REIHE INFORMATIK 7/95. Extremal Codes for Speed-Up of Distributed Parallel Arbitration

Size: px
Start display at page:

Download "REIHE INFORMATIK 7/95. Extremal Codes for Speed-Up of Distributed Parallel Arbitration"

Transcription

1 REIHE INFORMATIK 7/95 Extremal Codes for Speed-Up of Dstrbuted Parallel Arbtrato M. Makhaok, V. Cheravsky, R. Mäer, K.-H. Noffz Uverstät Mahem Semargebäude A5 D Mahem 1

2 EXTREMAL CODES FOR SPEED-UP OF DISTRIBUTED PARALLEL ARBITRATION M. Makhaok 1), V. Cheravsky 1), R. Mäer 2,3), K.-H. Noffz 2) 1) Isttute of Egeerg Cyberetcs, Academy of Sceces, Surgaov St. 6, Msk, Republc Belarus 2) Lehrstuhl für Iformatk V, Uverstät Mahem, A5, D Mahem, Germay 3) Iterdszpläres Zetrum für Wsseschaftlches Reche, Uverstät Hedelberg, Im Neuehemer Feld 368, D Hedelberg, Germay Idex terms: Buses, arbtrato prortes, extremal codes, dstrbuted parallel arbtrato, Futurebus+ Abstract. Ths paper descrbes a method that allows the speed up of parallel processes dstrbuted arbtrato schemes as used Futurebus+. It s based o specal arbtrato codes that decrease the maxmal arbtrato tme to a specfed value. Such codes ca be appled wth few, f ay, mor chages of the hardware. The geeral structure of these codes s gve. Itroducto. Oe mportat compoet of every multprocessor bus s the arbtrato system. May moder multprocessor buses use arbtrato systems prcpally based o the same dstrbuted parallel arbtrato scheme [1]. I ths scheme, the tme for acqurg bus mastershp depeds o the maxmal arbtrato tme requred. Ths tme s determed by the set of arbtrato prortes avalable [2-4]. Such a set wll be called a arbtrato code or smply a code below. The maxmal arbtrato tme s reduced, e.g., by usg bomal codes as dscussed [4] where these codes were troduced. From the dscusso t follows that the achevable speed-up depeds o the umber of arbtrato prortes the bomal code,.e., o ts capacty. However, t has ot bee cosdered whether the bomal codes have the maxmal possble capacty,.e., whether there are other codes wth equal or greater capacty that ca reduce the maxmal arbtrato tme to the same or a lower lmt. I the preset paper we derve the geeral structure of the codes of maxmal capacty that we wll call extremal codes. It wll be show that there exst dfferet extremal codes wth bomal codes as oe example oly. Arbtrato process. We cosder a dstrbuted parallel arbtrato process for a sglebus multprocessor system. To each processor a uque arbtrato prorty s assged that s a m-bt bary word. It wll be called a arbtrato word or smply a word below. If oe or more processors request bus mastershp, a arbtrato process s started. At Ths work has bee supported by the Deutsche Forschugsgemeschaft (DFG) uder grats Ma 1150/8-1 ad 436-WER ( /117/0). 2

3 ths tme each processor decdes whether t partcpates the arbtrato process or ot. The objectve of the arbtrato process s to fd the processor of hghest prorty amog all partcpatg oes. Arbtrato s doe a decetralzed way. Each processor has ts ow local arbtrato crcut whch s coected to m commo arbtrato les. O these les the logcal OR fucto s formed of the correspodg outputs of the local arbtrato crcuts. I Fg. 1 ths crcut s show wth outputs postve logc correspodg to expressos (1). I a real mplemetato egatve logc would be used,.e., the outputs would be verted so that the OR fucto could be realzed as a wred-or o opecollector les. It s mportat to dstgush betwee the state of a arbtrato le ad the state of the correspodg output of a arbtrato crcut. Such a output ca be chaged by a sgle processor by assertos or wthdrawals of bts of ts arbtrato word. Ths may or may ot fluece the state of the arbtrato le. Usually the arbtrato crcuts are realzed as combatoral logc ad the OR fuctos are computed as a wred-or usg ope-collector drvers. The output sgals deped o the word assged to the processor ad o the sgals curretly formed o the arbtrato les. Ay chage of the output sgals of a processor - assertos as well as wthdrawals - wll be called a swtch of the processor. If a processor s ot partcpatg the process t does ot assert ts output sgals,.e., keeps them the "zero" state. Whe the arbtrato process starts, all partcpatg processors assert all bts of ther words oto the arbtrato les. Smultaeously they compare ther ow bt values wth the curret state of the correspodg les. If a le s set to a logc "oe" ad the correspodg bt of the processor s word s zero, the processor swtches. Ths meas t determes the most sgfcat bt that s ot equal to the state of the correspodg arbtrato le ad wthdraws ths ad all ts less sgfcat bts from the les. If the codto s o loger true the processor swtches back to the prevous state. Let all processors have the same speed. The the arbtrato process for the j-th partcpatg processor ca be descrbed by the followg expressos 1. b 1,j (t+τ)=s(t)a 1,j, b 2,j (t+τ)=s(t)a 2,j (a 1,j Β 1 (t)), b 3,j (t+τ)=s(t)a 3,j (a 1,j Β 1 (t))(a 2,j Β 2 (t)), 1 If the local arbtrato crcuts of the processors takg part the arbtrato process have dfferet delays τ for swtchg ther sgals o the arbtarto les t s ecessary to take to accout results from [6] to mmze the maxmal arbtrato tme. 3

4 ... (1) b m,j (t+τ)=s(t)a m,j r=1,...,m-1 (a r,j Β r (t)), B (t)=v j P b,j (t), =1,...,m, where j P, P s the set of processors partcpatg the process, t s the curret tme (- <t< ), a,j s bt of the word whch s assged to processor j, (a,j {0,1}, a 1,j s the most sgfcat bt), b,j (t) s the output sgal asserted by processor j oto arbtrato le, B (t) s the sgal formed o arbtrato le, τ s the delay troduced by ay processor j, (0<τ< ), to assert ts sgals oto the arbtrato les or to wthdraw them from there, s(t) s the start sgal (s(t)=0 f t<0, ad s(t)=1 otherwse), m s the umber of bts the arbtrato words,.e., ther wdth. The frst m equaltes (1) express the dvdual output sgals of all processors. The last oe descrbes the curret state of the ope collector arbtrato les whch form the logc OR fucto of the sgals asserted. Durato of the arbtrato process. The arbtrato process s a sequece of swtches. From expressos (1) t follows that every processor ca execute at most m swtches oe arbtrato process. The followg example shows that m swtches ca deed occur. Let m=4, let three processors partcpate the process, ad let them have the words, 01, ad The processors eed a frst swtch to assert these umbers oto the arbtrato les. After the frst swtch, the les are the state The secod swtch s eeded to wthdraw the thrd bt of word, the forth bt of word 01, ad the last three bts of word The arbtrato les wll therefore be the state 00 after the secod swtch. Now the frst ad the secod processors assert ther last three bts aga. Thus, after the thrd swtch, the arbtrato les wll be the state 11. Fally, the wthdrawal of the last bt by the secod processor represets the fourth swtch. The word wll appear o the arbtrato les ad wll be stable. By comparg ths word wth ts ow arbtrato word each processor decdes ow f t has the maxmal prorty. The durato of the arbtrato process s proportoal to the umber of swtches executed. Sce we cosder a stuato such as Futurebus+, ether the delays of other modules or the modules partcpatg the process or ther arbtrato words are kow. Because the durato of the process caot be determed dyamcally as a 4

