Gradual Relaxation Techniques with Applications to Behavioral Synthesis *

Size: px
Start display at page:

Download "Gradual Relaxation Techniques with Applications to Behavioral Synthesis *"

Transcription

1 Gradual Relaxaton Technques wth Applcatons to Behavoral Synthess * Zhru Zhang, Ypng Fan, Modrag Potkonjak, Jason Cong Computer Scence Department, Unversty of Calforna, Los Angeles Los Angeles, CA 90095, USA {zhruz, fanyp, modrag, cong}@cs.ucla.edu ABSTRACT Heurstcs are wdely used for solvng computatonal ntractable synthess problems. However, untl now, there has been lmted effort to systematcally develop heurstcs that can be appled to a varety of synthess tasks. We focus on development of general optmzaton prncples so that they can be appled to a wde range of synthess problems. In partcular, we propose a new way to realze the most constranng prncple where at each step we gradually relax the constrants on the most constraned elements of the soluton. Ths basc optmzaton mechansm s augmented wth several new heurstc prncples: mnmal freedom reducton, negatve thnkng, calbraton, smultaneous step consderaton, and probablstc modelng. We have successfully appled these optmzaton prncples to a number of common behavoral synthess tasks. Specfcally, we demonstrate a systematc way to develop optmzaton algorthms for maxmum ndependent set, tme-constraned schedulng, and soft real-tme system schedulng. The effectveness of the approach and algorthms s valdated on extensve real-lfe benchmarks.. MOTIVATION We have two strategc objectves n ths paper: development of general optmzaton prncples and applcatons to a wde range of synthess problems. The optmzaton goal s to advance the state of the art n the desgn of heurstc algorthms. Heurstcs are wdely used for solvng computatonal ntractable synthess problems. However, untl now, there has been lmted effort to systematcally develop heurstcs that can be easly appled to a varety of synthess tasks. In ths paper, we propose a new heurstc optmzaton paradgm that can be appled on a broad spectrum of computatonally ntractable problems. Whle the tradtonal most constranng prncple always addresses the most constraned part of the problem frst, we employ the most constranng prncple where at each step we make a decson that maxmally relaxes the constrants on the most constraned elements of the soluton. Ths basc optmzaton mechansm s augmented wth several new heurstc nsghts: mnmal freedom reducton, negatve thnkng, calbraton, smultaneous step consderaton, and probablstc modelng. We call them the gradual relaxaton technques. The mnmal freedom reducton prncple ams to make the mnmal possble quantum of decson at each step. The ratonale * Ths research s partally supported by Natonal Scence Foundaton under award CCR s that after a small step s made, one can better evaluate ts mpacts and prevent the heurstc from followng a greedy mode of optmzaton to produce local optmal solutons. The man way to realze ths prncple s to use negatve thnkng,.e., to decde what the optmzaton process wll not do at the next step, nstead of what to do. The optons that stay at the end of ths process form a set of decsons to yeld a hgh qualty soluton. Calbraton s a step where the chances for optmzaton along a partcular drecton are evaluated and compared. If a partcular drecton s not promsng, all efforts along that lne are termnated or assgned a lower prorty. A typcal example s resource allocaton. If the desgn s such that the number of requred adders cannot be reduced below some bound, and there s room for mnmzaton of the number of multplers, the schedulng s conducted so that the operaton that has to be assgned to multplers preserves maxmum slack. Moreover, n some cases, t s advantageous to smultaneously consder several small steps appled on several parts of the problem n order to evaluate ther compound mpact. Fnally, n order to further ncrease the effectveness of the proposed heurstc, we also propose a new, more realstc technque for probablstc modelng durng the optmzaton process. Optmzaton technques can rarely be developed completely out of context. Even more dffcult s to evaluate ther effectveness, unless they are appled on mportant, real-lfe problems and compared wth exstng methods. Therefore, our second and actually man goal s to develop a system of effectve and fast optmzaton technques for common behavoral synthess tasks. Specfcally, we demonstrate a systematc way to develop optmzaton algorthms for maxmum ndependent set, tmeconstraned schedulng, and soft real-tme system schedulng. The problems are selected not only because of ther mportance, but also because they have a very dfferent nature. For example, n statc schedule one has to map all of the operatons to control steps whle n the MIS problem the goal s to select only a subset of nodes. Whle behavoral synthess s stll lookng to establsh tself, the ntuton s that ths wll make algorthms more accurate. Consder the smple motvatonal example shown n Fgure. We lst all the possble confguratons of a CDFG wth operatons scheduled to 5 avalable control steps. We denote prob(, t to be the probablty of assgnng the operaton n control step t. Accordng to tradtonal approaches, unform dstrbuton s assumed. For example, we have prob(, = prob(, = prob(, =0., However, the realstc probabltes should be prob(, = 0.6, prob(, = 0. and prob(, = 0., whch are sgnfcantly dfferent from the prevous ones.

2 c-step c-step c-step c-step4 c-step5 Fgure. A smple example of the non-unform dstrbuton. RELATED WORK In ths secton, we survey the related work along the followng four drectons: maxmally constraned mnmally constranng heurstcs, force-drected optmzaton, statc and soft real-tme system schedulng, and maxmum ndependent set. Btner and Rengold [] were the frst to propose the systematc use of the maxmally constraned varable selecton paradgm. They use t to gude the search wthn a generc backtrackng search for the optmal soluton. The frst to use the maxmally constraned rule as part of fast heurstc program was Brelaz for graph colorng []. Another popular famly of heurstcs s based on noton of slack [7][9]. Slack s usually defned as value proportonal to the cardnalty of a set of stll vable optons for assgnng a specfc varable from the problem formulaton. In CAD lterature, the most constraned least constranng paradgm has been wdely used, mostly under the name of force-drected heurstcs [0]. A more global pcture of the role of a maxmally constraned mnmally constranng approach as an effcent heurstcs s gven n []. Schedulng s a mandatory task n behavoral synthess and complaton [7][8]. In behavoral synthess, schedulng s closely related to resource allocaton, a task where the hardware requred for a specfc applcaton s allocated. We target two mportant and wdely requred types of schedulng: statc and soft real-tme systems. For the former, we assume a synchronous data flow model of computaton [5]. Our goal s to mnmze the requred hardware for a gven samplng rate. The prolferaton of multmeda applcatons and soft real-tme system s ganng rapdly ncreasng attenton. The bass for softreal system schedulng was formed n the work done n the Garca-Molna research group [][]. In multmeda research, soft real-tme system are most often used to model vdeo and WWW servers []. Real tme system communty has placed emphass on a sound formal defnton of requrements assocated wth soft real-tme systems [6]. In CAD and embedded system lterature notable soft real-tme related efforts nclude [8][7][] Maxmum Independent Set (MIS s one of the most popular generc NP-complete problems [8]. For example, t was among the frst set of problems proved to be NP-complete [8]. MIS s manly used n one of ts modfcatons as a set durng graph colorng. For example, t has been expermentally demonstrated that n many domans, fndng an MIS s suffcent to make colorng popular benchmarks both fast and provably optmal [4].. GENERIC TECHNIQUE In ths secton, we ntroduce the seven man new nsghts that can be used to buld a new generaton of effcent heurstcs for a varety of computatonally ntractable problems. The nsghts are as follows: ( Most constraned prncple. The tradtonal most constraned paradgm for developng heurstcs looks to frst resolve the components (e.g., nodes or varable that partcpated n the largest number of constrants or the constrants that are very strct. The ratonale s that t s better to resolve ths stuaton early, whle there s stll enough freedom to resolve them. The components are usually resolved n such a way that the dffculty of stll unresolved constrants s mnmally mpacted. In practce, the effectveness of these heurstcs s manly due to the second nsght. We propose to use a new nsght that resolves the components that partcpate n the same constrants as the most dffcult components. Note that these are not necessarly the most dffcult components themselves. The ratonale s that these components can often be resolved n such a way as to make the most dffcult constrants more relaxed. ( Mnmal freedom reducton. The key mpedment to effectveness of a heurstc s the often greedy behavor of optmzaton, where short term benefts are acheved at the expense of ncreased dffculty and cost n later stages of optmzaton. One way to avod such behavor s to make a small gradual atomc decson and evaluate ts ndvdual mpact before commttng to large decsons. ( Negatve thnkng. Ths dea s a specfc way to realze the prevous paradgm. Essentally, t s often more benefcal to state what wll not be consdered as opton nstead of resolvng a specfc component of the soluton. The smlar deas appeared n [6] for a specal case of schedulng and n [4] for the standard cell global routng (.e., the teratve deleton method. We generalze t other problems. For example, for the MIS problem, t means that we wll elmnate nodes as canddate for MIS one by one, untl the remanng set of nodes does not have any ncdent edges. Ths s n contrast to the standard procedure of selectng nodes for MIS one by one. (4 Compoundng Varables. At frst sght t seems that t s mpossble to apply the prevous two paradgms to problems where varables can be assgned only to bnary values. For example, n the SAT problem, the negatve decson to not assgn a varable to s that the varable s assgned to 0. Ths dffculty can be resolved, f we consder two or more varables smultaneously. For example, we may decde that varable x and x j wll not be smultaneously assgned to value. Therefore, after ths decson, there are stll three optons how these two varables can be smultaneously assgned. (5 Smultaneous steps consderaton. The prevous dea can be further generalzed even when there s no ntrnsc need to create compound varables. The practce ndcates that t s often advantageous to smultaneously consder a small negatve decson on a set of varables. (6 Calbraton. In many stuatons, one can conclude that the way a subset of varables s resolved wll not have an mpact on the qualty of the fnal soluton. In that stuaton, there s no need to preserve the optons for these varables. The soluton to the resource allocaton problem dscussed n Secton s a typcal example of calbraton.

