How To Calculate Backup From A Backup From An Oal To A Daa

Size: px
Start display at page:

Download "How To Calculate Backup From A Backup From An Oal To A Daa"

Transcription

1 6 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.4 No.7, July 04 Mahemacal Model of Daa Backup and Recovery Karel Burda The Faculy of Elecrcal Engneerng and Communcaon Brno Unversy of Technology, Brno, Czech Republc Summary The daa backup and daa recovery are well esablshed dscplnes whn he compuer scence already. Despe hs fac, however, here s currenly no suffcenly general mahemacal model for quanfcaon of backup sraeges. In hs paper, such mahemacal model s derved. The model nroduced s based on he assumpon ha he oal sze of he daa s consan. Usng hs model, we can quanavely evaluae he properes of dfferen ypes of backup sraeges. For he full, dfferenal and ncremenal backup sraeges, formulae are derved for he calculaon of he average oal backup sze and he average daa recovery sze n dependence on he probably of he change of daa uns. The quanave relaons obaned are dscussed. Key words: Daa backup, daa recovery, mahemacal model.. Inroducon Daa are one of he mos valuable asses of persons and organzaons. However, we can lose hese daa due o echncal falures, human errors, ec. Ths s he reason why daa are coped no backup daa reposores (so-called daa backup) n order o have he possbly of reconsrucng hese daa when he daa are los n he man daa reposory (so-called daa recovery). Wh regard o mahemacal models of backup processes, A. Frsch saes ha sudes performng numercal and quanave modelng backup processes are surprsngly rare (see [], p. 4). The models publshed so far allow only compuaons of full and ncremenal backups (e.g. [], p. 4) and he quanfcaon of he Amanda backup scheme ([], p or [], p. 5-9). Z. Kurmas and L. Chervenak have publshed a comparson of some backup sraeges [4], however, her evaluaon of backup sraeges s based on smulaons and herefore no analyc resuls are presened n hs paper. Oher research papers are orened on backup frequency and mng (e.g. [5] or [6]) and also do no provde any suable resuls for he evaluaon of general backup sraeges. Therefore, we can sae ha here s currenly no suffcenly general mahemacal model for quanfcaon of daa backup and recovery, despe he grea mporance of boh of hese procedures. Ths paper nroduces a mahemacal model of daa backup whch allows a quanave comparson of dfferen backup mehods. In he nex secon, he basc erms are formulaed and he process of daa backup and recovery s formalzed. The hrd secon s dedcaed o he schemes of daa recovery and he fourh secon deals wh he aomc backup (backups based on snapshos of daa uns). In he ffh secon, a mahemacal model for he quanfcaon of backups s derved and n he sxh secon, hs model s appled o dfferen daa backup mehods and he resuls obaned are dscussed.. Daa backup formalzaon The ermnology of he daa backup s no unfed and many auhors are usng he basc erms n dfferen and ofen conflcng ways [7]. Therefore, we specfy and formalze basc erms a frs. Daa D are a srucure of symbols, whch represens some nformaon. In he daa reposory, hese daa are arranged no ceran elemenary (.e. furher ndvsble) sequences of symbols (e.g. secors of he hard dsk). We refer o an elemenary sequence d of symbols as a daa un and denoe he x-h daa un as d x. In he course of me, he daa un d x can correspond o dfferen sequences of symbols, whch we call versons of he daa un d x. The daa un d x, whch does no represen any nformaon, s he so-called empy daa un. We wll formally denoe such a daa un as d x =. The model s based on he assumpon ha he oal sze of he daa D s consan and ha he sze of he meadaa (overhead of daa uns) s neglgble. We have already saed ha he symbol sequence of he daa un d x vares wh me. From he backup vewpon, he mporan evens are changng he daa uns and akng he backups. To smplfy descrpon, we assume ha he evens menoned can no come smulaneously and changng he daa uns and akng he backups are nsananeous,.e. he duraon of hese evens s equal o zero. When he daa un d x has been changed a he me, we denoe hs verson of he daa un as d x ( ). Ths verson s vald n he me nerval, ), where s he nsan of he nex change of he daa un d x. The quany d x ( ) can also represen a verson of he daa un whch s vald a he me. In our model, we assume ha f >, hen >. Manuscrp receved July 5, 04 Manuscrp revsed July 0, 04

2 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.4 No.7, July 04 7 We wll denoe he daa a he nsan as D( ) and assume ha hese daa conss of n daa uns d ( ) o d n ( ). Then, we can formally represen he daa as an array of n daa uns: n ) = d ( ), d ( ),, d ( ) = d ( ) () [ ] [ ]. D ( n x x= The bass of any arbrary ype of backup s he full backup. The full backup F( ) s a record of all daa uns whch were vald a he me,.e.: F( ) = D( ). () Full backups are large, because hey also conan daa uns whch have remaned unchanged snce he prevous backup. Therefore, hese backups are combned wh paral backups, whch conan only daa whch have been changed compared o some prevous backup. Le D( ) be he daa a he nsan and le D( ) be he daa a he nsan. Denoe by b x (, ) he change of he daa un d x a he nsan n conras o he nsan. I holds for hs change:, when d x has been deleed, b x(, ) = 0, when d x( ) = d x( ), () d x( ), oherwse. The frs lne expresses he suaon when he daa un d x has been deleed, he second lne represens he suaon when here has been no change n he daa un d x and he hrd lne corresponds o he suaon when he daa un d x has been eher creaed or modfed. Then, he paral backup P(, ) s a oal of changes of all daa uns whn he nerval (, : n, ) = b (, ) (4) [ ]. P ( x x= I generally holds ha b x (, ) b x (, ) and herefore P(, ) P(, ). For example, when D( ) = [d ( ), d ( ), d ( )] and D( ) = [d ( ), d ( ), ], hen P(, ) = [0, d ( ), ] bu P(, ) = [0, d ( ), d ( )]. When we know he daa D( ) and he paral backup P(, ), hen we can recover D( ),.e. daa a he nsan. Le us defne an operaon whn whch he daa D( ) are recovered from he daa D( ) and he paral backup P(, ). We refer o hs operaon as a daa revson. Formally, we defne he daa revson n he followng way: D ) P(, ) = D( ), (5) ( where he revson of each daa un s gven by he formula: d ( ) = d ( ) b (, ) = x x, when bx (, ) =, = d x( ), when bx (, ) = 0, bx (, ), oherwse. x (6) In he case of he frs lne, he backup conans he nformaon ha he daa un d x has been deleed n he course of he me nerval (,. For hs daa un, he resul of daa revson s an empy un. In he case of he second lne, he backup conans he nformaon ha here has been no change n he daa un d x n he gven nerval. For hs daa un, he resul of daa revson s he verson of he daa un a he nsan. The hrd lne corresponds o he suaon when he backup conans he changed verson of he daa un d x. In hs case, he resul of daa revson s he changed verson of he daa un. When holds for he daa revson n (5) ha <, hen we wll speak abou forward daa recovery. In hs case, we oban by daa revson a newer verson, D( ), from he older verson, D( ). When holds ha >, hen we wll speak abou backward daa recovery. In hs case, we oban by daa revson he older verson, D( ), from he newer verson, D( ). For our examples of backups, holds ha: D( ) = D( ) P(, ) = [ d( ), d( ), d( ) ] [ 0, d( ), ] = = [ d( ) 0, d( ) d( ), d( ) ] = [ d( ), d( ), ] and D( ) = D( ) P(, ) = [ d( ), d( ), ] [ 0, d( ), d( ) ] = = [ d( ) 0, d( ) d( ), d( ) ] = [ d( ), d( ), d( ) ]. When we do no need o dsngush beween he full backup F( ) and he paral backup P(, ), hen we wll generally denoe hese backups as B( ). We wll classfy paral backups no nerval and aomc backups. An nerval backup I(, ) s a paral backup P(, ) whch conans he laes verson of he daa uns ha have been changed n he nerval (,. Ths means ha f any daa un s changed several mes n he nerval (,, hen he gven nerval backup wll conan only he verson of he daa un whch was vald a he nsan. From hs, follows ha we can recover he sae of daa only for he nsans of akng backups. Ths ype of backup s hsorcally he oldes and hese backups have a smaller sze. The aomc backup s a modern mehod whch has been made possble by he so-called snapsho echnque (e.g. [8]). Wh hs echnque, when a new verson of a daa un s wren, he older verson of he gven daa un (he so-called snapsho) s reaned n he reposory. In he case of aomc backup, all snapshos are backed up and herefore we can recover he daa no he sae a an arbrary nsan. A drawback of hs ype of backup s he large sze of he aggregae of all snapshos. Before we explan he aomc backups n more deal, we wll frs dscuss he daa recovery schemes.