5 fucto of the partcpats ad ther delays, every module has always to assume that the maxmal umber of swtches wll occur ad that they wll be executed by the slowest modules,.e., the worst case. We, therefore, ca assume wthout loss of geeralty that the delays are detcal. Thus to get a correct result of the process for ay combato of m-bt arbtrato words, the partcpatg processors have to wat the tme m τ before they make ther decso. However, the maxmal umber of swtches ca be cosderably smaller tha m as show below. Deote by K the umber of words code K,.e., ts capacty. 1. If a code of wdth m has a capacty K <2 m, t s possble to reduce the maxmal umber of swtches that ca occur f ay combato of words from K wll partcpate the process. Cosder, e.g., a code of wdth 4 ad of capacty 15. If the words ad 01 are cluded the code K ad both partcpate the process, the worst case t wll requre four swtches as show above. However, f we exclude from K ether (the oly word possbly requrg four swtches) or 01 (the oly word creatg ths kd of coflct wth ), the the arbtrato process wll requre at most three swtches for all combatos of the 15 remag 4-bt arbtrato words. I ths case a speed-up of 25% s obtaed. 2. If a code of wdth m has the capacty K =2 m, the we caot leave out ay word because all of them are use. However, such a possblty exsts, f we use a addtoal arbtrato le,.e., a (m+1)-bt code stead of a m-bt code. Now we ca tur to the frst case ad - by approprate selecto of the words - reduce the maxmal umber of swtches by a factor of approxmately 2 as show [4]. Therefore, the followg problem arses. If the wdth of a code ad ts capacty are gve, how to select ts words from the set of all possble m-bt bary words so that the maxmal umber of swtches s mmzed. Extremal codes. To compare the umber of swtches dfferet arbtrato codes we troduce the followg otatos. a*b s the word that s the cocateato of words a ad b, a () s the word that s the cocateato of words a, {a*k} s the code formed by the left cocateato of word a wth every word from code K, {a* }={ *a}={a}, X (m) s the code cosstg of all possble words of wdth m. For example, 0 (3) = 000, X (2) ={11,,01,00}; ad f a=, b= 11, K={111,1} the a*b= 11, a (2) =, a*k={111,1}. Defto 1. We defe the depth d(k) of a arbtrato code K as the maxmal umber of swtches that ca occur f ay subset Z of K wll take part a arbtrato process. 5

6 Defto 2. We defe a code of wdth m ad depth as the extremal code Ex m f there s o code of the same wdth ad depth, ad greater capacty. Usg a code of smaller depth results a faster arbtrato process. Accordg to expressos (1) the depth ca vary from 1 to m. Therefore, our task s to fd for every depth the code K of maxmal capacty. Let us frst focus o the specal classes of extremal codes of depth =0,1,m. The structures of these codes are defed the followg Lemmas. Lemma 1. The code Ex 0 m has the capacty 1 ad cossts of the sgle zero word Ex 0 m={0 (m) }. (2) The process starts at t=0. For t<0 the arbtrato les are set to zero. Thus, the arbtrato process of the sgle word 0 (m) does ot requre ay swtches. Ay other word geerates at least oe swtch durg whch ts o-zero bt values are asserted oto the arbtrato les. Lemma 2. The code Ex 1 m has the capacty (m+1) ad s costructed by ay bt permutato appled to all words of,...,m 0 (m-) *1 (). (3) Proof. Let a 1 = a 1,1...a m,1, a 2 = a 1,2...a m,2, wth a 1 >a 2, be two words of a code Ex 1 m. If a,1 =0 ( [1,m]) the a,2 =0 also has to be true. Otherwse, the arbtrato process of {a 1,a 2 } could requre two swtches, the asserto of a 1 ad a 2 ad at least the wthdrawal of the bt a,2 =1. Ths s mpossble for a code of a depth equal to 1. Therefore, w a1 >w a2 f w a1 ad w a2 are the umber of o-zero bts the words a 1 ad a 2. Let us arrage all words of the code creasg order ad prove that ther umber s s equal to m+1. If w j s the umber of o-zero bts the word a j, the the expresso 0 w a1 <w a2 <...<w as m s vald. Obvously, the umber of o-zero bts the word a j s (j-1) ad the maxmal possble value of s s m+1. These codtos are oly satsfed for the codes Ex 1 m defed Lemma 2. For them the sgals B (t) are settled after the frst asserto of the maxmal word,.e., after the frst swtch. Fg. 2 shows examples of codes Ex 1 4 wth the correspodg permutato of the words defed by (3). Lemma 3. The code Ex m m has the capacty 2 m ad cossts of all possble m-bt words Ex m m=x (m). (4) 6

7 The code X (m) has the maxmal possble capacty accordg to ts defto ad ca cause a sequece of m swtches but ot more, as t was show [3,4]. Theorem. The capacty of ay extremal code Ex m s gve by the equato Ex m =Σ,..., C m, m 1, m 0 (5) where C m are the bomal coeffcets. Proof. Equalty (5) s obvously true for the cases =0,1,m that have bee cosdered the prevous Lemmas. We wll prove (5) for m>>1 by a double ducto wth parameters ad m. The hypothess s that (5) s true for the codes Ex m-1 ad Ex -1, Ex,..., Ex m-2. From ths we wll derve that equato (5) holds true for the code Ex m. Ths ducto starts from the cases =0,1,m cofrmed by the Lemmas, ad proves the Theorem up to ay gve m ad the progresso Ex 2 2,Ex 0 m Ex 2 3 Ex Ex 2 m; Ex 3 3,Ex 1 m Ex 3 4 Ex Ex 3 m;... Ex,Ex m Ex +1 Ex Ex m. Bass. Assume that the followg equatos are true. Ex m-1 =Σ,..., C m-1, (6) Ex k-1 =Σ,..., C k-1, k=,(+1),...,(m-1). (7) Iducto. We ted to prove (5) for the code Ex m for m>>1 assumg the truth of the bass. Frst we costruct a certa code R m ad cosder ts depth ad capacty. The code R m= j=-1,...,m E j (8) s the uo of the followg subcodes (Fg. 3a) E -1 ={1 (m-+1) *X (-1) }, E k ={1 (m-k) *0*Ex k-1}, k=,(+1),...,(m-1), E m ={0*Ex m-1}. 7

8 Obvously E p E q = for ay par of subcodes (p,q [-1,m]). Thus, the capacty of the code R m s the sum of the capactes of all subcodes R m = E -1 + E + E E m-1 + E m = = X (-1) + Ex -1 + Ex + + Ex m-2 + Ex m-1. Takg to accout the expressos (6) ad (7), assumed to be true as the bass, we ca derve (see Appedx) R m =2-1 +Σ,..., C m-1+σ k=,...,m-1 Σ,..., C k-1=σ,..., C m. (9) Below t wll be show that the depth of R m s. All words of E -1, E k, ad E m cosst of two parts, a detcal leadg part, ad a tralg part whch s dfferet for every word (Fg. 3a). Thus f oly words of oe of the subcodes E -1, E k, or E m partcpate the arbtrato, swtches happe oly the secod part. Therefore, the process takes o more tha -1,, or swtches correspodgly. Below we show that f words from dfferet subcodes partcpate the process the t caot have more tha swtches. Cosder the case whe words from E m ad E m-1 are partcpatg a arbtrato process smultaeously. The words from E m-1 are headed by the bts 1, ad the words of E m are headed by 0. Thus, after the frst swtch the logcal OR of the words of E m ad E m-1 appears o the bus. After the secod swtch all words of E m wll be wthdraw ad oly the words of E m-1 are left. I the worst case addtoal swtches are requred to determe the maxmum word amog them. The same argumet apples to ay other combato of subcodes from R m whch proves that the depth of R m s. The code R m has depth, wdth m, ad the capacty defed by (5). Let us show that there exsts o other code of the same depth ad wdth whch has a capacty greater tha R m,.e., R m s the extremal code. Assume the cotrary,.e., that a code S m of m-bt words wth d(s m)= ad S m > R m exsts. I order to costruct the code S m a form smlar to R m let us troduce the followg masks L -1 ={1 (m-+1) *X (-1) }, L k ={1 (m-k) *0*X (k-1) }, k=,(+1),...,(m-1), L m ={0*X (m-1) }. The uo of these masks s the code cosstg of all m-bt bary words X (m). So the code S m s a uo of subcodes (Fg. 3b): 8