3 (7 Probablstc modelng. Our motvatonal example ndcates ths assumpton s rarely close to realty, even when very small nstances are consdered. Therefore, we propose to model optons on how to resolve each varable n a nonunform way as a functon of all constrants that are mposed on that partcular varable. 4. WHEN IS GRADUAL RELAXATION MOST EFFECTIVE? It s well know that dfferent types of algorthms are best suted for dfferent types of optmzaton problems. For example, greedy algorthms are optmal and the fastest for the problems that have matrod structure [9]. Smlarly, n order to apply dynamc programmng, the problem of nterest has to have the property that the optmal soluton s composed of optmal sub-solutons. At the same tme, t s mportant to realze that the effectveness of a partcular algorthm also greatly depends on the structure of the nstance. For example, t s well known that the graph colorng problem for sparse graphs can be almost always solved optmally [5][5]. In general, however, t s not easy to dentfy classes of problems and nstances that can be solved usng a partcular technque effcently. Most often these nsghts are obtaned by comprehensve studes and statstcal analyss [0][]. Nevertheless, t s possble to obtan some mportant nsghts by consderng specfc characterstcs of the problem or the nstance. We frst focus our attenton on the type of nstances where the system of gradual relaxaton performs well. Force drected approaches work well when there exst sgnfcant dfference n slacks (.e., range of value that a partcular varable can be assgned to and strctness of constrants that are assocated wth each varable. The technque of gradual relaxaton through negatve thnkng provdes addtonal effectveness when a large number (or percentage of varables have sgnfcant slack and when varables have more complex nteractons among a large number of constrants. Ths observaton s a consequence of the fact that n these stuaton gradual relaxaton wll better reveal the mpact of each optmzaton decson. Compoundng varables and smultaneous steps consderaton technques are most effectve when each varable tself has a set of potental values wth small cardnalty. In partcular, t s effectve for bnary varables. Calbraton s most effectve for nstances where the fnal soluton only nvolves relatvely few occurrences of each type of resources. Fnally, probablstc modelng s most relevant for large and complex nstances. Furthermore, we expect that problems where each varable can be assgned to many values (e.g. colorng and schedulng are more amenable for optmzaton usng gradual relaxaton than problems that are defned on bnary varables (e.g. SAT. 5. DRIVER EXAMPLES In ths secton, we demonstrate how a properly selected subset of the proposed heurstc optmzaton prncples can be appled to common behavoral synthess tasks. 5. Maxmum Independent Set and Graph Colorng In ths subsecton, the MIS problem s used to demonstrate the most constranng, mnmal freedom reducton, and negatve thnkng prncples. In behavoral synthess, the resource sharng problem s usually transformed to the graph colorng problem for the correspondng resource conflct graph. Unfortunately, graph colorng s an NPcomplete problem n general [8], defned as: Problem: Graph k-colorng Gven: ( A graph G (V, E wth vertex set V and edge set E, ( A postve nteger K V. Objectve: To determne whether there exsts a colorng of G usng no more than k unque colors so that for every edge (u, v E, u and v have dfferent colors. The graph colorng problem s to fnd the mnmum number of colors needed to color a gven graph. An effcent heurstc to solve ths problem was presented n [4]. Its approach s to dvde the whole problem nto seral ndependent set problems,.e., n every step the algorthm searches a constraned ndependent set and assgns one color to ts vertces. The ndependent set search s bascally an teratve mprovement algorthm, whch randomly generates an ntal soluton and defnes the move as vertex excluson/ncluson. The maxmum ndependent set s also an NP-complete problem stated as: Problem: Maxmum Independent Set Gven: a graph G (V, E wth vertex set V and edge set E. Objectve: To fnd the maxmum-sze ndependent set n G. An ndependent set of G s a subset V V of vertces such that f u V and v V, then u and v are not adjacent. We focus on applyng the gradual relaxaton technque to solve the MIS problem. We observe that typcally n a real-lfe graph, the MIS sze s much smaller than the total graph sze. Thus the decson to choose a vertex nto an ndependent set mposes much more constrants to the other node than the decson to choose a vertex to be out of the ndependent set. Followng the mnmal freedom reducton prncple, we drop off vertces gradually from an ntal vertex set to obtan an ndependent set. The algorthm, descrbed n Fgure, s farly straghtforward. We start wth an ntal soluton set contanng all of the vertces. At each step, we select the maxmally constraned vertex and remove t from the current vertex set untl the soluton s legal. We use force to denote the cost functon. Note that a vertex wth a large number of neghbors s unlkely to be n the resultng ndependent set. Hence, we have the followng Level- heurstc for the force calculaton: force (v = Num_of_Neghbors (v However, by lookng forward one level, we have the observaton that for a vertex, f the number of ts neghbors neghbors s very large, t tends to be n the resultng set. Fgure shows ths stuaton. The black vertex has a maxmum degree, but t should be n the MIS wth the eght leaf vertces.