3 8 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.4 No.7, July 04. Daa recovery schemes The scheme of daa recovery defnes he procedure of daa reconsrucon from he backups acqured. In hs paper, we wll express daa recovery schemes by a graph. The nodes and of such a graph represen he backups B( ) and B( ) and he orened edge (, ) expresses ha we mus frs recover he daa from he backup B( ) and subsequenly revse hese daa by usng he daa from he backup B( ). We wll call hs relaon beween backups as a backup successon. In hs case, we wll call he backup B( ) a reference backup and he backup B( ) a subsequen backup. An example of he recovery scheme for fve consecuve backups B( ) o B( 5 ) s gven n Fg.. The backup B( ),.e. node, s a full backup whch conans all daa a he nsan. Smulaneously, hs backup s he reference backup for he paral backups B( ) and B( 4 ). These backups are reference backups for he paral backups B( ) and B( 5 ). The nerval backups B( ), B( ) and B( 5 ) conan changes of daa uns compared o he neares prevous backup. The nerval backup B( 4 ) conans he laes verson of he daa uns whch were changed whn he nerval (, recovery s very smple. The daa D( ) are smple o recover by wrng he full backup F( ) no he man reposory,.e.: D( ) = F( ). (7) Fg. : An example of he mlesone recovery scheme. Oher schemes are based on a combnaon of he full backup and parals backups. We wll refer o hese schemes as reference recovery schemes. The dfferenal and ncremenal recovery schemes are he mos frequenly used. The dfferenal recovery scheme s llusraed n Fg.. In hs ype of recovery scheme, he frs backup B( ) s he full backup and all he oher backups are nerval backups. The reference backup for each nerval backup s he frs backup. Formally, we can express hs scheme as B( ) = F( ) and B( ) = I(, ) where = o 5. The advanage over he prevous scheme s he smaller sze of he aggregae of backups bu he recovery of D( ) for > s more complcaed. In hs case, we frs wre he full backup F( ) no he man reposory a frs and hen we revse hese daa by he backup I(, ). Formally, we can express he daa recovery accordng o he dfferenal recovery scheme as: F( ), for =, D( ) = (8) F( ) I(, ), oherwse. 5 Fg. : An example of a recovery scheme. From Fg., s apparen ha f we wan o recover daa a he nsan, we mus frs wre he full backup B( ) no he man reposory, hen we mus revse hese daa by he backup B( ), and fnally, we mus sll revse he daa obaned by usng he backup B( ). Now, we can explan he recovery schemes more precsely. In hs paper, we focus on he mlesone, dfferenal and ncremenal recovery schemes because hese schemes are he mos wdely used. The above recovery schemes are assocaed wh he full, dfferenal and ncremenal backup sraeges [9]. For our example of fve backups, he aforesad schemes are llusraed n Fg. o Fg. 4. Fg. llusraes he mlesone recovery scheme. Ths scheme consss of full backups (he so-called mlesones) only,.e. B( ) = F( ). A drawback of hs recovery scheme s he large sze of backups (a large capacy of backup reposores s requred), bu he daa Fg.: An example of he dfferenal recovery scheme. Fg. 4 llusraes he ncremenal recovery scheme. In hs ype of recovery scheme, he frs backup B( ) s agan he full backup and all he oher backups are nerval backups. However, he reference backup for each nerval backup s he neares prevous backup. Formally, we can express hs scheme as B( ) = F( ) and B( ) = I(, ) where = o 5. The advanage over he wo prevous schemes s he smaller sze of he aggregae of backups; however, he recovery of D( ) for > s he mos complcaed on he average. Ths s because daa from he backup B( ) mus be sequenally revsed by all backups

4 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.4 No.7, July 04 9 B( ) o B( ). Formally, we can express he daa recovery accordng o he ncremenal recovery scheme as: F( ), for =, D( ) = F( ) I(, )... I( 4 4 5, ), oherwse. Fg. 4: An example of he ncremenal recovery scheme. All he reference schemes descrbed are forward daa recovery schemes,.e. we can recover he daa D( ) from he nal full backup F( ), where >. By analogy, here are also backward daa recovery schemes. In hese cases, he reference full backup conans he newes daa and he paral backups conan older versons of daa uns. If need be, we can reurn o a ceran former sae of he daa. The represenaves of backup sysems wh backward daa recovery are some backup sysems wh daa mrrorng. Daa mrrorng (e.g. [0]) s based on ha each change of any daa un n he man reposory s praccally nsanly execued n he backup reposory oo. Therefore, he daa n boh reposores are dencal and he backup reposory can be used n he case of he falure of he man reposory as a full-blown replacemen of he mpared reposory. When we back up older versons of daa uns, hen from he vald daa D( ), we can resore daa a he nsan, where <. 4. Aomc backup Now, we can reurn o he descrpon of he aomc backup. In he case of aomc backup, when a new verson of he daa un s wren, he older verson of hs daa un (he so-called snapsho) s reaned n he reposory. A backup program looks up and backs up snapshos whch have no been backed up ye. By usng he me daa, whch are kep abou each snapsho, we can arrange he versons of all daa uns accordng o her valdy n he me. Ths sequence allows us o recover daa a an arbrary nsan. From he backup vewpon, we can consder each snapsho as an ndvdual backup. Then we can consder he sequence of snapshos n me as a sequence of ndvdual backups, whch are organzed n he ncremenal recovery scheme. Each aomc backup always conans a sngle 5 (9) verson of a sngle daa un only. When he daa un d y s changed a he nsan, hen we can formally express he correspondng aomc backup A(, ) as: n, ) = b (, ) (0) [ ], A ( x x= where d y( ), for x = y, bx (, ) = () 0, oherwse. Of course, he operaon of daa revson also holds for aomc backups. We wll llusrae he aomc backup on a scenaro accordng o Table and Fg. 5. Tab. : A scenaro for he llusraon of he aomc backup. d ( ) d ( ) d ( ) Backups B d ( ) d ( ) d ( ) B( ) = F( ) = [d ( ), d ( ), d ( )] d ( ) d ( ) d ( ) B( ) = A(, ) = [0, d ( ), 0] d ( ) d ( ) d ( ) B( ) = A(, ) = [0, 0, d ( )] 4 d ( ) d ( 4 ) d ( ) B( 4 ) = A(, 4 ) = [0, d ( 4 ), 0] 5 d ( ) d ( 4 ) d ( ) 4 5 [d ( ), d ( ), d ( )] [0, d ( ), 0] [0, 0, d ( )] 4 [0, d ( 4 ), 0] Fg. 5: An example of he aomc backup. In hs scenaro, he daa conss of hree daa uns d x, where x {,, }. A he nsan, hree daa uns d ( ), d ( ) and d ( ) were placed no he reposory. The full backup F( ) was smulaneously creaed from hese versons of daa uns,.e. B( ) = F( ) = [d ( ), d ( ), d ( )] (see Fg. 5). A he nsan, he daa un d was changed from he verson d ( ) o he verson d ( ). A he nsan, he daa un d was changed from he verson d ( ) o he verson d ( ) whle a he nsan 4, he daa un d was changed from he verson d ( ) o he verson d ( 4 ). Le us suppose ha a backup program s acvaed a he nsan 5. A hs nsan, he vald daa uns are d ( ), d ( 4 ) and d ( ). Also suaed n he reposory are he snapshos d ( ), d ( ) and d ( ). Accordng o he me daa n he meadaa of snapshos,

5 0 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.4 No.7, July 04 he backup program can arrange hese snapshos n me. On hs bass, he backup program creaes he aomc backups B( ) = A(, ) = [0, d ( ), 0], B( ) = A(, ) = [0, 0, d ( )], and B( 4 ) = A(, 4 ) = [0, d ( 4 ), 0]. The snapshos can hen be deleed and he backups B( ), B( ), B( ), and B( 4 ) can be ulzed o recover he daa a an arbrary nsan n he nerval, 5 ). As regards he las lne of Table should be noed ha here was no daa change a he nsan 5 and herefore no aomc backup exss a hs momen. If he aggregae of aomc backups were overly large, hen we could ransform he aomc backups no a sngle nerval backup. In hs way, we can reduce he oal sze of backups because he nerval backup conans only he laes verson of each changed daa un. A he same me, however, we canno recover daa a an arbrary nsan. We can derve he nerval backup from aomc backups by usng he daa revson as follows: I, ) = A(, ) A(, )... A(, ). () ( For he above example, we can wre ha: I(, 4 ) = A(, ) A(, ) A(, 4) = = [ 0, d( ),0] [ 0,0, d( ) ] [ 0, d( 4 ),0] = = [ 0 0 0, d( ) 0 d( 4 ),0 d( ) 0] = = [ 0, d( 4 ), d( ) ] We see ha our aomc backups conan hree non-empy daa uns bu he nerval backup consss of wo nonempy daa uns only. We have reduced he sze of backups, bu now we can only recover he daa a he nsan 4. When we wan o have he possbly of recoverng daa from he pas, hen we can use a combnaon of daa mrrorng and snapshoong. The prncple s he followng. When he daa un d x n he man reposory s changed from he verson d x ( - ) o he verson d x ( ), hen hs change s also execued n he backup reposory and he older verson d x ( - ) s backed up as he aomc backup A(, - ). Aomc backups arranged accordng o me enable us o recover any pas daa. For example, when we need o recover he daa D( ), hen we execue hs recovery as: ) = D( ) A(, ) A(, )... A(, ). () D ( + An example of he recovery scheme whch s based on he combnaon of he daa mrrorng and snapshoong s n Fg. 6 and Table. Ths example corresponds o he scenaro from Table. For hs recovery scheme, he vald daa D( ) of he backup reposory are he reference full backup,.e. F( ) = D( ). A he nsan 5, holds ha F( 5 ) = [d ( ), d ( 4 ), d ( )]. Ths sae s smulaneously he sae a he nsan of he laes daa change,.e. a he nsan 4. Therefore, we can wre ha F( 5 ) = D( 5 ) = D( 4 ) = B( 4 ). The backup F( 5 )= B( 4 ) s he nal reference backup for he ncremenal recovery scheme wh subsequen aomc backups A( 4, ) = [0, d ( ), 0], A(, ) = [0, 0, d ( )], and A(, ) = [0, d ( ), 0]. Le us noe ha hs scheme s very smlar o he aomc ncremenal scheme n Fg. 5. The dfference beween hese schemes s her orenaon n me. In Fg. 5, we see he forward ncremenal scheme whle Fg. 6 llusraes he backward ncremenal scheme. We pon ou ha he backward recovery schemes need no be ncremenal aomc schemes only. As we menoned above, we can derve arbrary nerval backups from aomc backups by usng (). Ths enables us o consruc arbrary backward nerval schemes. Tab.: A scenaro llusrang he backward aomc backup. d ( ) d ( ) d ( ) Backups B d ( ) d ( ) d ( ) d ( ) d ( ) d ( ) B( ) = A(, ) = [0, d ( ), 0] d ( ) d ( ) d ( ) B( ) = A(, ) = [0, 0, d ( )] 4 d ( ) d ( 4 ) d ( ) B( ) = A( 4, ) = [0, d ( ), 0] 5 d ( ) d ( 4 ) d ( ) D( 5 ) = B( 4 ) = [d ( ), d ( 4 ), d ( )] [d ( ), d ( 4 ), d ( )] [0, d ( ), 0] [0, 0, d ( )] [0, d ( ), 0] Fg. 6: An example of he backward aomc backup. Now, we summarze he ermnology nroduced n hs paper. We wll classfy he backups as follows: full, paral, nerval, aomc. We wll sor he recovery schemes as follows: mlesone, reference, forward, backward. We wll classfy he reference recovery schemes as follows: dfferenal, ncremenal,