9 where S m= j=-1,...,m T j T -1 =L -1 S m={1 (m-+1) *T -1 } T k =L k S m={1 (m-k) *0*T k-1 }, k=,(+1),...,(m-1) T m =L m S m={0*t m-1 }. As before T p T q = ; p,q [-1,m]; for ay par of subcodes ad therefore S m = T -1 + T + T T m-1 + T m. To estmate S m, cosder the capactes of the dvdual subcodes. Sce subcode T -1 cossts of (-1)-bt words ts capacty does ot exceed 2-1, so T -1 = T = X(-1). The depth of the subcode T m-1 caot be greater tha. Otherwse the depth of T m ad therefore the depth of S m would exceed. Sce T m-1 cossts of (m-1)-bt words ts capacty has the upper lmt Ex m-1. For T m we therefore have T m = T m-1 Ex m-1. Sce the capacty of T -1 caot exceed that of E -1, ad sce the capacty of T m caot exceed that of E m, the equalty T k > E k has to hold true at least for oe k [,m-1] to meet the assumpto S m > R m. We defe s as the mmal value of the dex k for whch the equalty T k > E k holds. Sce T s = T s-1 > E s ad E s = Ex s-1, the depth d(t s )=d(t s-1 ) s greater tha (). Let us form the code D s+1 =T s+1 T s+2... T m. Cosder ths code the bts of colum (m-s+1) (see Fg. 3b). If D s+1 there exsts a word a j wth a o-zero bt (ms+1), the the arbtrato process for the code Z=T s {a j } ca requre more tha swtches. They are: the asserto of the words of Z (the frst swtch), the wthdrawal of the o-zero bt (m-s+1) of the word a j (the secod swtch), ad the settlemet of the maxmal word of the code T s-1 (a sequece of more tha () swtches). Ths s mpossble for the code S m of depth. So, all words of D s+1 as well as the words of T s must also have zeros colum (m-s+1). Let us form the code D* s by elmato of ths colum (m-s+1) from the code D s =T s D s+1. Obvously the depth d(d* s ) s stll equal to d(d s ). Furthermore, the depth d(d s ) as well as the depth of ay other subcode of S m caot be greater tha. The we have D* s < Ex m-1 for the capacty of the code D* s of (m-1)-bt words. Therefore, D s = D* s < Ex m-1. Now we have upper bouds for the subcodes of S m. Takg to accout the bass (6) ad (7), we ca fally estmate S m as S m = T -1 + T + T T s-1 + D s X (-1) + Ex -1 + Ex + Ex s-2 + Ex m-1 = =2-1 +Σ k=,...,s-1 Σ,..., C k-1+σ,..., C m-1 9

10 After comparg ths expresso wth (9), we drectly obta S m <Σ =1,...,m C m= R m () sce s<m. So the assumpto S m > R m leads to the cotradcto () whch proves that a code wth capacty greater tha R m caot exst. Let us prove ow that ay code S m of depth whch s dfferet from R m has a smaller capacty. Assume cotrast that S m = R m ad S m R m. If T k > Ex k-1 or d(t k)>() holds true for some k [,m-1] the the prevous argumet brgs us back to () whch cotradcts the assumpto. O the other had, T k < E k-1 for all k=,...,m-1 drectly yelds S m < R m. The oly alteratve left s T k = E k ad d(t k ) () for all k=,...,m-1. Ths s possble oly f T k-1 =Ex k-1 ad T k =E k. I the same way we have T m =E m, T -1 =E -1, ad therefore S m=r m. So, o code of depth ad wdth m ca have a hgher capacty tha determed by (5). Q.E.D. We proved that the capacty of the extremal codes s determed by expresso (5) for all m ad. Some values are gve Tab. 1. I [4] t was show that the capacty of the bomal codes of wdth m ad depth s also determed by the same expresso. Therefore, the bomal codes are extremal. Costructo of the Extremal Codes. Durg the proof of (5) t has bee show that ay code R m costructed accordace wth (8) has the depth ad the maxmum capacty. It also has bee prove that ay code dfferet from R m has a lower capacty. I vew of ths, expresso (8) represets the geeral structure of the extremal codes for ay ( 0,1,m). We ca, therefore, rewrte (8) as Ex m={0*ex m-1} {1 (m-+1) *X (-1) } =1,...,m- {1 () *0*Ex m--1}, 0,1,m. (11) Usg expressos (2)-(4) ad (11) the extremal codes of ay depth ad wdth ca be costructed. For eve ( 0,m) expresso (11) allows to costruct the code Ex m from the codes of depth 0, whch are determed uquely by Lemma 1. So for eve oly oe extremal code exsts. Sce the bomal code wth the same parameters m ad has the capacty gve by (5) t s ths extremal code. If s odd ( 1,m) the the codes Ex m are costructed from the codes of depth 1, whch are ot uque (Lemma 2). So t s possble to costruct dfferet extremal codes of the same odd. Here the bomal codes are oly oe example amog them. I Fg. 4 some examples of codes Ex 5 are show. The code Ex 3 5, e.g., s formed accordace wth (11) as

11 {0*Ex 3 4} {1 (3) *X (2) } {1 (2) *0*Ex 1 2} {1*0*Ex 1 3} where Ex 3 4={0*Ex 3 3} {11*X (2) } {*Ex 1 2} s also gve by (11). It s easy to determe the umber γ m of dfferet codes Ex m. From Lemmas 1-3 we have γ 0 m=γ m m=1 ad γ 1 m=m!. For other follows drectly from (11). γ m=γ m-1 γ m-2 γ m-3... γ -1 for m>>1 It allows to calculate γ m recursvely. Some values of γ m are gve Tab. 2. Applcato. Havg derved the expressos for the capactes of extremal codes we ca calculate the achevable speed-up. Cosder a system that uses a code K of wdth m, K <2 m. If the arbtrato words of ths code are chose wthout takg to accout ts depth, the the chaces are hgh to get a depth equal to m. Sce the system uses ot all possble bary words of wdth m, t s possble to replace the orgal code K by a extremal code order to reduce the maxmal umber of swtches. To acheve the maxmal speed-up oe has to fd a code Ex m of capacty Ex m K ad mmal depth. Now f K s replaced by Ex m, the maxmal umber of swtches s reduced from m to. Ths results a speed-up of the arbtrato process by factor of m/. Obvously the maxmal speed-up s s obtaed f the depth of the code K s mmal ad the umber of processors p s maxmal. For arbtrato the mmal s equal to 1. Here the maxmal umber of processors s the wdth of the arbtrato words plus 1. Thus, the hghest speed-up s obtaed wth a relatvely small umber of processors. Nowadays ths s the most relevat case. Cosder, e.g., a Futurebus+ system wth e processors. Sce Futurebus+ uses a prorty word of wdth m=8, t s possble to acheve a speed-up s=8 by applyg the extremal code of depth =1. For a hgher umber of processors the achevable speed-up decreases because codes Ex m of creasg depth are requred to assg arbtrato words to all processors. Tab. 3 shows the achevable speed-up for a system where all arbtrato words are used. I ths case the speed-up s obtaed by provdg addtoal arbtrato les. Ths allows to apply extremal codes of hgher wdth ad smaller depth. Note that the arbtrato tme ca be reduced by a factor of 2 by addg a sgle le oly [4]. Hgher speedups ca oly be obtaed usg addtoal arbtrato les. Note that after the maxmal word s fally asserted oto the arbtrato les swtches the local arbtrato crcuts of processors ca stll occur. Cosder, e.g., a arbtrato betwee the two words { 1111, 11 }. The arbtrato process takes oly oe swtch to assert the word After ths momet the processor havg the word 11 wth- 11

12 draws ts least sgfcat o-zero bt. Depedg o the techology used ths wthdrawal mght or mght ot fluece the arbtrato le whch s already set to a logc "oe". If, e.g., ope collector les are used to mplemet wred OR logc, the wthdrawal of a sgal from a le the logc "oe" state may result a "zero" gltch [5]. It s possble to avod ths gltch by usg the code =1,...,m {1 () *0 (m-) } as the Ex 1 m code. I ths case the code Ex m s uquely defed for ay m ad. For eve t cocdes wth the bomal code. Coclusos. I ths paper we derved the structure of codes whch reduce the maxmal arbtrato tme to a defed lmt. We obtaed a recursve formula that allows oe to costruct codes of hghest capacty for all possble umbers of swtches. No hgher speed-up ca be acheved by usg ay other code. It was show that, some cases, a cosderable umber of extremal codes exst. Bomal codes are oly oe represetatve of these codes. The recursve formula (11) that gves us the geeral structure of extremal codes s qute complcated. However, oe has to use t oly oce whe the arbtrato system s set up. For most arbtrato systems the codes ca be appled drectly wthout chages of the hardware. APPENDIX. Here we preset the detaled dervato of (9). The propertes m 1) C m = 2 m, m 0, m 2) = +1 C = Cm+1, m 0, ) C m-1 + Cm-1 = C m, m-2 0 are used below. For ay m ad (m>>1) we ca derve m j=-1 =2-1 + C j + m-2 j= C m-1 C j - =2-1 + j= m-2 j=-1 C j + C m-1 C j + C m-1 =2-1 + =2-1 + m-2 j= m-2 ( C j - j=0 j= j C j- j= C j + C j )+ 1, 2) C m-1 = C m-1 = 12

