Simple Compression Algorithm for Memoryless Laplacian Source Based on the Optimal Companding Technique

Size: px
Start display at page:

Download "Simple Compression Algorithm for Memoryless Laplacian Source Based on the Optimal Companding Technique"

Transcription

1 Smple Compresson Algorhm for Memoryless Laplacan Source Based on he Opmal Compandng Technque Zoran H Perć Marko D Pekovć Mlan R Dnčć ) Faculy of Elecronc Engneerng Aleksandra Medvedeva 4 perc@elfaknacyu ) Faculy of Scences and Mahemacs Všegradska dexerofns@gmalcom Absrac Ths paper has wo achevemens The frs am of hs paper s opmzaon of he lossy resson coder realzed as andng quanzer wh opmal resson law Ths opmzaon s acheved by opmzng mal amplude for ha opmal andng quanzer for Laplacan source Approxmae expresson n closed form for opmal mal amplude s found Alhough hs expresson s very smple and suable for praccal mplemenaon sasfy opmaly creron for Lloyd-Max quanzer (for R 6 bs/sample) In he second par of hs paper novel smple lossless resson mehod s presened Ths mehod s much smpler han Huffman mehod bu gves beer resuls Fnally a he end of he paper general resson mehod s obaned by onng opmal andng quanzer and lossless codng mehod Ths mehod s appled on he concree sll mage and good resuls are obaned Besdes sll mages hs mehod also could be used for resson speech and bomedcal sgnals Keywords: smple lossless resson algorhm andng quanzaon opmal mal amplude Inroducon Quanzers play an mporan role n every A/D converson sysem They are appled for he purpose of sorage and ransmsson of connual sgnals A vas of research was done n hs opc An effcen algorhm for desgn of he opmal quanzer for he source wh known dsrbuon was developed by Lloyd and Max (Max 960) Ths s an erave mehod whch gves he sequence of quanzers convergng o he opmal quanzer Lloyd-Max algorhm s effcen for he small number of quanzaon pons (usually for ) However for he large number of quanzaon pons s me consumng Also he realzaon lexy of he opmal quanzers wh many quanzaon levels s very hgh One soluon whch overcomes hese dffcules s he andng model (Jayan and oll 984; Gray 004) Quanzers based on he andng model known as andng quanzers have smple realzaon srucure and performances close o he opmal Also he desgn of such quanzers s more effcen han Lloyd-Max algorhm snce does no requre he erave mehod Ths dfference s very noable for some commonly used sources ncludng he Laplacan source We wll consder andng quanzer because of less lexy n desgn and mplemenaon Memoryless Laplacan source s commonly used model n many applcaons due o s smplcy and fac ha many parameers and characerscs can be found as he closed form relaons (ncludng he opmal resson funcon) Examples are speech sgnal mages vdeo sgnal bo-medcal mages (for example ued omography scanner) ec For ransmsson processng and sorng ha sgnals smple and fas resson algorhms are

2 desrable One soluon s gven n (Sarosolsky 007) for unform quanzer In hs paper we gve smpler soluon for non-unform quanzer In paper (Ramalho 00) lossless resson algorhm was gven ha provded only he addonal resson of he dgzed sgnal (PCM) bu whou provdng a qualy mprovemen Our algorhm gves qualy mprovemen as well as furher resson In paper (kolc and Perc 008) lossy resson and Lloyd-Max quanzer are appled In our paper smple algorhm for lossy and lossless resson s gven For lossy resson smpler quanzer s used whou deeroraon of performances Usng hs algorhm for speech sgnal resson mprovemen of 0 bs/sample s acheved n relaon o algorhm used n (kolc and Perc 008) In hs paper we wan o aclsh wo goals Our frs goal s o opmze performances of he andng quanzer for memoryless Laplacan source by opmzng mal amplude In he paper (a and euhoff 00) s gven one heursc soluon for he suppor regon of he opmal quanzer Also paper (a 004) deals wh he upper bound of hs suppor regon The heursc soluon for he mal sgnal amplude n he adng model s gven n (kolc and Perc 008) An nenson of hs paper s o generalze hese resuls by gvng several approxmae analycal expressons for he mal sgnal amplude and s numercal uaon As s shown one of he analycal expressons almos concdes wh he opmal value The second goal of hs paper s he consrucon of he lossless coder for Laplacan source wh smple realzaon srucure The well-known effcen algorhm for lossless codng of he nformaon sources wh known symbol probably s Huffman algorhm (Hankerson e al 004; Sayood 006) I requres very lex realzaon srucure and also s me consumng Hence we gve he smple codng algorhm for M -ary exended nformaon source and gve he opmzaon of he exenson order M Our algorhm has beer performances (smaller b-rae) and drascally smpler realzaon lexy han Huffman algorhm Our algorhm s also smpler han algorhm proposed n (Sarosolsky 007) because we have b n prefx and ( r ) bs or r bs n suffx ha s we have consan number of bs n prefx On he oher hand algorhm n (Sarosolsky 007) uses varable number of bs n prefx whch requres much lex coder and especally decoder Prefx defnes regon where symbol of M-ary source exenson falls ( M s a number of samples n frame and we can consder frame as M -ary source exenson) Suffx represens codeword for symbol of M -ary source exenson Fnally we have oned opmal andng quanzer and smple lossless codng n a generalzed resson mehod Ths generalzed mehod s appled on a sll mage and good resuls are obaned as shown n he paper Besde sll mages hs resson mehod can be appled on speech and bo-medcal sgnals Ths paper s organzed as follows Secon recalls some basc heory of quanzers and andng model In secon opmzaon of andng quanzer was done by fndng opmal value for he mal sgnal amplude We gave hree approxmae expressons n closed form for and also one numercal mehod for fndng whch s based on mnmzaon of approxmae expresson for dsoron In secon 4 we are he approxmaons wh numercally obaned mal sgnal amplude Secon 5 deals wh he consrucon and performance analyss of he smple lossless codng algorhm In Secon 6 we gave generalzed resson algorhm whch represens combnaon of lossy and lossless resson In hs secon applcaon of hs generalzed mehod on a sll mage s presened Secon 7 concludes he paper

3 Fxed rae scalar quanzers and andng echnque An -pon fxed rae scalar quanzer s characerzed by he se of real numbers called decson hresholds whch sasfy = < < L < < = + 0 and se y y called represenaon levels whch sasfy y α = ( ] for = Ses α α α form he paron of he se of real numbers R and are called quanzaon cells The quanzer s defned as many-o-one mappng Q : R R defned by Q( x) = y where x α In pracce npu sgnal value x s dscrezed (quanzed) o he value y Cells α α are called nner cells (or granular cells) whle α and α are ouer cells (or overload cells) In such way cells α α form granular whle cells α and α form an overload regon As fxed-rae and scalar are only ypes of quanzers consdered n he paper for brevy we wll ordnarly om hese adecves Suppose ha an npu sgnal s characerzed by connuous random varable X wh probably densy funcon (pdf) p( x ) In he res of he paper we wll suppose ha nformaon source s Laplacan source wh memoryless propery and zero mean value The pdf of hs source s gven by x p( x) = e σ σ The sources wh exponenal and Laplacan pdf are commonly encounered and he mehods for desgnng quanzers for hese sources are very smlar Whou loosng of generaly we can suppose ha σ = and las expresson becomes x p( x) = e Snce p( x ) s an even funcon an opmal quanzer Q mus sasfy = and y = y for arbrary = The qualy of he quanzer s measured by dsoron of resulng reproducon n arson o he orgnal Mosly used measure of dsoron s mean-squared error I s defned as ( ) ( ( )) ( ) ( ) = D Q = E X Q x = x y p x dx The -pon quanzer Q s opmal for he source X f here s no oher -pon quanzer Q such ha D( Q ) < D( Q) We also defne granular Dg ( Q ) and overload Dol ( Q ) dsoron by g ( ) = ( ) ( ) = D Q x y p x dx + ol ( ) = ( ) ( ) + ( ) ( ) g ol D D Q x y p x dx x y p x dx Obvously holds D( Q) = D ( Q) + D ( Q) Denoe by dsoron of an opmal -pon quanzer Consderable amoun of work has been focused on he desgn of opmal quanzers for resson sources n mage speech and oher applcaons As frs dscovered by Paner and De (Paner and De 95) for large holds D = c / Here c s he Paner-De consan + ( ) / c = p ( x) dx ()