6 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.4 No.7, July 04 combned. In he nex secon, we nroduce a mahemacal model for a quanave evaluaon of backups and recovery schemes. backup (.e. unl = + τ). We can formally express hs probably as a condonal probably: Q ( τ) = Pr( U > r + τ U > r). (6) 4 5. Mahemacal model r τ In hs secon, we deal wh a mahemacal model whch enables us o quanfy he sze of dfferen backups. Ths enables us o evaluae he properes of dfferen daa recovery schemes. We derve he model for he forward recovery schemes, bu he formulas obaned are also vald for he backwards recovery schemes. In he model, we assume ha all daa uns d x conss of he same number of symbols. We wll call hs number of symbols he sze of he daa un and denoe by d. We pessmscally suppose ha here are no empy daa uns n he man reposory,.e. all n daa uns always conan ceran daa. Then, for he oal amoun of he daa D sored n he man reposory, holds ha D = n d. The same formula evdenly holds for he sze of he full backup: F( ) = D. (4) Now, le us solve he sze of paral backups. Le us suppose ha he ( )-h change of he daa un d x occurred a he nsan - and he -h change of hs daa un occurred a he nsan. Denoe he me beween hese consecuve changes by u,.e. u = ( - ). In he model, he mes u are represened by he random varable U whose averaged value s Δ. We assume ha hs averaged value s he same for all daa uns. Then, he varable λ = /Δ represens he rae of changes of each daa un. We furher assume ha he probably dsrbuon of he random varable U s an exponenal dsrbuon. The cumulave dsrbuon funcon G of hs dsrbuon s: λu G( u) = Pr( U u) = e. (5) Now, we wll represen mahemacally he nerval backups. To hs purpose, we wll use he quanes n Fg. 7. In he Fgure, he nsans,,, and 4 are ndcaed. The nsans and 4 are nsans a whch a ceran daa un was changed. The nsans and are nsans a whch he daa backup was execued. We noe ha hese backups need no be adacen. The quany τ = ( ) s he me beween hese backups, he varable u = ( 4 ) s he me beween wo consecuve changes of he daa un (.e. a realzaon of he random varable U), and he quany r = ( ) s he me beween he change of he daa un and he followng backup. Now, we are neresed n he probably Q(τ), whch s he probably ha when he daa un has no been agan changed unl he frs backup (.e. unl ), hen hs daa un s no changed unl he nsan of he nex observed Fg.7: Quanes for he dervaon of he mahemacal model. The exponenal dsrbuon s a dsrbuon whou memory and herefore holds (e.g. [], p. 40): Pr( U > r + τ U > r) = Pr( U > τ). (7) I follows from (7) ha when he daa un s no changed n he me nerval r (he condon U > r), hen he probably ha hs daa un does no change n he me nerval (r + τ) s he same as he probably ha hs change does no occur n he me nerval τ. Then we have: Q( τ) = Pr( U > r + τ U > r) = Pr( U > τ) = λτ (8) = Pr( U τ) = G( τ) = e. The complemenary probably P(τ) = Q(τ) s, of course, a probably ha he daa un s changed n he me nerval τ. The daa D conss of n daa uns and herefore he nerval backup I( τ, ) on average consss of n P(τ) changed daa uns. Then he average sze I( τ, ) of hs backup s: λτ I( τ, ) = d n P( τ) = D ( e ). (9) Ths formula enables compung he average sze of each subsequen nerval backup B( ) = I( τ, ), whch s spaced he me nerval τ from s reference backup B( τ). Le us assume ha backups are execued perodcally wh he me nerval T. Then holds ha he me nerval beween wo dfferen backups τ = k T, where k =,,,... Now, le us defne he quany q: λt q = Q( T ) = e. (0) The quany q s he probably ha he daa un does no change n he me nerval T. The quany p = q s he probably ha he daa un changes n he me nerval T. For hs quany, holds: λt p = q = e. () Then he average sze of he nerval backup I( k T, ) s: λkt k I( k T, ) = D e = D q () u ( ) ( ).

7 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.4 No.7, July 04 Ths formula enables compung he average sze of each subsequen nerval backup B( ) = I( k T, ), whch s spaced he me nerval k T from s reference backup B( k T). Now, we deermne he oal average sze S of he aggregae of aomc backups from he me nerval T. We have already nroduced he noaon accordng o whch he rae of changes of each daa un s λ and he number of daa uns s n. Then, he oal number of changes N = n λ T. Each such change s recorded n one aomc backup of d symbols n sze. Therefore, he oal average sze S( T, ) of he aggregae of aomc backups from he me nerval T s: S( T, ) = d n λ T = D λ T = D ln. () p We noe ha he nverson of equaon () was used o express S( T, ) n he laes erm. By usng formulas (4), (), and (), we can quanfy he average sze of dfferen ypes of backup. Now, we can quanavely analyze he dfferen ypes of recovery scheme. 6. Dscusson We wll llusrae he ulzaon of our model on recovery schemes n Fgs o 4. The mlesone scheme s n Fg., he dfferenal scheme n Fg., he ncremenal scheme n Fg. 4, and he combned scheme n Fg.. We noe ha he schemes nroduced are comparable snce all hese schemes conss of M = 5 backups. We wll resrc ourselves o forward schemes only, bu he resuls obaned, of course, also hold for he equvalen backward recovery schemes. Wh all nerval recovery schemes, we assume ha he backups are execued wh he perod T. The sze of he daa D( ) and he backups B( ) wll be denoed D( ) and B( ), respecvely. For a comparson of he schemes, we wll ulze wo parameers. The frs parameer s he average oal backup sze C. We compue he value of hs parameer by summng he average szes of all backups n he gven scheme,.e.: C = M = B( ). (4) The parameer C represens he sorage space demands of he gven recovery scheme and herefore we ry o mnmze s value. A hgher value of he average oal backup sze means he backup reposory has a larger sorage capacy and herefore a hgher cos oo. The average recovery me s he nex creron for evaluang he recovery schemes. If we assume a consan rae of wrng daa from he backup reposory no he man reposory, hen he average recovery me depends on he average sze of backups ha we mus wre no he man reposory n order o recover he requred daa D( ). However, o recover dfferen daa,.e. D( ) o D( M ), we mus wre no he reposory he dfferen backups of dfferen szes. For example, n he recovery scheme accordng o Fg., we need o wre only he full backup B( ) o recover he daa D( ). However, o recover he daa D( 5 ), we need o wre no he reposory he backups B( ), B( 4 ), and B( 5 ). We wll call he se of backups needed for recoverng he daa D( ) he recovery daa and denoe her average sze by R. Now, le us reurn o he recovery schemes. We recall ha he node represens he backup a he nsan. We wll call he node of a full backup he roo. Le us denoe by V he se of nodes n he pah from he roo o he node. In our example, he pah for he node s gven by he node self and herefore V = {}. In he case of he node 5, he pah s gven by he sequence of nodes -4-5 and herefore V 5 = {, 4, 5}. Then, for R holds: R = B( ). (5) k k V For example, accordng o Fg., R 5 = B( ) + B( 4 ) + B( 5 ). By usng (5), we can compue he sze of he daa for recoverng any sae of D( ). We wll call he average value of all quanes R o R M he average daa recovery sze and denoe by R. Formally, we can wre: M R = R. (6) M = The average recovery me depends n drec proporon on he quany R and herefore we ry o mnmze he value of R. In such a case, he daa recovery wll be he fases on average. Now, we derve he formulas for C and R for all he schemes consdered. In he case of he mlesone scheme (Fg. ), we see ha he average oal backup sze equals he sum of he average szes of all M = 5 full backups,.e. holds: M C = F( ) = D M. = (7) The average szes of he recovery daa R are dencal for all D( ), where R = D. Therefore, for he average daa recovery sze R, smply holds: R = D. (8) In he case of he dfferenal recovery scheme (see Fg. ), we know ha B( ) = F( ) and B( ) = I(, ), where = o M. Then for he szes of sngle backups, s vald: F( ) = D, =, B( ) = (9) I(, ) = D ( q ), =,,... M.

8 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.4 No.7, July 04 By subsung hese quanes no (4), (5), and (6), we fnally oban: and q q C = D M q M (0) M D q q R = M. () M q In he case of he ncremenal recovery scheme (see Fg. 4), we know ha B( ) = F( ) and B( ) = I(, ), where = o M. Then for he szes of sngle backups, s vald: F( ) = D, =, B( ) = () I(, ) = D ( q), =,,... M. By subsung hese quanes no (4), (5), and (6), we fnally oban: C = D M M q () and [ ( ) ] D R = [ + M ( M ) q]. (4) For comparson, we addonally derve he quanes C and R for he combned recovery scheme n Fg.. The frs backup s full,.e. B( ) = D. The second, hrd and ffh backups are spaced one me nerval T from her reference backups and herefore we can wre ha B( ) = B( ) = B( 5 ) = D ( q). The fourh backup s spaced he me nerval T from s reference backup and herefore B( 4 ) = D ( q ). By subsung hese quanes no (4), (5), and (6), we oban: C = D 5 q q (5) and ( ) D R = ( 4 q q ). (6) 5 To compare C and R for he schemes consdered, we ulze he funcons C = f(p) and R = g(p), where p s he probably ha he daa un s changed n he me nerval T. The graphs of funcons C = f(p) n he mulples of he quany D are n Fg. 8 for he all schemes consdered. In he case of he mlesone recovery scheme, he value of C s he consan 5 D or generally M D. Now, le us analyze he reference recovery schemes. For p = 0, we can see ha C = D for all reference schemes. Ths s due o he fac ha here are no changes n he daa uns. All subsequen backups are empy and he value C = D s gven by he frs (.e. full) backup. The second common exreme of all reference schemes s he value of C when he rae of daa un changes λ. In hs case, all daa uns are always changed whn he nerval T and herefore he probably p =. Then, he oal average sze C of backups equals he value M D,.e. he sorage space demands of all reference schemes are dencal wh hose of he mlesone scheme. 5 C Mlesone [ D ] Dfferenal 4 Combned Incremenal p Fg. 8: The dependence of he average oal backup sze C on he probably p for dfferen recovery schemes. When we compare all schemes, we can see ha here are wo exremes. The frs exreme s he mlesone recovery scheme. For all p (0, ), hs scheme has he hghes demands on he sorage capacy of he backup reposory. The second exreme s he ncremenal recovery scheme where, by conras, hese sorage space demands are he lowes. Then, we can ake he sorage space demands of he mlesone recovery scheme as an upper bound: Cmax = D M (7) and he sorage space demands of he ncremenal recovery scheme as a lower bound: C = D M p (8) [( ) ] mn + The sorage space demands of he oher schemes are beween hese bounds. Fg. 9 llusraes he funcons R = g(p),.e. he dependence of he average daa recovery sze R on he probably p. We can see ha he lower bound of R s gven by he mlesone recovery scheme: R mn = D. (9) By conras, he upper bound of R s gven by he ncremenal recovery scheme: M R max = D p +. (40) From he wo fgures above, we can see ha he mlesone and ncremenal recovery schemes are boundary cases for boh he quany C and he quany R. The mlesone recovery scheme s he wors scheme from he vewpon of sorage space demands, bu he bes scheme from he vewpon of he recovery me. In he case of he

