Procrastination Scheduling for Fixed-Priority Tasks with Preemption Thresholds

Size: px
Start display at page:

Download "Procrastination Scheduling for Fixed-Priority Tasks with Preemption Thresholds"

Transcription

1 Procrastnaton Schedulng for Fxed-Prorty Tasks wth Preempton Thresholds XaoChuan He, Yan Ja Insttute of Network Technology and Informaton Securty School of Computer Scence Natonal Unversty of Defense Technology Changsha, Chna Abstract. Dynamc Voltage Scalng (DVS), whch adjusts the clock speed and supply voltage dynamcally, s an effectve technque n reducng the energy consumpton of embedded real-tme systems. However, the longer a job executes, the more energy n the leakage current the devce/processor consumes for the job. Procrastnaton schedulng, where task executon can be delayed to maxmze the duraton of dle ntervals by keepng the processor n a sleep/shutdown state even f there are pendng tasks wthn the tmng constrants mposed by performance requrements, has been proposed to mnmze leakage energy dran. Ths paper targets energy-effcent fxed-prorty wth preempton threshold schedulng for perodc real-tme tasks on a unprocessor DVS system wth non-neglgble leakage power consumpton. We propose a two-phase algorthm. In the frst phase, the executon speed,.e., the supply voltage of each task are determned by applyng off-lne algorthms, and n the second phase, the procrastnaton length of each task s derved by applyng on-lne smulated work-demand tme analyss, and thus the tme moment to turn on/off the system s determned on the fly. A seres of smulaton experments was evaluated for the performance of our algorthms. The results show that our proposed algorthms can derve energy-effcent schedules. 1 Introducton Low power utlzaton has been an mportant ssue for hardware manufacturng for next-generaton portable, scalable, and sophstcated embedded systems. To reduce the power consumpton wthout the sacrfce of performance, archtectural technques have been proposed to dynamcally trade the performance and power consumpton. Dynamc Voltage Scalng (DVS), whch adjusts the supply voltage and ts correspondng clock frequency dynamcally, s one of the most effectve low-power desgn technque for embedded real-tme systems. Snce the energy consumpton of CMOS crcuts has a quadratc dependency on the supply voltage, lowerng the supply voltage s one of the most effectve ways of reducng the energy consumpton. In many real-tme applcatons, average or worst-case task response tme s an mportant non-functonal desgn requrement of the system. For example, to mantan the system stablty, many embedded real-tme systems must complete the tasks before

2 ther deadlnes. For real-tme systems targetng commercal varable voltage mcroprocessors, snce lowerng the supply voltage also decreases the maxmum achevable clock speed [1], energy-effcent task schedulng s to reduce supply voltage dynamcally to the lowest possble level whle satsfyng the tasks tmng constrants. In the past decade, energy-effcent task schedulng wth varous deadlne constrants receved extensve attenton, especally for the mnmzaton of the energy consumpton of the dynamc voltage scalng part n a unprocessor envronment [2]. Recently, researchers have started explorng energy-effcent schedulng wth the consderatons of leakage current snce the power consumpton resultng from leakage current s comparable to the dynamc power dsspaton [3]. To reduce the energy consumpton resultng from leakage current, a system mght be turned off (to enter a dormant mode). For perodc real-tme tasks, Jejurkar et al. [4] and Lee et al. [5] proposed energy-effcent schedulng on a unprocessor by procrastnaton schedulng to decde when to turn off the system. Jejurkar and Gupta [3] then further consdered real-tme tasks that mght complete earler than ts worst-case estmaton by extendng the algorthms presented n [4]. Fxed-prorty preemptve (FPP) schedulng algorthms and fxed-prorty non-preemptve (FPNP) schedulng algorthms are two mportant classes of real-tme schedulng algorthms. To obtan the benefts of both FPP and FPNP algorthms, there are several other algorthms tryng to fll the gap between them. The fxed-prorty wth preempton threshold (FPPT) schedulng algorthm [6] s one of them. Under FPPT, each task has a par of prortes: regular prorty and preempton threshold, where the preempton threshold of a task s hgher than or equal to ts regular prorty. The preempton threshold represents the tasks runnng-tme preempton prorty level. It prevents the preempton of the task from other tasks, unless the preemptng tasks prorty s hgher than the preempton threshold of the current runnng task. Saksena and Wang have shown that task sets scheduled wth FPPT can have sgnfcant schedulablty mprovements over task set usng fxed prortes [6]. Ths paper consders energy-effcent FPPT schedulng of perodc real-tme tasks on a unprocessor whose dynamc voltage scalng porton mght be turned off for further energy savng. We further combne procrastnaton schedulng wth dynamc voltage scalng to mnmze the total statc and dynamc energy consumpton of the system. An on-lne algorthm was developed to calculate the respectve procrastnaton nterval for each task. A seres of smulaton experments was also evaluated for the performance of our algorthms. The results show that our proposed algorthms can derve energyeffcent schedules. The rest of ths paper s organzed as follows: Secton 2 defnes the leakage-aware energy-effcent FPPT schedulng problem n a unprocessor system. Prelmnary results are shown n Secton 2. The proposed algorthms are n Secton 3. Expermental results for the performance evaluaton of the proposed algorthms are presented n Secton 4. Secton 5 s the concluson.

3 2 System Model 2.1 Task Model Ths study deals wth the fxed prorty preemptve schedulng of tasks n a real-tme systems wth hard constrants,.e., systems n whch the respect of tme constrants s mandatory. The actvtes of the system are modeled by perodc tasks. The model of the system s defned by a task set T of cardnalty n, T = {τ 1, τ 2,..., τ n }. The j th job of task τ s denoted as J,j. The ndex, j, for jobs of a task s started from zero. A perodc task τ s characterzed by a 3-tuple (C, T, D ) where each request of τ, called nstance, has an executon CPU cycles (denoted as C ), and a relatve deadlne (denoted as D ). T tme unts separate two consecutve nstances of τ (hence T s the perod of the task). Gven a set T of n tasks, the hyper-perod of T, denoted by L, s defned so that L/T s an nteger for any task τ n T. The number of jobs n the hyper-perod of task τ s L/T. For example, L s the least common multple (LCM) of the perods of tasks n T when the perods of tasks are all nteger numbers. We focus on the case that all of the tasks arrve at tme 0. We also assocate wth each task τ a unque prorty {1, 2,..., n} such that contenton for resources s resolved n favor of the job wth the hghest prorty that s ready to run. The analyss presented n secton 3 uses the concept of busy and dle perods [7]. These are defned as follows: A level- busy perod s a contnuous tme nterval durng whch the notonal run-queue contans one or more tasks of actve prorty level or hgher. Smlarly, a level- dle perod s a tme nterval durng whch the run-queue s free of level or hgher prorty tasks. We note that the run queue may become momentarly free of level- tasks, when one tasks completes and another s released. Ths appears n our formulaton as an dle perod of zero length. 2.2 Power Consumpton and Executon Models We explore energy-effcent schedulng on a dynamc voltage scalng (DVS) processor. The power consumpton s contrbuted by the dynamc power consumpton resultng from the chargng and dschargng of gates on the CMOS crcuts and the statc power consumpton resultng from leakage current. The dynamc power consumpton P d of the dynamc voltage scalng part of the processor s a functon of the adopted processor speed f: P d = C eff V 2 DD f (1) f = αk (V DD V TH ) α V DD (2) where k s a devce related parameter, V TH s the threshold voltage, C eff s the effectve swtchng capactance per cycle and α ranges from 2 to 1.2 dependng on the devce technology. Snce power vares lnearly wth the clock speed and the square