13 = C m j=0 2 j + =o C m-1 = ) + C m-1+cm-1+1= C m-1 C m-1 = (Cm Cm-1 )+ 3) ) = C m +Cm-1 +Cm = C m +Cm+1= C m C m - C m 0 +1=. Refereces. [1] Borrll P. "A Comparso of 32-bt Buses", IEEE Mcro, Dec. 1985, pp [2] Taub D.M. "Corrected Settlg Tme of the Dstrbuted Parallel Arbter", IEE Proc. E, Vol. 139, No. 4, 1992, pp [3] Taub D.M. "Improved Cotrol Acqusto Scheme for the IEEE 896 Futurebus", IEEE Mcro, Vol. 7, No. 3, 1987, pp [4] Makhaok M., Cheravsky V., Mäer R., Stucky O.: "Bomal Codg Dstrbuted Arbtrato Systems", Electr. Lett., Vol. 25, No. 20, 1989, pp [5] Gustavso D.B., Theus J.: Wre-OR Logc o Trasmsso Les; IEEE Mcro, Vol. 3, No. 3, 1983, pp [6] Makhaok M., Mäer R.: "Sychrozato of Parallel Processes Dstrbuted Systems; Bougè L., Cosard M., Robert Y., Trystram D. (Eds.): Proc. CONPAR 92 - VAPP V, Lyos, Frace; Lect. Notes Comp. Sc. 634, Sprger, Berl (1992) pp

14 Captos. Fg. 1. Local arbtrato crcut. Fg. 2. Examples of extremal codes Ex 1 4. Fg. 3. The structure of the sets R m ad S m. Fg. 4. Examples of extremal codes Ex 5. Tab. 1. The capacty Ex m of extremal codes of depth ad wdth m. Tab. 2. The umber γ m of extremal codes of depth ad wdth m. Tab. 3. The speed-up s for dfferet word wdths m ad umber of processors p=2 m acheved by addg dm arbtrato les. 14

15 d m Table 1. m Table 2. m p dm Table 3. 15

16 s(t) b 1,j a 1,j B 1 b 2,j a 2,j B 2 b 3,j a 3,j B 3 b 4,j a 4,j B 4 wo/lost Fg. 1 16

17 ( )( 1243 )( 1324 ) ( 4321 ) Fg. 2 E -1 E E +1 E m-1 E m X (-1) Ex Ex Ex m Ex.. m-1 0 a) R m Fg. 3 T -1 T T s D s (m-s+1)-th colum b) S m T T -1 T s-1 D s 17

18 X (3) Ex Ex X (2) 1 Ex 2 1 Ex 3 X (2) 1 Ex 2 3 Ex 3 Ex Ex 5 5 Ex 5 4 Ex 5 3 Ex 5 2 Ex 5 1 Fg. 4 18

CHAPTER 4 RADICAL EXPRESSIONS

CHAPTER 4 RADICAL EXPRESSIONS 6 CHAPTER RADICAL EXPRESSIONS. The th Root of a Real Number A real umber a s called the th root of a real umber b f Thus, for example: s a square root of sce. s also a square root of sce ( ). s a cube

More information

Functions of Random Variables

Functions of Random Variables Fuctos of Radom Varables Chapter Fve Fuctos of Radom Varables 5. Itroducto A geeral egeerg aalyss model s show Fg. 5.. The model output (respose) cotas the performaces of a system or product, such as weght,

More information

PTAS for Bin-Packing

PTAS for Bin-Packing CS 663: Patter Matchg Algorthms Scrbe: Che Jag /9/00. Itroducto PTAS for B-Packg The B-Packg problem s NP-hard. If we use approxmato algorthms, the B-Packg problem could be solved polyomal tme. For example,

More information

Chapter 9 Jordan Block Matrices

Chapter 9 Jordan Block Matrices Chapter 9 Jorda Block atrces I ths chapter we wll solve the followg problem. Gve a lear operator T fd a bass R of F such that the matrx R (T) s as smple as possble. f course smple s a matter of taste.

More information

Bounds on the expected entropy and KL-divergence of sampled multinomial distributions. Brandon C. Roy

Bounds on the expected entropy and KL-divergence of sampled multinomial distributions. Brandon C. Roy Bouds o the expected etropy ad KL-dvergece of sampled multomal dstrbutos Brado C. Roy bcroy@meda.mt.edu Orgal: May 18, 2011 Revsed: Jue 6, 2011 Abstract Iformato theoretc quattes calculated from a sampled

More information

18.413: Error Correcting Codes Lab March 2, Lecture 8

18.413: Error Correcting Codes Lab March 2, Lecture 8 18.413: Error Correctg Codes Lab March 2, 2004 Lecturer: Dael A. Spelma Lecture 8 8.1 Vector Spaces A set C {0, 1} s a vector space f for x all C ad y C, x + y C, where we take addto to be compoet wse

More information

MATH 247/Winter Notes on the adjoint and on normal operators.

MATH 247/Winter Notes on the adjoint and on normal operators. MATH 47/Wter 00 Notes o the adjot ad o ormal operators I these otes, V s a fte dmesoal er product space over, wth gve er * product uv, T, S, T, are lear operators o V U, W are subspaces of V Whe we say

More information

Ideal multigrades with trigonometric coefficients

Ideal multigrades with trigonometric coefficients Ideal multgrades wth trgoometrc coeffcets Zarathustra Brady December 13, 010 1 The problem A (, k) multgrade s defed as a par of dstct sets of tegers such that (a 1,..., a ; b 1,..., b ) a j = =1 for all

More information

Third handout: On the Gini Index

Third handout: On the Gini Index Thrd hadout: O the dex Corrado, a tala statstca, proposed (, 9, 96) to measure absolute equalt va the mea dfferece whch s defed as ( / ) where refers to the total umber of dvduals socet. Assume that. The

More information

{ }{ ( )} (, ) = ( ) ( ) ( ) Chapter 14 Exercises in Sampling Theory. Exercise 1 (Simple random sampling): Solution:

{ }{ ( )} (, ) = ( ) ( ) ( ) Chapter 14 Exercises in Sampling Theory. Exercise 1 (Simple random sampling): Solution: Chapter 4 Exercses Samplg Theory Exercse (Smple radom samplg: Let there be two correlated radom varables X ad A sample of sze s draw from a populato by smple radom samplg wthout replacemet The observed

More information

2. Independence and Bernoulli Trials

2. Independence and Bernoulli Trials . Ideedece ad Beroull Trals Ideedece: Evets ad B are deedet f B B. - It s easy to show that, B deedet mles, B;, B are all deedet ars. For examle, ad so that B or B B B B B φ,.e., ad B are deedet evets.,

More information

Part 4b Asymptotic Results for MRR2 using PRESS. Recall that the PRESS statistic is a special type of cross validation procedure (see Allen (1971))

Part 4b Asymptotic Results for MRR2 using PRESS. Recall that the PRESS statistic is a special type of cross validation procedure (see Allen (1971)) art 4b Asymptotc Results for MRR usg RESS Recall that the RESS statstc s a specal type of cross valdato procedure (see Alle (97)) partcular to the regresso problem ad volves fdg Y $,, the estmate at the

More information

1 Onto functions and bijections Applications to Counting

1 Onto functions and bijections Applications to Counting 1 Oto fuctos ad bectos Applcatos to Coutg Now we move o to a ew topc. Defto 1.1 (Surecto. A fucto f : A B s sad to be surectve or oto f for each b B there s some a A so that f(a B. What are examples of

More information

A tighter lower bound on the circuit size of the hardest Boolean functions

A tighter lower bound on the circuit size of the hardest Boolean functions Electroc Colloquum o Computatoal Complexty, Report No. 86 2011) A tghter lower boud o the crcut sze of the hardest Boolea fuctos Masak Yamamoto Abstract I [IPL2005], Fradse ad Mlterse mproved bouds o the

More information

1 Mixed Quantum State. 2 Density Matrix. CS Density Matrices, von Neumann Entropy 3/7/07 Spring 2007 Lecture 13. ψ = α x x. ρ = p i ψ i ψ i.