4 The general mehod for desgn of an opmal -pon quanzer for he gven source X s Lloyd-Max algorhm (Max 960; Jayan and oll 984) Ths s erave process sars from (0) ( n) some nal quanzer Q and consrucs he sequence Q n = of quanzers whch converge o he opmal quanzer Q (values of hresholds ( n ) converge o he hresholds he opmal quanzer Q ) Due o he uaonal lexy of hs mehod s no suable for he desgn of opmal quanzers wh more han 8 levels Hence here are developed oher mehods for consrucon of nearly opmal quanzers for large number of quanzaon levels One of he such echnques whch s commonly used s andng echnque (Judell and Scharf 986) I forms he core of he ITU-T G7 PCM sandard recommended for codng speech sgnals Compandng echnque consss of he followng seps: Compress he npu sgnal x by applyng he ressor funcon c( x ) Apply he unform quanzer Q u on he ressed sgnal c( x ) Expand he quanzed verson of he ressed sgnal usng an nverse ressor funcon c ( x) The correspondng srucure of a nonunform quanzer conssng of a ressor a unform quanzer and expandor n cascade s called andor Opmal resson law s used n hs paper Compresson funcon for ha law n dependence of pdf s known bu opmal mal amplude for ha resson law has no deermned up o now In hs paper hs opmal mal amplude for opmal resson law wll be deermned There are several ways o defne he ressor funcon for opmal resson law Orgnally n (Judell and Scharf 986) and also n (a 004) s he funcon c : R ( ) defned by We wll use he smlar defnon as above x / p ( x) dx c0 ( x) = + / p ( x) dx x < x / p ( x) dx / p ( x) dx x > c( x; ) = + x [ ] Value s called mal sgnal amplude Snce an mage of he funcon c( x; ) s ( ) decson hresholds u and represenaon levels y u of he unform quanzer are defned by u = + ( = 0 ) yu = + = Hence decson hresholds and represenaon levels of equvalen non-unform quanzer Q( x) = c ( Q ( c( x))) can be deermned as he soluons of he followng equaons u c( ; ) = c( y ; ) = y u u By solvng he las equaons we oban he values and amplude of () y as he funcons of he mal 4

5 + ( )exp( ) = / log 0 = log / < + ( )exp( ) + ( + ) exp( ) y = / log y = log / < + + ( )exp( ) where log x represens naural logarhm of x Expressons for and y wll be used n furher consderaon and hence we wll gve hem explcly = log (5) + ( ) exp ( ) y = log (6) + ( ) exp ( ) For he ease of noaon we wll denoe he dsoron granular dsoron and overload dsoron by D D and D respecvely (we wll om he quanzer Q ) Usng () we oban ha g ol dsoron D s funcon of one parameer Same holds also for D g and fnd he opmal value (Perc e al 007) (4) D ol Our goal s o whch mnmzes dsoron D One approxmae soluon s gven n Esmaon of he opmal mal sgnal amplude Closed-form expresson for he exac dsoron D = D( ) as he funcon of he mal sgnal amplude D = D( ) can be obaned by combnng relaons () and (4) Obaned funconal dependence s very lcaed snce appears n bounds of negrals and also n funcons under negraons Hence he exac mnmzaon of D( ) s very lex (see secon 4) Our frs goal n hs paper s o fnd some approxmae closed-form expressons for opmal mal amplude Accordng o he (Perc and kolc 007) granular dsoron D g can be approxmaed by he value D % as follows g 9 / D g ( p ( x) dx) exp = = ( ) ( ) On he oher sde we can explcly ue overload dsoron by % (7) + (8) D = ( x y ) p( x) dx ol = e ( + + y y + y ) 5

6 Hence we wll consder an approxmae dsoron D% = D% g + Dol I s worh menonng ha bounds of he negral (7) are usually se o and respecvely and herefore mum sgnal amplude s equal o he suppor regon Maxmal error n hs approxmaon s obaned on he nerval ( ) Therefore n order o mprove he approxmaon we se he correspondng bound o and herefore lower bound of negral n (8) s also Dsorson 8x0-4 7x0-4 7x0-4 7x0-4 Exac value Approxmaon Relave error of dsorson 0% 08% 06% 04% 0% Maxmal sgnal amplude 00% Maxmal sgnal amplude Fgure a Graphs of D and D % (lef) and relave error (rgh) as he funcon of for = 8 Dsorson x0 - x0 - x0 - x0 - x0 - x0 - Exac value Approxmaon Relave error of dsorson 5% 4% % % % 0x Maxmal sgnal amplude 0% Maxmal sgnal amplude Fgure b Graphs of D and D % (lef) and relave error (rgh) as he funcon of for = 64 Dsorson 46x0-45x0-44x0-4x0-4x0-4x0 - Exac value Approxmaon Maxmal sgnal amplude Relave error of dsorson 5% 4% % % % 0% Maxmal sgnal amplude Fgure c Graphs of D and D % (lef) and relave error (rgh) as he funcon of for = Value D % s he funcon of and y and snce hese values are boh funcons of (expressons (5) and (6)) also D % can be expressed as he funcon of oe ha approxmaon of D by D % s very accurae especally for he large values of 6

7 On Fgure (a b c) we show graphs of D and D % for = 8 = 64 and = as he funcons of and graph of relave error D D% D We gve hree esmae expressons for he opmal mal sgnal amplude The frs esmaon s based on he fac ha he value of he las represenaon level deermned by (6) s no opmal n general For fxed an opmal value of he followng expresson y op + xp( x) dx = = + + p( x) dx y as y s gven by On he oher hand for he fxed y value of mnmzng D % s obaned as he soluon of he followng equaon D% = 0 (0) The soluon of he las equaon s gven by + ˆ = log () Value s deermned such ha nex condon holds ˆ + = y () Usng () and (6) we oban he followng equaon n erms of + log log + = + ( ) exp ( ) whose soluon s gven by () = log () + exp( / ) In he second approach we consder he smlar condon as () amely nsead of ˆ we wll pu he value log = ( ) gven by a (a 004) as he upper bound of he suppor regon In such way we oban he followng esmaon of he mal sgnal amplude () = log (4) exp( / ) Fnally n he hrd approach we wll search for he value of such ha ˆ = By solvng hs equaon we oban = log( + ) (5) I can be seen n he las secon ha quanzers correspondng o he mal sgnal ampludes () () and has almos opmal dsoron especally for he large values of ow we wll descrbe he erave mehod for he mnmzaon of he approxmae dsoron D % As we have already seen oal dsoron D % s he funcon of and y (accordng o he (9) 7