4 of the voltage, adjustng both can produce cubc power reductons, at least n theory. The statc power consumpton P s of the system comes from the leakage current of the processor, system I/O devces, and RAM. It mght be modeled as a nonnegatve constant, as n [8], or a lnear functon of the supply voltage (a sub-lnear functon of the executon speed) [9], [3], [4], [10]. The power consumpton of processor s denoted by P, whch s the sum of the dynamc and statc power consumpton. We consder systems n whch P(f) s a convex and ncreasng functon, and P(f)/f s a convex functon, smlarly to [11], [4]. Recent processors support multple varable voltage and frequency levels for energy effcent operaton of the system. Let the avalable frequences be {FLK 1, FLK 2,..., FLK s } n ncreasng order of frequency and the correspondng voltage levels be {v 1, v 2,..., v s }. We assume that the CPU speed f of task τ can be changed between a mnmum speed FLK 1 (mnmum supply voltage necessary to keep the system functonal) and a maxmum speed FLK s. In our framework, the voltage/speed changes take place only at context swtch tme and whle state savng nstructons execute. If not neglgble, the voltage change overhead can be ncorporated nto the worst-case workload of each task. The system could enter the dormant mode (or be turned off) whenever needed. The power consumpton of the system s treated as 0 when t s n the dormant mode [8] by scalng the statc power consumpton. We consder systems that could be turned on/off at nstant. When needed, turnng the system off mght further reduce the energy consumpton. The energy consumpton to turn off the system s assumed to be neglgble, but t mght requre addtonal energy to turn on the system [12]. We denote E sw as the energy of the swtchng overhead from the dormant mode to the actve mode. For the rest of ths paper, we say the system s dle at tme nstant t, f the processor does not execute any task at tme nstant t. When the system s actve and dle, the processor executes NOP nstructons and must be at processor speed FLK 1 to mnmze the energy consumpton. Let P I be the power consumpton when the system s dle and actve, where P I = P(FLK 1 ). 2.3 Crtcal Speed The crtcal speed ˆf s defned as the avalable speed of the processor to execute a cycle wth the mnmum energy consumpton. Because of the convexty of P(f), executng at a common speed for a CPU cycle mnmzes the energy consumpton. Hence, the energy consumpton to execute a CPU cycle at speed f s P(f)/f. Snce the power consumpton functon P(f) s a convex and ncreasng functon, where P(f)/f s merely a convex functon. P(f)/f s mnmzed when f s equal to f, wth d(p(f )/f ) df = 0. As a result, to mnmze the executon energy consumpton of T, we do not have to consder schedules that execute jobs at any lower speed than f snce we could execute jobs at speed f wth lower energy consumpton and less executon tme. If f s between FLK 1 and FLK s, we know that ˆf s f. If f s less than FLK 1, ˆf s set to FLK 1 to satsfy the hardware constrant. Smlarly, f f s greater than FLK s, ˆf s set to FLK s, and jobs are executed at FLK s to mnmze the energy consumpton. As a result, ˆf s mn{max{f, FLK 1 }, FLK s }. Executng a job of task τ at any speed less than ˆf

5 would ether consume more energy than that at ˆf wth more executon tme or volate the speed constrant. We assume that f could be obtaned effcently or pre-determned as a specfed parameter n the nput. 2.4 Problem defnton The problem consdered n ths paper s as follows: Defnton 1. (Leakage-Aware Energy-Effcent Schedulng for FPPT, LAEES-FPPT) Consder a set T of n ndependent tasks ready at tme 0. Each perodc task τ T s assocated wth a computaton requrement equal to C CPU-cycles and ts perod T, where the relatve deadlne of τ s equal to D. And each task τ T s assgned wth a unque prorty {1, 2,..., n} and a preempton threshold γ {1, 2,..., n} (γ ), where s used to compete for processor and gamma s used to protect τ from unnecessary task preemptons after τ starts. The power consumpton functon P(f) s a convex and ncreasng functon, whle P(f)/f s merely a convex functon. The processor s wth a dscrete spectrum of the avalable speeds n [FLK 1, FLK s ]. The energy of the swtchng overhead from the dormant mode to the actve mode of a system s E sw, and the power consumpton when the system s actve and dle s P I. The problem s to mnmze the energy consumpton n the hyper-perod L of tasks n T n the schedulng of fxed-prorty tasks wth preempton thresholds n T wthout mssng the tmng constrants. A schedule of a task set T s an assgnment of the avalable processor speeds for each correspondng task executon, where the job arrvals of each task τ T satsfy ts tmng constrant D. A schedule s feasble f no job msses ts deadlne. A schedule s optmal for the LAEES-FPPT problem, f t s feasble, and ts energy consumpton s the mnmum among all feasble schedules. For the rest of ths paper, let S be an optmal schedule for T. For a schedule, an dle nterval s a maxmal nterval when the system s dle, whle an executon nterval s a maxmal nterval when the processor executes some jobs. The system mght be turned off or be at the actve mode n an dle nterval, whle the system s actve n an executon nterval. For any set X, let X be the cardnalty of the set. For example, I S s the number of dle ntervals n schedule S n (0, L]. If the dle nterval s greater than E sw /P I, turnng off the system s worthwhle. Let t θ be the threshold dle nterval E sw /P I. If the dle nterval s greater than t θ, the longer the dle nterval s, the more the energy saved by turnng off the system. The energy consumpton of a schedule S, denoted as E(S), conssts of two parts: the executon energy consumpton φ(s) and the dle energy consumpton ε(s). The executon energy consumpton s the sum of the energy consumpton of the executons of jobs n S n the tme nterval (0, L]. The dle energy consumpton s the sum of the energy consumpton n the ntervals n (0, L] n whch the system does not execute any job. Let υ(t, S) be the speed at tme nstant t n schedule S. The executon energy consumpton φ(s) n E(S) s L P(υ(t, S))dt. The dle energy consumpton ε(s) n 0 E(S) s the summaton of E sw tmes the number of nstances that the system s turned from the dormant mode to the actve mode and P I tmes the total nterval length that the system s dle and actve n (0, L].

6 3 Proposed Algorthms Ths secton presents a two-phase algorthm for perodc real-tme tasks. The algorthms determne, n the frst phase, the executon speed,.e., the supply voltage, of each task, and n the second phase the moment to turn on/off the system on the fly. 3.1 An On-lne Procrastnaton Algorthm to Mnmze the Energy Leakage: LA-FPPT Let S e be the resultng FPPT schedule by applyng some off-lne dvs algorthms / [13]. For brevty, let C be the executon tme of a job of task τ n S e, C = C f opt. The frst phase of the proposed algorthm [13] decdes the executon speed of tasks n T to meet the tmng constrants and mnmze the executon energy consumpton. The second phase s to reduce the dle energy consumpton by turnng the system off on the fly. The dea behnd schedulng on the fly s to lengthen and aggregate the dle ntervals so that the resultng dle tme s long enough to turn off the system. The determnaton of dle ntervals can be done by procrastnatng the arrval tme of the next job to the system, as n [4],[14] for EDF schedulng, and n [9],[11] for fxed-prorty schedulng. In [9],[4] procrastnaton s done by computng the maxmum procrastnaton ntervals of all of the tasks n T based on the system utlzaton, whle the dle ntervals n [14],[11] are determned by procrastnatng the remanng jobs as late as possble. In ths secton, we proposes an on-lne smulated work-demand analyss approach to the determnaton of dle ntervals. If a job completes at tme nstant t, and the ready queue s empty, we have to decde whether the system should be turned off or dle. Let r (t) be the arrval tme of the next job of τ for any τ n T arrved after tme nstant t, t.e., r (t) = T. Let d (t) be the next deadlne on an nvocaton of task τ after T tme nstant t,.e., d (t) = r (t) + D. Our formulaton stems from consderng the schedulablty of each fxed-prorty task wth preempton thresholds at tme nstant t. We focus on fndng the maxmum (t), whch may be stolen at prorty level, durng the nterval [t, t+d (t)), whlst guaranteeng that task τ meets ts deadlne. (Note, Sπ max (t) amount of dle nterval, S max may not actually be avalable for dle due to the constrants on hard deadlne tasks wth prortes lower than. We return to ths pont later). To guarantee that task τ wll meet ts deadlne, we need to analyze the worst case scenaro from tme t onwards. We therefore assume that all tasks τ j are re-nvoked at ther earlest possble next release r j (t) and subsequently wth a perod of T j. (t), t s nstructve to vew the nterval [t, t + d (t)) as comprsng a number of level- busy and dle perods. Any level- dle tme between the completon of task τ and ts deadlne could be swapped for task τ s procrastnaton nterval Z wthout causng the deadlne to be mssed. Hence the maxmum procrastnaton nterval Z whch may be stolen s In attemptng to determne the maxmum guaranteed dle tme, S max equal to the total level- dle tme n the nterval. We use ths result to calculate S max (t). We frst derve equaton 3 usng technques gven n [15]. Although the ready queue s empty at tme nstant t, two components stll determne the extent of the busy perod under the nfluence of procrastnaton schedulng:

