Performance Modeling of Hierarchical Memories

Size: px
Start display at page:

Download "Performance Modeling of Hierarchical Memories"

Transcription

1 Performane Modelng of Herarhal Memores Marwan Sleman, Lester Lpsky, Kshor Konwar Department of omputer Sene and Engneerng Unversty of onnetut Storrs, T Emal: {marwan, lester, kshor}@engr.uonn.edu Abstrat As memory herarhy expands n modern omputng envronment from PU regsters and loal memory to mddle-ter, network storage, and nternet storage, the optmal goal of a omputer arhtet beomes to desgn a memory herarhy that maxmzes the overall desgn of hs mahne wth a mnmal ost. Ths requres dedng on the number, speed and sze of the herarhal layers. As the gap between proessor and memory speed s growng exponentally, t beomes more mportant to develop an analytal model to apture all these herarhal levels and optmze the memory aess tme to make a good utlzaton of both PU and memory. In ths paper we study the performane of systems wth mult-level herarhal memores by modelng ther aess tme whh helps the desgner optmze the ost and aess tme. We use a lnearalgebra queung theory approah to aheve our goal and we explan why prevous attempts faled to provde aurate models. Our model dffers from all the prevous related work by beng global and general and by usng some probablst equatons that show the nterdependene between the dfferent levels and by usng the P-K formula to dstngush between the memory aess tme and queung tme. Our approah s ndependent of the applaton usng the memory whle lassal approahes were program dependent. Moreover, our model aheves hgher levels of auray whle beng expandable to multple levels. Keywords: Memory Herarhy, Performane Analyss, Markov model, Queung, Aess Tme. INTRODUTION Beng able to make aurate estmates of how long a memory aess wll take to fnsh n a mult-level herarhal memory system s of prmary nterest n the performane ommunty. In suh a herarhal envronment, we have multple levels of memory startng from PU regsters and extendng to ahes, man RAM memory, loal dsks, and network storage [6]. As the memory herarhy extends to network and nternet storage passng by mddle-ter arhteture and ahng, the problem of optmzng the memory aess tme beomes more hallengng. Data s stored and held n eah level untl t s used or replaed. Eah memory level dffers from the others by ts sze and speed. As the memory beomes loser to the PU, t beomes faster but smaller and more expensve. Thus, when desgnng a memory herarhy, our am s to get the fastest desgn whle mantanng a ompromse between sze, speed and ost. Havng found that the prevous approahes n studyng memory aess tme were lmted wth the number of memory levels they an represent, depended on the applaton, and dd not provde wth hgh auray, we represented the herarhal memory by a M/G/ queue [] and we alulated the aess tme by usng lnear algebra queung theory. Then we onsdered the queung tme of the onseutve memory requests that an our n a database applaton for example and we showed the dfferene between the aess tme and the queung tme. Ths dfferene explans the ause of the nauray of prevous approahes n predtng the memory aess tme. We also study the behavor of the varane aess tme whh s hghly rtal beause t an dramatally affet the mss rato of the memory system and ts performane. The model we bult s more general than the lassal models beause t an take as nput dfferent nput parameters lke, aess tme, ht rato, ost, and memory request dstrbuton. The remanng of ths paper s organzed as follows: In seton, we present a lterature survey about prevous efforts related to the top and we explan our motvaton. We dsuss the PK formula method to alulate the memory request queung tme n seton 3. In seton 4, we present two ases for evaluatng our methods. Then, n seton 5, we show the alulaton results by plottng the values resultng from eah alternatve and we show the dfferene between the two. Fnally, n seton 6, we propose some tops for further nvestgaton, and we onlude n seton 7.. BAKGROUND AND MOTIVATION The herarhal memory aess tme was studed by several researhers but all the prevous approahes to model and optmze the aess tme were lmted. The lmtatons are the result ether from the dependene of the models on the applaton or from the lmtatons of the analytal model that an not represent the deep herarhes. For example, Balasubramonan et al. explan that the reent memory herarhy organzatons do not math the applatons requrements whh results n a degradaton n the performane []. Jn et al. develop a lmted analytal model that aptures only a two-level ahe [3], but we see n ther work a bg dsrepany between the predted and measured memory performane. Most researhers foused on two-level memory as shown n artles [8] and [9] and we don t see any work that foused enough on deep memory herarhes. None of the prevous researhers talked about the varane of the aess tme of the memory herarhy. Ths varane s of prmary mportane beause a hgh varane n the aess tme orresponds to a hgher mss rato and unexpeted delay -whh s undesrable. In a prevous work [4], we have shown that the aess tme for a herarhal memory wth an nfnte depth s power taled []. In ths paper, we expand our prevous work and work on optmzng the memory desgn by tryng to buld a memory wth a mnmal response tme and mantanng a mnmal ost. We also show analytally the ause of dsrepany between the analytal and measured values for prevous researhers. Our Model s based on Markov han analyss whh s ndependent of the dstrbuton of memory request that depends on the applaton. We onsder a herarhal memory that onssts of L levels and a lowest memory level m as shown n fgure. We model ths herarhy n a state dagram as shown n fgure. Eah physal memory level,, n fgure orresponds to two states n fgure : The upper state orresponds to the lookup tme whle the lower state orresponds to the memory aess tme. The frst state orresponds to the memory request from the proessor whle the last state orresponds to the lowest memory n the herarhy.

2 PU Level Level Level l m Aess Tme Fgure. Herarhal Memory Model: PU wth L ntermedate levels of memory and man memory m. T T T l Fgure. State dagram of a herarhal memory system wth L ntermedate levels: Intermedate memory levels are represented by two states. The fgure shows the ht rato h n every level and the aess tme T at every state. Ths herarhal memory system s a M/G/ queue. In order to buld the lnear algebra model for ths system [], we defne the followng terms: X s the random varable representng the system tme that orresponds to the total memory aess tme through all memory stages. P s the sub-stohast matrx that orresponds to the transton from one state to another one. p s the entrane vetor that orresponds to the state of the system when at the frst memory request. p s a row vetor of sze L +, where L s the number of ntermedate levels. ε s the unt olumn vetor of sze L +. M s the transton rate matrx; t orresponds to the rates of leavng eah state. M s a dagonal matrx of the same sze as P. and M. h I s the dentty matrx of the same dmenson as P B M(I P) V B - s the nverse of B. T T T l T m - h T -h (-) T - -h -h (l-) T l- h h l l l -h l 0 h h P hl hl T M 0 0 ε T 0 0 Tm We have shown n [4] that the memory aess tme s gven by the frst moment of V and t s ndependent of the dstrbuton of both the memory aess request (whh s dependent on the ompler) and the serve tme of the nodes (whh depend on the hardware spefatons of the memory levels). The mean memory aess tme s gven by: EX ( ) x p Vε () However the varane of the aess tme s dependent on the dstrbuton of the serve tme of the memory nodes. It s dependent on the frst and seond moments of V. For exponental dstrbutons, t s gven by: σ ex p V ε - (pvε ) () For non-exponental dstrbutons, the varane s gven by: σ X σ ex + p V T Γ ε Where Γ dag( v-, v-,, vl-) EX ( Where ) x v s the oeffent of varaton of state x n Fg.. In an applaton that has multple onseutve memory requests lke a database applaton for example, the memory requests wll be queued and must wat to get serve from a busy memory, so nether the prevous models nor the above model wll be suffent to predt the exat tme. Thus we use the P-K formula n the next paragraph to fnd the exat queung tme. 3. THE P-K FORMULA The Pollazek-Khnthne formula (alled P-K formula) [7] gves the expeted average number of ustomers n queue and n proess n M/G/ queues. The P-K formula was ombned wth Lttle s theorem [] to show that the mean tme spent by a ustomer n an M/G/ queue s gven by: x xρ T + ρ ρ (3)