1 Mixed Quantum State. 2 Density Matrix. CS Density Matrices, von Neumann Entropy 3/7/07 Spring 2007 Lecture 13. ψ = α x x. ρ = p i ψ i ψ i. CS 94- Desty Matrces, vo Neuma Etropy 3/7/07 Sprg 007 Lecture 3 I ths lecture, we wll dscuss the bascs of quatum formato theory I partcular, we wll dscuss mxed quatum states, desty matrces, vo Neuma etropy

More information

Arithmetic Mean and Geometric Mean

Arithmetic Mean and Geometric Mean Acta Mathematca Ntresa Vol, No, p 43 48 ISSN 453-6083 Arthmetc Mea ad Geometrc Mea Mare Varga a * Peter Mchalča b a Departmet of Mathematcs, Faculty of Natural Sceces, Costate the Phlosopher Uversty Ntra,

More information

Discrete Mathematics and Probability Theory Fall 2016 Seshia and Walrand DIS 10b

Discrete Mathematics and Probability Theory Fall 2016 Seshia and Walrand DIS 10b CS 70 Dscrete Mathematcs ad Probablty Theory Fall 206 Sesha ad Walrad DIS 0b. Wll I Get My Package? Seaky delvery guy of some compay s out delverg packages to customers. Not oly does he had a radom package

More information

(b) By independence, the probability that the string 1011 is received correctly is

(b) By independence, the probability that the string 1011 is received correctly is Soluto to Problem 1.31. (a) Let A be the evet that a 0 s trasmtted. Usg the total probablty theorem, the desred probablty s P(A)(1 ɛ ( 0)+ 1 P(A) ) (1 ɛ 1)=p(1 ɛ 0)+(1 p)(1 ɛ 1). (b) By depedece, the probablty

More information

Mu Sequences/Series Solutions National Convention 2014

Mu Sequences/Series Solutions National Convention 2014 Mu Sequeces/Seres Solutos Natoal Coveto 04 C 6 E A 6C A 6 B B 7 A D 7 D C 7 A B 8 A B 8 A C 8 E 4 B 9 B 4 E 9 B 4 C 9 E C 0 A A 0 D B 0 C C Usg basc propertes of arthmetc sequeces, we fd a ad bm m We eed

More information

ABOUT ONE APPROACH TO APPROXIMATION OF CONTINUOUS FUNCTION BY THREE-LAYERED NEURAL NETWORK

ABOUT ONE APPROACH TO APPROXIMATION OF CONTINUOUS FUNCTION BY THREE-LAYERED NEURAL NETWORK ABOUT ONE APPROACH TO APPROXIMATION OF CONTINUOUS FUNCTION BY THREE-LAYERED NEURAL NETWORK Ram Rzayev Cyberetc Isttute of the Natoal Scece Academy of Azerbaa Republc ramrza@yahoo.com Aygu Alasgarova Khazar

More information

CS286.2 Lecture 4: Dinur s Proof of the PCP Theorem

CS286.2 Lecture 4: Dinur s Proof of the PCP Theorem CS86. Lecture 4: Dur s Proof of the PCP Theorem Scrbe: Thom Bohdaowcz Prevously, we have prove a weak verso of the PCP theorem: NP PCP 1,1/ (r = poly, q = O(1)). Wth ths result we have the desred costat

More information

Chapter 8. Inferences about More Than Two Population Central Values

Chapter 8. Inferences about More Than Two Population Central Values Chapter 8. Ifereces about More Tha Two Populato Cetral Values Case tudy: Effect of Tmg of the Treatmet of Port-We tas wth Lasers ) To vestgate whether treatmet at a youg age would yeld better results tha

More information

Solving Constrained Flow-Shop Scheduling. Problems with Three Machines

Solving Constrained Flow-Shop Scheduling. Problems with Three Machines It J Cotemp Math Sceces, Vol 5, 2010, o 19, 921-929 Solvg Costraed Flow-Shop Schedulg Problems wth Three Maches P Pada ad P Rajedra Departmet of Mathematcs, School of Advaced Sceces, VIT Uversty, Vellore-632

More information

Lecture 07: Poles and Zeros

Lecture 07: Poles and Zeros Lecture 07: Poles ad Zeros Defto of poles ad zeros The trasfer fucto provdes a bass for determg mportat system respose characterstcs wthout solvg the complete dfferetal equato. As defed, the trasfer fucto

More information

MA 524 Homework 6 Solutions

MA 524 Homework 6 Solutions MA 524 Homework 6 Solutos. Sce S(, s the umber of ways to partto [] to k oempty blocks, ad c(, s the umber of ways to partto to k oempty blocks ad also the arrage each block to a cycle, we must have S(,

More information

Non-uniform Turán-type problems

Non-uniform Turán-type problems Joural of Combatoral Theory, Seres A 111 2005 106 110 wwwelsevercomlocatecta No-uform Turá-type problems DhruvMubay 1, Y Zhao 2 Departmet of Mathematcs, Statstcs, ad Computer Scece, Uversty of Illos at

More information

Chapter 5 Properties of a Random Sample

Chapter 5 Properties of a Random Sample Lecture 6 o BST 63: Statstcal Theory I Ku Zhag, /0/008 Revew for the prevous lecture Cocepts: t-dstrbuto, F-dstrbuto Theorems: Dstrbutos of sample mea ad sample varace, relatoshp betwee sample mea ad sample

More information

The Mathematical Appendix

The Mathematical Appendix The Mathematcal Appedx Defto A: If ( Λ, Ω, where ( λ λ λ whch the probablty dstrbutos,,..., Defto A. uppose that ( Λ,,..., s a expermet type, the σ-algebra o λ λ λ are defed s deoted by ( (,,...,, σ Ω.

More information

For combinatorial problems we might need to generate all permutations, combinations, or subsets of a set.

For combinatorial problems we might need to generate all permutations, combinations, or subsets of a set. Addtoal Decrease ad Coquer Algorthms For combatoral problems we mght eed to geerate all permutatos, combatos, or subsets of a set. Geeratg Permutatos If we have a set f elemets: { a 1, a 2, a 3, a } the

More information

Estimation of Stress- Strength Reliability model using finite mixture of exponential distributions

Estimation of Stress- Strength Reliability model using finite mixture of exponential distributions Iteratoal Joural of Computatoal Egeerg Research Vol, 0 Issue, Estmato of Stress- Stregth Relablty model usg fte mxture of expoetal dstrbutos K.Sadhya, T.S.Umamaheswar Departmet of Mathematcs, Lal Bhadur

More information

Derivation of 3-Point Block Method Formula for Solving First Order Stiff Ordinary Differential Equations

Derivation of 3-Point Block Method Formula for Solving First Order Stiff Ordinary Differential Equations Dervato of -Pot Block Method Formula for Solvg Frst Order Stff Ordary Dfferetal Equatos Kharul Hamd Kharul Auar, Kharl Iskadar Othma, Zara Bb Ibrahm Abstract Dervato of pot block method formula wth costat

More information

Assignment 5/MATH 247/Winter Due: Friday, February 19 in class (!) (answers will be posted right after class)

Assignment 5/MATH 247/Winter Due: Friday, February 19 in class (!) (answers will be posted right after class) Assgmet 5/MATH 7/Wter 00 Due: Frday, February 9 class (!) (aswers wll be posted rght after class) As usual, there are peces of text, before the questos [], [], themselves. Recall: For the quadratc form

More information

5 Short Proofs of Simplified Stirling s Approximation

5 Short Proofs of Simplified Stirling s Approximation 5 Short Proofs of Smplfed Strlg s Approxmato Ofr Gorodetsky, drtymaths.wordpress.com Jue, 20 0 Itroducto Strlg s approxmato s the followg (somewhat surprsg) approxmato of the factoral,, usg elemetary fuctos:

More information

Lecture 9: Tolerant Testing

Lecture 9: Tolerant Testing Lecture 9: Tolerat Testg Dael Kae Scrbe: Sakeerth Rao Aprl 4, 07 Abstract I ths lecture we prove a quas lear lower boud o the umber of samples eeded to do tolerat testg for L dstace. Tolerat Testg We have

More information

X X X E[ ] E X E X. is the ()m n where the ( i,)th. j element is the mean of the ( i,)th., then

X X X E[ ] E X E X. is the ()m n where the ( i,)th. j element is the mean of the ( i,)th., then Secto 5 Vectors of Radom Varables Whe workg wth several radom varables,,..., to arrage them vector form x, t s ofte coveet We ca the make use of matrx algebra to help us orgaze ad mapulate large umbers

More information

Introduction to local (nonparametric) density estimation. methods

Introduction to local (nonparametric) density estimation. methods Itroducto to local (oparametrc) desty estmato methods A slecture by Yu Lu for ECE 66 Sprg 014 1. Itroducto Ths slecture troduces two local desty estmato methods whch are Parze desty estmato ad k-earest