7 1. For the task τ k wth prorty π k < < γ k, τ k s released workload just before the start of busy perod 2. For the task τ j wth prorty π j > γ, τ j s released workload durng the busy perod The second component mples a recursve defnton. As the processng released ncreases monotoncally wth the length of the busy perod, a recurrence relaton can be used to fnd w (t): w m+1 (t) = S π (t) + max k,π k <<γ k C k + j,π j> ( w m π (t) x j (t) T j C j ) The term S π (t) represents the begnnng of level- dle tme from tme t. The recurrence relaton begns wth wπ 0 (t) = 0 and ends when wπ m+1 (t) = wπ m (t) or wπ m+1 (t) > d (t). Proof of convergence follows from analyss of smlar recurrence relatons by Audsley et al [15]. The fnal value of w π (t) defnes the length of the busy perod. Alternatvely, we may vew t+w π (t) as defnng the start of a level- dle tme. Gven the start of a level- dle tme, wthn the nterval [t, t + d (t)), the end of the dle tme, whch may be converted to procrastnaton nterval of task τ, occurs ether at the next release of a task τ j wth prorty π j > or at the end of the nterval. Equaton 4 gves the length, l (t, w π (t)), of the level- dle tme. l (t, w π (t)) = mn [ ] d (t) w( (t), ) wπ (t) r mn j(t) j,π T j γ j T j + r j (t) w π (t) where the term d (t) w π (t) means that the end of level- dle tme come about wπ (t) r at the end of [t, t + d (t)), the term j(t) T j + r j (t) descrbe the workload T j contrbuted by task τ j n the level- busy perod, whose length s denoted by w π (t). Combnng equatons 3 and 4, our method for determnng the maxmum dle tme, (t), proceeds as follows: S max 1. The dle tme whch may be derved, S π (t), s ntally set to zero 2. Equaton 3 s used to compute the end of a busy perod n the nterval [t, t + d (t)) 3. The end of the busy perod s used as the start of an dle perod by equaton 4 whch returns the length of contguous dle tme. 4. The dle tme, S π (t) s ncremented by the amount of dle tme found n step If the deadlne on task τ has been reached, then the maxmum dle tme whch can be derved s gven by S π (t). Otherwse, we repeat steps 2 to 5. The pseudo-codes of dynamc procrastnaton algorthm at tme nstant t when the ready queue s empty, and a job completes are shown n Algorthm 1. 4 Case Studes and Smulatons Secton 3 showed that our two-phase algorthm (EE-FPPT [13] + LA-PFFT) wll always render the controlled leakage current n CMOS crcuts and reduced energy consumptons that wll mantan the schedulablty of the workload. we use randomly-generated workloads to examne broad trends across a range of desgn ponts. (3) (4)

8 Algorthm 1 On-lne Algorthm to Mnmze Energy Leakage 1: procedure DYNAMIC PROCRASTINATION(t) a job completes at t and the ready queue s empty 2: sort T by ascendng prorty order 3: for ( = 1; n; n) do 4: r (t) t T T 5: d (t) r (t) + D 6: S π (t) 0 7: w m+1 (t) 0 (t) d (t)) do (t) 8: whle (w m+1 9: w m (t) w m+1 10: w m+1 (t) = S (t) + max k,π k < <γ k C k + 11: f (wπ m (t) = wπ m+1 (t)) then S π (t) S π (t) + l (t, wπ m (t)) wπ m+1 (t) wπ m+1 (t) + l (t,wπ m (t)) 12: end f j,π j ( ) w m π (t) r j(t) C j 13: end whle 14: Sπ max (t) S (t) 15: revse the arrval tme r (t) of job J,t by settng r (t) r (t) + Sπ max 16: end for 17: f ( mn r (t) t > t θ ) then τ T 18: turn the system off at tme t and turn on at mn r (t) τ T 19: else 20: reman on the actve mode 21: end f 22: end procedure T j (t) We nvestgate workload characterstcs that affect the energy savng capablty attanable through LA-FPPT. We now smulate and analyze randomly generated systems of tasks to better understand our approaches. The power consumpton functon of the system speed f was set as P(f) = f The normalzed total energy was adopted as the performance metrcs. The normalzed total energy of an algorthm for an nput nstance s the energy consumpton of the derved soluton n (0, L] dvded by the energy consumpton by applyng the orgnal FPPT schedulng wthout processor slowdown, procrastnaton and by turnng off the system when the dle nterval s long enough. We tred two dfferent expermental settngs. The frst experment nvestgate separately the effect of the swtchng overhead E sw, the system utlzaton on the lmted energy consumpton acheved by our methods. To cover a wde range of desgn ponts, 20,000 real-tme task sets wth 10 tasks each were randomly generated. These were created so 1000 have a utlzaton of 50%, 1000 have 52% utlzaton, and so on up to 90%. For each group of task sets who hold the same utlzaton, those were created so 20 have a E sw of 0.03, 20 have 0.04, and so on up to The second one focused on

9 the mpact of the number of tasks and E sw (0.17), another 20,000 real-tme task sets wth system utlzaton 67% each were randomly generated too. Those were created so 1000 nclude 5 ndependent tasks, 1000 nclude 6 ndependent tasks, and so on up to 25 tasks. Task perods s assgned randomly n the range [1, 100] wth a unform probablty dstrbuton functon. Moreover, task deadlnes were set equal to ther respectve perods (for smplcty, though not necessary). Tasks WCETs were set to ncur the requred overall system utlzaton. All 40,000 real-tme task sets generated were schedulable wth a fully preemptve polcy. Fg. 1. Experment I results for power savng of our approaches Average Normalzed Energy Average Normalzed Energy EE-FPPT EE-FPPT + LA-FPPT 0.4 EE-FPPT EE-FPPT + LA-FPPT E sw Utlzaton (a) Energy consumpton produced for LA- (b) Energy consumpton rses wth system FPPT rses wth large E sw, but keep almost constant for EE-FPPT, wth system systems. utlzaton, but soars up for hgh-utlzaton utlzaton = 0.67 Fg. 2. Experment II results for power savng of our approaches EE-FPPT EE-FPPT + LA-FPPT Utlzaton = 30% Utlzaton = 40% Utlzaton = 50% Utlzaton = 60% Average Normalzed Energy Percent of Systems Number of Tasks Normalzed Energy Consumpton (a) Energy consumpton declnes wth the (b) EE-FPPT + LA-FPPT dramatcally accomplshes the energy savngs, even for ncrement of the number of tasks, on the condton that system utlzaton = 0.67 hgh-utlzaton systems. and E sw = 0.17

10 Usng the MPTA, the total energy produced by each system was computed and normalzed to the energy requred by the orgnal verson of the system. The average normalzed energy were then plotted as a functon of E sw, the system utlzaton and the number of tasks n turn. The results are shown n fgures 2(a), 2(b) and 3(a) respectvely. In Fgure 2(a), the more the swtchng overhead E sw was, the more the normalzed energy consumpton was for schedules derved from Algorthms LA-FPPT. When E sw s relatvely small (E sw 0.18 ), the energy consumpton from leakage current n CMOS crcuts stll have ltter nfluence on the total energy consumpton, thus the more E sw, the less normalzed energy consumpton. In Fgure 2(b), our algorthms (EE-FPPT + LA-FPPT) outperformed orgnal Algorthm FPPT when the system utlzaton was greater than When the system utlzaton was large enough, procrastnaton mght create two (or more) dle ntervals to turn the system off, but the orgnal FPPT schedule mght make the system dle for a short nterval and turn the system off for a longer nterval. As a result, the energy consumpton of procrastnaton schedules mght consume more energy than the orgnal FPPT schedule when the system utlzaton s large enough. Moreover, when the utlzaton for task executon s large, the mprovement on dle energy consumpton s margnal snce task executon domnates the total energy consumpton. In Fgure 3(a), for all the smulated algorthms, the normalzed energy consumpton decreased for small number of tasks wth n 12, and was steady for n > 12. Ths s because the resultng utlzaton of a task was large when n was small n the expermental setup, and, hence, there was only lttle room for procrastnaton to save energy. For task sets wth n > 12, the maxmum procrastnaton nterval was domnated by tasks wth small perods, and, hence, the mprovement became margnal. Another nterestng property s the dstrbuton of the 20,000 systems of Experment I among the dfferent normalzed power consumpton levels. Fgure 3(b) show ths dstrbuton for the overall system utlzaton levels of 30%, 40%, 50%, and 60%, respectvely. As can be seen, the workloads scheduled wth the fxed-prorty schemes depend on the system utlzaton level to some extent. 5 Conclusons In ths paper we dscuss the energy-effcent schedulng problem of perodc realtme tasks by applyng FPPT polcy on a unprocessor dynamc voltage scalng system that can go nto the dormant mode for energy effcency. We propose a two-phase schedulng algorthm. In the frst phase, the executon speed,.e., the supply voltage, of each task s determned by applyng off-lne algorthms. In the second phase, the tme moment to turn on/off the system s determned on the fly. Theoretcal analyss shows that our 1 proposed algorthms could derve schedulng solutons wth at most max{, 2} (U bd ) 2 tmes of the energy consumpton of optmal solutons, where the term U bd represents the breakdown utlzaton [16] of a task set. A seres of smulaton experments was evaluated to demonstrate the performance of the proposed algorthms. Our expermental results show that our approaches can accomplsh dramatc energy savngs as the same tme keep the schedulablty of task set.