3 Where, s oeffent of varaton, ρ s the utlzaton fator, ρ λ x, and λ s the arrval rate. σ ex, x In ths paper we use equaton (3) to predt the queung tme for our herarhal memory system shown n fgure whh beomes as shown n fgure 3. The queung ours when we have a system wth multple onseutve memory requests that an not be aessed by the sequental memory at the same tme, so the memory requests are buffered n a queue; ths an be the ase of a shared memory on a parallel mahne, a smple database aess applaton, or a ppelned PU. In the last two ases the queung auses a bottlenek and affets the performane of the system beause t nreases the PI n a ppelned proessor [9] and nreases the query exeuton tme n a database applaton [0]. - h T -h (-) T - -h -h (l-) T l- l -h l dfferent mnma. We suppose that we have a herarhal memory system we are buldng and we assume that the system has a ost. Przybylsk [] used Agarwal s ahe mss model [] to show that the ahe mss rato s nversely proportonal to ts sze; so the ht rato of eah memory level, whh s the probablst omplement of the mss rato, s proportonal too to the nverse of ts sze and thus ts ost. But extendng ths observaton to multlevel memory s a lttle omplated and requres more alulatons; so we defne the followng parameters for the system n Fg.: a s the probablty of fndng data n the ntermedate memory level. S s the sze of eah ntermedate memory level. s the ost per unt of sze of eah ntermedate memory level. β a, where β s a onstant. S The total ost of the L-levels herarhal system beomes: L S (4) 4.. TWO-LEVEL ASH MEMORY SYSTEM We frst onsder the -level ash memory system shown n fgure 4. λ h h h l m T - h T 3 3 -h T T T l Fgure 3. Queung dagram of a herarhal memory system wth L ntermedate levels: Now the arrvng memory requests arrve wth rate λ and are queued before the PU. The varane of the queung tme of the model shown n fgure 3 s the same as that of the model shown n the prevous paragraph whh s shown n equaton (). It s obvous from both equatons () and (3) that the mean memory aess tmes depend on several parameters nludng ht rato h and aess tme T at eah memory level. 4. SAMPLE ASES FOR TEST AND EVALUATION In order to show the dfferene between the mean memory aess tme n equaton () and the mean queung tme for memory nput/output requests n equaton (3), we apply equatons () and (3) to several ases then ompare the results for eah ase. Our am s to show that the mean aess tme s not the same for both methods and has l Fgure 4. State dagram of a two-level ash memory: In ths fgure we dstngush between the memory ht rato and the probablty of fndng data n eah level. We name the frst level M and the seond level M. We defne the followng terms: S s the sze of memory level, M h h T T4 4 M M h s the ost per unt of sze of memory level, M S s the sze of memory level, M s the ost per unt of sze of memory level, M Y s the random varable representng the probablty of fndng data n a memory level a h M/ Y M) a h M) a h M M) h M) hh h( h) h( h) h M/ Y M) hh h h m

4 Where, h and h are respetvely the ht ratos of memory level and memory level. The ost of ths system s gven by: S S + S (5) 4.. THREE-LEVEL ASH MEMORY SYSTEM Then we onsder the 3-level ash memory system shown n Fg.5. T - h T 3 -h T 5 - l- h h h 3 -h 3 m h H H H a b HbH( Ha) h H ahh b H( Hb) h3 HH b The ost of the memory system s gven by: S S + S + S (6) a b 3 It s lear from equatons () and (3) and from the above dervatons n ths seton that the mean tme s a funton of the ht rato and sze of eah ntermedate level, so to optmze the mean aess tme, we wll have to optmze () and (3) versus these parameters. For the two systems we show here, we suppose that we have a onstant ost and we try to optmze the ntermedate memory ost and sze to get the fastest possble desgn as shown n the next seton. l T T 4 T 6 M M M 3 Fgure 5. State dagram of a three-level ash memory: In ths fgure we dstngush between the memory ht rato and the probablty of fndng data n eah level. We name the frst level M, the seond level M, and the thrd one M 3. Agan we defne the followng terms: S s the sze of memory level, M h s the ht rato at memory level, M, h Y M Y M a Pr( / ) s the ost per unt of sze of memory level, M s the ost per unt of sze of memory level, M b s the ost per unt of sze of memory level 3, M 3 Y s the random varable representng the probablty of fndng data n a memory level 5. ALULATIONS AND RESULTS Now that we have the analytal model to alulate the memory aess tme and queung tme, we wrote a Matlab ode to mplement our equatons and to verfy that what we mentoned s aurate. We have arred out an exhaustve set of program runs over several parameters. Sne the results are onsstent wth eah other we present only a few here. We frst onsder the -Level memory system n fgure 4 and we assume that t has an arbtrary fxed ost. To study the mean response tme versus S, the sze of memory level, we assume that S s n an arbtrary nterval (4.4<S<7.8). So S, the sze of memory level, wll be gven by dret dervaton from Equaton (5): S S We plot both the mean memory tme, E(x), and the queung tme, E(T), versus the sze of the level memory n fgure 6. We remark that they have dfferent mnma -whh onfrms our assumpton about the dfferene between them. We also defne the followng probabltes: H H a b 3 3 H M ) a 3 3 We an proof by a smple alulaton that the ht ratos are gven by: Fgure 6. Mean memory aess tme E(X) and mean queung tme E(T) versus the sze S of the Level memory, M, for a -Level herarhal memory system. E(x) has ts mnmum for S 6.4, whle E(T) has ts mnmum for S 6.8 Then we plot the varane of the memory aess tme obtaned from equaton () for the two-level memory system n fgure 7. The behavor of the varane s as mportant as that of the mean

5 memory tme and queung tme beause t auses devaton from the mean tme. Devaton from the eah sde of the mean tme s undesrable: fnshng too late s obvously undesrable beause t an dramatally affet the performane (lke nreasng the PI n a ppelned proessor for example), but fnshng too early s also undesrable beause we waste our memory resoures. We remark here that, whle the memory tme has a onvex behavor, the varane dereases as we nrease the memory sze and ths s normal beause t depends on the seond moment of the memory aess tme [4] whh nreases faster as the memory sze nreases. shown n fgure 5. Now the system s more omplated beause we have more varables. Here too, we assume that the system has an arbtrary ost and we assume that S and S 3, the szes of the seond and thrd memory levels, are n arbtrary ntervals (<S <6 and 6<S 3 <8). So, S, the sze of memory level, wll be gven by dret dervaton from Equaton (6): bs S3 S a We plot both the mean memory tme, E(x), and the queung tme, E(T), versus the szes of the memory levels and 3 n fgure 9. We remark here too that both surfaes are smlar and they have dfferent mnma. Fgure 7. Varane of the memory aess tme versus the sze S of the Level memory, M, for a -Level herarhal memory system. The varane dereases as the memory sze nreases. To emphasze more on the dfferene between E(X) and E(T), we alulate the dfferene between the value of the mnmum of E(T) and the value of X at the same value of S. We all ths dfferene DffT. We plot DffT versus the nput rate λ n fgure 8. We selet the values of λ to keep the system utlzaton ρ fator between zero and one []. We remark n fgure 8 that the dfferene s more sgnfant as the traff nput rate and system utlzaton nreases. Ths value goes up to.5 % of the mnmum value of the mean tme for a utlzaton fator lose to. Fgure 9. Mean memory aess tme E(X) and mean queung tme E(T) versus the szes of memory levels and 3 for a 3-Level herarhal memory system. Both surfaes look onave. E(x) has ts mnmum of.59 for S 4.5 and S 3 6, whle E(T) has ts mnmum of 3.68 for S 4.75 and S 3 6. The plot of the dfferene between the value of the mnmum of E (T) and the value of X versus λ n fgure 0 shows here a more sgnfant dfferene equal to % of the mnmal value of the memory aess tme. Ths dfferene explans the dfferene between the predted and measured performane that Jn et al. get n ther paper about performane predton on shared memory programs (3) n effet the authors there alulate the mean aess tme whle the values they measure are those of the queung tme, ths s why they get a 0% dfferene! Fgure 8. Dfferene between Mn(T) and the value of X for the same value of S for dfferent values of nput rate λ, for a -Level herarhal memory system. To show that our results are ndependent of the number of level n a herarhal memory, we repeat what we dd for the -level ash memory to the 3-level ash memory Fgure 0. Dfferene between Mn(T) and the value of X for the same value of S for dfferent values of nput rate λ, for a 3-Level herarhal memory system. Fnally to ompare the performane of the two memory systems we plot the varane of the memory aess tme for the three-level memory system n fgure. We remark n fgure that the