9 4 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.4 No.7, July 04 ncremenal recovery scheme, he exac oppose holds. Then, we can ake he dfferenal scheme and combned recovery schemes as a compromse beween he wo exremes above. R [ D ] Incremenal Mlesone Combned Dfferenal p Fg. 9: The dependence of he average daa recovery sze R on he probably p for dfferen recovery schemes. The las hng n hs dscusson s comparng he szes of he nerval and aomc backups. From (), we know ha he average sze I of he nerval backup whn he me nerval T s: I( T, ) = D ( q) = D p. (4) From (), we know ha he oal average sze S of all aomc backups from he me nerval T s: S( T, ) = D ln. (4) p Fg. 0 llusraes he dependence of S and I on he probably p. We can easly prove ha holds ha S( T, ) / I( T, ),.e. he average sze I of he nerval backup whn he me nerval T s no greaer han he oal average sze S of all aomc backups whn he same me nerval T. Ths s due o ha he aggregae of aomc backups conans all versons of daa uns and no only he laes versons. 5 S, I [ D ] p Fg. 0: Dependence of S and I on he probably p. I S From he Fgure, we can see ha he dfferences beween S and I are no sgnfcan for small values of p. However, hese dfferences are defnely sgnfcan for greaer values of p. For example, he rao S / I equals approxmaely.4 for p = 0.5. Then, n he case of he aomc backup, we need a backup reposory wh a capacy whch s 40 percen greaer han n he case of he nerval backup. In he case of p 0.8, he rao S / I equals,.e. we need a backup reposory wh a capacy whch s wo mes greaer han n he case of he nerval backup. The advanage of he aomc backup,.e. he possbly of recoverng he daa a any me nsan, s pad for by ncreased requremens for he sorage capacy of he backup reposory. 7. Concluson In he paper, he ermnology and mahemacal apparaus for daa backup and recovery s exended. The core of he paper s a mahemacal model of he daa backup and recovery. The mahemacal model enables us o compue he average sze of an arbrary backup (eq. and ) from he probably p ha he daa un s changed n he me nerval T. I enables us o deermne he average oal backup sze C (eq. 4) for any recovery scheme and also he average daa recovery sze R (eq. 6). The model proposed allows us o compare dfferen recovery schemes (e.g. Fg. 8 and 9) and exends he heory of he daa backup and recovery. The model also allows us o compare nerval and aomc backups (see eq., and Fg. 0). The model nroduced s based on he assumpons ha he probably p s he same for all daa uns and ha he daa un change s an even whch has no nfluence on he changes of oher daa uns,.e. hese changes are muually ndependen evens. Furher assumpons are ha he oal sze of he daa s consan and ha no daa un s empy. The assumpons nroduced are no general and herefore our nex goal s o creae a more general model. In any case, however, he model descrbed s suable for heorec purposes a leas. References [] A. Frsch: Sysem Backup: Mehodologes, Algorhms and Effcency Models. In J. Bergsra, M. Burgess: Handbook of Nework and Sysem Admnsraon. Elsever, Amserdam 007. [] S. Nelson: Pro Daa Backup and Recovery. Apress, New York 0. [] A. Frsch: Essenal Sysem Admnsraon. O Relly Meda, Sebasopol 00.

10 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.4 No.7, July 04 5 [4] Z. Kurmas, A. L. Chervenak: Evaluang Backup Algorhms. In IEEE Symposum on Mass Sorage Sysems. 000, pp [5] M. Burgess, T. Rean: A rsk analyss of dsk backup or reposory manenance. Scence of Compuer Programmng. 64 (007), pp. -. DOI: 0.06/.scco [6] C. Qan, Y. Huang, X. Zhao, Tosho Nakagawa: X: Opmal Backup Inerval for a Daabase Sysem wh Full and Perodc Incremenal Backup. Journal of Compuers. 5 (00), pp [7] Teradacyl: Backup Terms and Defnons. Teradacyl LLC. Rereved Aprl 7, 04 from hp:// ermnology.hml [8] N. Garrmella: Undersandng and explong snapsho echnology for daa proecon. Par : Snapsho echnology overvew. IBM developerworks. (Aprl 6, 006). Rereved Aprl 7, 04 from hp:// snapsm/ndex.hml?ca=da- [9] P. de Guse: Enerprse Sysems Backup and Recovery. CRC Press, Boca Raon 008. [0] M. Lone: Msson-Crcal Nework Plannng. Arech House, London 00. [] L. Lakaos, L. Szedl, M. Telek: Inroducon o Queueng Sysems wh Telecommuncaon Applcaons. Sprnger, New York 0. Karel Burda receved he M.S. and PhD. degrees n Elecrcal Engneerng from he Lpovsky Mkulas Mlary Academy n 98 and 988, respecvely. Durng , he was a lecurer n wo mlary academes. A presen, he works a Brno Unversy of Technology. Hs curren research neress nclude he secury of nformaon sysems and crypology.

Spline. Computer Graphics. B-splines. B-Splines (for basis splines) Generating a curve. Basis Functions. Lecture 14 Curves and Surfaces II

Spline. Computer Graphics. B-splines. B-Splines (for basis splines) Generating a curve. Basis Functions. Lecture 14 Curves and Surfaces II Lecure 4 Curves and Surfaces II Splne A long flexble srps of meal used by drafspersons o lay ou he surfaces of arplanes, cars and shps Ducks weghs aached o he splnes were used o pull he splne n dfferen

More information

The Virtual Machine Resource Allocation based on Service Features in Cloud Computing Environment

The Virtual Machine Resource Allocation based on Service Features in Cloud Computing Environment Send Orders for Reprns o reprns@benhamscence.ae The Open Cybernecs & Sysemcs Journal, 2015, 9, 639-647 639 Open Access The Vrual Machne Resource Allocaon based on Servce Feaures n Cloud Compung Envronmen

More information

Revision: June 12, 2010 215 E Main Suite D Pullman, WA 99163 (509) 334 6306 Voice and Fax

Revision: June 12, 2010 215 E Main Suite D Pullman, WA 99163 (509) 334 6306 Voice and Fax .3: Inucors Reson: June, 5 E Man Sue D Pullman, WA 9963 59 334 636 Voce an Fax Oerew We connue our suy of energy sorage elemens wh a scusson of nucors. Inucors, lke ressors an capacors, are passe wo-ermnal

More information

Capacity Planning. Operations Planning

Capacity Planning. Operations Planning Operaons Plannng Capacy Plannng Sales and Operaons Plannng Forecasng Capacy plannng Invenory opmzaon How much capacy assgned o each producon un? Realsc capacy esmaes Sraegc level Moderaely long me horzon

More information

Methodology of the CBOE S&P 500 PutWrite Index (PUT SM ) (with supplemental information regarding the CBOE S&P 500 PutWrite T-W Index (PWT SM ))

Methodology of the CBOE S&P 500 PutWrite Index (PUT SM ) (with supplemental information regarding the CBOE S&P 500 PutWrite T-W Index (PWT SM )) ehodology of he CBOE S&P 500 PuWre Index (PUT S ) (wh supplemenal nformaon regardng he CBOE S&P 500 PuWre T-W Index (PWT S )) The CBOE S&P 500 PuWre Index (cker symbol PUT ) racks he value of a passve

More information

MORE ON TVM, "SIX FUNCTIONS OF A DOLLAR", FINANCIAL MECHANICS. Copyright 2004, S. Malpezzi

MORE ON TVM, SIX FUNCTIONS OF A DOLLAR, FINANCIAL MECHANICS. Copyright 2004, S. Malpezzi MORE ON VM, "SIX FUNCIONS OF A DOLLAR", FINANCIAL MECHANICS Copyrgh 2004, S. Malpezz I wan everyone o be very clear on boh he "rees" (our basc fnancal funcons) and he "fores" (he dea of he cash flow model).

More information

Lecture 40 Induction. Review Inductors Self-induction RL circuits Energy stored in a Magnetic Field

Lecture 40 Induction. Review Inductors Self-induction RL circuits Energy stored in a Magnetic Field ecure 4 nducon evew nducors Self-nducon crcus nergy sored n a Magnec Feld 1 evew nducon end nergy Transfers mf Bv Mechancal energy ransform n elecrc and hen n hermal energy P Fv B v evew eformulaon of

More information

An Architecture to Support Distributed Data Mining Services in E-Commerce Environments

An Architecture to Support Distributed Data Mining Services in E-Commerce Environments An Archecure o Suppor Dsrbued Daa Mnng Servces n E-Commerce Envronmens S. Krshnaswamy 1, A. Zaslavsky 1, S.W. Loke 2 School of Compuer Scence & Sofware Engneerng, Monash Unversy 1 900 Dandenong Road, Caulfeld

More information

Pedro M. Castro Iiro Harjunkoski Ignacio E. Grossmann. Lisbon, Portugal Ladenburg, Germany Pittsburgh, USA

Pedro M. Castro Iiro Harjunkoski Ignacio E. Grossmann. Lisbon, Portugal Ladenburg, Germany Pittsburgh, USA Pedro M. Casro Iro Harjunkosk Ignaco E. Grossmann Lsbon Porugal Ladenburg Germany Psburgh USA 1 Process operaons are ofen subjec o energy consrans Heang and coolng ules elecrcal power Avalably Prce Challengng

More information

HEURISTIC ALGORITHM FOR SINGLE RESOURCE CONSTRAINED PROJECT SCHEDULING PROBLEM BASED ON THE DYNAMIC PROGRAMMING

HEURISTIC ALGORITHM FOR SINGLE RESOURCE CONSTRAINED PROJECT SCHEDULING PROBLEM BASED ON THE DYNAMIC PROGRAMMING Yugoslav Journal o Operaons Research Volume 19 (2009) Number 2, 281-298 DOI:10.2298/YUJOR0902281S HEURISTIC ALGORITHM FOR SINGLE RESOURCE CONSTRAINED PROJECT SCHEDULING PROBLEM BASED ON THE DYNAMIC PROGRAMMING