4 S V whle S s not a ndependent set update force for every vertex select the vertex v wth maxmal force remove v from S return S Fgure. The gradual relaxaton algorthm for MIS To descrbe ths look-ahead force, we defne the Level- heurstc as follows force (v = u Neghbors (v ( / Num_of_Neghbors (u The force update s n O( V tme for the Level- heurstc and O( V for the Level- heurstc. In every step of the algorthm, we remove one vertex. Hence the tme complexty s O( V and O( V. Fgure. An example of the Level- heurstc 5. Tme-Constraned Schedulng for Behavoral Synthess Problem: Tme-Constraned Schedulng Gven: ( A CDFG G (V, E. ( A tmng constrant T. Objectve: to schedule the operatons of V nto T control steps such that the resource usage s mnmzed. In ths subsecton we show how to use the negatve thnkng prncple to mprove the wdely accepted force-drected schedulng (FDS algorthm [0]. 5.. Enhancement by Negatve Thnkng The essence of FDS algorthm s to reduce the resource usage by evenly dstrbutng the operatons to the avalable control steps. The FDS algorthm uses the unform probablty that an operaton wll be scheduled wthn ts ALAP-ASAP range. At each step of FDS algorthm, the tme frame of the most constraned operaton wth least forces wll be reduced to a sngle control step. That s a sgnfcant step wth potentally hgh mpact. Followng the most constranng and negatve thnkng prncples, we frst select the operaton that has optons to be scheduled to the most congested tme slot. After that, we prune only one control step, the most congested one, from the operaton s tme frame. Therefore, we always keep a global pcture before schedulng an operaton by gradually shrnkng the tme frames and postpone the decson to a later stage when more accurate estmates are avalable. 5.. Effcency Improvement Snce the above method tends to ncrease the tmng complexty, we employ the Gradual Tme-Frame Reducton (GTFR technque proposed n [6] to speedup the process. Suppose the tme frame of the operaton n s denoted as [a n, b n ]. In each teraton, GTFR only computes for each node n the forces of assgnng n to a n and b n nstead of all ts feasble control steps. The assgnment ncurrng the most force wll be dentfed and the correspondng control step wll be removed. As ponted out n [6], the tme complexty of ths algorthm s O(T N, where T s the tmng constrant and N s the number of operatons. It s the same as that of the basc force-drected schedulng. Although the removal of each control step ncreases the number of teraton to O(TN, the tentatve assgnments has to be calculated reduce to, nstead of O(T tmes. 5. Soft Real-Tme Schedulng In ths subsecton, we use soft real-tme schedulng to llustrate the effectveness of mnmal freedom reducton and probablstc modelng prncples. Problem: Soft Real-Tme Schedulng Gven: ( A set of Ψ={τ,τ, τ n } tasks and each task τ =(a, d, e s characterzed by an arrval tme a, a deadlne d and an executon tme e. ( A sngle processor P. ( A tmng constrant T. Assumptons: All tasks are non-preemptve and ndependent. Objectve: To schedule a subset of tasks n Ψ on processor P wthn the avalable tme T so that the number of tasks scheduled s maxmzed. 5.. Basc Defntons The tme frame of a task s the tme nterval wthn whch t can be scheduled. Tme frame of task τ s denoted as tme-frame=[a, d ]. The moblty of a task determnes the maxmum number of dfferent schedules for t. We denote the moblty of a task τ as moblty(τ = tme-frame(τ -e +. We denote the probablty of the task τ at a tme slot t as prob(τ, t. It can be derved by the equaton shown n the left sde of Fgure 4. The probablty dstrbuton s non-unform as shown by the trapezum-shape curve n the rght sde of Fgure 4. (t - a + moblty( τ (d - t + prob (τ, t = moblty( τ e moblty( τ t - a + < e d - t + < e Otherwse prob(τ, t e / moblty(τ a a +e d -e Fgure 4. Defnton and the curve of Prob (τ ι, t A dstrbuton graph s defned to capture the lkelhood of the potental compettons among dfferent tasks for a sngle tme slot. The dstrbuton at the tme slot t s computed by DG( t prob( τ, t. Two tasks τ and τ j are sad to be n = conflct when ther tme frames overlap. It s denoted by conflct(τ,τ j =(tme-frame(τ tme-frame(τ j. = n 5.. Soluton Overvew Observe that the less the conflcts exst, the more chances we may have to schedule a larger number of the tasks. We use a two-step heurstc to solve the problem. In the frst step, the tasks are scheduled to mnmze the number of conflcts. Snce ths ntal soluton may not be feasble, the legalzaton s performed n the second step to get the best legal schedule n terms of the number of scheduled tasks. t d

5 Soft-Real-Tme-Schedulng (Ψ /*step: conflct mnmzaton*/ whle there exsts unlocked task n Ψ compute the dstrbuton graph for each unlocked task τ a-force( self-force-arrval(τ d-force( self-force-deadlne(τ choose task τ k whch ncurs the maxmal force //shrnk the tme frame of τ k f a-force(k d-force(k then Tme-Frame(τ k [a k+,d k] else Tme-Frame(τ k [a k,d k-] f Tme-Frame(τ k =e k then set τ k locked /*step: legalzaton*/ sort all τ n the non-decreasng order of a V {v.. v n} for to n for j + to n f all the predcatons hold then add an edge <v, v j> nto G schedule T by traversng the longest path n G Fgure 5. Soft real-tme schedulng algorthm 5.. Conflct Mnmzaton In order to mnmze the conflcts, we agan apply the mnmal freedom reducton technque to gradually shrnk the tme frames. The force concept s also employed to balance the conflcts along the avalable tme slots. We only consder the self-force of relatng a task τ to ts arrval tme and ts deadlne. They are denoted by self-force-arrval(τ and self-force-deadlne(τ, respectvely. self-force-arrval(τ = a + e d prob( τ, t prob( τ, t DG ( t( + DG( t(0 t = a self-force-dealne(τ = d t = d e + tasks( t t = a + e t = a tasks( t d + e prob ( τ, t prob ( τ, t DG ( t ( + DG ( t ( 0 tasks ( t tasks ( t At each step, the assgnment of a task to a tme slot whch results n hghest force wll be rejected, and the tme frame of the correspondng task τ wll be shrunk accordngly. When the sze of a task s tme frame s reduced to ts executon tme, ths task wll be locked (marked as sem-scheduled. The frst step algorthm ends when all tasks are locked wth the hope that they are evenly dstrbuted. The frst part of Fgure 5 llustrates the algorthm Legalzaton After the conflct mnmzaton, all tasks have been semscheduled, namely, the tme frame of each task s shrunk exactly to ts executon tme. A compatblty graph G can be bult from the ntal soluton. Defne G=(V, E where V={v n}. An edge <τ,τ j > E f and only f all the followng three predcatons hold: ( conflct(τ,τ j : τ andτ j are compatble ( a a j : τ arrves earler than τ j ( k <v,v k > E <v k,v j > E: edge <v, v j > s rredundant. It can be easly shown that G s a drected acyclc graph and we can get the best legal schedule by traversng the longest path n G. A task τ s scheduled n ts tme frame f v appears n the longest path, otherwse t s dropped. Note that the longest path n a DAG can be optmally found n lnear tme. The algorthm s llustrated n the second part of Fgure Expermental Results 5.4. Maxmum Independent Set To evaluate our MIS algorthm, we take a set of DIMACS benchmark graphs for the Clque problem challenge. Snce the maxmum clque of a graph s the maxmum ndependent set n the complementary graph, we complemented these benchmarks and apply the algorthm dscussed n Secton 5., as well as Krovsk s algorthm [4] for comparson. Our programs are mplemented wth C++/UNIX envronment. Krovsk s program s the open source from hs personal webste. The experments are run wth a Sun Blade 000 workstaton wth 750 MHz frequency. The expermental results are shown n Table 4. The frst column lsts the name of each benchmark. The second and thrd columns are the correspondng vertex and edge numbers. In column Opt, we lst the optmal soluton f known. The next three columns are ndependent set szes produced by three algorthms. Columns A and B are the results from our algorthm wth the Level- and Level- heurstc respectvely. Column C lsts the results from Krovsk s teratve mprovement algorthm. In most of the cases, our Level- heurstc produced better solutons than the Level- heurstc. Compared wth Krovsk s algorthm, our Level- algorthm acheves only margnally worse result but wth a dramatc mprovement n runtme. As shown n Table 5, on average, our Level- algorthm s more than 50 tmes faster than the Level- algorthm, whch s more than 0 tmes faster than the teratve mprovement algorthm Behavoral Synthess Schedulng Table. Schedulng results comparson under crtcal-path tme constrant Node Cycles ALU ALU* MULT MULT* FFT 7 7 WANG MCM HONDA DIR CHEM_PLANT AIRCRAFT_S _u5ml _u5ml FEIG_DCT GSM AIRCRAFT_L Total Improvement % - 4% We mplemented both the force-drected algorthm and our gradual relaxaton schedulng approach n C++/STL. The selected DFG benchmarks were taken from [] wth several DCT algorthms (WANG, DIR, FEIG_DCT, and a set of real-tme