8 (7) and (8)) Moreover relaons (5) and (6) explcly deermne hese wo quanes as he funcon of he opmal mal sgnal amplude Hence we oban closed form analycal expresson for oal dsoron D % as he funcon of number of quanzaon levels and opmal mal sgnal amplude e D% = D% ( ) An opmal value % can be obaned by applyng ewon erave mehod ddd % ( ) + = (6) d Dd % ( ) () () Esmaes and can be used as he sarng pons We can oban D% ( ) as he closed-form expresson and hen perform he symbolc dfferenaon o oban closed-form expressons for he frs and second dervave Smlar procedure (ewon mehod) can be repeaed for he numercal uaon of mnmum of he exac dsoron D( ) denoed by Snce n (6) we need second dervave of D( ) produced expressons wll be oo large Insead of ha s more suable o use some non-graden opmzaon mehod for mnmzaon of D( ) A good choce s Smplex mehod (Chong and Zak 00) due o s smplcy and ease of mplemenaon Recall ha exac expresson for D( ) s obaned by combnng equaons (4) wh equaon () The procedure for uaon of can be descrbed as follows: frs we fnd he exac expresson for D( ) subsung (4) n () and hen apply Smplex opmzaon mehod From he fgure we can see ha boh D and D % has only one global mnmum and hence boh ewon erave mehod and non-graden opmzaon mehods wll converge o % and respecvely A he end of hs secon we recall ha s he exac opmal mal amplude obaned by mnmzaon of he exac dsoron D bu ha mnmzaon procedure s very lcaed Because of ha we found four approxmae values for he opmal mal amplude: % () () % was obaned by mnmzaon of approxmae expresson for dsoron D % bu () ha mnmzaon procedure s also very lcaed On he oher hand expressons for are very smple and n closed-form and because of ha hey are very useful n praccal () applcaons As would be seen from he nex secon especally () and are very closed o 4 umercal examples In hs secon we wll are he approxmae values opmal value () respecvely whle % and () () () and % wh he exac s calculaed usng expressons () (4) and (5) s found by mnmzaon of D % and D as was descrbed n prevous secon Mnmzaon procedures for D and D % are mplemened n he symbolc programmng package MATHEMATICA (see for example (Wolfram 00)) The performance of a quanzer s ofen specfed n erms of sgnal-o-nose rao ( SRQ ) whch s drecly obaned from he dsoron D usng he followng relaon 8

9 σ SRQ = 0 log 0 D oe ha SRQ s descendng funcon of D and hence he quanzer s beer when SRQ has hgher value In Table we show he mal sgnal amplude approxmae values () () and % and exac value for dfferen values of number of quanzaon levels I can be seen ha values and are very close () () % Table : () () % and for dfferen values of Table gves he opmal SRQ of andng based quanzer ( SRQ o he ) as well as he dfferences beween SRQ value correspondng o he () () and % ( SRQ SRQ SRQ and SRQ respecvely) and SRQ ( ) ( ) SRQ = SRQ SRQ for = and SRQ = SRQ SRQ correspondng () () We denoed Table : SRQ () SRQ () SRQ SRQ SRQ * * * * * * * * * *0-6 99* SRQ () SRQ () SRQ SRQ and SRQ for dfferen values of 9

10 I s evden from Table ha andng quanzer wh mal sgnal amplude has SRQ almos equal o he mum possble value SRQ of he quanzer based on he andng model (dfference s arable wh he machne precson) Ths esmaon s even beer han % obaned by mnmzaon of he approxmae dsoron D % Oher wo esmaes have also close values of SRQ o he opmal Therefore we can conclude ha esmaon s very accurae approxmaon of he opmal mal sgnal amplude and can be effcenly used n he praccal applcaons oe ha for 64 all consdered andng models (wh mal sgnal amplude () equal o % and sasfes he soppng creron of he Lloyd-Max algorhm Accordng o he (Gray 004) hs soppng creron can be expressed as follows D D δ = < D where D s he dsoron of he opmal -level quanzer Comparng SRQ values soppng creron can be expressed as SRQ SRQ < 0 0 SRQ where SRQ = 0 log 0( / D ) Ths approves ha quanzers based on he andng model wh opmzed mal sgnal amplude are close o he opmal quanzers On he oher hand le us remember ha n pracce andng quanzers has much smpler srucure han opmal quanzers especally for he large number of quanzaon levels Hence hey are very suable for he praccal applcaons Opmzaon of he lossy algorhm based on opmal resson law was done n hs and prevous secon Also we gave some smple approxmae expressons n closed form for he mal amplude whch sasfed sop creron of Max-Lloyd algorhm The bes approxmae expresson s whch s very smple and gves neglgble error The am was o choose approxmae expresson wh mnmal error n arson o opmal soluon Imporance of he good choce of can be seen from Fgure : for small values for (eg = ) error of dsoron can be greaer han 5% Opmal andng quanzer presened n hs and prevous secon wll be used as a par of generalzed resson algorhm whch wll be presened n secon 6 5 ovel smple lossless resson mehod In hs secon novel smple lossless resson mehod wll be presened Ths s our second goal n hs paper Suppose ha he oupu of he -level quanzer are he ndces = 0 In he res of hs secon we wll suppose ha = r where r s posve neger Hence all ndces = 0 can be coded by he r -b codeword and average b-rae (average number of bs per sample) s equal o R = r We wll develop he smple resson algorhm whch mproves ha b-rae bu akng no accoun he properes of he source X Snce X has Laplacan dsrbuon mos probable ndces are mddle ones e ndces from he segmen I = / 4 / 4 The correspondng range for he source sgnal s [ x x ] where 0

11 x log exp = + Snce I consss of / ndces requred number of bs for he represenaon of each ndex I s r ow we wll descrbe he srucure of he coder Frs ndces are grouped no he frames conssng of M ndces Frs b of each frame s conrol b If all ndces n he group belongs o I hey are all coded wh r bs each and he frs b s se o Oherwse frs b s se o 0 and all ndces are coded wh r bs each Denoe by p he probably ha oupu ndex belongs o I (e ha sample of he sgnal belongs o ( x x ) ) I s equal o p = exp( x ) ow he probably ha M consecuve M ndces belongs o I s equal o p Hence we can ue average b-rae as follows M M R = p ( r ) + ( p ) r + (7) M where he las erm corresponds o he frs b of he frame For example by akng M = = 8 and = we oban x = 4540 p = and R = We wll are hs resul wh he enropy H of he source and average b-rae R h obaned by usng Huffman algorhm (Hankerson e al 004) Le us remember ha las wo quanes are gven by H = P log Rh = P log = P = P The lengh of source exenson M can be opmzed usng relaon (6) For = 8 can be shown ha opmal value of M s M = Also holds H = 6 88 and R = 6847 Le us noce ha H < R < H + and R < Rh Smlarly holds for he oher values of Hence we can conclude ha our lossless coder gves beer performances han Huffman algorhm I s worh menonng ha our coder s drascally smpler for he realzaon han Huffman code Hence s suable for mplemenaons where smplcy of he coder srucure s mporan h 6 Generalzed resson algorhm combnaon of lossy and lossless resson Generalzed resson algorhm consss of lossy and lossless resson algorhm Opmzaon of lossy algorhm wh opmal resson funcon was done by opmzaon of n secons and 4 An smple lossless resson algorhm was proposed n secon 5 where also opmzaon of exended nformaon source was done by opmzng M where M s he order of he source exenson In hs secon we are gvng he algorhms descrbng he lee srucure of our coder conssng of lossy coder (andng model based quanzer) only wh he presened lossless coder I s assumed ha sack S has lengh M and s empy on algorhm sar Also assume ha s equal o and hs value s preued Algorhm (Coder) Inpu Sgnal sample x

12 Sep Compue y = c( x ) Sep Apply unform quanzer on y and fnd he ndex of he cell where y belongs Sep Pu on he op of he sack S Sep 4 If S s no full go o he Sep and connue wh he nex sample Oherwse connue Sep 5 Take M ndces from he sack S Sep 6 If I for every = M hen code every ndex wh r bs and produce he bnary codeword C = ( c c( r ) M of he lengh ( r ) M Transm C Oherwse code every ndex wh r bs and produce he bnary codeword C = ( c c rm ) of he lengh rm Transm 0C quanzer Reurn A he end of hs secon we presen praccal example of usage of generalzed resson mehod for resson of a sll mage Fgure (lef) presens orgnal mage and Fgure (rgh) presens mage afer resson We can see ha ressed mage s vsually very close o he orgnal mage Frs we found dfferences of he consecuve mage samples because hey have Laplacan dsrbuon and hen appled resson mehod on hose dfferences To preven error propagaon besde dfferences we also ransm one orgnal mage sample on every 8 samples We use andng quanzer wh = 64 levels o quanzng dfferences of he samples and for hs quanzer SRQ = 808dB We use lossless algorhm wh M = Average b-rae for ressed mage s R = 5470 bs / pxel and for orgnal mage s R = 8 bs / pxel so he resson rao s R / R = 46 org M Algorhm (Decoder) Sep Take one b c from he npu Sep If c = ake (r )M bs from he npu and decode ndces M each from r bs Oherwse ake rm bs from he npu and decode ndces M each from r bs Sep For each = o M perform: Sep Se y o he value of represenaon level of -h cell of unform x = c ( y) r org r