More information

PerfCenter: A Methodology and Tool for Performance Analysis of Application Hosting Centers

PerfCenter: A Methodology and Tool for Performance Analysis of Application Hosting Centers PerfCener: A Mehodology and Tool for Performance Analyss of Applcaon Hosng Ceners Rukma P. Verlekar, Varsha Ape, Prakhar Goyal, Bhavsh Aggarwal Dep. of Compuer Scence and Engneerng Indan Insue of Technology

More information

GUIDANCE STATEMENT ON CALCULATION METHODOLOGY

GUIDANCE STATEMENT ON CALCULATION METHODOLOGY GUIDANCE STATEMENT ON CALCULATION METHODOLOGY Adopon Dae: 9/28/0 Effecve Dae: //20 Reroacve Applcaon: No Requred www.gpssandards.org 204 CFA Insue Gudance Saemen on Calculaon Mehodology GIPS GUIDANCE STATEMENT

More information

Linear Extension Cube Attack on Stream Ciphers Abstract: Keywords: 1. Introduction

Linear Extension Cube Attack on Stream Ciphers Abstract: Keywords: 1. Introduction Lnear Exenson Cube Aack on Sream Cphers Lren Dng Yongjuan Wang Zhufeng L (Language Engneerng Deparmen, Luo yang Unversy for Foregn Language, Luo yang cy, He nan Provnce, 47003, P. R. Chna) Absrac: Basng

More information

An Anti-spam Filter Combination Framework for Text-and-Image Emails through Incremental Learning

An Anti-spam Filter Combination Framework for Text-and-Image Emails through Incremental Learning An An-spam Fler Combnaon Framework for Tex-and-Image Emals hrough Incremenal Learnng 1 Byungk Byun, 1 Chn-Hu Lee, 2 Seve Webb, 2 Danesh Iran, and 2 Calon Pu 1 School of Elecrcal & Compuer Engr. Georga

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

APPLICATION OF CHAOS THEORY TO ANALYSIS OF COMPUTER NETWORK TRAFFIC Liudvikas Kaklauskas, Leonidas Sakalauskas

APPLICATION OF CHAOS THEORY TO ANALYSIS OF COMPUTER NETWORK TRAFFIC Liudvikas Kaklauskas, Leonidas Sakalauskas The XIII Inernaonal Conference Appled Sochasc Models and Daa Analyss (ASMDA-2009) June 30-July 3 2009 Vlnus LITHUANIA ISBN 978-9955-28-463-5 L. Sakalauskas C. Skadas and E. K. Zavadskas (Eds.): ASMDA-2009

More information

Genetic Algorithm with Range Selection Mechanism for Dynamic Multiservice Load Balancing in Cloud-Based Multimedia System

Genetic Algorithm with Range Selection Mechanism for Dynamic Multiservice Load Balancing in Cloud-Based Multimedia System ISSN : 2347-8446 (Onlne) Inernaonal Journal of Advanced Research n Genec Algorhm wh Range Selecon Mechansm for Dynamc Mulservce Load Balancng n Cloud-Based Mulmeda Sysem I Mchael Sadgun Rao Kona, II K.Purushoama

More information

12/7/2011. Procedures to be Covered. Time Series Analysis Using Statgraphics Centurion. Time Series Analysis. Example #1 U.S.

12/7/2011. Procedures to be Covered. Time Series Analysis Using Statgraphics Centurion. Time Series Analysis. Example #1 U.S. Tme Seres Analyss Usng Sagraphcs Cenuron Nel W. Polhemus, CTO, SaPon Technologes, Inc. Procedures o be Covered Descrpve Mehods (me sequence plos, auocorrelaon funcons, perodograms) Smoohng Seasonal Decomposon

More information

MULTI-WORKDAY ERGONOMIC WORKFORCE SCHEDULING WITH DAYS OFF

MULTI-WORKDAY ERGONOMIC WORKFORCE SCHEDULING WITH DAYS OFF Proceedngs of he 4h Inernaonal Conference on Engneerng, Projec, and Producon Managemen (EPPM 203) MULTI-WORKDAY ERGONOMIC WORKFORCE SCHEDULING WITH DAYS OFF Tar Raanamanee and Suebsak Nanhavanj School

More information

The Rules of the Settlement Guarantee Fund. 1. These Rules, hereinafter referred to as "the Rules", define the procedures for the formation

The Rules of the Settlement Guarantee Fund. 1. These Rules, hereinafter referred to as the Rules, define the procedures for the formation Vald as of May 31, 2010 The Rules of he Selemen Guaranee Fund 1 1. These Rules, herenafer referred o as "he Rules", defne he procedures for he formaon and use of he Selemen Guaranee Fund, as defned n Arcle

More information

Anomaly Detection in Network Traffic Using Selected Methods of Time Series Analysis

Anomaly Detection in Network Traffic Using Selected Methods of Time Series Analysis I. J. Compuer Nework and Informaon Secury, 2015, 9, 10-18 Publshed Onlne Augus 2015 n MECS (hp://www.mecs-press.org/) DOI: 10.5815/jcns.2015.09.02 Anomaly Deecon n Nework Traffc Usng Seleced Mehods of

More information

Estimating intrinsic currency values

Estimating intrinsic currency values Cung edge Foregn exchange Esmang nrnsc currency values Forex marke praconers consanly alk abou he srenghenng or weakenng of ndvdual currences. In hs arcle, Jan Chen and Paul Dous presen a new mehodology

More information

CLoud computing has recently emerged as a new

CLoud computing has recently emerged as a new 1 A Framework of Prce Bddng Confguraons for Resource Usage n Cloud Compung Kenl L, Member, IEEE, Chubo Lu, Keqn L, Fellow, IEEE, and Alber Y. Zomaya, Fellow, IEEE Absrac In hs paper, we focus on prce bddng

More information

Analysis of intelligent road network, paradigm shift and new applications

Analysis of intelligent road network, paradigm shift and new applications CONFERENCE ABOUT THE STATUS AND FUTURE OF THE EDUCATIONAL AND R&D SERVICES FOR THE VEHICLE INDUSTRY Analyss of nellgen road nework, paradgm shf and new applcaons Péer Tamás "Smarer Transpor" - IT for co-operave

More information

Ground rules. Guide to the calculation methods of the FTSE Actuaries UK Gilts Index Series v1.9

Ground rules. Guide to the calculation methods of the FTSE Actuaries UK Gilts Index Series v1.9 Ground rules Gude o he calculaon mehods of he FTSE Acuares UK Gls Index Seres v1.9 fserussell.com Ocober 2015 Conens 1.0 Inroducon... 4 1.1 Scope... 4 1.2 FTSE Russell... 5 1.3 Overvew of he calculaons...

More information

Network Effects on Standard Software Markets: A Simulation Model to examine Pricing Strategies

Network Effects on Standard Software Markets: A Simulation Model to examine Pricing Strategies Nework Effecs on Sandard Sofware Markes Page Nework Effecs on Sandard Sofware Markes: A Smulaon Model o examne Prcng Sraeges Peer Buxmann Absrac Ths paper examnes sraeges of sandard sofware vendors, n

More information

The Multi-shift Vehicle Routing Problem with Overtime

The Multi-shift Vehicle Routing Problem with Overtime The Mul-shf Vehcle Roung Problem wh Overme Yngao Ren, Maged Dessouy, and Fernando Ordóñez Danel J. Epsen Deparmen of Indusral and Sysems Engneerng Unversy of Souhern Calforna 3715 McClnoc Ave, Los Angeles,

More information

RESOLUTION OF THE LINEAR FRACTIONAL GOAL PROGRAMMING PROBLEM

RESOLUTION OF THE LINEAR FRACTIONAL GOAL PROGRAMMING PROBLEM Revsa Elecrónca de Comuncacones y Trabajos de ASEPUMA. Rec@ Volumen Págnas 7 a 40. RESOLUTION OF THE LINEAR FRACTIONAL GOAL PROGRAMMING PROBLEM RAFAEL CABALLERO rafael.caballero@uma.es Unversdad de Málaga

More information

Social security, education, retirement and growth*

Social security, education, retirement and growth* Hacenda P úblca Espa ñola / Revsa de Econom ía P úblca, 198-(3/2011): 9-36 2011, Insuo de Esudos Fscales Socal secury, educaon, reremen and growh* CRUZ A. ECHEVARR ÍA AMAIA IZA** Unversdad del Pa ís Vasco

More information

Boosting for Learning Multiple Classes with Imbalanced Class Distribution

Boosting for Learning Multiple Classes with Imbalanced Class Distribution Boosng for Learnng Mulple Classes wh Imbalanced Class Dsrbuon Yanmn Sun Deparmen of Elecrcal and Compuer Engneerng Unversy of Waerloo Waerloo, Onaro, Canada y8sun@engmal.uwaerloo.ca Mohamed S. Kamel Deparmen

More information

DEPARTMENT OF ECONOMETRICS AND BUSINESS STATISTICS. Exponential Smoothing for Inventory Control: Means and Variances of Lead-Time Demand

DEPARTMENT OF ECONOMETRICS AND BUSINESS STATISTICS. Exponential Smoothing for Inventory Control: Means and Variances of Lead-Time Demand ISSN 440-77X ISBN 0 736 094 X AUSTRALIA DEPARTMENT OF ECONOMETRICS AND BUSINESS STATISTICS Exponenal Smoohng for Invenory Conrol: Means and Varances of Lead-Tme Demand Ralph D. Snyder, Anne B. Koehler,

More information

Proceedings of the 2008 Winter Simulation Conference S. J. Mason, R. R. Hill, L. Mönch, O. Rose, T. Jefferson, J. W. Fowler eds.