More information

arxiv:math/ v1 [math.gm] 8 Dec 2005

arxiv:math/ v1 [math.gm] 8 Dec 2005 arxv:math/05272v [math.gm] 8 Dec 2005 A GENERALIZATION OF AN INEQUALITY FROM IMO 2005 NIKOLAI NIKOLOV The preset paper was spred by the thrd problem from the IMO 2005. A specal award was gve to Yure Boreko

More information

CHAPTER VI Statistical Analysis of Experimental Data

CHAPTER VI Statistical Analysis of Experimental Data Chapter VI Statstcal Aalyss of Expermetal Data CHAPTER VI Statstcal Aalyss of Expermetal Data Measuremets do ot lead to a uque value. Ths s a result of the multtude of errors (maly radom errors) that ca

More information

Lecture 3. Sampling, sampling distributions, and parameter estimation

Lecture 3. Sampling, sampling distributions, and parameter estimation Lecture 3 Samplg, samplg dstrbutos, ad parameter estmato Samplg Defto Populato s defed as the collecto of all the possble observatos of terest. The collecto of observatos we take from the populato s called

More information

Department of Agricultural Economics. PhD Qualifier Examination. August 2011

Department of Agricultural Economics. PhD Qualifier Examination. August 2011 Departmet of Agrcultural Ecoomcs PhD Qualfer Examato August 0 Istructos: The exam cossts of sx questos You must aswer all questos If you eed a assumpto to complete a questo, state the assumpto clearly

More information

Lecture 2 - What are component and system reliability and how it can be improved?

Lecture 2 - What are component and system reliability and how it can be improved? Lecture 2 - What are compoet ad system relablty ad how t ca be mproved? Relablty s a measure of the qualty of the product over the log ru. The cocept of relablty s a exteded tme perod over whch the expected

More information

Rademacher Complexity. Examples

Rademacher Complexity. Examples Algorthmc Foudatos of Learg Lecture 3 Rademacher Complexty. Examples Lecturer: Patrck Rebesch Verso: October 16th 018 3.1 Itroducto I the last lecture we troduced the oto of Rademacher complexty ad showed

More information

UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS

UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS Exam: ECON430 Statstcs Date of exam: Frday, December 8, 07 Grades are gve: Jauary 4, 08 Tme for exam: 0900 am 00 oo The problem set covers 5 pages Resources allowed:

More information

Chapter 3 Sampling For Proportions and Percentages

Chapter 3 Sampling For Proportions and Percentages Chapter 3 Samplg For Proportos ad Percetages I may stuatos, the characterstc uder study o whch the observatos are collected are qualtatve ature For example, the resposes of customers may marketg surveys

More information

2006 Jamie Trahan, Autar Kaw, Kevin Martin University of South Florida United States of America

2006 Jamie Trahan, Autar Kaw, Kevin Martin University of South Florida United States of America SOLUTION OF SYSTEMS OF SIMULTANEOUS LINEAR EQUATIONS Gauss-Sedel Method 006 Jame Traha, Autar Kaw, Kev Mart Uversty of South Florda Uted States of Amerca kaw@eg.usf.edu Itroducto Ths worksheet demostrates

More information

A New Measure of Probabilistic Entropy. and its Properties

A New Measure of Probabilistic Entropy. and its Properties Appled Mathematcal Sceces, Vol. 4, 200, o. 28, 387-394 A New Measure of Probablstc Etropy ad ts Propertes Rajeesh Kumar Departmet of Mathematcs Kurukshetra Uversty Kurukshetra, Ida rajeesh_kuk@redffmal.com

More information

ECONOMETRIC THEORY. MODULE VIII Lecture - 26 Heteroskedasticity

ECONOMETRIC THEORY. MODULE VIII Lecture - 26 Heteroskedasticity ECONOMETRIC THEORY MODULE VIII Lecture - 6 Heteroskedastcty Dr. Shalabh Departmet of Mathematcs ad Statstcs Ida Isttute of Techology Kapur . Breusch Paga test Ths test ca be appled whe the replcated data

More information

Simple Linear Regression

Simple Linear Regression Statstcal Methods I (EST 75) Page 139 Smple Lear Regresso Smple regresso applcatos are used to ft a model descrbg a lear relatoshp betwee two varables. The aspects of least squares regresso ad correlato

More information

The internal structure of natural numbers, one method for the definition of large prime numbers, and a factorization test

The internal structure of natural numbers, one method for the definition of large prime numbers, and a factorization test Fal verso The teral structure of atural umbers oe method for the defto of large prme umbers ad a factorzato test Emmaul Maousos APM Isttute for the Advacemet of Physcs ad Mathematcs 3 Poulou str. 53 Athes

More information

Strong Convergence of Weighted Averaged Approximants of Asymptotically Nonexpansive Mappings in Banach Spaces without Uniform Convexity

Strong Convergence of Weighted Averaged Approximants of Asymptotically Nonexpansive Mappings in Banach Spaces without Uniform Convexity BULLETIN of the MALAYSIAN MATHEMATICAL SCIENCES SOCIETY Bull. Malays. Math. Sc. Soc. () 7 (004), 5 35 Strog Covergece of Weghted Averaged Appromats of Asymptotcally Noepasve Mappgs Baach Spaces wthout

More information

Q-analogue of a Linear Transformation Preserving Log-concavity

Q-analogue of a Linear Transformation Preserving Log-concavity Iteratoal Joural of Algebra, Vol. 1, 2007, o. 2, 87-94 Q-aalogue of a Lear Trasformato Preservg Log-cocavty Daozhog Luo Departmet of Mathematcs, Huaqao Uversty Quazhou, Fua 362021, P. R. Cha ldzblue@163.com

More information

UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS

UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS Postpoed exam: ECON430 Statstcs Date of exam: Jauary 0, 0 Tme for exam: 09:00 a.m. :00 oo The problem set covers 5 pages Resources allowed: All wrtte ad prted

More information

best estimate (mean) for X uncertainty or error in the measurement (systematic, random or statistical) best

best estimate (mean) for X uncertainty or error in the measurement (systematic, random or statistical) best Error Aalyss Preamble Wheever a measuremet s made, the result followg from that measuremet s always subject to ucertaty The ucertaty ca be reduced by makg several measuremets of the same quatty or by mprovg

More information

Feature Selection: Part 2. 1 Greedy Algorithms (continued from the last lecture)

Feature Selection: Part 2. 1 Greedy Algorithms (continued from the last lecture) CSE 546: Mache Learg Lecture 6 Feature Selecto: Part 2 Istructor: Sham Kakade Greedy Algorthms (cotued from the last lecture) There are varety of greedy algorthms ad umerous amg covetos for these algorthms.

More information

9 U-STATISTICS. Eh =(m!) 1 Eh(X (1),..., X (m ) ) i.i.d

9 U-STATISTICS. Eh =(m!) 1 Eh(X (1),..., X (m ) ) i.i.d 9 U-STATISTICS Suppose,,..., are P P..d. wth CDF F. Our goal s to estmate the expectato t (P)=Eh(,,..., m ). Note that ths expectato requres more tha oe cotrast to E, E, or Eh( ). Oe example s E or P((,

More information

22 Nonparametric Methods.

22 Nonparametric Methods. 22 oparametrc Methods. I parametrc models oe assumes apror that the dstrbutos have a specfc form wth oe or more ukow parameters ad oe tres to fd the best or atleast reasoably effcet procedures that aswer

More information

A BASIS OF THE GROUP OF PRIMITIVE ALMOST PYTHAGOREAN TRIPLES

A BASIS OF THE GROUP OF PRIMITIVE ALMOST PYTHAGOREAN TRIPLES Joural of Algebra Number Theory: Advaces ad Applcatos Volume 6 Number 6 Pages 5-7 Avalable at http://scetfcadvaces.co. DOI: http://dx.do.org/.864/ataa_77 A BASIS OF THE GROUP OF PRIMITIVE ALMOST PYTHAGOREAN

More information

F. Inequalities. HKAL Pure Mathematics. 進佳數學團隊 Dr. Herbert Lam 林康榮博士. [Solution] Example Basic properties

F. Inequalities. HKAL Pure Mathematics. 進佳數學團隊 Dr. Herbert Lam 林康榮博士. [Solution] Example Basic properties 進佳數學團隊 Dr. Herbert Lam 林康榮博士 HKAL Pure Mathematcs F. Ieualtes. Basc propertes Theorem Let a, b, c be real umbers. () If a b ad b c, the a c. () If a b ad c 0, the ac bc, but f a b ad c 0, the ac bc. Theorem

More information

Lecture 3 Probability review (cont d)

