Simulation-Based Lean Six-Sigma And Design For Six-Sigma

Basem El-Haik Raid Al-Aomar Copyright 2006 by John Wiley & Sons, Inc. All rights reserved. P blished by John Wiley & Sons, Inc., !obo"en, #e$ Jersey. P blished si% lt&neo sly in C&n&d&. #o p&rt o' this p blic&tion %&y be reprod ced, stored in & retriev&l syste%, or tr&ns%itted in &ny 'or% or by &ny %e&ns, electronic, %ech&nic&l, photocopying, recording, sc&nning, or other$ise, e(cept &s per%itted nder Section )0* or )0+ o' the ),*6 -nited St&tes Copyright Act, $itho t either the prior $ritten per%ission o' the P blisher, or & thori.&tion thro gh p&y%ent o' the &ppropri&te per/copy 'ee to the Copyright Cle&r&nce Center, Inc., 222 0ose$ood 1rive, 1&nvers, 2A 0),23, 4,*+5 *60/+700, '&( 4,*+5 *60/77*0, or on the $eb &t $$$.copyright.co%. 0e8 ests to the P blisher 'or per%ission sho ld be &ddressed to the Per%issions 1ep&rt%ent, John Wiley & Sons, Inc., ))) 0iver Street, !obo"en, #J 0*030, 420)5 *7+/60)), '&( 420)5 *7+/600+, or online &t http9::$$$.$iley.co%:go:per%ission. ;i%it o' ;i&bility:1iscl&i%er o' W&rr&nty9 While the p blisher &nd & thor h&ve sed their best e''orts in prep&ring this boo", they %&"e no represent&tions or $&rr&nties $ith respect to the &cc r&cy or co%pleteness o' the contents o' this boo" &nd speci'ic&lly discl&i% &ny i%plied $&rr&nties o' %erch&nt&bility or 'itness 'or & p&rtic l&r p rpose. #o $&rr&nty %&y be cre&ted or e(tended by s&les represent&tives or $ritten s&les %&teri&ls. <he &dvice &nd str&tegies cont&ined herein %&y not be s it&ble 'or yo r sit &tion. =o sho ld cons lt $ith & pro'ession&l $here &ppropri&te. #either the p blisher nor & thor sh&ll be li&ble 'or &ny loss o' pro'it or &ny other co%%erci&l d&%&ges, incl ding b t not li%ited to speci&l, incident&l, conse8 enti&l, or other d&%&ges. >or gener&l in'or%&tion on o r other prod cts &nd services or 'or technic&l s pport, ple&se cont&ct o r C sto%er C&re 1ep&rt%ent $ithin the -nited St&tes &t 4+005 *62/2,*7, o tside the -nited St&tes &t 43)*5 6*2/3,,3 or '&( 43)*5 6*2/7002. Wiley &lso p blishes its boo"s in & v&riety o' electronic 'or%&ts. So%e content th&t &ppe&rs in print %&y not be &v&il&ble in electronic 'or%&ts. >or %ore in'or%&tion &bo t Wiley prod cts, visit o r $eb site &t $$$.$iley.co%. Library of Congress Cataloging-in-Publication Data: ?l/!&i", @&se%. Si% l&tion/b&sed le&n si(/sig%& &nd design 'or si(/sig%& : by @&se% ?l/!&i", 0&id Al/Ao%&r. p. c%. A Wiley/Interscience p blic&tion.A Incl des bibliogr&phic&l re'erences &nd inde(. IS@#/)39 ,*+/0/7*)/6,7,0/+ IS@#/)09 0/7*)/6,7,0/+ ). Si( sig%& 4B &lity control st&nd&rd5 2. <ot&l 8 &lity %&n&ge%ent. 3. Prod ction engineering. I. Al/Ao%&r, 0&id. II. <itle. <S)66.?3+3 2006 62Cdc2266+.6 200607+))2

)0 , + * 6 6 7 3 2 ) To our parents, families, and friends for their continuous support Basem and Raid Fall 2006