11 References 1. Takayasu Sakura, A.R.N.: Alpha-power law mosfet model and ts applcatons to cmos nverterdelay and other formulas. IEEE Journal of Sold-State Crcuts 25(2) (1990) Padmanabhan Plla, K.G.S.: Real-tme dynamc voltage scalng for low-power embedded operatng systems. In: 18th ACM Symposum on Operatng System Prncples. Volume 35., Chateau Lake Louse, Banff, Alberta, Canada, ACM (2001) Ravndra Jejurkar, R.K.G.: Dynamc slack reclamaton wth procrastnaton schedulng n real-tme embedded systems. In Kahng, W.H.J.J., Martn, G., B., A., eds.: 42nd Desgn Automaton Conference, San Dego, CA, USA, ACM (2005) Ravndra Jejurkar, Crstano Perera, R.K.G.: Leakage aware dynamc voltage scalng for real-tme embedded systems. In Kahng, S.M., Fx, L., B., A., eds.: 41th Desgn Automaton Conference, San Dego, CA, USA, ACM (2004) Yann-Hang Lee, Krshna P. Reddy, C.M.K.: Schedulng technques for reducng leakage power n hard real-tme systems. In: 15th Euromcro Conference on Real-Tme Systems (ECRTS 2003), Porto, Portugal, IEEE Computer Socety (2003) Manas Saksena, Y.W.: Scalable real-tme system desgn usng preempton thresholds. In: 21st IEEE Real-Tme Systems Symposum. (2000) Lehoczky, J.P.: Fxed prorty schedulng of perodc task sets wth arbtrary deadlnes. In: IEEE Real-Tme Systems Symposum, Lake Buena Vsta, Florda, USA, IEEE Computer Socety Press (1990) Rubn Xu, Daka Zhu, C.R.R.G.M.D.M.: Energy-effcent polces for embedded clusters. In Gupta, Y.P., Rajv, eds.: 2005 ACM SIGPLAN/SIGBED Conference on Languages, Complers, and Tools for Embedded Systems, Chcago, Illnos, USA, ACM (2005) 9. Ravndra Jejurkar, R.K.G.: Procrastnaton schedulng n fxed prorty real-tme systems. In: 2004 ACM SIGPLAN/SIGBED Conference on Languages, Complers, and Tools for Embedded Systems, Washngton, DC, USA, ACM (2004) Ravndra Jejurkar, R.K.G.: Dynamc voltage scalng for systemwde energy mnmzaton n real-tme embedded systems. In Roy, R.V.J., Cho, K., Twar, V., Kaushk, eds.: 2004 Internatonal Symposum on Low Power Electroncs and Desgn, Newport Beach, Calforna, USA, ACM (2004) Gang Quan, Lnwe Nu, X.S.H.B.M.: Fxed prorty schedulng for reducng overall energy on varable voltage processors. In: 25th IEEE Real-Tme Systems Symposum, Lsbon, Portugal, IEEE Computer Socety (2004) Sandy Iran, Sandeep K. Shukla, R.K.G.: Algorthms for power savngs. In: Fourteenth Annual ACM-SIAM Symposum on Dscrete Algorthms, ACM (2003) XaoChuan He, Y.J.: Energy-effcent schedulng fxed-prorty tasks wth preempton thresholds on varable voltage processors. In Gaudot, K.L., Jesshope, C.R., Jn, H., Jean-Luc, eds.: Network and Parallel Computng, IFIP Internatonal Conference (NPC 2007). Volume Lecture Notes n Computer Scence., Dalan, Chna, Sprnger (2007) Lnwe Nu, G.Q.: Reducng both dynamc and leakage energy consumpton for hard realtme systems. In Mahlke, M.J.I., Zhao, W., Lavagno, L., A., S., eds.: 2004 Internatonal Conference on Complers, Archtecture, and Synthess for Embedded Systems, Washngton DC, USA, ACM (2004) Nel C. Audsley, Alan Burns, M.R.A.J.W.: Applyng new schedulng theory to statc prorty pre-emptve schedulng. Software Engneerng Journal 8(5) (1993) John P. Lehoczky, Lu Sha, Y.D.: The rate monotonc schedulng algorthm: Exact characterzaton and average case behavor. In: IEEE Real-Tme Systems Symposum (1989)

Real-Time Systems. Multiprocessor scheduling. Multiprocessor scheduling. Multiprocessor scheduling

Real-Time Systems. Multiprocessor scheduling. Multiprocessor scheduling. Multiprocessor scheduling Real-Tme Systems Multprocessor schedulng Specfcaton Implementaton Verfcaton Multprocessor schedulng -- -- Global schedulng How are tasks assgned to processors? Statc assgnment The processor(s) used for

More information

Embedded Systems. 4. Aperiodic and Periodic Tasks

Embedded Systems. 4. Aperiodic and Periodic Tasks Embedded Systems 4. Aperodc and Perodc Tasks Lothar Thele 4-1 Contents of Course 1. Embedded Systems Introducton 2. Software Introducton 7. System Components 10. Models 3. Real-Tme Models 4. Perodc/Aperodc

More information

Energy-Efficient Scheduling Fixed-Priority tasks with Preemption Thresholds on Variable Voltage Processors

Energy-Efficient Scheduling Fixed-Priority tasks with Preemption Thresholds on Variable Voltage Processors Energy-Effcent Schedulng Fxed-Prorty tasks wth Preempton Thresholds on Varable Voltage Processors XaoChuan He, Yan Ja Insttute of Network Technology and Informaton Securty School of Computer Scence Natonal

More information

Resource Allocation with a Budget Constraint for Computing Independent Tasks in the Cloud

Resource Allocation with a Budget Constraint for Computing Independent Tasks in the Cloud Resource Allocaton wth a Budget Constrant for Computng Independent Tasks n the Cloud Wemng Sh and Bo Hong School of Electrcal and Computer Engneerng Georga Insttute of Technology, USA 2nd IEEE Internatonal

More information

Improved Worst-Case Response-Time Calculations by Upper-Bound Conditions

Improved Worst-Case Response-Time Calculations by Upper-Bound Conditions Improved Worst-Case Response-Tme Calculatons by Upper-Bound Condtons Vctor Pollex, Steffen Kollmann, Karsten Albers and Frank Slomka Ulm Unversty Insttute of Embedded Systems/Real-Tme Systems {frstname.lastname}@un-ulm.de

More information

Two Methods to Release a New Real-time Task

Two Methods to Release a New Real-time Task Two Methods to Release a New Real-tme Task Abstract Guangmng Qan 1, Xanghua Chen 2 College of Mathematcs and Computer Scence Hunan Normal Unversty Changsha, 410081, Chna qqyy@hunnu.edu.cn Gang Yao 3 Sebel

More information

Single-Facility Scheduling over Long Time Horizons by Logic-based Benders Decomposition