Lecture 3 Probability review (cont d) STATS 00: Itroducto to Statstcal Iferece Autum 06 Lecture 3 Probablty revew (cot d) 3. Jot dstrbutos If radom varables X,..., X k are depedet, the ther dstrbuto may be specfed by specfyg the dvdual dstrbuto

More information

Investigation of Partially Conditional RP Model with Response Error. Ed Stanek

Investigation of Partially Conditional RP Model with Response Error. Ed Stanek Partally Codtoal Radom Permutato Model 7- vestgato of Partally Codtoal RP Model wth Respose Error TRODUCTO Ed Staek We explore the predctor that wll result a smple radom sample wth respose error whe a

More information

Lecture 7. Confidence Intervals and Hypothesis Tests in the Simple CLR Model

Lecture 7. Confidence Intervals and Hypothesis Tests in the Simple CLR Model Lecture 7. Cofdece Itervals ad Hypothess Tests the Smple CLR Model I lecture 6 we troduced the Classcal Lear Regresso (CLR) model that s the radom expermet of whch the data Y,,, K, are the outcomes. The

More information

C-1: Aerodynamics of Airfoils 1 C-2: Aerodynamics of Airfoils 2 C-3: Panel Methods C-4: Thin Airfoil Theory

C-1: Aerodynamics of Airfoils 1 C-2: Aerodynamics of Airfoils 2 C-3: Panel Methods C-4: Thin Airfoil Theory ROAD MAP... AE301 Aerodyamcs I UNIT C: 2-D Arfols C-1: Aerodyamcs of Arfols 1 C-2: Aerodyamcs of Arfols 2 C-3: Pael Methods C-4: Th Arfol Theory AE301 Aerodyamcs I Ut C-3: Lst of Subects Problem Solutos?

More information

Newton s Power Flow algorithm

Newton s Power Flow algorithm Power Egeerg - Egll Beedt Hresso ewto s Power Flow algorthm Power Egeerg - Egll Beedt Hresso The ewto s Method of Power Flow 2 Calculatos. For the referece bus #, we set : V = p.u. ad δ = 0 For all other

More information

Multiple Regression. More than 2 variables! Grade on Final. Multiple Regression 11/21/2012. Exam 2 Grades. Exam 2 Re-grades

Multiple Regression. More than 2 variables! Grade on Final. Multiple Regression 11/21/2012. Exam 2 Grades. Exam 2 Re-grades STAT 101 Dr. Kar Lock Morga 11/20/12 Exam 2 Grades Multple Regresso SECTIONS 9.2, 10.1, 10.2 Multple explaatory varables (10.1) Parttog varablty R 2, ANOVA (9.2) Codtos resdual plot (10.2) Trasformatos

More information

Analysis of Lagrange Interpolation Formula

Analysis of Lagrange Interpolation Formula P IJISET - Iteratoal Joural of Iovatve Scece, Egeerg & Techology, Vol. Issue, December 4. www.jset.com ISS 348 7968 Aalyss of Lagrage Iterpolato Formula Vjay Dahya PDepartmet of MathematcsMaharaja Surajmal

More information

Generalization of the Dissimilarity Measure of Fuzzy Sets

Generalization of the Dissimilarity Measure of Fuzzy Sets Iteratoal Mathematcal Forum 2 2007 o. 68 3395-3400 Geeralzato of the Dssmlarty Measure of Fuzzy Sets Faramarz Faghh Boformatcs Laboratory Naobotechology Research Ceter vesa Research Isttute CECR Tehra

More information

Assignment 7/MATH 247/Winter, 2010 Due: Friday, March 19. Powers of a square matrix

Assignment 7/MATH 247/Winter, 2010 Due: Friday, March 19. Powers of a square matrix Assgmet 7/MATH 47/Wter, 00 Due: Frday, March 9 Powers o a square matrx Gve a square matrx A, ts powers A or large, or eve arbtrary, teger expoets ca be calculated by dagoalzg A -- that s possble (!) Namely,

More information

TESTS BASED ON MAXIMUM LIKELIHOOD

TESTS BASED ON MAXIMUM LIKELIHOOD ESE 5 Toy E. Smth. The Basc Example. TESTS BASED ON MAXIMUM LIKELIHOOD To llustrate the propertes of maxmum lkelhood estmates ad tests, we cosder the smplest possble case of estmatg the mea of the ormal

More information

Bayes (Naïve or not) Classifiers: Generative Approach

Bayes (Naïve or not) Classifiers: Generative Approach Logstc regresso Bayes (Naïve or ot) Classfers: Geeratve Approach What do we mea by Geeratve approach: Lear p(y), p(x y) ad the apply bayes rule to compute p(y x) for makg predctos Ths s essetally makg

More information

ENGI 4421 Propagation of Error Page 8-01

ENGI 4421 Propagation of Error Page 8-01 ENGI 441 Propagato of Error Page 8-01 Propagato of Error [Navd Chapter 3; ot Devore] Ay realstc measuremet procedure cotas error. Ay calculatos based o that measuremet wll therefore also cota a error.

More information

Econometric Methods. Review of Estimation

Econometric Methods. Review of Estimation Ecoometrc Methods Revew of Estmato Estmatg the populato mea Radom samplg Pot ad terval estmators Lear estmators Ubased estmators Lear Ubased Estmators (LUEs) Effcecy (mmum varace) ad Best Lear Ubased Estmators

More information

NP!= P. By Liu Ran. Table of Contents. The P versus NP problem is a major unsolved problem in computer

NP!= P. By Liu Ran. Table of Contents. The P versus NP problem is a major unsolved problem in computer NP!= P By Lu Ra Table of Cotets. Itroduce 2. Prelmary theorem 3. Proof 4. Expla 5. Cocluso. Itroduce The P versus NP problem s a major usolved problem computer scece. Iformally, t asks whether a computer

More information

å 1 13 Practice Final Examination Solutions - = CS109 Dec 5, 2018

å 1 13 Practice Final Examination Solutions - = CS109 Dec 5, 2018 Chrs Pech Fal Practce CS09 Dec 5, 08 Practce Fal Examato Solutos. Aswer: 4/5 8/7. There are multle ways to obta ths aswer; here are two: The frst commo method s to sum over all ossbltes for the rak of

More information

Investigating Cellular Automata

Investigating Cellular Automata Researcher: Taylor Dupuy Advsor: Aaro Wootto Semester: Fall 4 Ivestgatg Cellular Automata A Overvew of Cellular Automata: Cellular Automata are smple computer programs that geerate rows of black ad whte

More information

THE ROYAL STATISTICAL SOCIETY 2016 EXAMINATIONS SOLUTIONS HIGHER CERTIFICATE MODULE 5

THE ROYAL STATISTICAL SOCIETY 2016 EXAMINATIONS SOLUTIONS HIGHER CERTIFICATE MODULE 5 THE ROYAL STATISTICAL SOCIETY 06 EAMINATIONS SOLUTIONS HIGHER CERTIFICATE MODULE 5 The Socety s provdg these solutos to assst cadtes preparg for the examatos 07. The solutos are teded as learg ads ad should

More information

Laboratory I.10 It All Adds Up

Laboratory I.10 It All Adds Up Laboratory I. It All Adds Up Goals The studet wll work wth Rema sums ad evaluate them usg Derve. The studet wll see applcatos of tegrals as accumulatos of chages. The studet wll revew curve fttg sklls.