6 DSP programs. For the expermental results lsted n Table, we used the crtcal path as the tme constrant. The second and the thrd columns lst the sze of DFG and ts crtcal path length, respectvely. The followng columns lst the number of requred functon unt ALU and MULT for dfferent algorthms. Columns named ALU and MULT are the results from the force-drected algorthm, and columns ALU* and MULT* are from our algorthm. For small examples (FFT and WANG, these two algorthms produced the same results. As the DFG sze and parallelsm became larger, our algorthm consstently outperformed the forcedrected algorthm. On average, our algorthm acheved about % resource reducton compared wth the force-drected algorthm. Table. Schedulng results comparson under tme constrant wth.5 tmes of the crtcal path Node Cycles ALU ALU* MULTMULT* FFT 7 0 WANG MCM HONDA DIR CHEM_PLANT AIRCRAFT_S _u5ml _u5ml FEIG_DCT GSM AIRCRAFT_L Total Improvement % - 8% Another set of expermental results n Table s produced when we set the tmng constrant as the.5 tmes of the crtcal path length. Our algorthm acheves 5% ALU reducton and 8% MULT reducton compared wth the force-drected algorthm. Ths ndcates that when we relax the tme constrant, our gradual relaxaton technque performs even better wth more slack on each operaton node Soft Real-Tme Schedulng Table. Soft real-tme schedulng results T Task No. Perod EDF. No. Busy Utl. p_ts p_ts p_ts p4_ts p_ts p_ts p_ts p4_ts Average A set of benchmarks from [6] was appled to evaluate our soft real-tme schedulng algorthm. As shown n Table, every cell n the frst column s a task set executed n a partcular processor. The second and thrd columns are the task numbers and schedulng perods. The fourth column lsts resultng scheduled task numbers from the Earlest-Deadlne Frst (EDF schedulng The full schedulablty s not achevable n our testcases. Otherwse, EDF would generate optmal results. [4] for comparson. The next three columns lst resultng scheduled task number (No., processor busy tme (Busy, and utlzaton rato (Utl. from our algorthm. On average, our algorthm schedules.5 more tasks than EDF, whch s beng used n many real-tme applcatons. Our algorthm acheved about 86% utlzaton rato for ths benchmark set. 6. CONCLUSIONS We ntroduce a sute of new heurstc prncples for desgn of algorthms for computatonal ntractable synthess tasks. The effectveness of these prncples s demonstrated on three mportant behavoral synthess problems: maxmal ndependent set, statc tme-constraned schedulng, and synthess of soft realtme systems. 7. REFERENCES [] B. Adelberg, H. Garca-Molna; and B. Kao, Emulatng Soft Real- Tme Schedulng Usng Tradtonal Operatng System Schedulers, n Proceedngs of Real-Tme Systems Symposum, pp. 9-98, Dec [] J. Btner and E. Rengold, Backtrack Programmng Technques, Communcatons of the ACM, vol. 8(, pp , Nov [] D. Brelaz, New Methods to Color the Vertces of a Graph, Communcatons of the ACM, vol. (4, pp. 5-56, Apr [4] J. Cong and Patrck H. Madden, Performance Drven Global Routng for Standard Cell Desgn, n Proceedngs of Internatonal Symposum on Physcal Desgn, pp. 7-80, Apr [5] O. Coudert, Exact Colorng of Real-Lfe Graphs s Easy, n Proceedngs of the 4th Desgn Automaton Conference, pp. - 6, Jun [6] P. D'Argeno, J.-P. Katoen, and E. Brnksma, Specfcaton and Analyss of Soft Real-Tme Systems: Quantty and Qualty, n Proceedngs of the 0th IEEE Real-Tme Systems Symposum, pp. 04-4, Dec [7] R. I. Davs, K. W. Tndell, and A. Burns, Schedulng Slack Tme n Fxed Prorty Pre-emptve Systems, n Proceedngs of Real-Tme Systems Symposum, pp. -, Dec. 99. [8] M. R. Garey and D. S. Johnson, Computers and Intractablty: A Gude to the Theory of NP-Completeness, San Francsco: Freeman, 979. [9] M. Goldwasser, Patence s a Vrtue: The Effect of Slack on Compettveness for Admsson Control, to appear n Journal of Schedulng, 00. [0] D. S. Johnson, C. R. Aragon, L. A. McGeoch, and C. Schevon, Optmzaton by Smulated Annealng: An Expermental Evaluaton; Part, Graph Parttonng, Operatons Research, vol. 7(6, pp , 989 [] D. S. Johnson, C. R. Aragon, L. A. McGeoch, and C. Schevon, Optmzaton by Smulated Annealng: Part II, Graph Colorng and Number Parttonng, Operatons Research, vol. 9(, pp , 99. [] M. Jones, D. Rosu, and M.-C. Rosu, CPU Reservatons and Tme Constrants: Effcent, Predctable Schedulng of Independent Actvtes, n Proceedngs of the 6th ACM Symposum on Operatng System Prncples, vol. (5, pp. 98-, Dec [] B. Kao and H. Garca-Molna, Subtask Deadlne Assgnment for Complex Dstrbuted Soft Real-Tme Tasks, n Proceedngs of Internatonal Conference on Dstrbuted Computng Systems, pp. 7-8, Jun [4] D. Krovsk and M. Potkonjak, Effcent Colorng of a Large Spectrum of Graphs, n Proceedngs of the 5th Desgn Automaton Conference, pp. 47-4, Jun. 998.

7 [5] E. A. Lee and D. G. Messerschmtt, Synchronous Data Flow, n Proceedngs of the IEEE, vol. 75(9, pp. 5-45, Sep [6] C. Lee, M. Potkonjak, and W. H. Wolf, Synthess of Hard Real- Tme Applcaton Specfc Systems, Desgn Automaton for Embedded Systems, vol. 4(4, pp. 5-4, Oct [7] M. C. McFarland, A. C. Parker, and R. Camposano, The Hgh- Level Synthess of Dgtal Systems, n Proceedngs of the IEEE, vol. 78(, pp. 0-8, Feb [8] G. D. Mchel, "Synthess and Optmzaton of Dgtal Crcuts," McGraw-Hll, 994. [9] C. Papadmtrou and K. Stegltz, Combnatoral Optmzaton: Algorthms and Complexty, Prentce Hall, 98. [0] P. G. Pauln and J. P. Knght, Force-Drected Schedulng for the Behavoral Synthess of ASICs, n IEEE Trans. on Computer-Aded Desgn of Integrated Crcuts and Systems, vol. 8(6, pp , Jun [] J. Pearl, Heurstcs: Intellgent Search Strateges for Computer Problem Solvng, Addson-Wesley, 984. [] K. Rchter, D. Zegenben, M. Jersak, and R. Ernst, Model Composton for Schedulng Analyss n Platform Desgn, n Proceedngs of the 9th Desgn Automaton Conference, pp. 87-9, Jun. 00. [] M. B. Srvastava and M. Potkonjak, Optmum and Heurstc Transformaton Technques for Smultaneous Optmzaton of Latency and Throughput, n IEEE Trans. on Very Large Scale Integraton Systems, vol. (, pp. -9, Mar [4] J. A. Stankovc, M. Spur, M. D. Natale, and G. C. Buttazzo, Implcatons of Classcal Schedulng Results for Real-Tme Systems, IEEE Computer, vol. 8(6, pp. 6-5, Jun [5] J. S. Turner, Almost All k-colorable Graphs are Easy to Color, Journal of Algorthms, vol. 9, pp. 6-8, Mar [6] W. F. J. Verhaegh, P. E. R. Lppens, E. H. L. Aarts, J. H. M. Korst, J. L. van Meerbergen, and A. van der Werf, Improved Force-Drected Schedulng n Hgh-Throughput Dgtal Sgnal Processng, n IEEE Trans. on Computer-Aded Desgn of Integrated Crcuts and Systems, vol. 4(8, pp , Aug [7] D. Verkest, P. Yang, C. Wong, and P. Marchal, Optmsaton Problems for Dynamc Concurrent Task-Based Systems, n Proceedngs of the 00 Internatonal Conference on Computer- Aded Desgn, pp , Nov. 00. [8] D. Zegenben, J. Uerpmann, and R. Ernst, Dynamc Response Tme Optmzaton for SDF Graphs, n Proceedngs of the 000 Internatonal Conference on Computer-Aded Desgn, pp. 5-4, Nov Table 4. Soluton qualty comparson of three MIS algorthms Example V Edges Opt A B C brock00_ ? c-fat c-fat c-fat c-fat c-fat hammng hammng ? 4 8 hammng hammng johnson ? johnson MANN_a ? MANN_a ? MANN_a ? MANN_a ? p_hat ? p_hat ? p_hat ? p_hat ? p_hat ? p_hat p_hat p_hat p_hat ? p_hat ? p_hat ? sanr00_ sanr00_ ? sanr400_ ? 0 sanr400_ ? 7 0 Table 5. Runtme comparson of three MIS algorthms (sec Example V Edges Tme A Tme B Tme C brock00_ c-fat c-fat c-fat c-fat c-fat hammng hammng hammng hammng johnson johnson MANN_a MANN_a MANN_a MANN_a p_hat p_hat p_hat p_hat p_hat p_hat p_hat p_hat p_hat p_hat p_hat sanr00_ sanr00_ sanr400_ sanr400_ Total