6 standard devaton s more senstve to the upper level memory and t dereases faster as we nrease Sb, the sze of the upper level. If we ompare fgure to fgure 7, we remark the standard devaton s hgher n fgure whh means that addng one more level to the herarhy nreases the varane and ths observaton s very mportant beause the omputer arhtet must take t nto onsderaton when desgnng systems senstve to the varane. the levels and add more levels to the herarhal memory system: Inreasng the sze of a memory redues the varane; however addng more levels nreases t. Our observaton helps the desgner dede whether to use bgger levels of memores or use more levels n hs desgn. Our analytal model shown n equaton (3), s a unversal model and an represent deep memory herarhes that extend beyond the onept of loal mahne storage to network storage. Ths model uses Markov han analyss and an take dfferent types of memory request dstrbutons. Our model s also deal for optmzaton beause t an take dfferent nputs lke the ht rato, sze, ost and speed parameters of the ntermedate memory level. Ths flexblty of takng dfferent parameters makes t easy to expand and apture any level of herarhy. Fgure. Varane of the memory aess tme versus the szes of the ntermedate Levels for a 3-Level herarhal memory system. The varane dereases as the memory szes nreases. 6. FUTURE WORK Ths paper s part of a larger work examnng performane of herarhal memory systems wth both analytal and smulaton tehnques. There are many tops we are ether already nvestgatng or hope to nvestgate soon. We ntend to valdate our performane model by omparng the predted aess tmes aganst exeuton tmes measured on real mahnes by usng benhmarks. We ntend to nlude models that aount for loaltes and workng sets. In addton to the mean and varane, we are plannng to study the more performane metrs and measures of the memory aess tme for both the exponental and non exponental ases. These measures nlude, but are not lmted to, the onfdene nterval and the relablty of the aess tme n addton to the rato of the lag between the target tme and atual tme to the varane. We wll work on proofng the onvexty of the mean tmes obtaned n equatons () and (3) and applyng several optmzaton tehnques on equaton (3) to optmze the desgn of mult-level herarhal memores to ome out wth the fastest possble ost-effetve system. 7. ONLUSION We have developed an analytal model to evaluate the mean and varane of the aess tme for memory requests n a herarhal mult-level memory envronment. We have shown and explaned the dfferene between the mean memory aess tme and the memory requests queung tme. Ths dfferene explans the dsrepany between the analytal values and the pratal values obtaned by prevous researhers. We have also shown the behavor of the varane of the aess tme as we nrease REFERENES [] Lester Lpsky, Queueng Theory - A Lnear Algebra Approah, Maxwell Mamllan Internatonal publshng group, 99. [] Rajeev Balasubramonan, Davd Albonesz, Alper Buyuktosunoglu, and Sandhya Dwarkadas, Dynam Memory Herarhy Performane Optmzaton, 7 th nternatonal symposum on omputer arhteture, June 000. [3] Ruomng Jn, Gagan Agrawal, Performane Predton for Random Wrte Redutons: A ase Study n Modelng Shared Memory Programs, Proeedngs of the 00 AM SIGMETRIS nternatonal onferene on Measurement and modelng of omputer systems, Marna Del Rey, alforna, pages: 7 8, 00. [4] Kshor M. Konwar Lester Lpsky Marwan Sleman, Moments of Memory Aess Tme for Systems Wth Herarhal Memores, st Internatonal onferene on omputers and Ther Applatons (ATA- 006), Seattle WA, Marh 006. [5] rstna Hrstea, Danel Lenosk, and John Keen. Measurng Memory Herarhy Performane of ahe-oherent Multproessors Usng Mro Benhmarks. Proeedngs of S 97, 997. [6] Jon Wllam Togo, The Holy Gral of Network Storage Management, Prente Hall, 004 [7] Danel P Heyman and Matthew J Sobel, Stohast Models n Operatons Researh: Stohast Proesses and Operatng haratersts, ourer Dover Publatons, 004 [8] A. Smth, ahe Memores, omputng Surveys, 4(3): p , 98. [9] A. Smth, Dsk ahe-mss rato analyss and desgn onsderatons. AM Transaton on omputer Systems, 3(3), p 6-03, 985. [0] I. MaIntyre and B. Press, The Effet of ahe on the Performane of a Mult-Threaded Ppelned RIS Proessor, the Engneerng Insttute of anada, 99. [] S. Manegold, P. Bonz, and M. Kersten, Gener Database ost Models for Herarhal Memory Systems, Proeedngs of the 8th VLDB onferene, Hong Kong, hna, 00. [] S. Przybylsk, ahe and Memory Desgn: A Performane-Dreted Approah, Morgan Kaufmann Publshers, 990. [] A. Agarwal, Analyss of ahe Performane for Operatng Systems and Multprogrammng, Ph.D. thess, Stanford unversty, May 987. [3] R. Jn and G. Agrawal, Performane Predton for Random Wrte Redutons: A ase Study n Modelng Shared Memory Programs, Proeedngs of the 00 AM SIGMETRIS nternatonal onferene on Measurement and modelng of omputer systems, Marna Del Rey, alforna, p 7-8

JSM Survey Research Methods Section. Is it MAR or NMAR? Michail Sverchkov

JSM Survey Research Methods Section. Is it MAR or NMAR? Michail Sverchkov JSM 2013 - Survey Researh Methods Seton Is t MAR or NMAR? Mhal Sverhkov Bureau of Labor Statsts 2 Massahusetts Avenue, NE, Sute 1950, Washngton, DC. 20212, Sverhkov.Mhael@bls.gov Abstrat Most methods that

More information

technische universiteit eindhoven Analysis of one product /one location inventory control models prof.dr. A.G. de Kok 1

technische universiteit eindhoven Analysis of one product /one location inventory control models prof.dr. A.G. de Kok 1 TU/e tehnshe unverstet endhoven Analyss of one produt /one loaton nventory ontrol models prof.dr. A.G. de Kok Aknowledgements: I would lke to thank Leonard Fortun for translatng ths ourse materal nto Englsh

More information

The corresponding link function is the complementary log-log link The logistic model is comparable with the probit model if

The corresponding link function is the complementary log-log link The logistic model is comparable with the probit model if SK300 and SK400 Lnk funtons for bnomal GLMs Autumn 08 We motvate the dsusson by the beetle eample GLMs for bnomal and multnomal data Covers the followng materal from hapters 5 and 6: Seton 5.6., 5.6.3,

More information

Machine Learning: and 15781, 2003 Assignment 4

Machine Learning: and 15781, 2003 Assignment 4 ahne Learnng: 070 and 578, 003 Assgnment 4. VC Dmenson 30 onts Consder the spae of nstane X orrespondng to all ponts n the D x, plane. Gve the VC dmenson of the followng hpothess spaes. No explanaton requred.

