/04 Accelerators and Detectors 1 hysics, RWTH, WS 2003/04, T.Hebbeker 2003-12-03 1. Accelerators and Detectors In the following, we concentrate on the three machines SPS, Tevatron and LHC with

More information

Monopolist s mark-up and the elasticity of substitution

Monopolist s mark-up and the elasticity of substitution Croatian Oerational Research Review 377 CRORR 8(7), 377 39 Monoolist s mark-u and the elasticity of substitution Ilko Vrankić, Mira Kran, and Tomislav Herceg Deartment of Economic Theory, Faculty of Economics

More information

Paper C Exact Volume Balance Versus Exact Mass Balance in Compositional Reservoir Simulation

Paper C Exact Volume Balance Versus Exact Mass Balance in Compositional Reservoir Simulation Paer C Exact Volume Balance Versus Exact Mass Balance in Comositional Reservoir Simulation Submitted to Comutational Geosciences, December 2005. Exact Volume Balance Versus Exact Mass Balance in Comositional

More information

Preconditioning techniques for Newton s method for the incompressible Navier Stokes equations

Preconditioning techniques for Newton s method for the incompressible Navier Stokes equations Preconditioning techniques for Newton s method for the incomressible Navier Stokes equations H. C. ELMAN 1, D. LOGHIN 2 and A. J. WATHEN 3 1 Deartment of Comuter Science, University of Maryland, College

More information

Evaluating Circuit Reliability Under Probabilistic Gate-Level Fault Models

Evaluating Circuit Reliability Under Probabilistic Gate-Level Fault Models Evaluating Circuit Reliability Under Probabilistic Gate-Level Fault Models Ketan N. Patel, Igor L. Markov and John P. Hayes University of Michigan, Ann Arbor 48109-2122 {knatel,imarkov,jhayes}@eecs.umich.edu

More information

Topic 7: Using identity types

Topic 7: Using identity types Toic 7: Using identity tyes June 10, 2014 Now we would like to learn how to use identity tyes and how to do some actual mathematics with them. By now we have essentially introduced all inference rules

More information

Delay Performance of Threshold Policies for Dynamic Spectrum Access

Delay Performance of Threshold Policies for Dynamic Spectrum Access IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL., NO. 7, JULY 83 Delay Performance of Threshold Policies for Dynamic Sectrum Access Rong-Rong Chen and Xin Liu Abstract In this aer, we analyze the delay

More information

Computer arithmetic. Intensive Computation. Annalisa Massini 2017/2018

Computer arithmetic. Intensive Computation. Annalisa Massini 2017/2018 Comuter arithmetic Intensive Comutation Annalisa Massini 7/8 Intensive Comutation - 7/8 References Comuter Architecture - A Quantitative Aroach Hennessy Patterson Aendix J Intensive Comutation - 7/8 3

More information

Information collection on a graph

Information collection on a graph Information collection on a grah Ilya O. Ryzhov Warren Powell October 25, 2009 Abstract We derive a knowledge gradient olicy for an otimal learning roblem on a grah, in which we use sequential measurements

More information

Temperature, current and doping dependence of non-ideality factor for pnp and npn punch-through structures

Temperature, current and doping dependence of non-ideality factor for pnp and npn punch-through structures Indian Journal of Pure & Alied Physics Vol. 44, December 2006,. 953-958 Temerature, current and doing deendence of non-ideality factor for n and nn unch-through structures Khurshed Ahmad Shah & S S Islam

More information

On Code Design for Simultaneous Energy and Information Transfer

On Code Design for Simultaneous Energy and Information Transfer On Code Design for Simultaneous Energy and Information Transfer Anshoo Tandon Electrical and Comuter Engineering National University of Singaore Email: anshoo@nus.edu.sg Mehul Motani Electrical and Comuter

More information

PER-PATCH METRIC LEARNING FOR ROBUST IMAGE MATCHING. Sezer Karaoglu, Ivo Everts, Jan C. van Gemert, and Theo Gevers