Parallelism for Nested Loops with Non-uniform and Flow Dependences

Parallelism for Nested Loops with Non-uniform and Flow Dependences Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr

More information

An Optimal Algorithm for Prufer Codes *

An Optimal Algorithm for Prufer Codes * J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,

More information

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields

A mathematical programming approach to the analysis, design and scheduling of offshore oilfields 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 A mathematcal programmng approach to the analyss, desgn and

More information

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz

Compiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz Compler Desgn Sprng 2014 Regster Allocaton Sample Exercses and Solutons Prof. Pedro C. Dnz USC / Informaton Scences Insttute 4676 Admralty Way, Sute 1001 Marna del Rey, Calforna 90292 pedro@s.edu Regster

More information

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique

The Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique //00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy

More information

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation

An Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation 17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 An Iteratve Soluton Approach to Process Plant Layout usng Mxed

More information

TPL-Aware Displacement-driven Detailed Placement Refinement with Coloring Constraints

TPL-Aware Displacement-driven Detailed Placement Refinement with Coloring Constraints TPL-ware Dsplacement-drven Detaled Placement Refnement wth Colorng Constrants Tao Ln Iowa State Unversty tln@astate.edu Chrs Chu Iowa State Unversty cnchu@astate.edu BSTRCT To mnmze the effect of process

More information

Support Vector Machines

Support Vector Machines /9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.

More information

DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT

DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT DESIGNING TRANSMISSION SCHEDULES FOR WIRELESS AD HOC NETWORKS TO MAXIMIZE NETWORK THROUGHPUT Bran J. Wolf, Joseph L. Hammond, and Harlan B. Russell Dept. of Electrcal and Computer Engneerng, Clemson Unversty,

More information

Meta-heuristics for Multidimensional Knapsack Problems

Meta-heuristics for Multidimensional Knapsack Problems 2012 4th Internatonal Conference on Computer Research and Development IPCSIT vol.39 (2012) (2012) IACSIT Press, Sngapore Meta-heurstcs for Multdmensonal Knapsack Problems Zhbao Man + Computer Scence Department,

More information

Load Balancing for Hex-Cell Interconnection Network

Load Balancing for Hex-Cell Interconnection Network Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,

More information

AADL : about scheduling analysis

AADL : about scheduling analysis AADL : about schedulng analyss Schedulng analyss, what s t? Embedded real-tme crtcal systems have temporal constrants to meet (e.g. deadlne). Many systems are bult wth operatng systems provdng multtaskng

More information

An Entropy-Based Approach to Integrated Information Needs Assessment

An Entropy-Based Approach to Integrated Information Needs Assessment Dstrbuton Statement A: Approved for publc release; dstrbuton s unlmted. An Entropy-Based Approach to ntegrated nformaton Needs Assessment June 8, 2004 Wllam J. Farrell Lockheed Martn Advanced Technology

More information

Comparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments

Comparison of Heuristics for Scheduling Independent Tasks on Heterogeneous Distributed Environments Comparson of Heurstcs for Schedulng Independent Tasks on Heterogeneous Dstrbuted Envronments Hesam Izakan¹, Ath Abraham², Senor Member, IEEE, Václav Snášel³ ¹ Islamc Azad Unversty, Ramsar Branch, Ramsar,

More information

The Codesign Challenge

The Codesign Challenge ECE 4530 Codesgn Challenge Fall 2007 Hardware/Software Codesgn The Codesgn Challenge Objectves In the codesgn challenge, your task s to accelerate a gven software reference mplementaton as fast as possble.

More information

For instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1)

For instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1) Secton 1.2 Subsets and the Boolean operatons on sets If every element of the set A s an element of the set B, we say that A s a subset of B, or that A s contaned n B, or that B contans A, and we wrte A

More information

A Facet Generation Procedure. for solving 0/1 integer programs

A Facet Generation Procedure. for solving 0/1 integer programs A Facet Generaton Procedure for solvng 0/ nteger programs by Gyana R. Parja IBM Corporaton, Poughkeepse, NY 260 Radu Gaddov Emery Worldwde Arlnes, Vandala, Oho 45377 and Wlbert E. Wlhelm Teas A&M Unversty,

More information

Parallel matrix-vector multiplication

Parallel matrix-vector multiplication Appendx A Parallel matrx-vector multplcaton The reduced transton matrx of the three-dmensonal cage model for gel electrophoress, descrbed n secton 3.2, becomes excessvely large for polymer lengths more

More information

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour 6.854 Advanced Algorthms Petar Maymounkov Problem Set 11 (November 23, 2005) Wth: Benjamn Rossman, Oren Wemann, and Pouya Kheradpour Problem 1. We reduce vertex cover to MAX-SAT wth weghts, such that the

More information

Module Management Tool in Software Development Organizations

Module Management Tool in Software Development Organizations Journal of Computer Scence (5): 8-, 7 ISSN 59-66 7 Scence Publcatons Management Tool n Software Development Organzatons Ahmad A. Al-Rababah and Mohammad A. Al-Rababah Faculty of IT, Al-Ahlyyah Amman Unversty,

More information

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization

Problem Definitions and Evaluation Criteria for Computational Expensive Optimization Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty

More information

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana

More information

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration

Improvement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,

More information

Real-time Fault-tolerant Scheduling Algorithm for Distributed Computing Systems

Real-time Fault-tolerant Scheduling Algorithm for Distributed Computing Systems Real-tme Fault-tolerant Schedulng Algorthm for Dstrbuted Computng Systems Yun Lng, Y Ouyang College of Computer Scence and Informaton Engneerng Zheang Gongshang Unversty Postal code: 310018 P.R.CHINA {ylng,

More information

5 The Primal-Dual Method

5 The Primal-Dual Method 5 The Prmal-Dual Method Orgnally desgned as a method for solvng lnear programs, where t reduces weghted optmzaton problems to smpler combnatoral ones, the prmal-dual method (PDM) has receved much attenton

More information

Mathematics 256 a course in differential equations for engineering students

Mathematics 256 a course in differential equations for engineering students Mathematcs 56 a course n dfferental equatons for engneerng students Chapter 5. More effcent methods of numercal soluton Euler s method s qute neffcent. Because the error s essentally proportonal to the

More information

Solving two-person zero-sum game by Matlab

Solving two-person zero-sum game by Matlab Appled Mechancs and Materals Onlne: 2011-02-02 ISSN: 1662-7482, Vols. 50-51, pp 262-265 do:10.4028/www.scentfc.net/amm.50-51.262 2011 Trans Tech Publcatons, Swtzerland Solvng two-person zero-sum game by

More information

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms

Course Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms Course Introducton Course Topcs Exams, abs, Proects A quc loo at a few algorthms 1 Advanced Data Structures and Algorthms Descrpton: We are gong to dscuss algorthm complexty analyss, algorthm desgn technques

More information

GSLM Operations Research II Fall 13/14

GSLM Operations Research II Fall 13/14 GSLM 58 Operatons Research II Fall /4 6. Separable Programmng Consder a general NLP mn f(x) s.t. g j (x) b j j =. m. Defnton 6.. The NLP s a separable program f ts objectve functon and all constrants are

More information

Repeater Insertion for Two-Terminal Nets in Three-Dimensional Integrated Circuits

Repeater Insertion for Two-Terminal Nets in Three-Dimensional Integrated Circuits Repeater Inserton for Two-Termnal Nets n Three-Dmensonal Integrated Crcuts Hu Xu, Vasls F. Pavlds, and Govann De Mchel LSI - EPFL, CH-5, Swtzerland, {hu.xu,vasleos.pavlds,govann.demchel}@epfl.ch Abstract.