More information

Controller Design for Networked Control Systems in Multiple-packet Transmission with Random Delays

Controller Design for Networked Control Systems in Multiple-packet Transmission with Random Delays Appled Mehans and Materals Onlne: 03-0- ISSN: 66-748, Vols. 78-80, pp 60-604 do:0.408/www.sentf.net/amm.78-80.60 03 rans eh Publatons, Swtzerland H Controller Desgn for Networed Control Systems n Multple-paet

More information

Queueing Networks II Network Performance

Queueing Networks II Network Performance Queueng Networks II Network Performance Davd Tpper Assocate Professor Graduate Telecommuncatons and Networkng Program Unversty of Pttsburgh Sldes 6 Networks of Queues Many communcaton systems must be modeled

More information

FAULT DETECTION AND IDENTIFICATION BASED ON FULLY-DECOUPLED PARITY EQUATION

FAULT DETECTION AND IDENTIFICATION BASED ON FULLY-DECOUPLED PARITY EQUATION Control 4, Unversty of Bath, UK, September 4 FAUL DEECION AND IDENIFICAION BASED ON FULLY-DECOUPLED PARIY EQUAION C. W. Chan, Hua Song, and Hong-Yue Zhang he Unversty of Hong Kong, Hong Kong, Chna, Emal:

More information

Complement of an Extended Fuzzy Set

Complement of an Extended Fuzzy Set Internatonal Journal of Computer pplatons (0975 8887) Complement of an Extended Fuzzy Set Trdv Jyot Neog Researh Sholar epartment of Mathemats CMJ Unversty, Shllong, Meghalaya usmanta Kumar Sut ssstant

More information

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS Avalable onlne at http://sck.org J. Math. Comput. Sc. 3 (3), No., 6-3 ISSN: 97-537 COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

More information

Performance Modeling of Hierarchical Memories

Performance Modeling of Hierarchical Memories Performne Modelng of Herrl Memores Mrwn Slemn, Lester Lpsky, Ksor Konwr Deprtment of omputer Sene nd Engneerng Unversty of onnetut Storrs, T 0669-55 Eml: {mrwn, lester, ksor}@engr.uonn.edu Abstrt As te

More information

BINARY LAMBDA-SET FUNCTION AND RELIABILITY OF AIRLINE

BINARY LAMBDA-SET FUNCTION AND RELIABILITY OF AIRLINE BINARY LAMBDA-SET FUNTION AND RELIABILITY OF AIRLINE Y. Paramonov, S. Tretyakov, M. Hauka Ra Tehnal Unversty, Aeronautal Insttute, Ra, Latva e-mal: yur.paramonov@mal.om serejs.tretjakovs@mal.om mars.hauka@mal.om

More information

STK4900/ Lecture 4 Program. Counterfactuals and causal effects. Example (cf. practical exercise 10)

STK4900/ Lecture 4 Program. Counterfactuals and causal effects. Example (cf. practical exercise 10) STK4900/9900 - Leture 4 Program 1. Counterfatuals and ausal effets 2. Confoundng 3. Interaton 4. More on ANOVA Setons 4.1, 4.4, 4.6 Supplementary materal on ANOVA Example (f. pratal exerse 10) How does

More information

3D Numerical Analysis for Impedance Calculation and High Performance Consideration of Linear Induction Motor for Rail-guided Transportation

3D Numerical Analysis for Impedance Calculation and High Performance Consideration of Linear Induction Motor for Rail-guided Transportation ADVANCED ELECTROMAGNETICS SYMPOSIUM, AES 13, 19 MARCH 13, SHARJAH UNITED ARAB EMIRATES 3D Numeral Analss for Impedane Calulaton and Hgh Performane Consderaton of Lnear Induton Motor for Ral-guded Transportaton

More information

Simulated Power of the Discrete Cramér-von Mises Goodness-of-Fit Tests

Simulated Power of the Discrete Cramér-von Mises Goodness-of-Fit Tests Smulated of the Cramér-von Mses Goodness-of-Ft Tests Steele, M., Chaselng, J. and 3 Hurst, C. School of Mathematcal and Physcal Scences, James Cook Unversty, Australan School of Envronmental Studes, Grffth

More information

Analysis of Discrete Time Queues (Section 4.6)

Analysis of Discrete Time Queues (Section 4.6) Analyss of Dscrete Tme Queues (Secton 4.6) Copyrght 2002, Sanjay K. Bose Tme axs dvded nto slots slot slot boundares Arrvals can only occur at slot boundares Servce to a job can only start at a slot boundary

More information

Structure and Drive Paul A. Jensen Copyright July 20, 2003

Structure and Drive Paul A. Jensen Copyright July 20, 2003 Structure and Drve Paul A. Jensen Copyrght July 20, 2003 A system s made up of several operatons wth flow passng between them. The structure of the system descrbes the flow paths from nputs to outputs.

More information

Department of Statistics University of Toronto STA305H1S / 1004 HS Design and Analysis of Experiments Term Test - Winter Solution

Department of Statistics University of Toronto STA305H1S / 1004 HS Design and Analysis of Experiments Term Test - Winter Solution Department of Statstcs Unversty of Toronto STA35HS / HS Desgn and Analyss of Experments Term Test - Wnter - Soluton February, Last Name: Frst Name: Student Number: Instructons: Tme: hours. Ads: a non-programmable

More information

Modeling Mobility-Assisted Data Collection in Wireless Sensor Networks

Modeling Mobility-Assisted Data Collection in Wireless Sensor Networks Modelng Moblty-Asssted Data Colleton n Wreless Sensor Networks Hsham M. Almasaed and Ahmed E. Kamal Dept. of Eletral and Computer Eng., Iowa State Unversty, Ames, IA 11, USA E-mal:{hsham,kamal}@astate.edu

More information

Instance-Based Learning and Clustering

Instance-Based Learning and Clustering Instane-Based Learnng and Clusterng R&N 04, a bt of 03 Dfferent knds of Indutve Learnng Supervsed learnng Bas dea: Learn an approxmaton for a funton y=f(x based on labelled examples { (x,y, (x,y,, (x n,y

More information

Clustering. CS4780/5780 Machine Learning Fall Thorsten Joachims Cornell University

Clustering. CS4780/5780 Machine Learning Fall Thorsten Joachims Cornell University Clusterng CS4780/5780 Mahne Learnng Fall 2012 Thorsten Joahms Cornell Unversty Readng: Mannng/Raghavan/Shuetze, Chapters 16 (not 16.3) and 17 (http://nlp.stanford.edu/ir-book/) Outlne Supervsed vs. Unsupervsed

More information

Phase Transition in Collective Motion

Phase Transition in Collective Motion Phase Transton n Colletve Moton Hefe Hu May 4, 2008 Abstrat There has been a hgh nterest n studyng the olletve behavor of organsms n reent years. When the densty of lvng systems s nreased, a phase transton