Single-Facility Scheduling over Long Time Horizons by Logic-based Benders Decomposition Sngle-Faclty Schedulng over Long Tme Horzons by Logc-based Benders Decomposton Elvn Coban and J. N. Hooker Tepper School of Busness, Carnege Mellon Unversty ecoban@andrew.cmu.edu, john@hooker.tepper.cmu.edu

More information

Simultaneous Optimization of Berth Allocation, Quay Crane Assignment and Quay Crane Scheduling Problems in Container Terminals

Simultaneous Optimization of Berth Allocation, Quay Crane Assignment and Quay Crane Scheduling Problems in Container Terminals Smultaneous Optmzaton of Berth Allocaton, Quay Crane Assgnment and Quay Crane Schedulng Problems n Contaner Termnals Necat Aras, Yavuz Türkoğulları, Z. Caner Taşkın, Kuban Altınel Abstract In ths work,

More information

The Minimum Universal Cost Flow in an Infeasible Flow Network

The Minimum Universal Cost Flow in an Infeasible Flow Network Journal of Scences, Islamc Republc of Iran 17(2): 175-180 (2006) Unversty of Tehran, ISSN 1016-1104 http://jscencesutacr The Mnmum Unversal Cost Flow n an Infeasble Flow Network H Saleh Fathabad * M Bagheran

More information

Effective Power Optimization combining Placement, Sizing, and Multi-Vt techniques

Effective Power Optimization combining Placement, Sizing, and Multi-Vt techniques Effectve Power Optmzaton combnng Placement, Szng, and Mult-Vt technques Tao Luo, Davd Newmark*, and Davd Z Pan Department of Electrcal and Computer Engneerng, Unversty of Texas at Austn *Advanced Mcro

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

Winter 2008 CS567 Stochastic Linear/Integer Programming Guest Lecturer: Xu, Huan

Winter 2008 CS567 Stochastic Linear/Integer Programming Guest Lecturer: Xu, Huan Wnter 2008 CS567 Stochastc Lnear/Integer Programmng Guest Lecturer: Xu, Huan Class 2: More Modelng Examples 1 Capacty Expanson Capacty expanson models optmal choces of the tmng and levels of nvestments

More information

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur Module 3 LOSSY IMAGE COMPRESSION SYSTEMS Verson ECE IIT, Kharagpur Lesson 6 Theory of Quantzaton Verson ECE IIT, Kharagpur Instructonal Objectves At the end of ths lesson, the students should be able to:

More information

On the Multicriteria Integer Network Flow Problem

On the Multicriteria Integer Network Flow Problem BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 5, No 2 Sofa 2005 On the Multcrtera Integer Network Flow Problem Vassl Vasslev, Marana Nkolova, Maryana Vassleva Insttute of

More information

EEL 6266 Power System Operation and Control. Chapter 3 Economic Dispatch Using Dynamic Programming

EEL 6266 Power System Operation and Control. Chapter 3 Economic Dispatch Using Dynamic Programming EEL 6266 Power System Operaton and Control Chapter 3 Economc Dspatch Usng Dynamc Programmng Pecewse Lnear Cost Functons Common practce many utltes prefer to represent ther generator cost functons as sngle-

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

Overhead-Aware Compositional Analysis of Real-Time Systems

Overhead-Aware Compositional Analysis of Real-Time Systems Overhead-Aware ompostonal Analyss of Real-Tme Systems Lnh T.X. Phan, Meng Xu, Jaewoo Lee, nsup Lee, Oleg Sokolsky PRESE enter Department of omputer and nformaton Scence Unversty of Pennsylvana ompostonal

More information

Quantifying the Sub-optimality of Uniprocessor Fixed Priority Pre-emptive Scheduling for Sporadic Tasksets with Arbitrary Deadlines

Quantifying the Sub-optimality of Uniprocessor Fixed Priority Pre-emptive Scheduling for Sporadic Tasksets with Arbitrary Deadlines Quantfyng the Sub-optmalty of Unprocessor Fxed Prorty Pre-emptve Schedulng for Sporadc Tasksets wth Arbtrary Deadlnes Robert Davs, Sanjoy Baruah, Thomas Rothvoss, Alan Burns To cte ths verson: Robert Davs,

More information

College of Computer & Information Science Fall 2009 Northeastern University 20 October 2009

College of Computer & Information Science Fall 2009 Northeastern University 20 October 2009 College of Computer & Informaton Scence Fall 2009 Northeastern Unversty 20 October 2009 CS7880: Algorthmc Power Tools Scrbe: Jan Wen and Laura Poplawsk Lecture Outlne: Prmal-dual schema Network Desgn:

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

Last Time. Priority-based scheduling. Schedulable utilization Rate monotonic rule: Keep utilization below 69% Static priorities Dynamic priorities

Last Time. Priority-based scheduling. Schedulable utilization Rate monotonic rule: Keep utilization below 69% Static priorities Dynamic priorities Last Tme Prorty-based schedulng Statc prortes Dynamc prortes Schedulable utlzaton Rate monotonc rule: Keep utlzaton below 69% Today Response tme analyss Blockng terms Prorty nverson And solutons Release

More information

Limited Preemptive Scheduling for Real-Time Systems: a Survey

Limited Preemptive Scheduling for Real-Time Systems: a Survey Lmted Preemptve Schedulng for Real-Tme Systems: a Survey Gorgo C. Buttazzo, Fellow Member, IEEE, Marko Bertogna, Senor Member, IEEE, and Gang Yao Abstract The queston whether preemptve algorthms are better

More information

Minimizing Energy Consumption of MPI Programs in Realistic Environment

Minimizing Energy Consumption of MPI Programs in Realistic Environment Mnmzng Energy Consumpton of MPI Programs n Realstc Envronment Amna Guermouche, Ncolas Trquenaux, Benoît Pradelle and Wllam Jalby Unversté de Versalles Sant-Quentn-en-Yvelnes arxv:1502.06733v2 [cs.dc] 25

More information

Improving the Sensitivity of Deadlines with a Specific Asynchronous Scenario for Harmonic Periodic Tasks scheduled by FP

Improving the Sensitivity of Deadlines with a Specific Asynchronous Scenario for Harmonic Periodic Tasks scheduled by FP Improvng the Senstvty of Deadlnes wth a Specfc Asynchronous Scenaro for Harmonc Perodc Tasks scheduled by FP P. Meumeu Yoms, Y. Sorel, D. de Rauglaudre AOSTE Project-team INRIA Pars-Rocquencourt Le Chesnay,

More information

Coarse-Grain MTCMOS Sleep

Coarse-Grain MTCMOS Sleep Coarse-Gran MTCMOS Sleep Transstor Szng Usng Delay Budgetng Ehsan Pakbazna and Massoud Pedram Unversty of Southern Calforna Dept. of Electrcal Engneerng DATE-08 Munch, Germany Leakage n CMOS Technology

More information

AN EXTENDIBLE APPROACH FOR ANALYSING FIXED PRIORITY HARD REAL-TIME TASKS

AN EXTENDIBLE APPROACH FOR ANALYSING FIXED PRIORITY HARD REAL-TIME TASKS AN EXENDIBLE APPROACH FOR ANALYSING FIXED PRIORIY HARD REAL-IME ASKS K. W. ndell 1 Department of Computer Scence, Unversty of York, England YO1 5DD ABSRAC As the real-tme computng ndustry moves away from

More information

Minimisation of the Average Response Time in a Cluster of Servers

Minimisation of the Average Response Time in a Cluster of Servers Mnmsaton of the Average Response Tme n a Cluster of Servers Valery Naumov Abstract: In ths paper, we consder task assgnment problem n a cluster of servers. We show that optmal statc task assgnment s tantamount

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

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

Kernel Methods and SVMs Extension

Kernel Methods and SVMs Extension Kernel Methods and SVMs Extenson The purpose of ths document s to revew materal covered n Machne Learnng 1 Supervsed Learnng regardng support vector machnes (SVMs). Ths document also provdes a general

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

Improving the Quality of Control of Periodic Tasks Scheduled by FP with an Asynchronous Approach