13 7 Concluson Fgure : Orgnal mage (lef) and ressed mage (rgh) Ths paper provdes he smple srucure coder for memoryless Laplacan source We used andng model based quanzer and perform he opmzaon of he mal sgnal amplude There are derved analycal esmaes of he opmal mal sgnal amplude as well as one numercal esmae (based on he mnmzaon of approxmae dsoron) and exac numercal mehod for s uaon I s shown ha one of he esmaes has almos opmal dsoron Due o he exac smple analycal expresson hs esmae s very suable for he praccal applcaons Also we develop smple lossless coder algorhm for Laplacan source and ared wh he Huffman code and source enropy Man advanage of our mehod s smple realzaon srucure and less b-rae ared wh he Huffman mehod Generally our coder gves he very smple realzaon srucure and performances close o opmal and hence s very useful n praccal applcaons such as speech sgnals mages bomedcal mages ec References Chong EKP SH Zak (00) An Inroducon o Opmzaon John Wley & Sons nc ew York Chcheser Wenhem Brsbane Sngapore Torono Gray R (004) Quanzaon and Daa Compresson Lecure oes Sanford Unversy Hankerson D GA Harrs PD Johnson Jr (004) Inroducon nformaon heory and daa resson nd ed CHAPMA & HALL/CRC Jayan S P oll (984) Dgal Codng of Waveforms Prence-Hall ew Jersey Chaper 4 pp 9 9 Judell L Scharf (986) A smple dervaon of Lloyd s classcal resul for he opmum scalar quanzer IEEE Transacons on Informaon Theory () 6 8 Max J (960) Quanzng for mnmum dsoron IRE Transacons on Informaon Theory Vol IT-6 7- a S DL euhoff (ov 00) On he Suppor of MSE-Opmal Fxed-Rae Scalar Quanzers IEEE Transacons on Informaon Theory 47(7) a S (May 004) On he Suppor of Fxed-Rae Mnmum Mean-Squared Error Scalar Quanzers for Laplacan Source IEEE Transacons on Informaon Theory 50(5) kolc J Z Perc (008) Lloyd-Max s algorhm mplemenaon n speech codng algorhm based on forward adapve echnque Informaca 9() Paner PF W De (Jan 95) Quanzaon dsoron n pulse coun modulaon wh nonunform spacng of levels Proc IRE Perć ZH JR kolć DM Pokraac (007) Esmaon of he suppor regon for Laplacan source scalar quanzers Journal of Elecrcal Engneerng 58() 47 5 Perć ZH JR kolć (007) An effecve mehod for nalzaon of Lloyd-Max s algorhm of opmal scalar quanzaon for Laplacan source Informaca 8() -0 Ramalho M (00) Ramalho G7 Lossless (RGL) Codec Whepaper Cysco Sysems Inc Sayood K (006) Inroducon o daa Compresson rd ed Elsever Inc Sarosolsky R (007) Smple fas and adapve lossless mage resson algorhm Sofware-Pracce and Experence Wolfram S (00 ) Mahemaca Book 5h ed Wolfram Meda

14 Zoran H Perć was born n s Serba n 964 He receved he B Sc degree n elecroncs and elecommuncaons from he Faculy of Elecronc Engneerng š Serba Yugoslava n 989 and M Sc degree n elecommuncaons from he Unversy of š n 994 He receved he Ph D degree from he Unversy of š also n 999 He s currenly Professor a he Deparmen of Telecommuncaons and vcedean of he Faculy of Elecronc Engneerng Unversy of š Serba Hs curren research neress nclude he nformaon heory source and channel codng and sgnal processng He s parculary workng on scalar and vecor quanzaon echnques n speech and mage codng He was auhor and coauhor n over 00 papers n dgal communcaons Dr Perc has been a Revewer for IEEE Transacons on Informaon Theory He s member Edoral Board of Journal Elecroncs and Elecrcal Engneerng Marko D Pekovć was born n š Serba n 984 He graduaed mahemacs and uer scence a he Faculy of Scences and Mahemacs š Serba n 006 and elecommuncaons a Faculy of Elecronc Engneerng š Serba n 007 He receved PhD degree n uer scence from Unversy of š n 008 Currenly he s he research asssan a he Faculy of Scences and Mahemacs š Serba Hs research neress nclude he source and channel codng generalzed nverses of marces Hankel deermnans and opmzaon mehods He s he auhor of abou 0 papers ( of hem n peer-revewed nernaonal ournals) Dr Pekovć has been Revewer for Journal of Compuaonal and Appled Mahemacs Compuers and Mahemacs Wh Applcaons and Inernaonal Journal of Compuer Mahemacs He s suppored by Mnsry of Scence Republc of Serba Gran o 440 Mlan R Dnčć was born n š Serba n 98 He receved he B Sc degree n elecommuncaons from he Faculy of Elecronc Engneerng s n 007 He s currenly on docoral sudes on he same faculy and he s scholar of Mnsry of Scence Republc of Serba Hs curren research neress nclude source codng and quanzaon of speech sgnals and mages 4

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS R&RATA # Vol.) 8, March FURTHER AALYSIS OF COFIDECE ITERVALS FOR LARGE CLIET/SERVER COMPUTER ETWORKS Vyacheslav Abramov School of Mahemacal Scences, Monash Unversy, Buldng 8, Level 4, Clayon Campus, Wellngon

More information

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Journal of Appled Mahemacs and Compuaonal Mechancs 3, (), 45-5 HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Sansław Kukla, Urszula Sedlecka Insue of Mahemacs,

More information

On One Analytic Method of. Constructing Program Controls

On One Analytic Method of. Constructing Program Controls Appled Mahemacal Scences, Vol. 9, 05, no. 8, 409-407 HIKARI Ld, www.m-hkar.com hp://dx.do.org/0.988/ams.05.54349 On One Analyc Mehod of Consrucng Program Conrols A. N. Kvko, S. V. Chsyakov and Yu. E. Balyna

More information

( ) () we define the interaction representation by the unitary transformation () = ()

( ) () we define the interaction representation by the unitary transformation () = () Hgher Order Perurbaon Theory Mchael Fowler 3/7/6 The neracon Represenaon Recall ha n he frs par of hs course sequence, we dscussed he chrödnger and Hesenberg represenaons of quanum mechancs here n he chrödnger

More information

Variants of Pegasos. December 11, 2009

Variants of Pegasos. December 11, 2009 Inroducon Varans of Pegasos SooWoong Ryu bshboy@sanford.edu December, 009 Youngsoo Cho yc344@sanford.edu Developng a new SVM algorhm s ongong research opc. Among many exng SVM algorhms, we wll focus on

More information

An introduction to Support Vector Machine

An introduction to Support Vector Machine An nroducon o Suppor Vecor Machne 報告者 : 黃立德 References: Smon Haykn, "Neural Neworks: a comprehensve foundaon, second edon, 999, Chaper 2,6 Nello Chrsann, John Shawe-Tayer, An Inroducon o Suppor Vecor Machnes,

More information

On computing differential transform of nonlinear non-autonomous functions and its applications

On computing differential transform of nonlinear non-autonomous functions and its applications On compung dfferenal ransform of nonlnear non-auonomous funcons and s applcaons Essam. R. El-Zahar, and Abdelhalm Ebad Deparmen of Mahemacs, Faculy of Scences and Humanes, Prnce Saam Bn Abdulazz Unversy,

More information

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005 Dynamc Team Decson Theory EECS 558 Proec Shruvandana Sharma and Davd Shuman December 0, 005 Oulne Inroducon o Team Decson Theory Decomposon of he Dynamc Team Decson Problem Equvalence of Sac and Dynamc