More information

Chapter 13: Multiple Regression

Chapter 13: Multiple Regression Chapter 13: Multple Regresson 13.1 Developng the multple-regresson Model The general model can be descrbed as: It smplfes for two ndependent varables: The sample ft parameter b 0, b 1, and b are used to

More information

Fork-Join program response time on multiprocessors with exchangeable join *

Fork-Join program response time on multiprocessors with exchangeable join * Wang et al. / J Zhejang Unv SCIECE A 26 7(6):927-936 927 Journal of Zhejang Unversty SCIECE A ISS 9-395 (Prnt); ISS 862-775 (Onlne) www.zju.edu.n/jzus; www.sprngerlnk.om E-mal: jzus@zju.edu.n Fork-Jon

More information

Heuristic Replica Placement Algorithms in Content Distribution Networks

Heuristic Replica Placement Algorithms in Content Distribution Networks 46 JOURAL OF ETWORK VOL 6 O 3 ARH 20 Heurst Repla Plaement Algorthms n ontent Dstrbuton etwors Jng un Graduate Unversty of hnese Aademy of enes Beng hna Emal: sas@yahoon uxang Gao Wenguo Yang and Zhpeng

More information

Exact Inference: Introduction. Exact Inference: Introduction. Exact Inference: Introduction. Exact Inference: Introduction.

Exact Inference: Introduction. Exact Inference: Introduction. Exact Inference: Introduction. Exact Inference: Introduction. Exat nferene: ntroduton Exat nferene: ntroduton Usng a ayesan network to ompute probabltes s alled nferene n general nferene nvolves queres of the form: E=e E = The evdene varables = The query varables

More information

Lecture Notes on Linear Regression

Lecture Notes on Linear Regression Lecture Notes on Lnear Regresson Feng L fl@sdueducn Shandong Unversty, Chna Lnear Regresson Problem In regresson problem, we am at predct a contnuous target value gven an nput feature vector We assume

More information

DOAEstimationforCoherentSourcesinBeamspace UsingSpatialSmoothing

DOAEstimationforCoherentSourcesinBeamspace UsingSpatialSmoothing DOAEstmatonorCoherentSouresneamspae UsngSpatalSmoothng YnYang,ChunruWan,ChaoSun,QngWang ShooloEletralandEletronEngneerng NanangehnologalUnverst,Sngapore,639798 InsttuteoAoustEngneerng NorthwesternPoltehnalUnverst,X

More information

CS-433: Simulation and Modeling Modeling and Probability Review

CS-433: Simulation and Modeling Modeling and Probability Review CS-433: Smulaton and Modelng Modelng and Probablty Revew Exercse 1. (Probablty of Smple Events) Exercse 1.1 The owner of a camera shop receves a shpment of fve cameras from a camera manufacturer. Unknown

More information

Prediction of the reliability of genomic breeding values for crossbred performance

Prediction of the reliability of genomic breeding values for crossbred performance Vandenplas et al. Genet Sel Evol 217 49:43 DOI 1.1186/s12711-17-318-1 Genets Seleton Evoluton RESERCH RTICLE Open ess Predton of the relablty of genom breedng values for rossbred performane Jéréme Vandenplas

More information

Credit Card Pricing and Impact of Adverse Selection

Credit Card Pricing and Impact of Adverse Selection Credt Card Prcng and Impact of Adverse Selecton Bo Huang and Lyn C. Thomas Unversty of Southampton Contents Background Aucton model of credt card solctaton - Errors n probablty of beng Good - Errors n

More information

Logic effort and gate sizing

Logic effort and gate sizing EEN454 Dgtal Integrated rcut Desgn Logc effort and gate szng EEN 454 Introducton hp desgners face a bewlderng arra of choces What s the best crcut topolog for a functon? How man stages of logc gve least

More information

Analytical calculation of adiabatic processes in real gases

Analytical calculation of adiabatic processes in real gases Journal of Physs: Conferene Seres PAPER OPEN ACCESS Analytal alulaton of adabat roesses n real gases o te ths artle: I B Amarskaja et al 016 J. Phys.: Conf. Ser. 754 11003 Related ontent - Shortuts to

More information

Statistical Evaluation of WATFLOOD

Statistical Evaluation of WATFLOOD tatstcal Evaluaton of WATFLD By: Angela MacLean, Dept. of Cvl & Envronmental Engneerng, Unversty of Waterloo, n. ctober, 005 The statstcs program assocated wth WATFLD uses spl.csv fle that s produced wth

More information

The optimal delay of the second test is therefore approximately 210 hours earlier than =2.

The optimal delay of the second test is therefore approximately 210 hours earlier than =2. THE IEC 61508 FORMULAS 223 The optmal delay of the second test s therefore approxmately 210 hours earler than =2. 8.4 The IEC 61508 Formulas IEC 61508-6 provdes approxmaton formulas for the PF for smple

More information

PHYSICS 212 MIDTERM II 19 February 2003

PHYSICS 212 MIDTERM II 19 February 2003 PHYSICS 1 MIDERM II 19 Feruary 003 Exam s losed ook, losed notes. Use only your formula sheet. Wrte all work and answers n exam ooklets. he aks of pages wll not e graded unless you so request on the front

More information

Outline. Clustering: Similarity-Based Clustering. Supervised Learning vs. Unsupervised Learning. Clustering. Applications of Clustering

Outline. Clustering: Similarity-Based Clustering. Supervised Learning vs. Unsupervised Learning. Clustering. Applications of Clustering Clusterng: Smlarty-Based Clusterng CS4780/5780 Mahne Learnng Fall 2013 Thorsten Joahms Cornell Unversty Supervsed vs. Unsupervsed Learnng Herarhal Clusterng Herarhal Agglomeratve Clusterng (HAC) Non-Herarhal

More information

UNIVERSITY OF TORONTO Faculty of Arts and Science. December 2005 Examinations STA437H1F/STA1005HF. Duration - 3 hours

UNIVERSITY OF TORONTO Faculty of Arts and Science. December 2005 Examinations STA437H1F/STA1005HF. Duration - 3 hours UNIVERSITY OF TORONTO Faculty of Arts and Scence December 005 Examnatons STA47HF/STA005HF Duraton - hours AIDS ALLOWED: (to be suppled by the student) Non-programmable calculator One handwrtten 8.5'' x

More information

Prediction of Solid Paraffin Precipitation Using Solid Phase Equation of State

Prediction of Solid Paraffin Precipitation Using Solid Phase Equation of State Predton of old Paraffn Preptaton Usng old Phase Equaton of tate Proeedngs of European Congress of Chemal Engneerng (ECCE-6) Copenhagen, 16- eptember 7 Predton of old Paraffn Preptaton Usng old Phase Equaton

More information

Improving the Performance of Fading Channel Simulators Using New Parameterization Method

Improving the Performance of Fading Channel Simulators Using New Parameterization Method Internatonal Journal of Eletrons and Eletral Engneerng Vol. 4, No. 5, Otober 06 Improvng the Performane of Fadng Channel Smulators Usng New Parameterzaton Method Omar Alzoub and Moheldn Wanakh Department

More information

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity

LINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 30 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 2 Remedes for multcollnearty Varous technques have

More information

Hopfield networks and Boltzmann machines. Geoffrey Hinton et al. Presented by Tambet Matiisen