Improving the Quality of Control of Periodic Tasks Scheduled by FP with an Asynchronous Approach Improvng the Qualty of Control of Perodc Tasks Scheduled by FP wth an Asynchronous Approach P. Meumeu Yoms, L. George, Y. Sorel, D. de Rauglaudre AOSTE Project-team INRIA Pars-Rocquencourt Le Chesnay,

More information

Variability-Driven Module Selection with Joint Design Time Optimization and Post-Silicon Tuning

Variability-Driven Module Selection with Joint Design Time Optimization and Post-Silicon Tuning Asa and South Pacfc Desgn Automaton Conference 2008 Varablty-Drven Module Selecton wth Jont Desgn Tme Optmzaton and Post-Slcon Tunng Feng Wang, Xaoxa Wu, Yuan Xe The Pennsylvana State Unversty Department

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

An Admission Control Algorithm in Cloud Computing Systems

An Admission Control Algorithm in Cloud Computing Systems An Admsson Control Algorthm n Cloud Computng Systems Authors: Frank Yeong-Sung Ln Department of Informaton Management Natonal Tawan Unversty Tape, Tawan, R.O.C. ysln@m.ntu.edu.tw Yngje Lan Management Scence

More information

Clock-Gating and Its Application to Low Power Design of Sequential Circuits

Clock-Gating and Its Application to Low Power Design of Sequential Circuits Clock-Gatng and Its Applcaton to Low Power Desgn of Sequental Crcuts ng WU Department of Electrcal Engneerng-Systems, Unversty of Southern Calforna Los Angeles, CA 989, USA, Phone: (23)74-448 Massoud PEDRAM

More information

Chapter - 2. Distribution System Power Flow Analysis

Chapter - 2. Distribution System Power Flow Analysis Chapter - 2 Dstrbuton System Power Flow Analyss CHAPTER - 2 Radal Dstrbuton System Load Flow 2.1 Introducton Load flow s an mportant tool [66] for analyzng electrcal power system network performance. Load

More information

Energy-Efficient Primary/Backup Scheduling Techniques for Heterogeneous Multicore Systems

Energy-Efficient Primary/Backup Scheduling Techniques for Heterogeneous Multicore Systems Energy-Effcent Prmary/Backup Schedulng Technques for Heterogeneous Multcore Systems Abhshek Roy, Hakan Aydn epartment of Computer Scence George Mason Unversty Farfax, Vrgna 22030 aroy6@gmu.edu, aydn@cs.gmu.edu

More information

PreDVS: Preemptive Dynamic Voltage Scaling for Real-time Systems using Approximation Scheme

PreDVS: Preemptive Dynamic Voltage Scaling for Real-time Systems using Approximation Scheme PreDVS: Preemptve Dynamc Voltage Scalng for Real-tme Systems usng Approxmaton Scheme Wexun Wang and Prabhat Mshra Department of Computer and Informaton Scence and Engneerng Unversty of Florda, Ganesvlle,

More information

Annexes. EC.1. Cycle-base move illustration. EC.2. Problem Instances

Annexes. EC.1. Cycle-base move illustration. EC.2. Problem Instances ec Annexes Ths Annex frst llustrates a cycle-based move n the dynamc-block generaton tabu search. It then dsplays the characterstcs of the nstance sets, followed by detaled results of the parametercalbraton

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

Fixed-Priority Multiprocessor Scheduling with Liu & Layland s Utilization Bound

Fixed-Priority Multiprocessor Scheduling with Liu & Layland s Utilization Bound Fxed-Prorty Multprocessor Schedulng wth Lu & Layland s Utlzaton Bound Nan Guan, Martn Stgge, Wang Y and Ge Yu Department of Informaton Technology, Uppsala Unversty, Sweden Department of Computer Scence

More information

Clock-Driven Scheduling (in-depth) Cyclic Schedules: General Structure

Clock-Driven Scheduling (in-depth) Cyclic Schedules: General Structure CPSC-663: Real-me Systems n-depth Precompute statc schedule o-lne e.g. at desgn tme: can aord expensve algorthms. Idle tmes can be used or aperodc jobs. Possble mplementaton: able-drven Schedulng table

More information

Quantifying the Sub-optimality of Uniprocessor Fixed Priority Non-Pre-emptive Scheduling

Quantifying the Sub-optimality of Uniprocessor Fixed Priority Non-Pre-emptive Scheduling Quantfyng the Sub-optmalty of Unprocessor Fxed Prorty Non-Pre-emptve Schedulng Robert I Davs Real-Tme Systems Research Group, Department of Computer Scence, Unversty of York, York, UK robdavs@csyorkacuk

More information

Two-Phase Low-Energy N-Modular Redundancy for Hard Real-Time Multi-Core Systems

Two-Phase Low-Energy N-Modular Redundancy for Hard Real-Time Multi-Core Systems 1 Two-Phase Low-Energy N-Modular Redundancy for Hard Real-Tme Mult-Core Systems Mohammad Saleh, Alreza Ejlal, and Bashr M. Al-Hashm, Fellow, IEEE Abstract Ths paper proposes an N-modular redundancy (NMR)

More information

Lecture 14: Bandits with Budget Constraints

Lecture 14: Bandits with Budget Constraints IEOR 8100-001: Learnng and Optmzaton for Sequental Decson Makng 03/07/16 Lecture 14: andts wth udget Constrants Instructor: Shpra Agrawal Scrbed by: Zhpeng Lu 1 Problem defnton In the regular Mult-armed

More information

1 Derivation of Rate Equations from Single-Cell Conductance (Hodgkin-Huxley-like) Equations

1 Derivation of Rate Equations from Single-Cell Conductance (Hodgkin-Huxley-like) Equations Physcs 171/271 -Davd Klenfeld - Fall 2005 (revsed Wnter 2011) 1 Dervaton of Rate Equatons from Sngle-Cell Conductance (Hodgkn-Huxley-lke) Equatons We consder a network of many neurons, each of whch obeys

More information

4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA

4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA 4 Analyss of Varance (ANOVA) 5 ANOVA 51 Introducton ANOVA ANOVA s a way to estmate and test the means of multple populatons We wll start wth one-way ANOVA If the populatons ncluded n the study are selected

More information

A FAST HEURISTIC FOR TASKS ASSIGNMENT IN MANYCORE SYSTEMS WITH VOLTAGE-FREQUENCY ISLANDS

A FAST HEURISTIC FOR TASKS ASSIGNMENT IN MANYCORE SYSTEMS WITH VOLTAGE-FREQUENCY ISLANDS Shervn Haamn A FAST HEURISTIC FOR TASKS ASSIGNMENT IN MANYCORE SYSTEMS WITH VOLTAGE-FREQUENCY ISLANDS INTRODUCTION Increasng computatons n applcatons has led to faster processng. o Use more cores n a chp

More information

Lab 2e Thermal System Response and Effective Heat Transfer Coefficient

Lab 2e Thermal System Response and Effective Heat Transfer Coefficient 58:080 Expermental Engneerng 1 OBJECTIVE Lab 2e Thermal System Response and Effectve Heat Transfer Coeffcent Warnng: though the experment has educatonal objectves (to learn about bolng heat transfer, etc.),

More information

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE Analytcal soluton s usually not possble when exctaton vares arbtrarly wth tme or f the system s nonlnear. Such problems can be solved by numercal tmesteppng

More information

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number

More information

On the correction of the h-index for career length

On the correction of the h-index for career length 1 On the correcton of the h-ndex for career length by L. Egghe Unverstet Hasselt (UHasselt), Campus Depenbeek, Agoralaan, B-3590 Depenbeek, Belgum 1 and Unverstet Antwerpen (UA), IBW, Stadscampus, Venusstraat

More information

FUZZY GOAL PROGRAMMING VS ORDINARY FUZZY PROGRAMMING APPROACH FOR MULTI OBJECTIVE PROGRAMMING PROBLEM

FUZZY GOAL PROGRAMMING VS ORDINARY FUZZY PROGRAMMING APPROACH FOR MULTI OBJECTIVE PROGRAMMING PROBLEM Internatonal Conference on Ceramcs, Bkaner, Inda Internatonal Journal of Modern Physcs: Conference Seres Vol. 22 (2013) 757 761 World Scentfc Publshng Company DOI: 10.1142/S2010194513010982 FUZZY GOAL