PER-PATCH METRIC LEARNING FOR ROBUST IMAGE MATCHING. Sezer Karaoglu, Ivo Everts, Jan C. van Gemert, and Theo Gevers PER-PATCH METRIC LEARNING FOR ROBUST IMAGE MATCHING Sezer Karaoglu, Ivo Everts, Jan C. van Gemert, and Theo Gevers Intelligent Systems Lab, Amsterdam, University of Amsterdam, 1098 XH Amsterdam, The Netherlands

More information

s v 0 q 0 v 1 q 1 v 2 (q 2) v 3 q 3 v 4

s v 0 q 0 v 1 q 1 v 2 (q 2) v 3 q 3 v 4 Discrete Adative Transmission for Fading Channels Lang Lin Λ, Roy D. Yates, Predrag Sasojevic WINLAB, Rutgers University 7 Brett Rd., NJ- fllin, ryates, sasojevg@winlab.rutgers.edu Abstract In this work

More information

A generalization of Amdahl's law and relative conditions of parallelism

A generalization of Amdahl's law and relative conditions of parallelism A generalization of Amdahl's law and relative conditions of arallelism Author: Gianluca Argentini, New Technologies and Models, Riello Grou, Legnago (VR), Italy. E-mail: gianluca.argentini@riellogrou.com

More information

Analysis of some entrance probabilities for killed birth-death processes

Analysis of some entrance probabilities for killed birth-death processes Analysis of some entrance robabilities for killed birth-death rocesses Master s Thesis O.J.G. van der Velde Suervisor: Dr. F.M. Sieksma July 5, 207 Mathematical Institute, Leiden University Contents Introduction

More information

Analysis of Pressure Transient Response for an Injector under Hydraulic Stimulation at the Salak Geothermal Field, Indonesia

Analysis of Pressure Transient Response for an Injector under Hydraulic Stimulation at the Salak Geothermal Field, Indonesia roceedings World Geothermal Congress 00 Bali, Indonesia, 5-9 Aril 00 Analysis of ressure Transient Resonse for an Injector under Hydraulic Stimulation at the Salak Geothermal Field, Indonesia Jorge A.

More information

Estimation of Separable Representations in Psychophysical Experiments

Estimation of Separable Representations in Psychophysical Experiments Estimation of Searable Reresentations in Psychohysical Exeriments Michele Bernasconi (mbernasconi@eco.uninsubria.it) Christine Choirat (cchoirat@eco.uninsubria.it) Raffaello Seri (rseri@eco.uninsubria.it)

More information

Determining Momentum and Energy Corrections for g1c Using Kinematic Fitting

Determining Momentum and Energy Corrections for g1c Using Kinematic Fitting CLAS-NOTE 4-17 Determining Momentum and Energy Corrections for g1c Using Kinematic Fitting Mike Williams, Doug Alegate and Curtis A. Meyer Carnegie Mellon University June 7, 24 Abstract We have used the

More information

Bayesian Spatially Varying Coefficient Models in the Presence of Collinearity

Bayesian Spatially Varying Coefficient Models in the Presence of Collinearity Bayesian Satially Varying Coefficient Models in the Presence of Collinearity David C. Wheeler 1, Catherine A. Calder 1 he Ohio State University 1 Abstract he belief that relationshis between exlanatory

More information

Modeling Volume Changes in Porous Electrodes

Modeling Volume Changes in Porous Electrodes Journal of The Electrochemical Society, 53 A79-A86 2006 003-465/2005/53/A79/8/$20.00 The Electrochemical Society, Inc. Modeling olume Changes in Porous Electrodes Parthasarathy M. Gomadam*,a,z John W.

More information

The non-stochastic multi-armed bandit problem

The non-stochastic multi-armed bandit problem Submitted for journal ublication. The non-stochastic multi-armed bandit roblem Peter Auer Institute for Theoretical Comuter Science Graz University of Technology A-8010 Graz (Austria) auer@igi.tu-graz.ac.at

More information

Efficient Approximations for Call Admission Control Performance Evaluations in Multi-Service Networks

Efficient Approximations for Call Admission Control Performance Evaluations in Multi-Service Networks Efficient Aroximations for Call Admission Control Performance Evaluations in Multi-Service Networks Emre A. Yavuz, and Victor C. M. Leung Deartment of Electrical and Comuter Engineering The University