More information

Summary of the lecture in Biostatistics

Summary of the lecture in Biostatistics Summary of the lecture Bostatstcs Probablty Desty Fucto For a cotuos radom varable, a probablty desty fucto s a fucto such that: 0 dx a b) b a dx A probablty desty fucto provdes a smple descrpto of the

More information

Objectives of Multiple Regression

Objectives of Multiple Regression Obectves of Multple Regresso Establsh the lear equato that best predcts values of a depedet varable Y usg more tha oe eplaator varable from a large set of potetal predctors {,,... k }. Fd that subset of

More information

AN UPPER BOUND FOR THE PERMANENT VERSUS DETERMINANT PROBLEM BRUNO GRENET

AN UPPER BOUND FOR THE PERMANENT VERSUS DETERMINANT PROBLEM BRUNO GRENET AN UPPER BOUND FOR THE PERMANENT VERSUS DETERMINANT PROBLEM BRUNO GRENET Abstract. The Permaet versus Determat problem s the followg: Gve a matrx X of determates over a feld of characterstc dfferet from

More information

Simulation Output Analysis

Simulation Output Analysis Smulato Output Aalyss Summary Examples Parameter Estmato Sample Mea ad Varace Pot ad Iterval Estmato ermatg ad o-ermatg Smulato Mea Square Errors Example: Sgle Server Queueg System x(t) S 4 S 4 S 3 S 5

More information

7.0 Equality Contraints: Lagrange Multipliers

7.0 Equality Contraints: Lagrange Multipliers Systes Optzato 7.0 Equalty Cotrats: Lagrage Multplers Cosder the zato of a o-lear fucto subject to equalty costrats: g f() R ( ) 0 ( ) (7.) where the g ( ) are possbly also olear fuctos, ad < otherwse

More information

ESS Line Fitting

ESS Line Fitting ESS 5 014 17. Le Fttg A very commo problem data aalyss s lookg for relatoshpetwee dfferet parameters ad fttg les or surfaces to data. The smplest example s fttg a straght le ad we wll dscuss that here

More information

SPECIAL CONSIDERATIONS FOR VOLUMETRIC Z-TEST FOR PROPORTIONS

SPECIAL CONSIDERATIONS FOR VOLUMETRIC Z-TEST FOR PROPORTIONS SPECIAL CONSIDERAIONS FOR VOLUMERIC Z-ES FOR PROPORIONS Oe s stctve reacto to the questo of whether two percetages are sgfcatly dfferet from each other s to treat them as f they were proportos whch the

More information

QR Factorization and Singular Value Decomposition COS 323

QR Factorization and Singular Value Decomposition COS 323 QR Factorzato ad Sgular Value Decomposto COS 33 Why Yet Aother Method? How do we solve least-squares wthout currg codto-squarg effect of ormal equatos (A T A A T b) whe A s sgular, fat, or otherwse poorly-specfed?

More information

MULTIDIMENSIONAL HETEROGENEOUS VARIABLE PREDICTION BASED ON EXPERTS STATEMENTS. Gennadiy Lbov, Maxim Gerasimov

MULTIDIMENSIONAL HETEROGENEOUS VARIABLE PREDICTION BASED ON EXPERTS STATEMENTS. Gennadiy Lbov, Maxim Gerasimov Iteratoal Boo Seres "Iformato Scece ad Computg" 97 MULTIIMNSIONAL HTROGNOUS VARIABL PRICTION BAS ON PRTS STATMNTS Geady Lbov Maxm Gerasmov Abstract: I the wors [ ] we proposed a approach of formg a cosesus

More information

Block-Based Compact Thermal Modeling of Semiconductor Integrated Circuits

Block-Based Compact Thermal Modeling of Semiconductor Integrated Circuits Block-Based Compact hermal Modelg of Semcoductor Itegrated Crcuts Master s hess Defese Caddate: Jg Ba Commttee Members: Dr. Mg-Cheg Cheg Dr. Daqg Hou Dr. Robert Schllg July 27, 2009 Outle Itroducto Backgroud

More information

Maps on Triangular Matrix Algebras

Maps on Triangular Matrix Algebras Maps o ragular Matrx lgebras HMED RMZI SOUROUR Departmet of Mathematcs ad Statstcs Uversty of Vctora Vctora, BC V8W 3P4 CND sourour@mathuvcca bstract We surveys results about somorphsms, Jorda somorphsms,

More information

MA/CSSE 473 Day 27. Dynamic programming

MA/CSSE 473 Day 27. Dynamic programming MA/CSSE 473 Day 7 Dyamc Programmg Bomal Coeffcets Warshall's algorthm (Optmal BSTs) Studet questos? Dyamc programmg Used for problems wth recursve solutos ad overlappg subproblems Typcally, we save (memoze)

More information

8.1 Hashing Algorithms

8.1 Hashing Algorithms CS787: Advaced Algorthms Scrbe: Mayak Maheshwar, Chrs Hrchs Lecturer: Shuch Chawla Topc: Hashg ad NP-Completeess Date: September 21 2007 Prevously we looked at applcatos of radomzed algorthms, ad bega

More information

THE ROYAL STATISTICAL SOCIETY HIGHER CERTIFICATE

THE ROYAL STATISTICAL SOCIETY HIGHER CERTIFICATE THE ROYAL STATISTICAL SOCIETY 00 EXAMINATIONS SOLUTIONS HIGHER CERTIFICATE PAPER I STATISTICAL THEORY The Socety provdes these solutos to assst caddates preparg for the examatos future years ad for the

More information

On the Primitive Classes of K * KHALED S. FELALI Department of Mathematical Sciences, Umm Al-Qura University, Makkah Al-Mukarramah, Saudi Arabia

On the Primitive Classes of K * KHALED S. FELALI Department of Mathematical Sciences, Umm Al-Qura University, Makkah Al-Mukarramah, Saudi Arabia JKAU: Sc., O vol. the Prmtve, pp. 55-62 Classes (49 of A.H. K (BU) / 999 A.D.) * 55 O the Prmtve Classes of K * (BU) KHALED S. FELALI Departmet of Mathematcal Sceces, Umm Al-Qura Uversty, Makkah Al-Mukarramah,

More information

d dt d d dt dt Also recall that by Taylor series, / 2 (enables use of sin instead of cos-see p.27 of A&F) dsin

d dt d d dt dt Also recall that by Taylor series, / 2 (enables use of sin instead of cos-see p.27 of A&F) dsin Learzato of the Swg Equato We wll cover sectos.5.-.6 ad begg of Secto 3.3 these otes. 1. Sgle mache-fte bus case Cosder a sgle mache coected to a fte bus, as show Fg. 1 below. E y1 V=1./_ Fg. 1 The admttace

More information

Bounds for the Connective Eccentric Index

Bounds for the Connective Eccentric Index It. J. Cotemp. Math. Sceces, Vol. 7, 0, o. 44, 6-66 Bouds for the Coectve Eccetrc Idex Nlaja De Departmet of Basc Scece, Humates ad Socal Scece (Mathematcs Calcutta Isttute of Egeerg ad Maagemet Kolkata,

More information

Module 7: Probability and Statistics

Module 7: Probability and Statistics Lecture 4: Goodess of ft tests. Itroducto Module 7: Probablty ad Statstcs I the prevous two lectures, the cocepts, steps ad applcatos of Hypotheses testg were dscussed. Hypotheses testg may be used to

More information

Lecture 1. (Part II) The number of ways of partitioning n distinct objects into k distinct groups containing n 1,

Lecture 1. (Part II) The number of ways of partitioning n distinct objects into k distinct groups containing n 1, Lecture (Part II) Materals Covered Ths Lecture: Chapter 2 (2.6 --- 2.0) The umber of ways of parttog dstct obects to dstct groups cotag, 2,, obects, respectvely, where each obect appears exactly oe group

More information

Lecture 02: Bounding tail distributions of a random variable

Lecture 02: Bounding tail distributions of a random variable CSCI-B609: A Theorst s Toolkt, Fall 206 Aug 25 Lecture 02: Boudg tal dstrbutos of a radom varable Lecturer: Yua Zhou Scrbe: Yua Xe & Yua Zhou Let us cosder the ubased co flps aga. I.e. let the outcome

More information

Physics 114 Exam 2 Fall Name:

Physics 114 Exam 2 Fall Name: Physcs 114 Exam Fall 015 Name: For gradg purposes (do ot wrte here): Questo 1. 1... 3. 3. Problem Aswer each of the followg questos. Pots for each questo are dcated red. Uless otherwse dcated, the amout

More information

Ordinary Least Squares Regression. Simple Regression. Algebra and Assumptions.

Ordinary Least Squares Regression. Simple Regression. Algebra and Assumptions. Ordary Least Squares egresso. Smple egresso. Algebra ad Assumptos. I ths part of the course we are gog to study a techque for aalysg the lear relatoshp betwee two varables Y ad X. We have pars of observatos

More information

Unit 9. The Tangent Bundle

Unit 9. The Tangent Bundle Ut 9. The Taget Budle ========================================================================================== ---------- The taget sace of a submafold of R, detfcato of taget vectors wth dervatos at

More information

NP!= P. By Liu Ran. Table of Contents. The P vs. NP problem is a major unsolved problem in computer

NP!= P. By Liu Ran. Table of Contents. The P vs. NP problem is a major unsolved problem in computer NP!= P By Lu Ra Table of Cotets. Itroduce 2. Strategy 3. Prelmary theorem 4. Proof 5. Expla 6. Cocluso. Itroduce The P vs. NP problem s a major usolved problem computer scece. Iformally, t asks whether

More information