More information

X- Chart Using ANOM Approach

X- Chart Using ANOM Approach ISSN 1684-8403 Journal of Statstcs Volume 17, 010, pp. 3-3 Abstract X- Chart Usng ANOM Approach Gullapall Chakravarth 1 and Chaluvad Venkateswara Rao Control lmts for ndvdual measurements (X) chart are

More information

Configuration Management in Multi-Context Reconfigurable Systems for Simultaneous Performance and Power Optimizations*

Configuration Management in Multi-Context Reconfigurable Systems for Simultaneous Performance and Power Optimizations* Confguraton Management n Mult-Context Reconfgurable Systems for Smultaneous Performance and Power Optmzatons* Rafael Maestre, Mlagros Fernandez Departamento de Arqutectura de Computadores y Automátca Unversdad

More information

Verification by testing

Verification by testing Real-Tme Systems Specfcaton Implementaton System models Executon-tme analyss Verfcaton Verfcaton by testng Dad? How do they know how much weght a brdge can handle? They drve bgger and bgger trucks over

More information

Greedy Technique - Definition

Greedy Technique - Definition Greedy Technque Greedy Technque - Defnton The greedy method s a general algorthm desgn paradgm, bult on the follong elements: confguratons: dfferent choces, collectons, or values to fnd objectve functon:

More information

Maintaining temporal validity of real-time data on non-continuously executing resources

Maintaining temporal validity of real-time data on non-continuously executing resources Mantanng temporal valdty of real-tme data on non-contnuously executng resources Tan Ba, Hong Lu and Juan Yang Hunan Insttute of Scence and Technology, College of Computer Scence, 44, Yueyang, Chna Wuhan

More information

Wishing you all a Total Quality New Year!

Wishing you all a Total Quality New Year! Total Qualty Management and Sx Sgma Post Graduate Program 214-15 Sesson 4 Vnay Kumar Kalakband Assstant Professor Operatons & Systems Area 1 Wshng you all a Total Qualty New Year! Hope you acheve Sx sgma

More information

Scheduling with Integer Time Budgeting for Low-Power Optimization

Scheduling with Integer Time Budgeting for Low-Power Optimization Schedlng wth Integer Tme Bdgetng for Low-Power Optmzaton We Jang, Zhr Zhang, Modrag Potkonjak and Jason Cong Compter Scence Department Unversty of Calforna, Los Angeles Spported by NSF, SRC. Otlne Introdcton

More information

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION

CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION 24 CHAPTER 2 PROPOSED IMPROVED PARTICLE SWARM OPTIMIZATION The present chapter proposes an IPSO approach for multprocessor task schedulng problem wth two classfcatons, namely, statc ndependent tasks and

More information

A GENETIC ALGORITHM FOR PROCESS SCHEDULING IN DISTRIBUTED OPERATING SYSTEMS CONSIDERING LOAD BALANCING

A GENETIC ALGORITHM FOR PROCESS SCHEDULING IN DISTRIBUTED OPERATING SYSTEMS CONSIDERING LOAD BALANCING A GENETIC ALGORITHM FOR PROCESS SCHEDULING IN DISTRIBUTED OPERATING SYSTEMS CONSIDERING LOAD BALANCING M. Nkravan and M. H. Kashan Department of Electrcal Computer Islamc Azad Unversty, Shahrar Shahreqods

More information

Determining the Optimal Bandwidth Based on Multi-criterion Fusion

Determining the Optimal Bandwidth Based on Multi-criterion Fusion Proceedngs of 01 4th Internatonal Conference on Machne Learnng and Computng IPCSIT vol. 5 (01) (01) IACSIT Press, Sngapore Determnng the Optmal Bandwdth Based on Mult-crteron Fuson Ha-L Lang 1+, Xan-Mn

More information

Simulation Based Analysis of FAST TCP using OMNET++

Simulation Based Analysis of FAST TCP using OMNET++ Smulaton Based Analyss of FAST TCP usng OMNET++ Umar ul Hassan 04030038@lums.edu.pk Md Term Report CS678 Topcs n Internet Research Sprng, 2006 Introducton Internet traffc s doublng roughly every 3 months

More information

An Investigation into Server Parameter Selection for Hierarchical Fixed Priority Pre-emptive Systems

An Investigation into Server Parameter Selection for Hierarchical Fixed Priority Pre-emptive Systems An Investgaton nto Server Parameter Selecton for Herarchcal Fxed Prorty Pre-emptve Systems R.I. Davs and A. Burns Real-Tme Systems Research Group, Department of omputer Scence, Unversty of York, YO10 5DD,

More information

ARTICLE IN PRESS. Signal Processing: Image Communication

ARTICLE IN PRESS. Signal Processing: Image Communication Sgnal Processng: Image Communcaton 23 (2008) 754 768 Contents lsts avalable at ScenceDrect Sgnal Processng: Image Communcaton journal homepage: www.elsever.com/locate/mage Dstrbuted meda rate allocaton

More information

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data

A Fast Content-Based Multimedia Retrieval Technique Using Compressed Data A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,

More information

Problem Set 3 Solutions

Problem Set 3 Solutions Introducton to Algorthms October 4, 2002 Massachusetts Insttute of Technology 6046J/18410J Professors Erk Demane and Shaf Goldwasser Handout 14 Problem Set 3 Solutons (Exercses were not to be turned n,

More information

Reliability and Energy-aware Cache Reconfiguration for Embedded Systems

Reliability and Energy-aware Cache Reconfiguration for Embedded Systems Relablty and Energy-aware Cache Reconfguraton for Embedded Systems Yuanwen Huang and Prabhat Mshra Department of Computer and Informaton Scence and Engneerng Unversty of Florda, Ganesvlle FL 326-62, USA

More information

A New Approach For the Ranking of Fuzzy Sets With Different Heights

A New Approach For the Ranking of Fuzzy Sets With Different Heights New pproach For the ankng of Fuzzy Sets Wth Dfferent Heghts Pushpnder Sngh School of Mathematcs Computer pplcatons Thapar Unversty, Patala-7 00 Inda pushpndersnl@gmalcom STCT ankng of fuzzy sets plays

More information

A Binarization Algorithm specialized on Document Images and Photos

A Binarization Algorithm specialized on Document Images and Photos A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a

More information

Load-Balanced Anycast Routing

Load-Balanced Anycast Routing Load-Balanced Anycast Routng Chng-Yu Ln, Jung-Hua Lo, and Sy-Yen Kuo Department of Electrcal Engneerng atonal Tawan Unversty, Tape, Tawan sykuo@cc.ee.ntu.edu.tw Abstract For fault-tolerance and load-balance

More information

A Saturation Binary Neural Network for Crossbar Switching Problem

A Saturation Binary Neural Network for Crossbar Switching Problem A Saturaton Bnary Neural Network for Crossbar Swtchng Problem Cu Zhang 1, L-Qng Zhao 2, and Rong-Long Wang 2 1 Department of Autocontrol, Laonng Insttute of Scence and Technology, Benx, Chna bxlkyzhangcu@163.com

More information

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr)

Helsinki University Of Technology, Systems Analysis Laboratory Mat Independent research projects in applied mathematics (3 cr) Helsnk Unversty Of Technology, Systems Analyss Laboratory Mat-2.08 Independent research projects n appled mathematcs (3 cr) "! #$&% Antt Laukkanen 506 R ajlaukka@cc.hut.f 2 Introducton...3 2 Multattrbute

More information

Lecture 7 Real Time Task Scheduling. Forrest Brewer

Lecture 7 Real Time Task Scheduling. Forrest Brewer Lecture 7 Real Tme Task Schedulng Forrest Brewer Real Tme ANSI defnes real tme as A Real tme process s a process whch delvers the results of processng n a gven tme span A data may requre processng at a

More information

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS

A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Proceedngs of the Wnter Smulaton Conference M E Kuhl, N M Steger, F B Armstrong, and J A Jones, eds A MOVING MESH APPROACH FOR SIMULATION BUDGET ALLOCATION ON CONTINUOUS DOMAINS Mark W Brantley Chun-Hung