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

The Study of Teaching-learning-based Optimization Algorithm

The Study of Teaching-learning-based Optimization Algorithm Advanced Scence and Technology Letters Vol. (AST 06), pp.05- http://dx.do.org/0.57/astl.06. The Study of Teachng-learnng-based Optmzaton Algorthm u Sun, Yan fu, Lele Kong, Haolang Q,, Helongang Insttute

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

Outline. Communication. Bellman Ford Algorithm. Bellman Ford Example. Bellman Ford Shortest Path [1]

Outline. Communication. Bellman Ford Algorithm. Bellman Ford Example. Bellman Ford Shortest Path [1] DYNAMIC SHORTEST PATH SEARCH AND SYNCHRONIZED TASK SWITCHING Jay Wagenpfel, Adran Trachte 2 Outlne Shortest Communcaton Path Searchng Bellmann Ford algorthm Algorthm for dynamc case Modfcatons to our algorthm

More information

Complete subgraphs in multipartite graphs

Complete subgraphs in multipartite graphs Complete subgraphs n multpartte graphs FLORIAN PFENDER Unverstät Rostock, Insttut für Mathematk D-18057 Rostock, Germany Floran.Pfender@un-rostock.de Abstract Turán s Theorem states that every graph G

More information

Interconnect Optimization for Deep-Submicron and Giga-Hertz ICs

Interconnect Optimization for Deep-Submicron and Giga-Hertz ICs Interconnect Optmzaton for Deep-Submcron and Gga-Hertz ICs Le He http://cadlab.cs.ucla.edu/~hele UCLA Computer Scence Department Los Angeles, CA 90095 Outlne Background and overvew LR-based STIS optmzaton

More information

Global Sensitivity. Tuesday 20 th February, 2018

Global Sensitivity. Tuesday 20 th February, 2018 Global Senstvty Tuesday 2 th February, 28 ) Local Senstvty Most senstvty analyses [] are based on local estmates of senstvty, typcally by expandng the response n a Taylor seres about some specfc values

More information

ECE559VV Project Report

ECE559VV Project Report ECE559VV Project Report (Supplementary Notes Loc Xuan Bu I. MAX SUM-RATE SCHEDULING: THE UPLINK CASE We have seen (n the presentaton that, for downlnk (broadcast channels, the strategy maxmzng the sum-rate

More information

Convexity preserving interpolation by splines of arbitrary degree

Convexity preserving interpolation by splines of arbitrary degree Computer Scence Journal of Moldova, vol.18, no.1(52), 2010 Convexty preservng nterpolaton by splnes of arbtrary degree Igor Verlan Abstract In the present paper an algorthm of C 2 nterpolaton of dscrete

More information

A Single-Machine Deteriorating Job Scheduling Problem of Minimizing the Makespan with Release Times

A Single-Machine Deteriorating Job Scheduling Problem of Minimizing the Makespan with Release Times A Sngle-Machne Deteroratng Job Schedulng Problem of Mnmzng the Makespan wth Release Tmes Wen-Chung Lee, Chn-Cha Wu, and Yu-Hsang Chung Abstract In ths paper, we study a sngle-machne deteroratng ob schedulng

More information

NP-Completeness : Proofs

NP-Completeness : Proofs NP-Completeness : Proofs Proof Methods A method to show a decson problem Π NP-complete s as follows. (1) Show Π NP. (2) Choose an NP-complete problem Π. (3) Show Π Π. A method to show an optmzaton problem

More information

Notes prepared by Prof Mrs) M.J. Gholba Class M.Sc Part(I) Information Technology

Notes prepared by Prof Mrs) M.J. Gholba Class M.Sc Part(I) Information Technology Inverse transformatons Generaton of random observatons from gven dstrbutons Assume that random numbers,,, are readly avalable, where each tself s a random varable whch s unformly dstrbuted over the range(,).

More information

find (x): given element x, return the canonical element of the set containing x;

find (x): given element x, return the canonical element of the set containing x; COS 43 Sprng, 009 Dsjont Set Unon Problem: Mantan a collecton of dsjont sets. Two operatons: fnd the set contanng a gven element; unte two sets nto one (destructvely). Approach: Canoncal element method:

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

O-line Temporary Tasks Assignment. Abstract. In this paper we consider the temporary tasks assignment

O-line Temporary Tasks Assignment. Abstract. In this paper we consider the temporary tasks assignment O-lne Temporary Tasks Assgnment Yoss Azar and Oded Regev Dept. of Computer Scence, Tel-Avv Unversty, Tel-Avv, 69978, Israel. azar@math.tau.ac.l??? Dept. of Computer Scence, Tel-Avv Unversty, Tel-Avv, 69978,

More information

Assortment Optimization under MNL

Assortment Optimization under MNL Assortment Optmzaton under MNL Haotan Song Aprl 30, 2017 1 Introducton The assortment optmzaton problem ams to fnd the revenue-maxmzng assortment of products to offer when the prces of products are 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

Amiri s Supply Chain Model. System Engineering b Department of Mathematics and Statistics c Odette School of Business

Amiri s Supply Chain Model. System Engineering b Department of Mathematics and Statistics c Odette School of Business Amr s Supply Chan Model by S. Ashtab a,, R.J. Caron b E. Selvarajah c a Department of Industral Manufacturng System Engneerng b Department of Mathematcs Statstcs c Odette School of Busness Unversty of

More information

Lecture 4. Instructor: Haipeng Luo

Lecture 4. Instructor: Haipeng Luo Lecture 4 Instructor: Hapeng Luo In the followng lectures, we focus on the expert problem and study more adaptve algorthms. Although Hedge s proven to be worst-case optmal, one may wonder how well t would

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

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

Uncertainty in measurements of power and energy on power networks

Uncertainty in measurements of power and energy on power networks Uncertanty n measurements of power and energy on power networks E. Manov, N. Kolev Department of Measurement and Instrumentaton, Techncal Unversty Sofa, bul. Klment Ohrdsk No8, bl., 000 Sofa, Bulgara Tel./fax:

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

Application of B-Spline to Numerical Solution of a System of Singularly Perturbed Problems

Application of B-Spline to Numerical Solution of a System of Singularly Perturbed Problems Mathematca Aeterna, Vol. 1, 011, no. 06, 405 415 Applcaton of B-Splne to Numercal Soluton of a System of Sngularly Perturbed Problems Yogesh Gupta Department of Mathematcs Unted College of Engneerng &

More information

A SEPARABLE APPROXIMATION DYNAMIC PROGRAMMING ALGORITHM FOR ECONOMIC DISPATCH WITH TRANSMISSION LOSSES. Pierre HANSEN, Nenad MLADENOVI]

A SEPARABLE APPROXIMATION DYNAMIC PROGRAMMING ALGORITHM FOR ECONOMIC DISPATCH WITH TRANSMISSION LOSSES. Pierre HANSEN, Nenad MLADENOVI] Yugoslav Journal of Operatons Research (00) umber 57-66 A SEPARABLE APPROXIMATIO DYAMIC PROGRAMMIG ALGORITHM FOR ECOOMIC DISPATCH WITH TRASMISSIO LOSSES Perre HASE enad MLADEOVI] GERAD and Ecole des Hautes

More information

LOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin

LOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin Proceedngs of the 007 Wnter Smulaton Conference S G Henderson, B Bller, M-H Hseh, J Shortle, J D Tew, and R R Barton, eds LOW BIAS INTEGRATED PATH ESTIMATORS James M Calvn Department of Computer Scence

More information

Real-Time Operating Systems M. 11. Real-Time: Periodic Task Scheduling

Real-Time Operating Systems M. 11. Real-Time: Periodic Task Scheduling Real-Tme Operatng Systems M 11. Real-Tme: Perodc Task Schedulng Notce The course materal ncludes sldes downloaded from:! http://codex.cs.yale.edu/av/os-book/! and! (sldes by Slberschatz, Galvn, and Gagne,

More information

Computer Control: Task Synchronisation in Dynamic Priority Scheduling

Computer Control: Task Synchronisation in Dynamic Priority Scheduling Computer Control: Task Synchronsaton n Dynamc Prorty Schedulng Sérgo Adrano Fernandes Lopes Department of Industral Electroncs Engneerng School Unversty of Mnho Campus de Azurém 4800 Gumarães - PORTUGAL