Summary Of Cha ter Contents

<he boo" is org&ni.ed into three p&rts9 P&rt I, Ch&pters ), 2, &nd 3, h&s si(/sig%& ' nd&%ent&ls &s its the%e. P&rt II, Ch&pters 7, 6, 6, &nd *, covers si% l&tion ' nd&%ent&ls. P&rt III, begins in Ch&pters +, ,, &nd )0 by 'oc sing on si% l&tion/b&sed le&n si(/sig%& 43S/;SS5 &nd design 'or si(/sig%& 43S/1>SS5 &nd provides sever&l c&se st dies. In Ch&pter )) $e present ele%ents o' s ccess' l deploy%ent o' 3S %ethods. >ollo$ing is & s %%&ry o' the ch&pters.

!art "
In Ch&pter ) $e introd ce the concepts o' process &nd customer satisfaction &nd highlight ho$ c sto%ers e(perience & prod ct or & service &s $ell &s its delivery process. <his ch&pter concentr&tes on the service side &nd de'ines v&rio s ele%ents o' & generic process &s $ell &s tools s ch &s process %&pping. We &lso de%ysti'y the concept o' transaction &nd event discreteness. In this ch&pter $e e(pl&in $h&t si(/sig%& is &nd ho$ it h&s evolved over ti%e. We e(pl&in th&t it is & process/b&sed %ethodology &nd introd ce the re&der to process %odeling $ith & high/level overvie$ o' process %&pping, v&l e stre&% %&pping &nd v&l e &n&lysis &s $ell &s b siness process %&n&ge%ent syste%s. <he critic&lity &nd &pplic&tion o' %e&s re%ent syste%s &n&lysis is introd ced. <he 12AIC %ethodology &nd ho$ it incorpor&tes these concepts into & ro&d %&p %ethod is &lso e(pl&ined. Ch&pter 2 covers the le&n si( sig%& 4;SS5 concept &nd disc sses topics rel&ted to the integr&tion o' si(/sig%& &nd le&n %&n '&ct ring. <he 'oc s in this ch&pter is on the det&ils o' the ;SS &ppro&ch, the enh&nce%ents %&de to si(/sig%& 12AIC tollg&tes, le&n %&n '&ct ring concepts &nd &i%s, v&l e stre&% %&pping, &nd le&n %&n '&ct ring techni8 es. <he ch&pter &lso highlights the synergy &nd bene'its o' i%ple%enting &n ;SS syste% &s & 'o nd&tion 'or the proposed 3S/ ;SS &ppro&ch. Ch&pter 3 is &n introd ction to & high/level 1>SS process. <he 1>SS &ppro&ch introd ced helps design te&%s 'r&%e their proDect $ith 'in&nci&l, c lt r&l, &nd str&tegic i%plic&tions to the b siness. In this ch&pter $e 'or% &nd integr&te sever&l str&tegic, t&ctic&l, &nd synergistic %ethodologies to enh&nce service 1>SS c&p&bilities &nd to deliver & bro&d set o' opti%i.ed sol tions. We highlights &nd present the service 1>SS ph&ses9 identi'y, ch&r&cteri.e, opti%i.e, &nd veri'y.