Hopfield networks and Boltzmann machines. Geoffrey Hinton et al. Presented by Tambet Matiisen Hopfeld networks and Boltzmann machnes Geoffrey Hnton et al. Presented by Tambet Matsen 18.11.2014 Hopfeld network Bnary unts Symmetrcal connectons http://www.nnwj.de/hopfeld-net.html Energy functon The

More information

The Feynman path integral

The Feynman path integral The Feynman path ntegral Aprl 3, 205 Hesenberg and Schrödnger pctures The Schrödnger wave functon places the tme dependence of a physcal system n the state, ψ, t, where the state s a vector n Hlbert space

More information

Turbulence classification of load data by the frequency and severity of wind gusts. Oscar Moñux, DEWI GmbH Kevin Bleibler, DEWI GmbH

Turbulence classification of load data by the frequency and severity of wind gusts. Oscar Moñux, DEWI GmbH Kevin Bleibler, DEWI GmbH Turbulence classfcaton of load data by the frequency and severty of wnd gusts Introducton Oscar Moñux, DEWI GmbH Kevn Blebler, DEWI GmbH Durng the wnd turbne developng process, one of the most mportant

More information

Computation of Higher Order Moments from Two Multinomial Overdispersion Likelihood Models

Computation of Higher Order Moments from Two Multinomial Overdispersion Likelihood Models Computaton of Hgher Order Moments from Two Multnomal Overdsperson Lkelhood Models BY J. T. NEWCOMER, N. K. NEERCHAL Department of Mathematcs and Statstcs, Unversty of Maryland, Baltmore County, Baltmore,

More information

Module 9. Lecture 6. Duality in Assignment Problems

Module 9. Lecture 6. Duality in Assignment Problems Module 9 1 Lecture 6 Dualty n Assgnment Problems In ths lecture we attempt to answer few other mportant questons posed n earler lecture for (AP) and see how some of them can be explaned through the concept

More information

Design and Optimization of Fuzzy Controller for Inverse Pendulum System Using Genetic Algorithm

Design and Optimization of Fuzzy Controller for Inverse Pendulum System Using Genetic Algorithm Desgn and Optmzaton of Fuzzy Controller for Inverse Pendulum System Usng Genetc Algorthm H. Mehraban A. Ashoor Unversty of Tehran Unversty of Tehran h.mehraban@ece.ut.ac.r a.ashoor@ece.ut.ac.r Abstract:

More information

Stanford University CS359G: Graph Partitioning and Expanders Handout 4 Luca Trevisan January 13, 2011

Stanford University CS359G: Graph Partitioning and Expanders Handout 4 Luca Trevisan January 13, 2011 Stanford Unversty CS359G: Graph Parttonng and Expanders Handout 4 Luca Trevsan January 3, 0 Lecture 4 In whch we prove the dffcult drecton of Cheeger s nequalty. As n the past lectures, consder an undrected

More information

Brander and Lewis (1986) Link the relationship between financial and product sides of a firm.

Brander and Lewis (1986) Link the relationship between financial and product sides of a firm. Brander and Lews (1986) Lnk the relatonshp between fnanal and produt sdes of a frm. The way a frm fnanes ts nvestment: (1) Debt: Borrowng from banks, n bond market, et. Debt holders have prorty over a

More information

The calculation of ternary vapor-liquid system equilibrium by using P-R equation of state

The calculation of ternary vapor-liquid system equilibrium by using P-R equation of state The alulaton of ternary vapor-lqud syste equlbru by usng P-R equaton of state Y Lu, Janzhong Yn *, Rune Lu, Wenhua Sh and We We Shool of Cheal Engneerng, Dalan Unversty of Tehnology, Dalan 11601, P.R.Chna

More information

Horizontal mergers for buyer power. Abstract

Horizontal mergers for buyer power. Abstract Horzontal mergers for buyer power Ramon Faul-Oller Unverstat d'alaant Llus Bru Unverstat de les Illes Balears Abstrat Salant et al. (1983) showed n a Cournot settng that horzontal mergers are unproftable

More information

C/CS/Phy191 Problem Set 3 Solutions Out: Oct 1, 2008., where ( 00. ), so the overall state of the system is ) ( ( ( ( 00 ± 11 ), Φ ± = 1

C/CS/Phy191 Problem Set 3 Solutions Out: Oct 1, 2008., where ( 00. ), so the overall state of the system is ) ( ( ( ( 00 ± 11 ), Φ ± = 1 C/CS/Phy9 Problem Set 3 Solutons Out: Oct, 8 Suppose you have two qubts n some arbtrary entangled state ψ You apply the teleportaton protocol to each of the qubts separately What s the resultng state obtaned

More information

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur Analyss of Varance and Desgn of Exerments-I MODULE III LECTURE - 2 EXPERIMENTAL DESIGN MODELS Dr. Shalabh Deartment of Mathematcs and Statstcs Indan Insttute of Technology Kanur 2 We consder the models

More information

18.1 Introduction and Recap

18.1 Introduction and Recap CS787: Advanced Algorthms Scrbe: Pryananda Shenoy and Shjn Kong Lecturer: Shuch Chawla Topc: Streamng Algorthmscontnued) Date: 0/26/2007 We contnue talng about streamng algorthms n ths lecture, ncludng

More information

Maximizing Overlap of Large Primary Sampling Units in Repeated Sampling: A comparison of Ernst s Method with Ohlsson s Method

Maximizing Overlap of Large Primary Sampling Units in Repeated Sampling: A comparison of Ernst s Method with Ohlsson s Method Maxmzng Overlap of Large Prmary Samplng Unts n Repeated Samplng: A comparson of Ernst s Method wth Ohlsson s Method Red Rottach and Padrac Murphy 1 U.S. Census Bureau 4600 Slver Hll Road, Washngton DC

More information

Comparison of Regression Lines

Comparison of Regression Lines STATGRAPHICS Rev. 9/13/2013 Comparson of Regresson Lnes Summary... 1 Data Input... 3 Analyss Summary... 4 Plot of Ftted Model... 6 Condtonal Sums of Squares... 6 Analyss Optons... 7 Forecasts... 8 Confdence

More information

Lecture 4: November 17, Part 1 Single Buffer Management

Lecture 4: November 17, Part 1 Single Buffer Management Lecturer: Ad Rosén Algorthms for the anagement of Networs Fall 2003-2004 Lecture 4: November 7, 2003 Scrbe: Guy Grebla Part Sngle Buffer anagement In the prevous lecture we taled about the Combned Input

More information

Newsvendor Bounds and Heuristics for Serial Supply Chains with Regular and Expedited Shipping

Newsvendor Bounds and Heuristics for Serial Supply Chains with Regular and Expedited Shipping Newsvendor Bounds and Heursts for Seral Supply Chans wth egular and xpedted Shppng Sean X. Zhou, Xul Chao 2 Department of Systems ngneerng and ngneerng Management, The Chnese Unversty of Hong Kong, Shatn,

More information

Problem Set 9 Solutions

Problem Set 9 Solutions Desgn and Analyss of Algorthms May 4, 2015 Massachusetts Insttute of Technology 6.046J/18.410J Profs. Erk Demane, Srn Devadas, and Nancy Lynch Problem Set 9 Solutons Problem Set 9 Solutons Ths problem

More information

Correlation and Regression without Sums of Squares. (Kendall's Tau) Rudy A. Gideon ABSTRACT

Correlation and Regression without Sums of Squares. (Kendall's Tau) Rudy A. Gideon ABSTRACT Correlaton and Regson wthout Sums of Squa (Kendall's Tau) Rud A. Gdeon ABSTRACT Ths short pee provdes an ntroduton to the use of Kendall's τ n orrelaton and smple lnear regson. The error estmate also uses