Proceedings of the 2008 Winter Simulation Conference S. J. Mason, R. R. Hill, L. Mönch, O. Rose, T. Jefferson, J. W. Fowler eds. Proceedngs of he 008 Wner Smulaon Conference S. J. Mason, R. R. Hll, L. Mönch, O. Rose, T. Jefferson, J. W. Fowler eds. DEMAND FORECAST OF SEMICONDUCTOR PRODUCTS BASED ON TECHNOLOGY DIFFUSION Chen-Fu Chen,

More information

Index Mathematics Methodology

Index Mathematics Methodology Index Mahemacs Mehodology S&P Dow Jones Indces: Index Mehodology Ocober 2015 Table of Conens Inroducon 4 Dfferen Varees of Indces 4 The Index Dvsor 5 Capalzaon Weghed Indces 6 Defnon 6 Adjusmens o Share

More information

An Ensemble Data Mining and FLANN Combining Short-term Load Forecasting System for Abnormal Days

An Ensemble Data Mining and FLANN Combining Short-term Load Forecasting System for Abnormal Days JOURNAL OF SOFTWARE, VOL. 6, NO. 6, JUNE 0 96 An Ensemble Daa Mnng and FLANN Combnng Shor-erm Load Forecasng Sysem for Abnormal Days Mng L College of Auomaon, Guangdong Unversy of Technology, Guangzhou,

More information

Multiple Periodic Preventive Maintenance for Used Equipment under Lease

Multiple Periodic Preventive Maintenance for Used Equipment under Lease Mulle Perodc Prevenve Manenance or Used Equmen under ease Paarasaya Boonyaha, Jarumon Jauronnaee, Member, IAENG Absrac Ugradng acon revenve manenance are alernaves o reduce he used equmen alures rae whch

More information

Both human traders and algorithmic

Both human traders and algorithmic Shuhao Chen s a Ph.D. canddae n sascs a Rugers Unversy n Pscaaway, NJ. bhmchen@sa.rugers.edu Rong Chen s a professor of Rugers Unversy n Pscaaway, NJ and Peng Unversy, n Bejng, Chna. rongchen@sa.rugers.edu

More information

A 3D Model Retrieval System Using The Derivative Elevation And 3D-ART

A 3D Model Retrieval System Using The Derivative Elevation And 3D-ART 3 Model Rereal Sysem Usng he erae leaon nd 3-R Jau-Lng Shh* ng-yen Huang Yu-hen Wang eparmen of ompuer Scence and Informaon ngneerng hung Hua Unersy Hsnchu awan RO -mal: sjl@chueduw bsrac In recen years

More information

SPC-based Inventory Control Policy to Improve Supply Chain Dynamics

SPC-based Inventory Control Policy to Improve Supply Chain Dynamics Francesco Cosanno e al. / Inernaonal Journal of Engneerng and Technology (IJET) SPC-based Invenory Conrol Polcy o Improve Supply Chan ynamcs Francesco Cosanno #, Gulo Gravo #, Ahmed Shaban #3,*, Massmo

More information

Insurance. By Mark Dorfman, Alexander Kling, and Jochen Russ. Abstract

Insurance. By Mark Dorfman, Alexander Kling, and Jochen Russ. Abstract he Impac Of Deflaon On Insurance Companes Offerng Parcpang fe Insurance y Mar Dorfman, lexander Klng, and Jochen Russ bsrac We presen a smple model n whch he mpac of a deflaonary economy on lfe nsurers

More information

(Im)possibility of Safe Exchange Mechanism Design

(Im)possibility of Safe Exchange Mechanism Design (Im)possbly of Safe Exchange Mechansm Desgn Tuomas Sandholm Compuer Scence Deparmen Carnege Mellon Unversy 5 Forbes Avenue Psburgh, PA 15213 sandholm@cs.cmu.edu XaoFeng Wang Deparmen of Elecrcal and Compuer

More information

How Much Life Insurance is Enough?

How Much Life Insurance is Enough? How Much Lfe Insurance s Enough? Uly-Based pproach By LJ Rossouw BSTRCT The paper ams o nvesgae how much lfe nsurance proecon cover a uly maxmsng ndvdual should buy. Ths queson s relevan n he nsurance

More information

THE USE IN BANKS OF VALUE AT RISK METHOD IN MARKET RISK MANAGEMENT. Ioan TRENCA *

THE USE IN BANKS OF VALUE AT RISK METHOD IN MARKET RISK MANAGEMENT. Ioan TRENCA * ANALELE ŞTIINłIFICE ALE UNIVERSITĂłII ALEXANDRU IOAN CUZA DIN IAŞI Tomul LVI ŞnŃe Economce 009 THE USE IN BANKS OF VALUE AT RISK METHOD IN MARKET RISK MANAGEMENT Ioan TRENCA * Absrac In sophscaed marke

More information

A Background Layer Model for Object Tracking through Occlusion

A Background Layer Model for Object Tracking through Occlusion A Background Layer Model for Obec Trackng hrough Occluson Yue Zhou and Ha Tao Deparmen of Compuer Engneerng Unversy of Calforna, Sana Cruz, CA 95064 {zhou,ao}@soe.ucsc.edu Absrac Moon layer esmaon has

More information

INTERNATIONAL JOURNAL OF STRATEGIC MANAGEMENT

INTERNATIONAL JOURNAL OF STRATEGIC MANAGEMENT IJSM, Volume, Number, 0 ISSN: 555-4 INTERNATIONAL JOURNAL OF STRATEGIC MANAGEMENT SPONSORED BY: Angelo Sae Unversy San Angelo, Texas, USA www.angelo.edu Managng Edors: Professor Alan S. Khade, Ph.D. Calforna

More information

Cost- and Energy-Aware Load Distribution Across Data Centers

Cost- and Energy-Aware Load Distribution Across Data Centers - and Energy-Aware Load Dsrbuon Across Daa Ceners Ken Le, Rcardo Banchn, Margare Maronos, and Thu D. Nguyen Rugers Unversy Prnceon Unversy Inroducon Today, many large organzaons operae mulple daa ceners.

More information

Prices of Credit Default Swaps and the Term Structure of Credit Risk

Prices of Credit Default Swaps and the Term Structure of Credit Risk Prces of Cred Defaul Swaps and he Term Srucure of Cred Rsk by Mary Elzabeh Desrosers A Professonal Maser s Projec Submed o he Faculy of he WORCESTER POLYTECHNIC INSTITUTE n paral fulfllmen of he requremens

More information

Optimization of Nurse Scheduling Problem with a Two-Stage Mathematical Programming Model

Optimization of Nurse Scheduling Problem with a Two-Stage Mathematical Programming Model Asa Pacfc Managemen Revew 15(4) (2010) 503-516 Opmzaon of Nurse Schedulng Problem wh a Two-Sage Mahemacal Programmng Model Chang-Chun Tsa a,*, Cheng-Jung Lee b a Deparmen of Busness Admnsraon, Trans World

More information

SHIPPING ECONOMIC ANALYSIS FOR ULTRA LARGE CONTAINERSHIP

SHIPPING ECONOMIC ANALYSIS FOR ULTRA LARGE CONTAINERSHIP Journal of he Easern Asa Socey for Transporaon Sudes, Vol. 6, pp. 936-951, 2005 SHIPPING ECONOMIC ANALYSIS FOR ULTRA LARGE CONTAINERSHIP Chaug-Ing HSU Professor Deparen of Transporaon Technology and Manageen

More information

Kalman filtering as a performance monitoring technique for a propensity scorecard

Kalman filtering as a performance monitoring technique for a propensity scorecard Kalman flerng as a performance monorng echnque for a propensy scorecard Kaarzyna Bjak * Unversy of Souhampon, Souhampon, UK, and Buro Informacj Kredyowej S.A., Warsaw, Poland Absrac Propensy scorecards

More information

A binary powering Schur algorithm for computing primary matrix roots

A binary powering Schur algorithm for computing primary matrix roots Numercal Algorhms manuscr No. (wll be nsered by he edor) A bnary owerng Schur algorhm for comung rmary marx roos Federco Greco Bruno Iannazzo Receved: dae / Acceed: dae Absrac An algorhm for comung rmary

More information

Case Study on Web Service Composition Based on Multi-Agent System

Case Study on Web Service Composition Based on Multi-Agent System 900 JOURNAL OF SOFTWARE, VOL. 8, NO. 4, APRIL 2013 Case Sudy on Web Servce Composon Based on Mul-Agen Sysem Shanlang Pan Deparmen of Compuer Scence and Technology, Nngbo Unversy, Chna PanShanLang@gmal.com

More information

Using Cellular Automata for Improving KNN Based Spam Filtering

Using Cellular Automata for Improving KNN Based Spam Filtering The Inernaonal Arab Journal of Informaon Technology, Vol. 11, No. 4, July 2014 345 Usng Cellular Auomaa for Improvng NN Based Spam Flerng Faha Bargou, Bouzane Beldjlal, and Baghdad Aman Compuer Scence

More information

Structural jump-diffusion model for pricing collateralized debt obligations tranches

Structural jump-diffusion model for pricing collateralized debt obligations tranches Appl. Mah. J. Chnese Unv. 010, 54): 40-48 Srucural jump-dffuson model for prcng collaeralzed deb oblgaons ranches YANG Ru-cheng Absrac. Ths paper consders he prcng problem of collaeralzed deb oblgaons

More information

Testing techniques and forecasting ability of FX Options Implied Risk Neutral Densities. Oren Tapiero

Testing techniques and forecasting ability of FX Options Implied Risk Neutral Densities. Oren Tapiero Tesng echnques and forecasng ably of FX Opons Impled Rsk Neural Denses Oren Tapero 1 Table of Conens Absrac 3 Inroducon 4 I. The Daa 7 1. Opon Selecon Crerons 7. Use of mpled spo raes nsead of quoed spo

More information

Time Series. A thesis. Submitted to the. Edith Cowan University. Perth, Western Australia. David Sheung Chi Fung. In Fulfillment of the Requirements

Time Series. A thesis. Submitted to the. Edith Cowan University. Perth, Western Australia. David Sheung Chi Fung. In Fulfillment of the Requirements Mehods for he Esmaon of Mssng Values n Tme Seres A hess Submed o he Faculy of Communcaons, ealh and Scence Edh Cowan Unversy Perh, Wesern Ausrala By Davd Sheung Ch Fung In Fulfllmen of he Requremens For

More information

A Heuristic Solution Method to a Stochastic Vehicle Routing Problem

A Heuristic Solution Method to a Stochastic Vehicle Routing Problem A Heursc Soluon Mehod o a Sochasc Vehcle Roung Problem Lars M. Hvaum Unversy of Bergen, Bergen, Norway. larsmh@.ub.no Arne Løkkeangen Molde Unversy College, 6411 Molde, Norway. Arne.Lokkeangen@hmolde.no

More information

Market-Clearing Electricity Prices and Energy Uplift

Market-Clearing Electricity Prices and Energy Uplift Marke-Clearng Elecrcy Prces and Energy Uplf Paul R. Grbk, Wllam W. Hogan, and Susan L. Pope December 31, 2007 Elecrcy marke models requre energy prces for balancng, spo and shor-erm forward ransacons.