!art ""
In Ch&pter 7 $e introd ce the b&sic concepts o' si% l&tion %odeling $ith & 'oc s on process %odeling &nd ti%e/b&sed per'or%&nce %e&s re%ent.We &lso cl&ri'y the role o' si% l&tion st dies in serving the incre&sing needs o' co%p&nies th&t see" contin o s i%prove%ent &nd opti%&lity in prod ction &nd b siness processes. <o this end, $e provide &n introd ction to the concept, ter%inology, &nd types o' %odels, &long $ith & description o' si% l&tion t&(ono%y &nd & D sti'ic&tion 'or tili.ing si% l&tion tools in & v&riety o' re&l/$orld &pplic&tions. S ch & b&c"gro nd is essenti&l to est&blishing & b&sic nderst&nding o' $h&t si% l&tion is &ll &bo t &nd to nderst&nding the "ey si% l&tion role in si% l&tion/b&sed si(/sig%& st dies &nd &pplic&tions. In Ch&pter 6 $e present the det&ils &nd %ech&nics o' 1?S th&t &re essenti&l to providing & 'le(ible pl&t'or% 'or se o' & si% l&tion/b&sed si(/sig%& &ppro&ch9 ele%ents o' syste% %odeling, events &ctiv&tion, r&ndo% n %ber gener&tion, ti%e &dv&nce%ent, &ni%&tion, &nd &cc % l&ting st&tistics. In &ddition to po$er' l 1?S %ech&nics, $e &ddress ho$ '&st co%p t&tions on tod&yEs high/speed processors, &long $ith the gro$ing gr&phics c&p&bility, contrib te to the e''ectiveness &nd vis &li.&tion c&p&bility o' 1?S %odels. We th s provide & deeper nderst&nding o' the 1?S

process, co%ponents, &nd %ech&nis%s. ?(&%ples o' %&n &l &nd co%p ter si% l&tions &re sed to cl&ri'y the 1?S ' nction&lity. Ch&pter 6 is 'oc sed on &n&ly.ing the v&rio s &spects o' the si% l&tion process &nd the set o' techni8 es &nd steps 'ollo$ed $hen cond cting & si% l&tion st dy. <he ele%ents o' the si% l&tion process disc ssed incl de proDect scoping, concept &l %odeling, d&t& collection, %odel b ilding, %odel &n&lyses, &nd doc %ent&tion. <he ch&pter &lso highlights the lin"&ge o' these si% l&tion pr&ctices to the over&ll scope o' & si(/sig%& proDect, to the si% l&tion so't$&re sed, &nd to the process 'ollo$ed by & si(/sig%& te&% in c&rrying o t & co%plete si% l&tion st dy. >in&lly, $e present the i%plic&tions o' si% l&tion on si(/sig%& &pplic&tions to syste% design, proble% solving, &nd contin o s i%prove%ent. In Ch&pter * $e disc ss the %&in iss es involved in &n&ly.ing si% l&tion o tp ts9 disting ishing bet$een ter%in&ting &nd ste&dy/st&te si% l&tion, nderst&nding the stoch&stic n&t re o' si% l&tion o tco%es, deter%ining si% l&tion r n controls 4i.e., $&r%/ p period, r n length, &nd n %ber o' replic&tions5, &nd selecting the &ppropri&te %odel o tp t &n&lysis %ethod. We &lso disc ss the %&in %ethods o' o tp t &n&lysis, incl ding st&tistic&l &n&lysis, e(peri%ent&l design, &nd opti%i.&tion.