More information

1 Matrix representations of canonical matrices

1 Matrix representations of canonical matrices 1 Matrx representatons of canoncal matrces 2-d rotaton around the orgn: ( ) cos θ sn θ R 0 = sn θ cos θ 3-d rotaton around the x-axs: R x = 1 0 0 0 cos θ sn θ 0 sn θ cos θ 3-d rotaton around the y-axs:

More information

Introduction to Continuous-Time Markov Chains and Queueing Theory

Introduction to Continuous-Time Markov Chains and Queueing Theory Introducton to Contnuous-Tme Markov Chans and Queueng Theory From DTMC to CTMC p p 1 p 12 1 2 k-1 k p k-1,k p k-1,k k+1 p 1 p 21 p k,k-1 p k,k-1 DTMC 1. Transtons at dscrete tme steps n=,1,2, 2. Past doesn

More information

Erratum: A Generalized Path Integral Control Approach to Reinforcement Learning

Erratum: A Generalized Path Integral Control Approach to Reinforcement Learning Journal of Machne Learnng Research 00-9 Submtted /0; Publshed 7/ Erratum: A Generalzed Path Integral Control Approach to Renforcement Learnng Evangelos ATheodorou Jonas Buchl Stefan Schaal Department of

More information

Implementation of α-qss Stiff Integration Methods for Solving the Detailed Combustion Chemistry

Implementation of α-qss Stiff Integration Methods for Solving the Detailed Combustion Chemistry Proeedngs of the World Congress on Engneerng 2007 Vol II Implementaton of α-qss Stff Integraton Methods for Solvng the Detaled Combuston Chemstry Shafq R. Quresh and Robert Prosser Abstrat Implt methods

More information

Midterm Examination. Regression and Forecasting Models

Midterm Examination. Regression and Forecasting Models IOMS Department Regresson and Forecastng Models Professor Wllam Greene Phone: 22.998.0876 Offce: KMC 7-90 Home page: people.stern.nyu.edu/wgreene Emal: wgreene@stern.nyu.edu Course web page: people.stern.nyu.edu/wgreene/regresson/outlne.htm

More information

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems Numercal Analyss by Dr. Anta Pal Assstant Professor Department of Mathematcs Natonal Insttute of Technology Durgapur Durgapur-713209 emal: anta.bue@gmal.com 1 . Chapter 5 Soluton of System of Lnear Equatons

More information

Chapter Newton s Method

Chapter Newton s Method Chapter 9. Newton s Method After readng ths chapter, you should be able to:. Understand how Newton s method s dfferent from the Golden Secton Search method. Understand how Newton s method works 3. Solve

More information

MAE140 - Linear Circuits - Winter 16 Final, March 16, 2016

MAE140 - Linear Circuits - Winter 16 Final, March 16, 2016 ME140 - Lnear rcuts - Wnter 16 Fnal, March 16, 2016 Instructons () The exam s open book. You may use your class notes and textbook. You may use a hand calculator wth no communcaton capabltes. () You have

More information

FREQUENCY DISTRIBUTIONS Page 1 of The idea of a frequency distribution for sets of observations will be introduced,

FREQUENCY DISTRIBUTIONS Page 1 of The idea of a frequency distribution for sets of observations will be introduced, FREQUENCY DISTRIBUTIONS Page 1 of 6 I. Introducton 1. The dea of a frequency dstrbuton for sets of observatons wll be ntroduced, together wth some of the mechancs for constructng dstrbutons of data. Then

More information

Lecture 2: Prelude to the big shrink

Lecture 2: Prelude to the big shrink Lecture 2: Prelude to the bg shrnk Last tme A slght detour wth vsualzaton tools (hey, t was the frst day... why not start out wth somethng pretty to look at?) Then, we consdered a smple 120a-style regresson

More information

Interval Valued Neutrosophic Soft Topological Spaces

Interval Valued Neutrosophic Soft Topological Spaces 8 Interval Valued Neutrosoph Soft Topologal njan Mukherjee Mthun Datta Florentn Smarandah Department of Mathemats Trpura Unversty Suryamannagar gartala-7990 Trpura Indamal: anjan00_m@yahooon Department

More information

Lecture 4: Universal Hash Functions/Streaming Cont d

Lecture 4: Universal Hash Functions/Streaming Cont d CSE 5: Desgn and Analyss of Algorthms I Sprng 06 Lecture 4: Unversal Hash Functons/Streamng Cont d Lecturer: Shayan Oves Gharan Aprl 6th Scrbe: Jacob Schreber Dsclamer: These notes have not been subjected

More information

Gravity Drainage Prior to Cake Filtration

Gravity Drainage Prior to Cake Filtration 1 Gravty Dranage Pror to ake Fltraton Sott A. Wells and Gregory K. Savage Department of vl Engneerng Portland State Unversty Portland, Oregon 97207-0751 Voe (503) 725-4276 Fax (503) 725-4298 ttp://www.e.pdx.edu/~wellss

More information

Section 8.3 Polar Form of Complex Numbers

Section 8.3 Polar Form of Complex Numbers 80 Chapter 8 Secton 8 Polar Form of Complex Numbers From prevous classes, you may have encountered magnary numbers the square roots of negatve numbers and, more generally, complex numbers whch are the

More information

x = , so that calculated

x = , so that calculated Stat 4, secton Sngle Factor ANOVA notes by Tm Plachowsk n chapter 8 we conducted hypothess tests n whch we compared a sngle sample s mean or proporton to some hypotheszed value Chapter 9 expanded ths to

More information

On the Interplay of Dynamic Voltage Scaling and Dynamic Power Management in Real-Time Embedded Applications

On the Interplay of Dynamic Voltage Scaling and Dynamic Power Management in Real-Time Embedded Applications On the Interplay of Dynam Voltage Salng and Dynam Power Management n Real-Tme Embedded Applatons Vnay Devadas, Hakan Aydn Dept. of Computer Sene, George Mason Unversty Farfax, VA, USA {vdevadas,aydn}@s.gmu.edu

More information

For now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results.

For now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results. Neural Networks : Dervaton compled by Alvn Wan from Professor Jtendra Malk s lecture Ths type of computaton s called deep learnng and s the most popular method for many problems, such as computer vson

More information

1. Inference on Regression Parameters a. Finding Mean, s.d and covariance amongst estimates. 2. Confidence Intervals and Working Hotelling Bands

1. Inference on Regression Parameters a. Finding Mean, s.d and covariance amongst estimates. 2. Confidence Intervals and Working Hotelling Bands Content. Inference on Regresson Parameters a. Fndng Mean, s.d and covarance amongst estmates.. Confdence Intervals and Workng Hotellng Bands 3. Cochran s Theorem 4. General Lnear Testng 5. Measures of

More information

Lecture 10 Support Vector Machines II

Lecture 10 Support Vector Machines II Lecture 10 Support Vector Machnes II 22 February 2016 Taylor B. Arnold Yale Statstcs STAT 365/665 1/28 Notes: Problem 3 s posted and due ths upcomng Frday There was an early bug n the fake-test data; fxed

More information

Calculation of time complexity (3%)

Calculation of time complexity (3%) Problem 1. (30%) Calculaton of tme complexty (3%) Gven n ctes, usng exhaust search to see every result takes O(n!). Calculaton of tme needed to solve the problem (2%) 40 ctes:40! dfferent tours 40 add

More information

CHAPTER III Neural Networks as Associative Memory

CHAPTER III Neural Networks as Associative Memory CHAPTER III Neural Networs as Assocatve Memory Introducton One of the prmary functons of the bran s assocatve memory. We assocate the faces wth names, letters wth sounds, or we can recognze the people

More information

Grover s Algorithm + Quantum Zeno Effect + Vaidman

Grover s Algorithm + Quantum Zeno Effect + Vaidman Grover s Algorthm + Quantum Zeno Effect + Vadman CS 294-2 Bomb 10/12/04 Fall 2004 Lecture 11 Grover s algorthm Recall that Grover s algorthm for searchng over a space of sze wors as follows: consder the

More information

Negative Binomial Regression

Negative Binomial Regression STATGRAPHICS Rev. 9/16/2013 Negatve Bnomal Regresson Summary... 1 Data Input... 3 Statstcal Model... 3 Analyss Summary... 4 Analyss Optons... 7 Plot of Ftted Model... 8 Observed Versus Predcted... 10 Predctons...