More information

On Energy-Optimal Voltage Scheduling for Fixed-Priority Hard Real-Time Systems

On Energy-Optimal Voltage Scheduling for Fixed-Priority Hard Real-Time Systems On Energy-Optmal Voltage Schedulng for Fxed-Prorty Hard Real-Tme Systems HAN-SAEM YUN and JIHONG KIM Seoul Natonal Unversty We address the problem of energy-optmal voltage schedulng for fxed-prorty hard

More information

ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EQUATION

ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EQUATION Advanced Mathematcal Models & Applcatons Vol.3, No.3, 2018, pp.215-222 ON A DETERMINATION OF THE INITIAL FUNCTIONS FROM THE OBSERVED VALUES OF THE BOUNDARY FUNCTIONS FOR THE SECOND-ORDER HYPERBOLIC EUATION

More information

Pop-Click Noise Detection Using Inter-Frame Correlation for Improved Portable Auditory Sensing

Pop-Click Noise Detection Using Inter-Frame Correlation for Improved Portable Auditory Sensing Advanced Scence and Technology Letters, pp.164-168 http://dx.do.org/10.14257/astl.2013 Pop-Clc Nose Detecton Usng Inter-Frame Correlaton for Improved Portable Audtory Sensng Dong Yun Lee, Kwang Myung Jeon,

More information

Determining Transmission Losses Penalty Factor Using Adaptive Neuro Fuzzy Inference System (ANFIS) For Economic Dispatch Application

Determining Transmission Losses Penalty Factor Using Adaptive Neuro Fuzzy Inference System (ANFIS) For Economic Dispatch Application 7 Determnng Transmsson Losses Penalty Factor Usng Adaptve Neuro Fuzzy Inference System (ANFIS) For Economc Dspatch Applcaton Rony Seto Wbowo Maurdh Hery Purnomo Dod Prastanto Electrcal Engneerng Department,

More information

Schedulability Analysis of Task Sets with Upper- and Lower-Bound Temporal Constraints

Schedulability Analysis of Task Sets with Upper- and Lower-Bound Temporal Constraints Schedulablty Analyss of Task Sets wth Upper- and Lower-Bound Temporal Constrants The MIT Faculty has made ths artcle openly avalable. Please share how ths access benefts you. Your story matters. Ctaton

More information

Quantifying the Exact Sub-Optimality of Non-Preemptive Scheduling

Quantifying the Exact Sub-Optimality of Non-Preemptive Scheduling Quantfyng the Exact Sub-Optmalty of Non-Preemptve Schedulng Robert I. Davs 1, Abhlash Thekklakattl 2, Olver Gettngs 1, Radu Dobrn 2, and Saskumar Punnekkat 2 1 Real-Tme Systems Research Group, Unversty

More information

Analysis of Queuing Delay in Multimedia Gateway Call Routing

Analysis of Queuing Delay in Multimedia Gateway Call Routing Analyss of Queung Delay n Multmeda ateway Call Routng Qwe Huang UTtarcom Inc, 33 Wood Ave. outh Iseln, NJ 08830, U..A Errol Lloyd Computer Informaton cences Department, Unv. of Delaware, Newark, DE 976,

More information

Optimal Scheduling Algorithms to Minimize Total Flowtime on a Two-Machine Permutation Flowshop with Limited Waiting Times and Ready Times of Jobs

Optimal Scheduling Algorithms to Minimize Total Flowtime on a Two-Machine Permutation Flowshop with Limited Waiting Times and Ready Times of Jobs Optmal Schedulng Algorthms to Mnmze Total Flowtme on a Two-Machne Permutaton Flowshop wth Lmted Watng Tmes and Ready Tmes of Jobs Seong-Woo Cho Dept. Of Busness Admnstraton, Kyongg Unversty, Suwon-s, 443-760,

More information

Hongyi Miao, College of Science, Nanjing Forestry University, Nanjing ,China. (Received 20 June 2013, accepted 11 March 2014) I)ϕ (k)

Hongyi Miao, College of Science, Nanjing Forestry University, Nanjing ,China. (Received 20 June 2013, accepted 11 March 2014) I)ϕ (k) ISSN 1749-3889 (prnt), 1749-3897 (onlne) Internatonal Journal of Nonlnear Scence Vol.17(2014) No.2,pp.188-192 Modfed Block Jacob-Davdson Method for Solvng Large Sparse Egenproblems Hongy Mao, College of

More information

Energy and Feasibility Optimal Global Scheduling Framework on big.little platforms

Energy and Feasibility Optimal Global Scheduling Framework on big.little platforms Energy and Feasblty Optmal Global Schedulng Framework on bg.little platforms Hoon Sung Chwa, Jaebaek Seo, Hyuck Yoo Jnkyu Lee, Insk Shn Department of Computer Scence, KAIST, Republc of Korea Department

More information

COMPLETE BUFFER SHARING IN ATM NETWORKS UNDER BURSTY ARRIVALS

COMPLETE BUFFER SHARING IN ATM NETWORKS UNDER BURSTY ARRIVALS COMPLETE BUFFER SHARING WITH PUSHOUT THRESHOLDS IN ATM NETWORKS UNDER BURSTY ARRIVALS Ozgur Aras and Tugrul Dayar Abstract. Broadband Integrated Servces Dgtal Networks (B{ISDNs) are to support multple

More information

The Second Anti-Mathima on Game Theory

The Second Anti-Mathima on Game Theory The Second Ant-Mathma on Game Theory Ath. Kehagas December 1 2006 1 Introducton In ths note we wll examne the noton of game equlbrum for three types of games 1. 2-player 2-acton zero-sum games 2. 2-player

More information

Integrated approach in solving parallel machine scheduling and location (ScheLoc) problem

Integrated approach in solving parallel machine scheduling and location (ScheLoc) problem Internatonal Journal of Industral Engneerng Computatons 7 (2016) 573 584 Contents lsts avalable at GrowngScence Internatonal Journal of Industral Engneerng Computatons homepage: www.growngscence.com/ec

More information

9 Derivation of Rate Equations from Single-Cell Conductance (Hodgkin-Huxley-like) Equations

9 Derivation of Rate Equations from Single-Cell Conductance (Hodgkin-Huxley-like) Equations Physcs 171/271 - Chapter 9R -Davd Klenfeld - Fall 2005 9 Dervaton of Rate Equatons from Sngle-Cell Conductance (Hodgkn-Huxley-lke) Equatons We consder a network of many neurons, each of whch obeys a set

More information

8 Derivation of Network Rate Equations from Single- Cell Conductance Equations

8 Derivation of Network Rate Equations from Single- Cell Conductance Equations Physcs 178/278 - Davd Klenfeld - Wnter 2015 8 Dervaton of Network Rate Equatons from Sngle- Cell Conductance Equatons We consder a network of many neurons, each of whch obeys a set of conductancebased,

More information

Suggested solutions for the exam in SF2863 Systems Engineering. June 12,

Suggested solutions for the exam in SF2863 Systems Engineering. June 12, Suggested solutons for the exam n SF2863 Systems Engneerng. June 12, 2012 14.00 19.00 Examner: Per Enqvst, phone: 790 62 98 1. We can thnk of the farm as a Jackson network. The strawberry feld s modelled

More information

A PROBABILITY-DRIVEN SEARCH ALGORITHM FOR SOLVING MULTI-OBJECTIVE OPTIMIZATION PROBLEMS

A PROBABILITY-DRIVEN SEARCH ALGORITHM FOR SOLVING MULTI-OBJECTIVE OPTIMIZATION PROBLEMS HCMC Unversty of Pedagogy Thong Nguyen Huu et al. A PROBABILITY-DRIVEN SEARCH ALGORITHM FOR SOLVING MULTI-OBJECTIVE OPTIMIZATION PROBLEMS Thong Nguyen Huu and Hao Tran Van Department of mathematcs-nformaton,

More information

Solving Nonlinear Differential Equations by a Neural Network Method

Solving Nonlinear Differential Equations by a Neural Network Method Solvng Nonlnear Dfferental Equatons by a Neural Network Method Luce P. Aarts and Peter Van der Veer Delft Unversty of Technology, Faculty of Cvlengneerng and Geoscences, Secton of Cvlengneerng Informatcs,

More information