More information

Solution in semi infinite diffusion couples (error function analysis)

Solution in semi infinite diffusion couples (error function analysis) Soluon n sem nfne dffuson couples (error funcon analyss) Le us consder now he sem nfne dffuson couple of wo blocks wh concenraon of and I means ha, n a A- bnary sysem, s bondng beween wo blocks made of

More information

A Novel Efficient Stopping Criterion for BICM-ID System

A Novel Efficient Stopping Criterion for BICM-ID System A Novel Effcen Soppng Creron for BICM-ID Sysem Xao Yng, L Janpng Communcaon Unversy of Chna Absrac Ths paper devses a novel effcen soppng creron for b-nerleaved coded modulaon wh erave decodng (BICM-ID)

More information

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim Korean J. Mah. 19 (2011), No. 3, pp. 263 272 GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS Youngwoo Ahn and Kae Km Absrac. In he paper [1], an explc correspondence beween ceran

More information

Comb Filters. Comb Filters

Comb Filters. Comb Filters The smple flers dscussed so far are characered eher by a sngle passband and/or a sngle sopband There are applcaons where flers wh mulple passbands and sopbands are requred Thecomb fler s an example of

More information

Robustness Experiments with Two Variance Components

Robustness Experiments with Two Variance Components Naonal Insue of Sandards and Technology (NIST) Informaon Technology Laboraory (ITL) Sascal Engneerng Dvson (SED) Robusness Expermens wh Two Varance Componens by Ana Ivelsse Avlés avles@ns.gov Conference

More information

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS THE PREICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS INTROUCTION The wo dmensonal paral dfferenal equaons of second order can be used for he smulaon of compeve envronmen n busness The arcle presens he

More information

Cubic Bezier Homotopy Function for Solving Exponential Equations

Cubic Bezier Homotopy Function for Solving Exponential Equations Penerb Journal of Advanced Research n Compung and Applcaons ISSN (onlne: 46-97 Vol. 4, No.. Pages -8, 6 omoopy Funcon for Solvng Eponenal Equaons S. S. Raml *,,. Mohamad Nor,a, N. S. Saharzan,b and M.

More information

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy Arcle Inernaonal Journal of Modern Mahemacal Scences, 4, (): - Inernaonal Journal of Modern Mahemacal Scences Journal homepage: www.modernscenfcpress.com/journals/jmms.aspx ISSN: 66-86X Florda, USA Approxmae

More information

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model Probablsc Model for Tme-seres Daa: Hdden Markov Model Hrosh Mamsuka Bonformacs Cener Kyoo Unversy Oulne Three Problems for probablsc models n machne learnng. Compung lkelhood 2. Learnng 3. Parsng (predcon

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm H ( q, p, ) = q p L( q, q, ) H p = q H q = p H = L Equvalen o Lagrangan formalsm Smpler, bu

More information

Sampling Procedure of the Sum of two Binary Markov Process Realizations

Sampling Procedure of the Sum of two Binary Markov Process Realizations Samplng Procedure of he Sum of wo Bnary Markov Process Realzaons YURY GORITSKIY Dep. of Mahemacal Modelng of Moscow Power Insue (Techncal Unversy), Moscow, RUSSIA, E-mal: gorsky@yandex.ru VLADIMIR KAZAKOV

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 0 Canoncal Transformaons (Chaper 9) Wha We Dd Las Tme Hamlon s Prncple n he Hamlonan formalsm Dervaon was smple δi δ Addonal end-pon consrans pq H( q, p, ) d 0 δ q ( ) δq ( ) δ

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm Hqp (,,) = qp Lqq (,,) H p = q H q = p H L = Equvalen o Lagrangan formalsm Smpler, bu wce as

More information

Linear Response Theory: The connection between QFT and experiments

Linear Response Theory: The connection between QFT and experiments Phys540.nb 39 3 Lnear Response Theory: The connecon beween QFT and expermens 3.1. Basc conceps and deas Q: ow do we measure he conducvy of a meal? A: we frs nroduce a weak elecrc feld E, and hen measure

More information

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore.

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore. Ths documen s downloaded from DR-NTU, Nanyang Technologcal Unversy Lbrary, Sngapore. Tle A smplfed verb machng algorhm for word paron n vsual speech processng( Acceped verson ) Auhor(s) Foo, Say We; Yong,

More information

Scattering at an Interface: Oblique Incidence

Scattering at an Interface: Oblique Incidence Course Insrucor Dr. Raymond C. Rumpf Offce: A 337 Phone: (915) 747 6958 E Mal: rcrumpf@uep.edu EE 4347 Appled Elecromagnecs Topc 3g Scaerng a an Inerface: Oblque Incdence Scaerng These Oblque noes may

More information

Li An-Ping. Beijing , P.R.China

Li An-Ping. Beijing , P.R.China A New Type of Cpher: DICING_csb L An-Png Bejng 100085, P.R.Chna apl0001@sna.com Absrac: In hs paper, we wll propose a new ype of cpher named DICING_csb, whch s derved from our prevous sream cpher DICING.

More information

Chapter Lagrangian Interpolation

Chapter Lagrangian Interpolation Chaper 5.4 agrangan Inerpolaon Afer readng hs chaper you should be able o:. dere agrangan mehod of nerpolaon. sole problems usng agrangan mehod of nerpolaon and. use agrangan nerpolans o fnd deraes and

More information

Tight results for Next Fit and Worst Fit with resource augmentation

Tight results for Next Fit and Worst Fit with resource augmentation Tgh resuls for Nex F and Wors F wh resource augmenaon Joan Boyar Leah Epsen Asaf Levn Asrac I s well known ha he wo smple algorhms for he classc n packng prolem, NF and WF oh have an approxmaon rao of

More information

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5 TPG460 Reservor Smulaon 08 page of 5 DISCRETIZATIO OF THE FOW EQUATIOS As we already have seen, fne dfference appromaons of he paral dervaves appearng n he flow equaons may be obaned from Taylor seres

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4 CS434a/54a: Paern Recognon Prof. Olga Veksler Lecure 4 Oulne Normal Random Varable Properes Dscrmnan funcons Why Normal Random Varables? Analycally racable Works well when observaon comes form a corruped

More information

CS286.2 Lecture 14: Quantum de Finetti Theorems II

CS286.2 Lecture 14: Quantum de Finetti Theorems II CS286.2 Lecure 14: Quanum de Fne Theorems II Scrbe: Mara Okounkova 1 Saemen of he heorem Recall he las saemen of he quanum de Fne heorem from he prevous lecure. Theorem 1 Quanum de Fne). Le ρ Dens C 2

More information

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation Global Journal of Pure and Appled Mahemacs. ISSN 973-768 Volume 4, Number 6 (8), pp. 89-87 Research Inda Publcaons hp://www.rpublcaon.com Exsence and Unqueness Resuls for Random Impulsve Inegro-Dfferenal

More information

CHAPTER 10: LINEAR DISCRIMINATION

CHAPTER 10: LINEAR DISCRIMINATION CHAPER : LINEAR DISCRIMINAION Dscrmnan-based Classfcaon 3 In classfcaon h K classes (C,C,, C k ) We defned dscrmnan funcon g j (), j=,,,k hen gven an es eample, e chose (predced) s class label as C f g

More information

Lecture 11 SVM cont

Lecture 11 SVM cont Lecure SVM con. 0 008 Wha we have done so far We have esalshed ha we wan o fnd a lnear decson oundary whose margn s he larges We know how o measure he margn of a lnear decson oundary Tha s: he mnmum geomerc

More information

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes.