More information

An Improved Calibration Method for a Chopped Pyrgeometer

An Improved Calibration Method for a Chopped Pyrgeometer 96 JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY VOLUME 17 An Imroved Calibration Method for a Choed Pyrgeometer FRIEDRICH FERGG OtoLab, Ingenieurbüro, Munich, Germany PETER WENDLING Deutsches Forschungszentrum

More information

Adiabatic Shear Bands in Simple and Dipolar Plastic Materials

Adiabatic Shear Bands in Simple and Dipolar Plastic Materials Adiabatic Shear Bands in Simle and Diolar Plastic Materials T W \-1RIGHT us Army Ballistic Research Laboratory Aberdeen Proving Ground, MD 215 R C BATRA University of Missouri-Rolla Rolla, Missouri 6541

More information

A SIMPLE PLASTICITY MODEL FOR PREDICTING TRANSVERSE COMPOSITE RESPONSE AND FAILURE

A SIMPLE PLASTICITY MODEL FOR PREDICTING TRANSVERSE COMPOSITE RESPONSE AND FAILURE THE 19 TH INTERNATIONAL CONFERENCE ON COMPOSITE MATERIALS A SIMPLE PLASTICITY MODEL FOR PREDICTING TRANSVERSE COMPOSITE RESPONSE AND FAILURE K.W. Gan*, M.R. Wisnom, S.R. Hallett, G. Allegri Advanced Comosites

More information

Feedback-error control

Feedback-error control Chater 4 Feedback-error control 4.1 Introduction This chater exlains the feedback-error (FBE) control scheme originally described by Kawato [, 87, 8]. FBE is a widely used neural network based controller

More information

Availability and Maintainability. Piero Baraldi

Availability and Maintainability. Piero Baraldi Availability and Maintainability 1 Introduction: reliability and availability System tyes Non maintained systems: they cannot be reaired after a failure (a telecommunication satellite, a F1 engine, a vessel

More information

New Schedulability Test Conditions for Non-preemptive Scheduling on Multiprocessor Platforms

New Schedulability Test Conditions for Non-preemptive Scheduling on Multiprocessor Platforms New Schedulability Test Conditions for Non-reemtive Scheduling on Multirocessor Platforms Technical Reort May 2008 Nan Guan 1, Wang Yi 2, Zonghua Gu 3 and Ge Yu 1 1 Northeastern University, Shenyang, China

More information

8 STOCHASTIC PROCESSES

8 STOCHASTIC PROCESSES 8 STOCHASTIC PROCESSES The word stochastic is derived from the Greek στoχαστικoς, meaning to aim at a target. Stochastic rocesses involve state which changes in a random way. A Markov rocess is a articular

More information

Elementary Analysis in Q p

Elementary Analysis in Q p Elementary Analysis in Q Hannah Hutter, May Szedlák, Phili Wirth November 17, 2011 This reort follows very closely the book of Svetlana Katok 1. 1 Sequences and Series In this section we will see some

More information

An Adaptive Three-bus Power System Equivalent for Estimating Voltage Stability Margin from Synchronized Phasor Measurements

An Adaptive Three-bus Power System Equivalent for Estimating Voltage Stability Margin from Synchronized Phasor Measurements An Adative Three-bus Power System Equivalent for Estimating oltage Stability argin from Synchronized Phasor easurements Fengkai Hu, Kai Sun University of Tennessee Knoxville, TN, USA fengkaihu@utk.edu

More information

1. INTRODUCTION. Fn 2 = F j F j+1 (1.1)

1. INTRODUCTION. Fn 2 = F j F j+1 (1.1) CERTAIN CLASSES OF FINITE SUMS THAT INVOLVE GENERALIZED FIBONACCI AND LUCAS NUMBERS The beautiful identity R.S. Melham Deartment of Mathematical Sciences, University of Technology, Sydney PO Box 23, Broadway,