!art """
Ch&pter + cont&ins &ll the proDect ro&d %&ps 'or the 3S &nd 3S/;SS %ethodologies. <he 12AIC process &nd le&n principles &nd concepts &re integr&ted into & synergetic ro&d %&p in & 3S/;SS environ%ent. In Ch&pter + $e &lso present & 3S/1>SS proDect ro&d %&p, $hich highlights &t & high level the identi'y, ch&rc&teri.e, opti%i.e, &nd v&lid&te ph&ses over the seven develop%ent st&ges9 ide& cre&tion, voice o' the c sto%er &nd b siness, concept develop%ent, preli%in&ry design, design opti%i.&tion, veri'ic&tion, &nd l& nch re&diness. In this ch&pter the concept o' tollg&te is introd ced. We &lso highlight the %ost &ppropri&te 1>SS tools &nd %ethods by the 1>SS ph&se, indic&ting $here it is %ost &ppropri&te to st&rt tool se. <he ro&d %&ps &lso recogni.e the concept o' tollgates, design %ilestones $here 1>SS te&%s pd&te st&"eholders on develop%ent &nd &s" th&t & decision be %&de &s to $hether to &pprove going to the ne(t st&ge, recycling b&c" to &n e&rlier st&ge, or c&nceling the proDect &ltogether. In Ch&pter , $e disc ss the det&ils o' 3S/;SS &pplic&tion to re&l/$orld syste%s sing the ro&d %&p presented in Ch&pter +. <he 'oc s is on the pr&ctic&l &spects o' 3S/;SS se in reengineering tr&ns&ction&l processes in both services &nd %&n '&ct ring. <hese &spects &re disc ssed thro gh &n &pplic&tion o' the 3S/;SS &ppro&ch to & %&n '&ct ring c&se st dy. ?%ph&sis is pl&ced on tili.ing si% l&tion/b&sed &pplic&tion o' le&n techni8 es &nd si(/sig%& &n&lysis to i%prove & set o' process C<Bs, de'ined in ter%s o' ti%e/b&sed per'or%&nce to t&"e &dv&nt&ge o' si% l&tion %odeling 'or & syste%/level &pplic&tion o' si(/sig%& st&tistic&l tools sed in 12AIC &nd le&n techni8 es. In Ch&pter )0 $e develop & 3S/1>SS clinic c&se st dy sing the ro&d %&p in Ch&pter +. We sho$ the &pplic&tion o' sever&l tools &s they sp&n over the proDect ro&d %&p. <he c&se st dy &lso highlights %ost &ppropri&te 1>SS tools. It indic&tes $here it is %ost &ppropri&te to st&rt tool se s ch &s tr&ns'er ' nctions &nd 8 &lity ' nction deploy%ent. >ollo$ing the 1>SS ro&d %&p helps &cceler&te the process introd ction &nd &ligns bene'its 'or c sto%ers &nd st&"eholders.

Chapter 11 is a practical guide to successful 3S-DFSS and 3S-LSS development, deployment, and project execution ! generic frame"or# for successful project management and development is proposed $ased on the authors% experience &his frame"or# demystifies the am$iguity involved in project selection and specifies a method for project development 'e provide a practical guide for successful

development of 3S projects $ased on the road maps discussed in Chapter ( &he guide $egins $y discussing the uni)ue characteristics of 3S projects

!art ""# Simulation Fundamentals

Cha ter List
Ch&pter 79 @&sic Si% l&tion Concepts Ch&pter 69 1iscrete ?vent Si% l&tion Ch&pter 69 <he Si% l&tion Process Ch&pter *9 Si% l&tion An&lysis