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes. umercal negraon of he dffuson equaon (I) Fne dfference mehod. Spaal screaon. Inernal nodes. R L V For hermal conducon le s dscree he spaal doman no small fne spans, =,,: Balance of parcles for an nernal

More information

Relative controllability of nonlinear systems with delays in control

Relative controllability of nonlinear systems with delays in control Relave conrollably o nonlnear sysems wh delays n conrol Jerzy Klamka Insue o Conrol Engneerng, Slesan Techncal Unversy, 44- Glwce, Poland. phone/ax : 48 32 37227, {jklamka}@a.polsl.glwce.pl Keywor: Conrollably.

More information

e-journal Reliability: Theory& Applications No 2 (Vol.2) Vyacheslav Abramov

e-journal Reliability: Theory& Applications No 2 (Vol.2) Vyacheslav Abramov June 7 e-ournal Relably: Theory& Applcaons No (Vol. CONFIDENCE INTERVALS ASSOCIATED WITH PERFORMANCE ANALYSIS OF SYMMETRIC LARGE CLOSED CLIENT/SERVER COMPUTER NETWORKS Absrac Vyacheslav Abramov School

More information

Bandlimited channel. Intersymbol interference (ISI) This non-ideal communication channel is also called dispersive channel

Bandlimited channel. Intersymbol interference (ISI) This non-ideal communication channel is also called dispersive channel Inersymol nererence ISI ISI s a sgnal-dependen orm o nererence ha arses ecause o devaons n he requency response o a channel rom he deal channel. Example: Bandlmed channel Tme Doman Bandlmed channel Frequency

More information

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL Sco Wsdom, John Hershey 2, Jonahan Le Roux 2, and Shnj Waanabe 2 Deparmen o Elecrcal Engneerng, Unversy o Washngon, Seale, WA, USA

More information

Attribute Reduction Algorithm Based on Discernibility Matrix with Algebraic Method GAO Jing1,a, Ma Hui1, Han Zhidong2,b

Attribute Reduction Algorithm Based on Discernibility Matrix with Algebraic Method GAO Jing1,a, Ma Hui1, Han Zhidong2,b Inernaonal Indusral Informacs and Compuer Engneerng Conference (IIICEC 05) Arbue educon Algorhm Based on Dscernbly Marx wh Algebrac Mehod GAO Jng,a, Ma Hu, Han Zhdong,b Informaon School, Capal Unversy

More information

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany Herarchcal Markov Normal Mxure models wh Applcaons o Fnancal Asse Reurns Appendx: Proofs of Theorems and Condonal Poseror Dsrbuons John Geweke a and Gann Amsano b a Deparmens of Economcs and Sascs, Unversy

More information

P R = P 0. The system is shown on the next figure:

P R = P 0. The system is shown on the next figure: TPG460 Reservor Smulaon 08 page of INTRODUCTION TO RESERVOIR SIMULATION Analycal and numercal soluons of smple one-dmensonal, one-phase flow equaons As an nroducon o reservor smulaon, we wll revew he smples

More information

Chapter 6 DETECTION AND ESTIMATION: Model of digital communication system. Fundamental issues in digital communications are

Chapter 6 DETECTION AND ESTIMATION: Model of digital communication system. Fundamental issues in digital communications are Chaper 6 DEECIO AD EIMAIO: Fundamenal ssues n dgal communcaons are. Deecon and. Esmaon Deecon heory: I deals wh he desgn and evaluaon of decson makng processor ha observes he receved sgnal and guesses

More information

FTCS Solution to the Heat Equation

FTCS Solution to the Heat Equation FTCS Soluon o he Hea Equaon ME 448/548 Noes Gerald Reckenwald Porland Sae Unversy Deparmen of Mechancal Engneerng gerry@pdxedu ME 448/548: FTCS Soluon o he Hea Equaon Overvew Use he forward fne d erence

More information

SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β

SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β SARAJEVO JOURNAL OF MATHEMATICS Vol.3 (15) (2007), 137 143 SOME NOISELESS CODING THEOREMS OF INACCURACY MEASURE OF ORDER α AND TYPE β M. A. K. BAIG AND RAYEES AHMAD DAR Absrac. In hs paper, we propose

More information

Volatility Interpolation

Volatility Interpolation Volaly Inerpolaon Prelmnary Verson March 00 Jesper Andreasen and Bran Huge Danse Mares, Copenhagen wan.daddy@danseban.com brno@danseban.com Elecronc copy avalable a: hp://ssrn.com/absrac=69497 Inro Local

More information

3. OVERVIEW OF NUMERICAL METHODS

3. OVERVIEW OF NUMERICAL METHODS 3 OVERVIEW OF NUMERICAL METHODS 3 Inroducory remarks Ths chaper summarzes hose numercal echnques whose knowledge s ndspensable for he undersandng of he dfferen dscree elemen mehods: he Newon-Raphson-mehod,

More information

Chapter 6: AC Circuits

Chapter 6: AC Circuits Chaper 6: AC Crcus Chaper 6: Oulne Phasors and he AC Seady Sae AC Crcus A sable, lnear crcu operang n he seady sae wh snusodal excaon (.e., snusodal seady sae. Complee response forced response naural response.

More information

Fall 2010 Graduate Course on Dynamic Learning

Fall 2010 Graduate Course on Dynamic Learning Fall 200 Graduae Course on Dynamc Learnng Chaper 4: Parcle Flers Sepember 27, 200 Byoung-Tak Zhang School of Compuer Scence and Engneerng & Cognve Scence and Bran Scence Programs Seoul aonal Unversy hp://b.snu.ac.kr/~bzhang/

More information

WiH Wei He

WiH Wei He Sysem Idenfcaon of onlnear Sae-Space Space Baery odels WH We He wehe@calce.umd.edu Advsor: Dr. Chaochao Chen Deparmen of echancal Engneerng Unversy of aryland, College Par 1 Unversy of aryland Bacground

More information

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair

Performance Analysis for a Network having Standby Redundant Unit with Waiting in Repair TECHNI Inernaonal Journal of Compung Scence Communcaon Technologes VOL.5 NO. July 22 (ISSN 974-3375 erformance nalyss for a Nework havng Sby edundan Un wh ang n epar Jendra Sngh 2 abns orwal 2 Deparmen

More information

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment

EEL 6266 Power System Operation and Control. Chapter 5 Unit Commitment EEL 6266 Power Sysem Operaon and Conrol Chaper 5 Un Commmen Dynamc programmng chef advanage over enumeraon schemes s he reducon n he dmensonaly of he problem n a src prory order scheme, here are only N

More information

12d Model. Civil and Surveying Software. Drainage Analysis Module Detention/Retention Basins. Owen Thornton BE (Mech), 12d Model Programmer

12d Model. Civil and Surveying Software. Drainage Analysis Module Detention/Retention Basins. Owen Thornton BE (Mech), 12d Model Programmer d Model Cvl and Surveyng Soware Dranage Analyss Module Deenon/Reenon Basns Owen Thornon BE (Mech), d Model Programmer owen.hornon@d.com 4 January 007 Revsed: 04 Aprl 007 9 February 008 (8Cp) Ths documen

More information

Including the ordinary differential of distance with time as velocity makes a system of ordinary differential equations.

Including the ordinary differential of distance with time as velocity makes a system of ordinary differential equations. Soluons o Ordnary Derenal Equaons An ordnary derenal equaon has only one ndependen varable. A sysem o ordnary derenal equaons consss o several derenal equaons each wh he same ndependen varable. An eample

More information

Robust and Accurate Cancer Classification with Gene Expression Profiling

Robust and Accurate Cancer Classification with Gene Expression Profiling Robus and Accurae Cancer Classfcaon wh Gene Expresson Proflng (Compuaonal ysems Bology, 2005) Auhor: Hafeng L, Keshu Zhang, ao Jang Oulne Background LDA (lnear dscrmnan analyss) and small sample sze problem

More information

A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION

A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION S19 A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION by Xaojun YANG a,b, Yugu YANG a*, Carlo CATTANI c, and Mngzheng ZHU b a Sae Key Laboraory for Geomechancs and Deep Underground Engneerng, Chna Unversy

More information

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System Communcaons n Sascs Theory and Mehods, 34: 475 484, 2005 Copyrgh Taylor & Francs, Inc. ISSN: 0361-0926 prn/1532-415x onlne DOI: 10.1081/STA-200047430 Survval Analyss and Relably A Noe on he Mean Resdual

More information

Example: MOSFET Amplifier Distortion

Example: MOSFET Amplifier Distortion 4/25/2011 Example MSFET Amplfer Dsoron 1/9 Example: MSFET Amplfer Dsoron Recall hs crcu from a prevous handou: ( ) = I ( ) D D d 15.0 V RD = 5K v ( ) = V v ( ) D o v( ) - K = 2 0.25 ma/v V = 2.0 V 40V.