More information

CHAPTER 2 DECOMPOSITION OF GRAPHS

CHAPTER 2 DECOMPOSITION OF GRAPHS CHAPTER DECOMPOSITION OF GRAPHS. INTRODUCTION A graph H s called a Supersubdvson of a graph G f H s obtaned from G by replacng every edge uv of G by a bpartte graph,m (m may vary for each edge by dentfyng

More information

Smoothing Spline ANOVA for variable screening

Smoothing Spline ANOVA for variable screening Smoothng Splne ANOVA for varable screenng a useful tool for metamodels tranng and mult-objectve optmzaton L. Rcco, E. Rgon, A. Turco Outlne RSM Introducton Possble couplng Test case MOO MOO wth Game Theory

More information

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points; Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features

More information

Cluster Analysis of Electrical Behavior

Cluster Analysis of Electrical Behavior Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School

More information

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints

Sum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints Australan Journal of Basc and Appled Scences, 2(4): 1204-1208, 2008 ISSN 1991-8178 Sum of Lnear and Fractonal Multobjectve Programmng Problem under Fuzzy Rules Constrants 1 2 Sanjay Jan and Kalash Lachhwan

More information

Mixed-Criticality Scheduling on Multiprocessors using Task Grouping

Mixed-Criticality Scheduling on Multiprocessors using Task Grouping Mxed-Crtcalty Schedulng on Multprocessors usng Task Groupng Jankang Ren Lnh Th Xuan Phan School of Software Technology, Dalan Unversty of Technology, Chna Computer and Informaton Scence Department, Unversty

More information

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Learning the Kernel Parameters in Kernel Minimum Distance Classifier Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department

More information

EVALUATION OF THE PERFORMANCES OF ARTIFICIAL BEE COLONY AND INVASIVE WEED OPTIMIZATION ALGORITHMS ON THE MODIFIED BENCHMARK FUNCTIONS

EVALUATION OF THE PERFORMANCES OF ARTIFICIAL BEE COLONY AND INVASIVE WEED OPTIMIZATION ALGORITHMS ON THE MODIFIED BENCHMARK FUNCTIONS Academc Research Internatonal ISS-L: 3-9553, ISS: 3-9944 Vol., o. 3, May 0 EVALUATIO OF THE PERFORMACES OF ARTIFICIAL BEE COLOY AD IVASIVE WEED OPTIMIZATIO ALGORITHMS O THE MODIFIED BECHMARK FUCTIOS Dlay

More information

Complex Numbers. Now we also saw that if a and b were both positive then ab = a b. For a second let s forget that restriction and do the following.

Complex Numbers. Now we also saw that if a and b were both positive then ab = a b. For a second let s forget that restriction and do the following. Complex Numbers The last topc n ths secton s not really related to most of what we ve done n ths chapter, although t s somewhat related to the radcals secton as we wll see. We also won t need the materal

More information

Dynamic Voltage Scaling of Supply and Body Bias Exploiting Software Runtime Distribution

Dynamic Voltage Scaling of Supply and Body Bias Exploiting Software Runtime Distribution Dynamc Voltage Scalng of Supply and Body Bas Explotng Software Runtme Dstrbuton Sungpack Hong EE Department Stanford Unversty Sungjoo Yoo, Byeong Bn, Kyu-Myung Cho, Soo-Kwan Eo Samsung Electroncs Taehwan

More information

Optimal Scheduling of Capture Times in a Multiple Capture Imaging System

Optimal Scheduling of Capture Times in a Multiple Capture Imaging System Optmal Schedulng of Capture Tmes n a Multple Capture Imagng System Tng Chen and Abbas El Gamal Informaton Systems Laboratory Department of Electrcal Engneerng Stanford Unversty Stanford, Calforna 9435,

More information

Real-Time Systems. Real-Time Systems. Verification by testing. Verification by testing

Real-Time Systems. Real-Time Systems. Verification by testing. Verification by testing EDA222/DIT161 Real-Tme Systems, Chalmers/GU, 2014/2015 Lecture #8 Real-Tme Systems Real-Tme Systems Lecture #8 Specfcaton Professor Jan Jonsson Implementaton System models Executon-tme analyss Department

More information

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance

Tsinghua University at TAC 2009: Summarizing Multi-documents by Information Distance Tsnghua Unversty at TAC 2009: Summarzng Mult-documents by Informaton Dstance Chong Long, Mnle Huang, Xaoyan Zhu State Key Laboratory of Intellgent Technology and Systems, Tsnghua Natonal Laboratory for

More information

Petri Net Based Software Dependability Engineering

Petri Net Based Software Dependability Engineering Proc. RELECTRONIC 95, Budapest, pp. 181-186; October 1995 Petr Net Based Software Dependablty Engneerng Monka Hener Brandenburg Unversty of Technology Cottbus Computer Scence Insttute Postbox 101344 D-03013

More information

Report on On-line Graph Coloring

Report on On-line Graph Coloring 2003 Fall Semester Comp 670K Onlne Algorthm Report on LO Yuet Me (00086365) cndylo@ust.hk Abstract Onlne algorthm deals wth data that has no future nformaton. Lots of examples demonstrate that onlne algorthm

More information

TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z.

TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS. Muradaliyev A.Z. TECHNIQUE OF FORMATION HOMOGENEOUS SAMPLE SAME OBJECTS Muradalyev AZ Azerbajan Scentfc-Research and Desgn-Prospectng Insttute of Energetc AZ1012, Ave HZardab-94 E-mal:aydn_murad@yahoocom Importance of

More information

An Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices

An Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices Internatonal Mathematcal Forum, Vol 7, 2012, no 52, 2549-2554 An Applcaton of the Dulmage-Mendelsohn Decomposton to Sparse Null Space Bases of Full Row Rank Matrces Mostafa Khorramzadeh Department of Mathematcal

More information

Virtual Machine Migration based on Trust Measurement of Computer Node

Virtual Machine Migration based on Trust Measurement of Computer Node Appled Mechancs and Materals Onlne: 2014-04-04 ISSN: 1662-7482, Vols. 536-537, pp 678-682 do:10.4028/www.scentfc.net/amm.536-537.678 2014 Trans Tech Publcatons, Swtzerland Vrtual Machne Mgraton based on

More information

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching

A Fast Visual Tracking Algorithm Based on Circle Pixels Matching A Fast Vsual Trackng Algorthm Based on Crcle Pxels Matchng Zhqang Hou hou_zhq@sohu.com Chongzhao Han czhan@mal.xjtu.edu.cn Ln Zheng Abstract: A fast vsual trackng algorthm based on crcle pxels matchng

More information

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory

Virtual Memory. Background. No. 10. Virtual Memory: concept. Logical Memory Space (review) Demand Paging(1) Virtual Memory Background EECS. Operatng System Fundamentals No. Vrtual Memory Prof. Hu Jang Department of Electrcal Engneerng and Computer Scence, York Unversty Memory-management methods normally requres the entre process

More information

Control strategies for network efficiency and resilience with route choice

Control strategies for network efficiency and resilience with route choice Control strateges for networ effcency and reslence wth route choce Andy Chow Ru Sha Centre for Transport Studes Unversty College London, UK Centralsed strateges UK 1 Centralsed strateges Some effectve

More information

Programming in Fortran 90 : 2017/2018

Programming in Fortran 90 : 2017/2018 Programmng n Fortran 90 : 2017/2018 Programmng n Fortran 90 : 2017/2018 Exercse 1 : Evaluaton of functon dependng on nput Wrte a program who evaluate the functon f (x,y) for any two user specfed values

More information

LECTURE NOTES Duality Theory, Sensitivity Analysis, and Parametric Programming

LECTURE NOTES Duality Theory, Sensitivity Analysis, and Parametric Programming CEE 60 Davd Rosenberg p. LECTURE NOTES Dualty Theory, Senstvty Analyss, and Parametrc Programmng Learnng Objectves. Revew the prmal LP model formulaton 2. Formulate the Dual Problem of an LP problem (TUES)