More information

The Definition and Measurement of Productivity* Mark Rogers

The Definition and Measurement of Productivity* Mark Rogers The Defnon and Measuremen of Producvy* Mark Rogers Melbourne Insue of Appled Economc and Socal Research The Unversy of Melbourne Melbourne Insue Workng Paper No. 9/98 ISSN 1328-4991 ISBN 0 7325 0912 6

More information

Best estimate calculations of saving contracts by closed formulas Application to the ORSA

Best estimate calculations of saving contracts by closed formulas Application to the ORSA Bes esmae calculaons of savng conracs by closed formulas Applcaon o he ORSA - Franços BONNIN (Ala) - Frédérc LANCHE (Unversé Lyon 1, Laboraore SAF) - Marc JUILLARD (Wner & Assocés) 01.5 (verson modfée

More information

Scientific Ontology Construction Based on Interval Valued Fuzzy Theory under Web 2.0

Scientific Ontology Construction Based on Interval Valued Fuzzy Theory under Web 2.0 JOUNAL OF SOFTWAE, VOL. 8, NO. 8, AUGUST 2013 1835 Scenfc Onology Consrucon Based on Inerval Valued Fuzzy Theory under Web 2.0 Na Xue, Sulng Ja, Jnxng Hao and Qang Wang School of Economcs and Managemen,

More information

Finance and Economics Discussion Series Divisions of Research & Statistics and Monetary Affairs Federal Reserve Board, Washington, D.C.

Finance and Economics Discussion Series Divisions of Research & Statistics and Monetary Affairs Federal Reserve Board, Washington, D.C. Fnance and Economcs Dscusson Seres Dvsons of Research & Sascs and Moneary Affars Federal Reserve Board, Washngon, D.C. Prcng Counerpary Rs a he Trade Level and CVA Allocaons Mchael Pyhn and Dan Rosen 200-0

More information

Currency Exchange Rate Forecasting from News Headlines

Currency Exchange Rate Forecasting from News Headlines Currency Exchange Rae Forecasng from News Headlnes Desh Peramunelleke Raymond K. Wong School of Compuer Scence & Engneerng Unversy of New Souh Wales Sydney, NSW 2052, Ausrala deshp@cse.unsw.edu.au wong@cse.unsw.edu.au

More information

Evaluation of the Stochastic Modelling on Options

Evaluation of the Stochastic Modelling on Options Zhjuan Mao, Zhan Lang, Jnguo Lan, Hongkun Zhang / Inernaonal Journal of Engneerng Research and Applcaons (IJERA) ISSN: 48-96 www.jera.com Vol., Issue 3, May-Jun 0, pp.463-473 Evaluaon of he Sochasc Modellng

More information

A Hybrid Method for Forecasting Stock Market Trend Using Soft-Thresholding De-noise Model and SVM

A Hybrid Method for Forecasting Stock Market Trend Using Soft-Thresholding De-noise Model and SVM A Hybrd Mehod for Forecasng Sock Marke Trend Usng Sof-Thresholdng De-nose Model and SVM Xueshen Su, Qnghua Hu, Daren Yu, Zongxa Xe, and Zhongyng Q Harbn Insue of Technology, Harbn 150001, Chna Suxueshen@Gmal.com

More information

Distributed Load Balancing in a Multiple Server System by Shift-Invariant Protocol Sequences

Distributed Load Balancing in a Multiple Server System by Shift-Invariant Protocol Sequences 03 IEEE Wreess Communcaons and Neorkng Conference (WCNC): NETWORS Dsrbued Load Baancng n a Mupe Server Sysem by Shf-Invaran rooco Sequences Yupeng Zhang and Wng Shng Wong Deparmen of Informaon Engneerng

More information

A GENERALIZED FRAMEWORK FOR CREDIT RISK PORTFOLIO MODELS

A GENERALIZED FRAMEWORK FOR CREDIT RISK PORTFOLIO MODELS A GENERALIZED FRAMEWORK FOR CREDIT RISK PORTFOLIO MODELS H. UGUR KOYLUOGLU ANDREW HICKMAN Olver, Wyman & Company CSFP Capal, Inc. * 666 Ffh Avenue Eleven Madson Avenue New Yor, New Yor 10103 New Yor, New

More information

Analyzing Energy Use with Decomposition Methods

Analyzing Energy Use with Decomposition Methods nalyzng nergy Use wh Decomposon Mehods eve HNN nergy Technology Polcy Dvson eve.henen@ea.org nergy Tranng Week Pars 1 h prl 213 OCD/ 213 Dscusson nergy consumpon and energy effcency? How can energy consumpon

More information

Y2K* Stephanie Schmitt-Grohé. Rutgers Uni ersity, 75 Hamilton Street, New Brunswick, New Jersey 08901 E-mail: grohe@econ.rutgers.edu.

Y2K* Stephanie Schmitt-Grohé. Rutgers Uni ersity, 75 Hamilton Street, New Brunswick, New Jersey 08901 E-mail: grohe@econ.rutgers.edu. Revew of Economc Dynamcs 2, 850856 Ž 1999. Arcle ID redy.1999.0065, avalable onlne a hp:www.dealbrary.com on Y2K* Sephane Schm-Grohé Rugers Unersy, 75 Hamlon Sree, New Brunswc, New Jersey 08901 E-mal:

More information

Linear methods for regression and classification with functional data

Linear methods for regression and classification with functional data Lnear mehods for regresson and classfcaon wh funconal daa Glber Sapora Chare de Sasue Appluée & CEDRIC Conservaore Naonal des Ars e Méers 9 rue San Marn, case 44 754 Pars cedex 3, France sapora@cnam.fr

More information

Prot sharing: a stochastic control approach.

Prot sharing: a stochastic control approach. Pro sharng: a sochasc conrol approach. Donaen Hanau Aprl 2, 2009 ESC Rennes. 35065 Rennes, France. Absrac A majory of lfe nsurance conracs encompass a guaraneed neres rae and a parcpaon o earnngs of he

More information

Distribution Channel Strategy and Efficiency Performance of the Life insurance. Industry in Taiwan. Abstract

Distribution Channel Strategy and Efficiency Performance of the Life insurance. Industry in Taiwan. Abstract Dsrbuon Channel Sraegy and Effcency Performance of he Lfe nsurance Indusry n Tawan Absrac Changes n regulaons and laws he pas few decades have afeced Tawan s lfe nsurance ndusry and caused many nsurers

More information

ANALYSIS OF SOURCE LOCATION ALGORITHMS Part I: Overview and non-iterative methods

ANALYSIS OF SOURCE LOCATION ALGORITHMS Part I: Overview and non-iterative methods ANALYSIS OF SOURCE LOCATION ALGORITHMS Par I: Overvew and non-erave mehods MAOCHEN GE Pennsylvana Sae Unversy, Unversy Park PA 1680 Absrac Ths arcle and he accompanyng one dscuss he source locaon heores

More information

A Hybrid AANN-KPCA Approach to Sensor Data Validation

A Hybrid AANN-KPCA Approach to Sensor Data Validation Proceedngs of he 7h WSEAS Inernaonal Conference on Appled Informacs and Communcaons, Ahens, Greece, Augus 4-6, 7 85 A Hybrd AANN-KPCA Approach o Sensor Daa Valdaon REZA SHARIFI, REZA LANGARI Deparmen of

More information

Modèles financiers en temps continu

Modèles financiers en temps continu Modèles fnancers en emps connu Inroducon o dervave prcng by Mone Carlo 204-204 Dervave Prcng by Mone Carlo We consder a conngen clam of maury T e.g. an equy opon on one or several underlyng asses, whn

More information

HAND: Highly Available Dynamic Deployment Infrastructure for Globus Toolkit 4

HAND: Highly Available Dynamic Deployment Infrastructure for Globus Toolkit 4 HAND: Hghly Avalable Dynamc Deploymen Infrasrucure for Globus Toolk 4 L Q 1, Ha Jn 1, Ian Foser,3, Jarek Gawor 1 Huazhong Unversy of Scence and Technology, Wuhan, 430074, Chna quck@chnagrd.edu.cn; hjn@hus.edu.cn

More information

Applying the Theta Model to Short-Term Forecasts in Monthly Time Series

Applying the Theta Model to Short-Term Forecasts in Monthly Time Series Applyng he Thea Model o Shor-Term Forecass n Monhly Tme Seres Glson Adamczuk Olvera *, Marcelo Gonçalves Trenn +, Anselmo Chaves Neo ** * Deparmen of Mechancal Engneerng, Federal Technologcal Unversy of

More information

Optimization Design of the Multi-stage Inventory Management for Supply Chain

Optimization Design of the Multi-stage Inventory Management for Supply Chain Inernaonal Journal of u- an e- Servce, Scence an echnology Vol.8, o. 25, pp.35-366 hp://x.o.org/.4257/uness.25.8..34 Opmzaon Desgn of he ul-sage Invenory anagemen for Supply Chan Hongxng Deng*, ng Yuan

More information

A robust optimisation approach to project scheduling and resource allocation. Elodie Adida* and Pradnya Joshi

A robust optimisation approach to project scheduling and resource allocation. Elodie Adida* and Pradnya Joshi In. J. Servces Operaons and Informacs, Vol. 4, No. 2, 2009 169 A robus opmsaon approach o projec schedulng and resource allocaon Elode Adda* and Pradnya Josh Deparmen of Mechancal and Indusral Engneerng,

More information

Guidelines and Specification for the Construction and Maintenance of the. NASDAQ OMX Credit SEK Indexes

Guidelines and Specification for the Construction and Maintenance of the. NASDAQ OMX Credit SEK Indexes Gudelnes and Specfcaon for he Consrucon and Manenance of he NASDAQ OMX Cred SEK Indexes Verson as of Aprl 7h 2014 Conens Rules for he Consrucon and Manenance of he NASDAQ OMX Cred SEK Index seres... 3

More information

Event Based Project Scheduling Using Optimized Ant Colony Algorithm Vidya Sagar Ponnam #1, Dr.N.Geethanjali #2