More information

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s Ordnary Dfferenal Equaons n Neuroscence wh Malab eamples. Am - Gan undersandng of how o se up and solve ODE s Am Undersand how o se up an solve a smple eample of he Hebb rule n D Our goal a end of class

More information

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION INTERNATIONAL TRADE T. J. KEHOE UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 27 EXAMINATION Please answer wo of he hree quesons. You can consul class noes, workng papers, and arcles whle you are workng on he

More information

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria IOSR Journal of Mahemacs (IOSR-JM e-issn: 78-578, p-issn: 9-765X. Volume 0, Issue 4 Ver. IV (Jul-Aug. 04, PP 40-44 Mulple SolonSoluons for a (+-dmensonalhroa-sasuma shallow waer wave equaon UsngPanlevé-Bӓclund

More information

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!") i+1,q - [(!

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!) i+1,q - [(! ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL The frs hng o es n wo-way ANOVA: Is here neracon? "No neracon" means: The man effecs model would f. Ths n urn means: In he neracon plo (wh A on he horzonal

More information

Implementation of Quantized State Systems in MATLAB/Simulink

Implementation of Quantized State Systems in MATLAB/Simulink SNE T ECHNICAL N OTE Implemenaon of Quanzed Sae Sysems n MATLAB/Smulnk Parck Grabher, Mahas Rößler 2*, Bernhard Henzl 3 Ins. of Analyss and Scenfc Compung, Venna Unversy of Technology, Wedner Haupsraße

More information

Time-interval analysis of β decay. V. Horvat and J. C. Hardy

Time-interval analysis of β decay. V. Horvat and J. C. Hardy Tme-nerval analyss of β decay V. Horva and J. C. Hardy Work on he even analyss of β decay [1] connued and resuled n he developmen of a novel mehod of bea-decay me-nerval analyss ha produces hghly accurae

More information

Comparison of Differences between Power Means 1

Comparison of Differences between Power Means 1 In. Journal of Mah. Analyss, Vol. 7, 203, no., 5-55 Comparson of Dfferences beween Power Means Chang-An Tan, Guanghua Sh and Fe Zuo College of Mahemacs and Informaon Scence Henan Normal Unversy, 453007,

More information

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model BGC1: Survval and even hsory analyss Oslo, March-May 212 Monday May 7h and Tuesday May 8h The addve regresson model Ørnulf Borgan Deparmen of Mahemacs Unversy of Oslo Oulne of program: Recapulaon Counng

More information

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading Onlne Supplemen for Dynamc Mul-Technology Producon-Invenory Problem wh Emssons Tradng by We Zhang Zhongsheng Hua Yu Xa and Baofeng Huo Proof of Lemma For any ( qr ) Θ s easy o verfy ha he lnear programmng

More information

Advanced Machine Learning & Perception

Advanced Machine Learning & Perception Advanced Machne Learnng & Percepon Insrucor: Tony Jebara SVM Feaure & Kernel Selecon SVM Eensons Feaure Selecon (Flerng and Wrappng) SVM Feaure Selecon SVM Kernel Selecon SVM Eensons Classfcaon Feaure/Kernel

More information

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC CH.3. COMPATIBILITY EQUATIONS Connuum Mechancs Course (MMC) - ETSECCPB - UPC Overvew Compably Condons Compably Equaons of a Poenal Vecor Feld Compably Condons for Infnesmal Srans Inegraon of he Infnesmal

More information

Graduate Macroeconomics 2 Problem set 5. - Solutions

Graduate Macroeconomics 2 Problem set 5. - Solutions Graduae Macroeconomcs 2 Problem se. - Soluons Queson 1 To answer hs queson we need he frms frs order condons and he equaon ha deermnes he number of frms n equlbrum. The frms frs order condons are: F K

More information

2/20/2013. EE 101 Midterm 2 Review

2/20/2013. EE 101 Midterm 2 Review //3 EE Mderm eew //3 Volage-mplfer Model The npu ressance s he equalen ressance see when lookng no he npu ermnals of he amplfer. o s he oupu ressance. I causes he oupu olage o decrease as he load ressance

More information

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study)

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study) Inernaonal Mahemacal Forum, Vol. 8, 3, no., 7 - HIKARI Ld, www.m-hkar.com hp://dx.do.org/.988/mf.3.3488 New M-Esmaor Objecve Funcon n Smulaneous Equaons Model (A Comparave Sudy) Ahmed H. Youssef Professor

More information

Part II CONTINUOUS TIME STOCHASTIC PROCESSES

Part II CONTINUOUS TIME STOCHASTIC PROCESSES Par II CONTINUOUS TIME STOCHASTIC PROCESSES 4 Chaper 4 For an advanced analyss of he properes of he Wener process, see: Revus D and Yor M: Connuous marngales and Brownan Moon Karazas I and Shreve S E:

More information

FI 3103 Quantum Physics

FI 3103 Quantum Physics /9/4 FI 33 Quanum Physcs Aleander A. Iskandar Physcs of Magnesm and Phooncs Research Grou Insu Teknolog Bandung Basc Conces n Quanum Physcs Probably and Eecaon Value Hesenberg Uncerany Prncle Wave Funcon

More information

RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA

RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA Mchaela Chocholaá Unversy of Economcs Braslava, Slovaka Inroducon (1) one of he characersc feaures of sock reurns

More information

Tools for Analysis of Accelerated Life and Degradation Test Data

Tools for Analysis of Accelerated Life and Degradation Test Data Acceleraed Sress Tesng and Relably Tools for Analyss of Acceleraed Lfe and Degradaon Tes Daa Presened by: Reuel Smh Unversy of Maryland College Park smhrc@umd.edu Sepember-5-6 Sepember 28-30 206, Pensacola

More information

Genetic Algorithm in Parameter Estimation of Nonlinear Dynamic Systems

Genetic Algorithm in Parameter Estimation of Nonlinear Dynamic Systems Genec Algorhm n Parameer Esmaon of Nonlnear Dynamc Sysems E. Paeraks manos@egnaa.ee.auh.gr V. Perds perds@vergna.eng.auh.gr Ah. ehagas kehagas@egnaa.ee.auh.gr hp://skron.conrol.ee.auh.gr/kehagas/ndex.hm

More information

Introduction ( Week 1-2) Course introduction A brief introduction to molecular biology A brief introduction to sequence comparison Part I: Algorithms

Introduction ( Week 1-2) Course introduction A brief introduction to molecular biology A brief introduction to sequence comparison Part I: Algorithms Course organzaon Inroducon Wee -2) Course nroducon A bref nroducon o molecular bology A bref nroducon o sequence comparson Par I: Algorhms for Sequence Analyss Wee 3-8) Chaper -3, Models and heores» Probably

More information

Method of upper lower solutions for nonlinear system of fractional differential equations and applications

Method of upper lower solutions for nonlinear system of fractional differential equations and applications Malaya Journal of Maemak, Vol. 6, No. 3, 467-472, 218 hps://do.org/1.26637/mjm63/1 Mehod of upper lower soluons for nonlnear sysem of fraconal dfferenal equaons and applcaons D.B. Dhagude1 *, N.B. Jadhav2

More information

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue. Lnear Algebra Lecure # Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons

More information

Notes on the stability of dynamic systems and the use of Eigen Values.