More information

Homework Math 180: Introduction to GR Temple-Winter (3) Summarize the article:

Homework Math 180: Introduction to GR Temple-Winter (3) Summarize the article: Homework Math 80: Introduton to GR Temple-Wnter 208 (3) Summarze the artle: https://www.udas.edu/news/dongwthout-dark-energy/ (4) Assume only the transformaton laws for etors. Let X P = a = a α y = Y α

More information

Adaptive Multilayer Neural Network Control of Blood Pressure

Adaptive Multilayer Neural Network Control of Blood Pressure Proeedng of st Internatonal Symposum on Instrument Sene and Tenology. ISIST 99. P4-45. 999. (ord format fle: ISIST99.do) Adaptve Multlayer eural etwork ontrol of Blood Pressure Fe Juntao, Zang bo Department

More information

Development of a computer model for long-throated flumes based on manning equation and different side slopes of trapezoidal channels

Development of a computer model for long-throated flumes based on manning equation and different side slopes of trapezoidal channels Iran Agrultural Researh (08) 37() Development of a omputer model for long-throated flumes based on mannng equaton and dfferent sde slopes of trapezodal hannels M. Mahbod * and Sh. Zand-Parsa Department

More information

Lecture 3. Ax x i a i. i i

Lecture 3. Ax x i a i. i i 18.409 The Behavor of Algorthms n Practce 2/14/2 Lecturer: Dan Spelman Lecture 3 Scrbe: Arvnd Sankar 1 Largest sngular value In order to bound the condton number, we need an upper bound on the largest

More information

A Theorem of Mass Being Derived From Electrical Standing Waves (As Applied to Jean Louis Naudin's Test)

A Theorem of Mass Being Derived From Electrical Standing Waves (As Applied to Jean Louis Naudin's Test) A Theorem of Mass Beng Derved From Eletral Standng Waves (As Appled to Jean Lous Naudn's Test) - by - Jerry E Bayles Aprl 4, 000 Ths paper formalzes a onept presented n my book, "Eletrogravtaton As A Unfed

More information

A Robust Method for Calculating the Correlation Coefficient

A Robust Method for Calculating the Correlation Coefficient A Robust Method for Calculatng the Correlaton Coeffcent E.B. Nven and C. V. Deutsch Relatonshps between prmary and secondary data are frequently quantfed usng the correlaton coeffcent; however, the tradtonal

More information

Cokriging Partial Grades - Application to Block Modeling of Copper Deposits

Cokriging Partial Grades - Application to Block Modeling of Copper Deposits Cokrgng Partal Grades - Applcaton to Block Modelng of Copper Deposts Serge Séguret 1, Julo Benscell 2 and Pablo Carrasco 2 Abstract Ths work concerns mneral deposts made of geologcal bodes such as breccas

More information

ALGEBRAIC SCHUR COMPLEMENT APPROACH FOR A NON LINEAR 2D ADVECTION DIFFUSION EQUATION

ALGEBRAIC SCHUR COMPLEMENT APPROACH FOR A NON LINEAR 2D ADVECTION DIFFUSION EQUATION st Annual Internatonal Interdsplnary Conferene AIIC 03 4-6 Aprl Azores Portugal - Proeedngs- ALGEBRAIC SCHUR COMPLEMENT APPROACH FOR A NON LINEAR D ADVECTION DIFFUSION EQUATION Hassan Belhad Professor

More information

Using Artificial Neural Networks and Support Vector Regression to Model the Lyapunov Exponent

Using Artificial Neural Networks and Support Vector Regression to Model the Lyapunov Exponent Usng Artfal Neural Networks and Support Vetor Regresson to Model the Lyapunov Exponent Abstrat: Adam Maus* Aprl 3, 009 Fndng the salent patterns n haot data has been the holy gral of Chaos Theory. Examples

More information

Tracking with Kalman Filter

Tracking with Kalman Filter Trackng wth Kalman Flter Scott T. Acton Vrgna Image and Vdeo Analyss (VIVA), Charles L. Brown Department of Electrcal and Computer Engneerng Department of Bomedcal Engneerng Unversty of Vrgna, Charlottesvlle,

More information

Voltammetry. Bulk electrolysis: relatively large electrodes (on the order of cm 2 ) Voltammetry:

Voltammetry. Bulk electrolysis: relatively large electrodes (on the order of cm 2 ) Voltammetry: Voltammetry varety of eletroanalytal methods rely on the applaton of a potental funton to an eletrode wth the measurement of the resultng urrent n the ell. In ontrast wth bul eletrolyss methods, the objetve

More information

A Simple Inventory System

A Simple Inventory System A Smple Inventory System Lawrence M. Leems and Stephen K. Park, Dscrete-Event Smulaton: A Frst Course, Prentce Hall, 2006 Hu Chen Computer Scence Vrgna State Unversty Petersburg, Vrgna February 8, 2017

More information

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4)

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4) I. Classcal Assumptons Econ7 Appled Econometrcs Topc 3: Classcal Model (Studenmund, Chapter 4) We have defned OLS and studed some algebrac propertes of OLS. In ths topc we wll study statstcal propertes

More information

EEE 241: Linear Systems

EEE 241: Linear Systems EEE : Lnear Systems Summary #: Backpropagaton BACKPROPAGATION The perceptron rule as well as the Wdrow Hoff learnng were desgned to tran sngle layer networks. They suffer from the same dsadvantage: they

More information

GEL 446: Applied Environmental Geology

GEL 446: Applied Environmental Geology GE 446: ppled Envronmental Geology Watershed Delneaton and Geomorphology Watershed Geomorphology Watersheds are fundamental geospatal unts that provde a physal and oneptual framewor wdely used by sentsts,

More information

Numerical Heat and Mass Transfer

Numerical Heat and Mass Transfer Master degree n Mechancal Engneerng Numercal Heat and Mass Transfer 06-Fnte-Dfference Method (One-dmensonal, steady state heat conducton) Fausto Arpno f.arpno@uncas.t Introducton Why we use models and

More information

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix Lectures - Week 4 Matrx norms, Condtonng, Vector Spaces, Lnear Independence, Spannng sets and Bass, Null space and Range of a Matrx Matrx Norms Now we turn to assocatng a number to each matrx. We could

More information

Horizontal Mergers for Buyer Power

Horizontal Mergers for Buyer Power Horzontal Mergers for Buyer Power Lluís Bru a and Ramon Faulí-Oller b* Marh, 004 Abstrat: Salant et al. (1983) showed n a Cournot settng that horzontal mergers are unproftable beause outsders reat by nreasng

More information