More information

Performance Evaluation of Information Retrieval Systems

Performance Evaluation of Information Retrieval Systems Why System Evaluaton? Performance Evaluaton of Informaton Retreval Systems Many sldes n ths secton are adapted from Prof. Joydeep Ghosh (UT ECE) who n turn adapted them from Prof. Dk Lee (Unv. of Scence

More information

Chapter 1. Introduction

Chapter 1. Introduction Chapter 1 Introducton 1.1 Parallel Processng There s a contnual demand for greater computatonal speed from a computer system than s currently possble (.e. sequental systems). Areas need great computatonal

More information

Concurrent Apriori Data Mining Algorithms

Concurrent Apriori Data Mining Algorithms Concurrent Apror Data Mnng Algorthms Vassl Halatchev Department of Electrcal Engneerng and Computer Scence York Unversty, Toronto October 8, 2015 Outlne Why t s mportant Introducton to Assocaton Rule Mnng

More information

Test-Cost Modeling and Optimal Test-Flow Selection of 3D-Stacked ICs

Test-Cost Modeling and Optimal Test-Flow Selection of 3D-Stacked ICs Test-Cost Modelng and Optmal Test-Flow Selecton of 3D-Stacked ICs Mukesh Agrawal, Student Member, IEEE, and Krshnendu Chakrabarty, Fellow, IEEE Abstract Three-dmensonal (3D) ntegraton s an attractve technology

More information

y and the total sum of

y and the total sum of Lnear regresson Testng for non-lnearty In analytcal chemstry, lnear regresson s commonly used n the constructon of calbraton functons requred for analytcal technques such as gas chromatography, atomc absorpton

More information

Design and Analysis of Algorithms

Design and Analysis of Algorithms Desgn and Analyss of Algorthms Heaps and Heapsort Reference: CLRS Chapter 6 Topcs: Heaps Heapsort Prorty queue Huo Hongwe Recap and overvew The story so far... Inserton sort runnng tme of Θ(n 2 ); sorts

More information

Biostatistics 615/815

Biostatistics 615/815 The E-M Algorthm Bostatstcs 615/815 Lecture 17 Last Lecture: The Smplex Method General method for optmzaton Makes few assumptons about functon Crawls towards mnmum Some recommendatons Multple startng ponts

More information

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,

More information

Analysis of Collaborative Distributed Admission Control in x Networks

Analysis of Collaborative Distributed Admission Control in x Networks 1 Analyss of Collaboratve Dstrbuted Admsson Control n 82.11x Networks Thnh Nguyen, Member, IEEE, Ken Nguyen, Member, IEEE, Lnha He, Member, IEEE, Abstract Wth the recent surge of wreless home networks,

More information

Interconnect Optimization for High-Level Synthesis of SSA Form Programs

Interconnect Optimization for High-Level Synthesis of SSA Form Programs Interconnect Optmzaton for Hgh-Level Synthess of SSA Form Programs Phlp Brsk Aay K. Verma Paolo Ienne Processor Archtecture Laboratory Swss Federal Insttute of Technology (EPFL) Lausanne, Swtzerland {phlp.brsk,

More information

11. APPROXIMATION ALGORITHMS

11. APPROXIMATION ALGORITHMS Copng wth NP-completeness 11. APPROXIMATION ALGORITHMS load balancng center selecton prcng method: vertex cover LP roundng: vertex cover generalzed load balancng knapsack problem Q. Suppose I need to solve

More information

Explicit Formulas and Efficient Algorithm for Moment Computation of Coupled RC Trees with Lumped and Distributed Elements

Explicit Formulas and Efficient Algorithm for Moment Computation of Coupled RC Trees with Lumped and Distributed Elements Explct Formulas and Effcent Algorthm for Moment Computaton of Coupled RC Trees wth Lumped and Dstrbuted Elements Qngan Yu and Ernest S.Kuh Electroncs Research Lab. Unv. of Calforna at Berkeley Berkeley

More information

Fast Computation of Shortest Path for Visiting Segments in the Plane

Fast Computation of Shortest Path for Visiting Segments in the Plane Send Orders for Reprnts to reprnts@benthamscence.ae 4 The Open Cybernetcs & Systemcs Journal, 04, 8, 4-9 Open Access Fast Computaton of Shortest Path for Vstng Segments n the Plane Ljuan Wang,, Bo Jang

More information

Adaptive Resource Allocation Control with On-Line Search for Fair QoS Level

Adaptive Resource Allocation Control with On-Line Search for Fair QoS Level Adaptve Resource Allocaton Control wth On-Lne Search for Far QoS Level Fumko Harada, Toshmtsu Usho, Graduate School of Engneerng Scence Osaka Unversty {harada@hopf, usho@}sysesosaka-uacjp Yukkazu akamoto

More information

Constructing Minimum Connected Dominating Set: Algorithmic approach

Constructing Minimum Connected Dominating Set: Algorithmic approach Constructng Mnmum Connected Domnatng Set: Algorthmc approach G.N. Puroht and Usha Sharma Centre for Mathematcal Scences, Banasthal Unversty, Rajasthan 304022 usha.sharma94@yahoo.com Abstract: Connected

More information

Classifier Selection Based on Data Complexity Measures *

Classifier Selection Based on Data Complexity Measures * Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.

More information

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009.

Assignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009. Farrukh Jabeen Algorthms 51 Assgnment #2 Due Date: June 15, 29. Assgnment # 2 Chapter 3 Dscrete Fourer Transforms Implement the FFT for the DFT. Descrbed n sectons 3.1 and 3.2. Delverables: 1. Concse descrpton

More information

Reducing Frame Rate for Object Tracking

Reducing Frame Rate for Object Tracking Reducng Frame Rate for Object Trackng Pavel Korshunov 1 and We Tsang Oo 2 1 Natonal Unversty of Sngapore, Sngapore 11977, pavelkor@comp.nus.edu.sg 2 Natonal Unversty of Sngapore, Sngapore 11977, oowt@comp.nus.edu.sg

More information

Wavefront Reconstructor

Wavefront Reconstructor A Dstrbuted Smplex B-Splne Based Wavefront Reconstructor Coen de Vsser and Mchel Verhaegen 14-12-201212 2012 Delft Unversty of Technology Contents Introducton Wavefront reconstructon usng Smplex B-Splnes

More information

CMPS 10 Introduction to Computer Science Lecture Notes

CMPS 10 Introduction to Computer Science Lecture Notes CPS 0 Introducton to Computer Scence Lecture Notes Chapter : Algorthm Desgn How should we present algorthms? Natural languages lke Englsh, Spansh, or French whch are rch n nterpretaton and meanng are not

More information

A SYSTOLIC APPROACH TO LOOP PARTITIONING AND MAPPING INTO FIXED SIZE DISTRIBUTED MEMORY ARCHITECTURES

A SYSTOLIC APPROACH TO LOOP PARTITIONING AND MAPPING INTO FIXED SIZE DISTRIBUTED MEMORY ARCHITECTURES A SYSOLIC APPROACH O LOOP PARIIONING AND MAPPING INO FIXED SIZE DISRIBUED MEMORY ARCHIECURES Ioanns Drosts, Nektaros Kozrs, George Papakonstantnou and Panayots sanakas Natonal echncal Unversty of Athens

More information

Quality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation

Quality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation Intellgent Informaton Management, 013, 5, 191-195 Publshed Onlne November 013 (http://www.scrp.org/journal/m) http://dx.do.org/10.36/m.013.5601 Qualty Improvement Algorthm for Tetrahedral Mesh Based on

More information

Efficient Distributed File System (EDFS)

Efficient Distributed File System (EDFS) Effcent Dstrbuted Fle System (EDFS) (Sem-Centralzed) Debessay(Debsh) Fesehaye, Rahul Malk & Klara Naherstedt Unversty of Illnos-Urbana Champagn Contents Problem Statement, Related Work, EDFS Desgn Rate

More information

An efficient iterative source routing algorithm

An efficient iterative source routing algorithm An effcent teratve source routng algorthm Gang Cheng Ye Tan Nrwan Ansar Advanced Networng Lab Department of Electrcal Computer Engneerng New Jersey Insttute of Technology Newar NJ 7 {gc yt Ansar}@ntedu

More information