Notes on the stability of dynamic systems and the use of Eigen Values. Noes on he sabl of dnamc ssems and he use of Egen Values. Source: Macro II course noes, Dr. Davd Bessler s Tme Seres course noes, zarads (999) Ineremporal Macroeconomcs chaper 4 & Techncal ppend, and Hamlon

More information

ISSN MIT Publications

ISSN MIT Publications MIT Inernaonal Journal of Elecrcal and Insrumenaon Engneerng Vol. 1, No. 2, Aug 2011, pp 93-98 93 ISSN 2230-7656 MIT Publcaons A New Approach for Solvng Economc Load Dspach Problem Ansh Ahmad Dep. of Elecrcal

More information

Lecture 6: Learning for Control (Generalised Linear Regression)

Lecture 6: Learning for Control (Generalised Linear Regression) Lecure 6: Learnng for Conrol (Generalsed Lnear Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure 6: RLSC - Prof. Sehu Vjayakumar Lnear Regresson

More information

Supplementary Material to: IMU Preintegration on Manifold for E cient Visual-Inertial Maximum-a-Posteriori Estimation

Supplementary Material to: IMU Preintegration on Manifold for E cient Visual-Inertial Maximum-a-Posteriori Estimation Supplemenary Maeral o: IMU Prenegraon on Manfold for E cen Vsual-Ineral Maxmum-a-Poseror Esmaon echncal Repor G-IRIM-CP&R-05-00 Chrsan Forser, Luca Carlone, Fran Dellaer, and Davde Scaramuzza May 0, 05

More information

Epistemic Game Theory: Online Appendix

Epistemic Game Theory: Online Appendix Epsemc Game Theory: Onlne Appendx Edde Dekel Lucano Pomao Marcano Snscalch July 18, 2014 Prelmnares Fx a fne ype srucure T I, S, T, β I and a probably µ S T. Le T µ I, S, T µ, βµ I be a ype srucure ha

More information

Should Exact Index Numbers have Standard Errors? Theory and Application to Asian Growth

Should Exact Index Numbers have Standard Errors? Theory and Application to Asian Growth Should Exac Index umbers have Sandard Errors? Theory and Applcaon o Asan Growh Rober C. Feensra Marshall B. Rensdorf ovember 003 Proof of Proposon APPEDIX () Frs, we wll derve he convenonal Sao-Vara prce

More information

Fuzzy Set Theory in Modeling Uncertainty Data. via Interpolation Rational Bezier Surface Function

Fuzzy Set Theory in Modeling Uncertainty Data. via Interpolation Rational Bezier Surface Function Appled Mahemacal Scences, Vol. 7, 013, no. 45, 9 38 HIKARI Ld, www.m-hkar.com Fuzzy Se Theory n Modelng Uncerany Daa va Inerpolaon Raonal Bezer Surface Funcon Rozam Zakara Deparmen of Mahemacs, Faculy

More information

, t 1. Transitions - this one was easy, but in general the hardest part is choosing the which variables are state and control variables

, t 1. Transitions - this one was easy, but in general the hardest part is choosing the which variables are state and control variables Opmal Conrol Why Use I - verss calcls of varaons, opmal conrol More generaly More convenen wh consrans (e.g., can p consrans on he dervaves More nsghs no problem (a leas more apparen han hrogh calcls of

More information

On elements with index of the form 2 a 3 b in a parametric family of biquadratic elds

On elements with index of the form 2 a 3 b in a parametric family of biquadratic elds On elemens wh ndex of he form a 3 b n a paramerc famly of bquadrac elds Bora JadrevĆ Absrac In hs paper we gve some resuls abou prmve negral elemens p(c p n he famly of bcyclc bquadrac elds L c = Q ) c;

More information

Modeling and Solving of Multi-Product Inventory Lot-Sizing with Supplier Selection under Quantity Discounts

Modeling and Solving of Multi-Product Inventory Lot-Sizing with Supplier Selection under Quantity Discounts nernaonal ournal of Appled Engneerng Research SSN 0973-4562 Volume 13, Number 10 (2018) pp. 8708-8713 Modelng and Solvng of Mul-Produc nvenory Lo-Szng wh Suppler Selecon under Quany Dscouns Naapa anchanaruangrong

More information

Dual Approximate Dynamic Programming for Large Scale Hydro Valleys

Dual Approximate Dynamic Programming for Large Scale Hydro Valleys Dual Approxmae Dynamc Programmng for Large Scale Hydro Valleys Perre Carpener and Jean-Phlppe Chanceler 1 ENSTA ParsTech and ENPC ParsTech CMM Workshop, January 2016 1 Jon work wh J.-C. Alas, suppored

More information

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas)

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas) Lecure 8: The Lalace Transform (See Secons 88- and 47 n Boas) Recall ha our bg-cure goal s he analyss of he dfferenal equaon, ax bx cx F, where we emloy varous exansons for he drvng funcon F deendng on

More information

ON THE WEAK LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS

ON THE WEAK LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS ON THE WEA LIMITS OF SMOOTH MAPS FOR THE DIRICHLET ENERGY BETWEEN MANIFOLDS FENGBO HANG Absrac. We denfy all he weak sequenal lms of smooh maps n W (M N). In parcular, hs mples a necessary su cen opologcal

More information

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data Anne Chao Ncholas J Goell C seh lzabeh L ander K Ma Rober K Colwell and Aaron M llson 03 Rarefacon and erapolaon wh ll numbers: a framewor for samplng and esmaon n speces dversy sudes cology Monographs

More information

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue. Mah E-b Lecure #0 Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons are

More information

Department of Economics University of Toronto

Department of Economics University of Toronto Deparmen of Economcs Unversy of Torono ECO408F M.A. Economercs Lecure Noes on Heeroskedascy Heeroskedascy o Ths lecure nvolves lookng a modfcaons we need o make o deal wh he regresson model when some of

More information

( ) [ ] MAP Decision Rule

( ) [ ] MAP Decision Rule Announcemens Bayes Decson Theory wh Normal Dsrbuons HW0 due oday HW o be assgned soon Proec descrpon posed Bomercs CSE 90 Lecure 4 CSE90, Sprng 04 CSE90, Sprng 04 Key Probables 4 ω class label X feaure

More information

A DECOMPOSITION METHOD FOR SOLVING DIFFUSION EQUATIONS VIA LOCAL FRACTIONAL TIME DERIVATIVE

A DECOMPOSITION METHOD FOR SOLVING DIFFUSION EQUATIONS VIA LOCAL FRACTIONAL TIME DERIVATIVE S13 A DECOMPOSITION METHOD FOR SOLVING DIFFUSION EQUATIONS VIA LOCAL FRACTIONAL TIME DERIVATIVE by Hossen JAFARI a,b, Haleh TAJADODI c, and Sarah Jane JOHNSTON a a Deparen of Maheacal Scences, Unversy

More information

A NOVEL NETWORK METHOD DESIGNING MULTIRATE FILTER BANKS AND WAVELETS

A NOVEL NETWORK METHOD DESIGNING MULTIRATE FILTER BANKS AND WAVELETS A NOVEL NEWORK MEHOD DESIGNING MULIRAE FILER BANKS AND WAVELES Yng an Deparmen of Elecronc Engneerng and Informaon Scence Unversy of Scence and echnology of Chna Hefe 37, P. R. Chna E-mal: yan@usc.edu.cn

More information

CS 268: Packet Scheduling

CS 268: Packet Scheduling Pace Schedulng Decde when and wha pace o send on oupu ln - Usually mplemened a oupu nerface CS 68: Pace Schedulng flow Ion Soca March 9, 004 Classfer flow flow n Buffer managemen Scheduler soca@cs.bereley.edu

More information

Math 128b Project. Jude Yuen

Math 128b Project. Jude Yuen Mah 8b Proec Jude Yuen . Inroducon Le { Z } be a sequence of observed ndependen vecor varables. If he elemens of Z have a on normal dsrbuon hen { Z } has a mean vecor Z and a varancecovarance marx z. Geomercally

More information