Event Based Project Scheduling Using Optimized Ant Colony Algorithm Vidya Sagar Ponnam #1, Dr.N.Geethanjali #2 Inernaonal Journal of Compuer Trends and Technology (IJCTT) Volume 18 Number 6 Dec 2014 Even Based Projec Schedulng Usng Opmzed An Colony Algorhm Vdya Sagar Ponnam #1, Dr.N.Geehanjal #2 1 Research Scholar,

More information

Systematic risk measurement in the global banking stock market with time series analysis and CoVaR

Systematic risk measurement in the global banking stock market with time series analysis and CoVaR Invesmen Managemen and Fnancal Innovaons, Volume 1, Issue 1, 213 Tesuo Kurosak (USA, Young Shn Km (Germany Sysemac rsk measuremen n he global bankng sock marke wh meseres analyss and CoVaR Absrac Movaed

More information

A Real-time Adaptive Traffic Monitoring Approach for Multimedia Content Delivery in Wireless Environment *

A Real-time Adaptive Traffic Monitoring Approach for Multimedia Content Delivery in Wireless Environment * A Real-e Adapve Traffc Monorng Approach for Muleda Conen Delvery n Wreless Envronen * Boonl Adpa and DongSong Zhang Inforaon Syses Deparen Unversy of Maryland, Balore Couny Balore, MD, U.S.A. bdpa1@ubc.edu,

More information

The US Dollar Index Futures Contract

The US Dollar Index Futures Contract The S Dollar Inde uures Conrac I. Inroducon The S Dollar Inde uures Conrac Redfeld (986 and Eyan, Harpaz, and Krull (988 presen descrpons and prcng models for he S dollar nde (SDX fuures conrac. Ths arcle

More information

Load Balancing in Internet Using Adaptive Packet Scheduling and Bursty Traffic Splitting

Load Balancing in Internet Using Adaptive Packet Scheduling and Bursty Traffic Splitting 152 IJCSNS Inernaonal Journal of Compuer Scence and Nework Secury, VOL.8 No.1, Ocober 28 Load Balancng n Inerne Usng Adapve Packe Schedulng and Bursy Traffc Splng M. Azah Research Scholar, Anna Unversy,

More information

Return Persistence, Risk Dynamics and Momentum Exposures of Equity and Bond Mutual Funds

Return Persistence, Risk Dynamics and Momentum Exposures of Equity and Bond Mutual Funds Reurn Perssence, Rsk Dynamcs and Momenum Exposures of Equy and Bond Muual Funds Joop Hu, Marn Marens, and Therry Pos Ths Verson: 22-2-2008 Absrac To analyze perssence n muual fund performance, s common

More information

t φρ ls l ), l = o, w, g,

t φρ ls l ), l = o, w, g, Reservor Smulaon Lecure noe 6 Page 1 of 12 OIL-WATER SIMULATION - IMPES SOLUTION We have prevously lsed he mulphase flow equaons for one-dmensonal, horzonal flow n a layer of consan cross seconal area

More information

606 IEEE TRANSACTIONS ON MOBILE COMPUTING, VOL. 6, NO. 6, JUNE 2007. Impact of Human Mobility on Opportunistic Forwarding Algorithms

606 IEEE TRANSACTIONS ON MOBILE COMPUTING, VOL. 6, NO. 6, JUNE 2007. Impact of Human Mobility on Opportunistic Forwarding Algorithms 66 IEEE TRANSACTIONS ON MOBILE COMPUTING, VOL. 6, NO. 6, JUNE 27 Impac of Human Mobly on Opporunsc Forwardng Algorms Augusn Canreau, Pan Hu, Jon Crowcrof, Fellow, IEEE, Crsope Do, Rcard Gass, and James

More information

International Journal of Mathematical Archive-7(5), 2016, 193-198 Available online through www.ijma.info ISSN 2229 5046

International Journal of Mathematical Archive-7(5), 2016, 193-198 Available online through www.ijma.info ISSN 2229 5046 Inernaonal Journal of Mahemacal rchve-75), 06, 9-98 valable onlne hrough wwwjmanfo ISSN 9 506 NOTE ON FUZZY WEKLY OMPLETELY PRIME - IDELS IN TERNRY SEMIGROUPS U NGI REDDY *, Dr G SHOBHLTH Research scholar,

More information

CONTROLLER PERFORMANCE MONITORING AND DIAGNOSIS. INDUSTRIAL PERSPECTIVE

CONTROLLER PERFORMANCE MONITORING AND DIAGNOSIS. INDUSTRIAL PERSPECTIVE Copyrgh IFAC 5h Trennal World Congress, Barcelona, Span CONTROLLER PERFORMANCE MONITORING AND DIAGNOSIS. INDUSTRIAL PERSPECTIVE Derrck J. Kozub Shell Global Soluons USA Inc. Weshollow Technology Cener,

More information

OUTPUT, OUTCOME, AND QUALITY ADJUSTMENT IN MEASURING HEALTH AND EDUCATION SERVICES

OUTPUT, OUTCOME, AND QUALITY ADJUSTMENT IN MEASURING HEALTH AND EDUCATION SERVICES bs_bs_banner row_504 257..278 Revew of Income and Wealh Seres 58, Number 2, June 2012 DOI: 10.1111/j.1475-4991.2012.00504.x OUTPUT, OUTCOME, AND QUALITY ADJUSTMENT IN MEASURING HEALTH AND EDUCATION SERVICES

More information

Attribution Strategies and Return on Keyword Investment in Paid Search Advertising

Attribution Strategies and Return on Keyword Investment in Paid Search Advertising Arbuon Sraeges and Reurn on Keyword Invesmen n Pad Search Adversng by Hongshuang (Alce) L, P. K. Kannan, Sva Vswanahan and Abhshek Pan * December 15, 2015 * Honshuang (Alce) L s Asssan Professor of Markeng,

More information

THE IMPACT OF QUICK RESPONSE IN INVENTORY-BASED COMPETITION

THE IMPACT OF QUICK RESPONSE IN INVENTORY-BASED COMPETITION Workng Paper WP no 722 November, 2007 THE IMPACT OF QUICK RESPONSE IN INVENTORY-BASED COMPETITION Felpe Caro Vícor Marínez de Albénz 2 Professor, UCLA Anderson School of Managemen 2 Professor, Operaons

More information

Fixed Income Attribution. Remco van Eeuwijk, Managing Director Wilshire Associates Incorporated 15 February 2006

Fixed Income Attribution. Remco van Eeuwijk, Managing Director Wilshire Associates Incorporated 15 February 2006 Fxed Incoe Arbuon eco van Eeuwk Managng Drecor Wlshre Assocaes Incorporaed 5 February 2006 Agenda Inroducon Goal of Perforance Arbuon Invesen Processes and Arbuon Mehodologes Facor-based Perforance Arbuon

More information

MODEL-BASED APPROACH TO CHARACTERIZATION OF DIFFUSION PROCESSES VIA DISTRIBUTED CONTROL OF ACTUATED SENSOR NETWORKS

MODEL-BASED APPROACH TO CHARACTERIZATION OF DIFFUSION PROCESSES VIA DISTRIBUTED CONTROL OF ACTUATED SENSOR NETWORKS MODEL-BASED APPROACH TO CHARACTERIZATION OF DIFFUSION PROCESSES IA DISTRIBUTED CONTROL OF ACTUATED SENSOR NETWORKS Kevn L. Moore and YangQuan Chen Cener for Self-Organzng and Inellgen Sysems Uah Sae Unversy

More information

II. IMPACTS OF WIND POWER ON GRID OPERATIONS

II. IMPACTS OF WIND POWER ON GRID OPERATIONS IEEE Energy2030 Alana, Georga, USA 17-18 November 2008 Couplng Wnd Generaors wh eferrable Loads A. Papavaslou, and S. S. Oren UC Berkeley, eparmen of Indusral Engneerng and Operaons esearch, 4141 Echeverry

More information

Modelling Operational Risk in Financial Institutions using Hybrid Dynamic Bayesian Networks. Authors:

Modelling Operational Risk in Financial Institutions using Hybrid Dynamic Bayesian Networks. Authors: Modellng Operaonal Rsk n Fnancal Insuons usng Hybrd Dynamc Bayesan Neworks Auhors: Professor Marn Nel Deparmen of Compuer Scence, Queen Mary Unversy of London, Mle nd Road, London, 1 4NS, Uned Kngdom Phone:

More information

FRAMEWORK OF MEETING SCHEDULING IN COMPUTER SYSTEMS

FRAMEWORK OF MEETING SCHEDULING IN COMPUTER SYSTEMS FRAMEWORK OF MEEING CEDULING IN COMPUER YEM Goran Marnovc, Faculy of Elecrcal Engneerng, J.J. rossmayer Unversy of Ose, goran.marnovc@efos.hr ABRAC Developmen of compuer echnologes s a necessary bu no

More information

Selected Financial Formulae. Basic Time Value Formulae PV A FV A. FV Ad

Selected Financial Formulae. Basic Time Value Formulae PV A FV A. FV Ad Basc Tme Value e Fuure Value of a Sngle Sum PV( + Presen Value of a Sngle Sum PV ------------------ ( + Solve for for a Sngle Sum ln ------ PV -------------------- ln( + Solve for for a Sngle Sum ------

More information

Comparative Study of Multicast Authentication Schemes with Application to Wide-Area Measurement System

Comparative Study of Multicast Authentication Schemes with Application to Wide-Area Measurement System Comparave Sudy of Mulcas Auhencaon Schemes wh Applcaon o Wde-Area Measuremen Sysem Yee We Law Deparmen of EEE The Unversy of Melbourne Ausrala ywlaw@unmelbeduau Slaven Marusc Deparmen of EEE The Unversy

More information

Omar Shatnawi. Eks p l o a t a c j a i Ni e z a w o d n o s c Ma in t e n a n c e a n d Reliability Vo l.16, No. 4, 2014 585. 1.

Omar Shatnawi. Eks p l o a t a c j a i Ni e z a w o d n o s c Ma in t e n a n c e a n d Reliability Vo l.16, No. 4, 2014 585. 1. Arcle caon nfo: Shanaw O. Measurng commercal sofware operaonal relably: an nerdscplnary modellng approach. Esploaacja Nezawodnosc Manenance and Relably 014; 16 (4): 585 594. Omar Shanaw Measurng commercal

More information