More information

STABILITY ANALYSIS AND CONTROL OF STOCHASTIC DYNAMIC SYSTEMS USING POLYNOMIAL CHAOS. A Dissertation JAMES ROBERT FISHER

STABILITY ANALYSIS AND CONTROL OF STOCHASTIC DYNAMIC SYSTEMS USING POLYNOMIAL CHAOS. A Dissertation JAMES ROBERT FISHER STABILITY ANALYSIS AND CONTROL OF STOCHASTIC DYNAMIC SYSTEMS USING POLYNOMIAL CHAOS A Dissertation by JAMES ROBERT FISHER Submitted to the Office of Graduate Studies of Texas A&M University in artial fulfillment

More information

CONVOLVED SUBSAMPLING ESTIMATION WITH APPLICATIONS TO BLOCK BOOTSTRAP

CONVOLVED SUBSAMPLING ESTIMATION WITH APPLICATIONS TO BLOCK BOOTSTRAP Submitted to the Annals of Statistics arxiv: arxiv:1706.07237 CONVOLVED SUBSAMPLING ESTIMATION WITH APPLICATIONS TO BLOCK BOOTSTRAP By Johannes Tewes, Dimitris N. Politis and Daniel J. Nordman Ruhr-Universität

More information

CMSC 425: Lecture 4 Geometry and Geometric Programming

CMSC 425: Lecture 4 Geometry and Geometric Programming CMSC 425: Lecture 4 Geometry and Geometric Programming Geometry for Game Programming and Grahics: For the next few lectures, we will discuss some of the basic elements of geometry. There are many areas

More information

Robust Predictive Control of Input Constraints and Interference Suppression for Semi-Trailer System

Robust Predictive Control of Input Constraints and Interference Suppression for Semi-Trailer System Vol.7, No.7 (4),.37-38 htt://dx.doi.org/.457/ica.4.7.7.3 Robust Predictive Control of Inut Constraints and Interference Suression for Semi-Trailer System Zhao, Yang Electronic and Information Technology

More information

Optimal Random Access and Random Spectrum Sensing for an Energy Harvesting Cognitive Radio with and without Primary Feedback Leveraging

Optimal Random Access and Random Spectrum Sensing for an Energy Harvesting Cognitive Radio with and without Primary Feedback Leveraging Otimal Random Access and Random Sectrum Sensing for an Energy Harvesting Cognitive Radio with and without Primary Feedback Leveraging Ahmed El Shafie Wireless Intelligent Networks Center WINC, Nile University,

More information

A Study of Active Queue Management for Congestion Control

A Study of Active Queue Management for Congestion Control A Study of Active Queue Management for Congestion Control Victor Firoiu vfiroiu@nortelnetworks.com Nortel Networks 3 Federal St. illerica, MA 1821 USA Marty orden mborden@tollbridgetech.com Tollridge Technologies

More information

General Linear Model Introduction, Classes of Linear models and Estimation

General Linear Model Introduction, Classes of Linear models and Estimation Stat 740 General Linear Model Introduction, Classes of Linear models and Estimation An aim of scientific enquiry: To describe or to discover relationshis among events (variables) in the controlled (laboratory)

More information

On split sample and randomized confidence intervals for binomial proportions

On split sample and randomized confidence intervals for binomial proportions On slit samle and randomized confidence intervals for binomial roortions Måns Thulin Deartment of Mathematics, Usala University arxiv:1402.6536v1 [stat.me] 26 Feb 2014 Abstract Slit samle methods have

More information

Positive decomposition of transfer functions with multiple poles

Positive decomposition of transfer functions with multiple poles Positive decomosition of transfer functions with multile oles Béla Nagy 1, Máté Matolcsi 2, and Márta Szilvási 1 Deartment of Analysis, Technical University of Budaest (BME), H-1111, Budaest, Egry J. u.

More information

An Ant Colony Optimization Approach to the Probabilistic Traveling Salesman Problem

An Ant Colony Optimization Approach to the Probabilistic Traveling Salesman Problem An Ant Colony Otimization Aroach to the Probabilistic Traveling Salesman Problem Leonora Bianchi 1, Luca Maria Gambardella 1, and Marco Dorigo 2 1 IDSIA, Strada Cantonale Galleria 2, CH-6928 Manno, Switzerland

More information