Ch&pter 79 Basi$ Simulation Con$e ts

%&' "ntrodu$tion
Prod ction &nd b siness syste%s &re "ey b ilding bloc"s in the str ct re o' %odern ind stri&l societies. Co%p&nies &nd ind stri&l 'ir%s, thro gh $hich prod ction &nd b siness oper&tions &re s &lly per'or%ed, represent the %&Dor sector o' tod&yEs glob&l econo%y. <here'ore, in the l&st dec&de, co%p&nies h&ve %&de contin o s i%prove%ent in their prod ction &nd b siness syste%s & %ilestone in their str&tegic pl&nning 'or the ne$ %illenni %. It is s &lly &sserted th&t prod ction &nd b siness oper&tions h&ve the potenti&l to strengthen or $e&"en & co%p&nyEs co%petitive &bility. <o re%&in co%petitive, co%p&nies h&ve to %&int&in & high level o' per'or%&nce by %&int&ining high 8 &lity, lo$ cost, short %&n '&ct ring le&d ti%es, &nd & high level o' c sto%er s&tis'&ction. As & res lt o' 'ierce co%petition &nd decre&sing b siness s&'ety %&rgins, e''icient &nd rob st prod ction &nd b siness oper&tions h&ve beco%e & necessity 'or s rviv&l in the %&r"etpl&ce. 2&ny ind stri&l engineering s bDects, s ch &s oper&tions rese&rch, 8 &lity control, &nd si% l&tion, o''er rob st &nd e''icient design &nd proble%/solving tools $ith the lti%&te &i% o' per'or%&nce enh&nce%ent. ?(&%ples o' this per'or%&nce incl de the thro ghp t o' & '&ctory, the 8 &lity o' & prod ct, or the pro'it o' &n org&ni.&tion. Si% l&tion %odeling &s &n ind stri&l engineering tool 'or syste% design &nd i%prove%ent h&s ndergone tre%endo s develop%ent in the p&st dec&de. <his c&n be pict red thro gh the gro$ing c&p&bilities o' si% l&tion so't$&re tools &nd the &pplic&tion o' si% l&tion sol tions to & v&riety o' re&l/$orld proble%s in di''erent b siness &ren&s. With the &id o' si% l&tion, co%p&nies h&ve been &ble to design e''icient prod ction &nd b siness syste%s, v&lid&te &nd tr&de o'' proposed design sol tion &ltern&tives, tro bleshoot potenti&l proble%s, i%prove syste%s per'or%&nce %etrics, &nd conse8 ently, c t cost, %eet t&rgets, &nd boost s&les &nd pro'its. In the l&st dec&de, si% l&tion %odeling h&s been pl&ying & %&Dor role in designing, &n&ly.ing, &nd opti%i.ing engineering syste%s in & $ide r&nge o' ind stri&l &nd b siness &pplic&tions. 0ecently, si(/sig%& pr&ctitioners st&rted to recogni.e the essenti&l si% l&tion role in syste% design, proble% solving, &nd contin o s i%prove%ent. Syste%/level si% l&tion in p&rtic l&r h&s beco%e & "ey si(/sig%& tool 'or representing &nd %e&s ring the ti%e/b&sed per'or%&nce o' re&l/ $orld stoch&stic prod ction &nd b siness syste%s. ?(&%ples incl de si(/sig%& st dies th&t &re 'oc sed on enh&ncing the prod ctivity, 8 &lity, inventory, 'lo$, e''iciency, &nd le&d ti%e in

prod ction &nd b siness syste%s. In these st dies, si% l&tion is p&rtic l&rly essenti&l 'or syste% represent&tion, per'or%&nce ev&l &tion, e(peri%ent&l design, $h&t/i' &n&lysis, &nd opti%i.&tion. With s ch c&p&bility, si% l&tion %odeling c&n be tili.ed &s &n i%port&nt tool in si(/sig%& &pplic&tions 'or both 1>SS &nd 12AIC. @e'ore presenting the det&ils o' si% l&tion tili.&tion in the 3S/;SS &nd 3S/1>SS &ppro&ches, $e provide &n introd ctory to the b&sic concepts o' si% l&tion %odeling, $ith & 'oc s on process %odeling &nd ti%e/b&sed per'or%&nce %e&s re%ent.We &lso cl&ri'y the role o' si% l&tion st dies in serving the incre&sing needs o' co%p&nies th&t see" contin o s i%prove%ent &nd opti%&lity to prod ction &nd b siness processes. <o this end, $e provide &n introd ction to the concept, ter%inology, &nd types o' %odels, together $ith & description o' si% l&tion t&(ono%y &nd & D sti'ic&tion 'or tili.ing si% l&tion tools in & v&riety o' re&l/$orld &pplic&tions. S ch & b&c"gro nd is essenti&l to est&blishing & b&sic nderst&nding o' $h&t si% l&tion is &ll &bo t &nd to nderst&nding the "ey si% l&tion role in si% l&tion/b&sed si(/sig%& st dies &nd &pplic&tions.

%&( System )odeling

System modeling &s & ter% incl des t$o i%port&nt co%%only sed concepts9 system &nd modeling. It is i%per&tive to cl&ri'y s ch concepts be'ore &tte%pting to 'oc s on their relev&nce to the Fsi% l&tionA topic. !ence, in this section $e introd ce these t$o concepts &nd provide & generic cl&ssi'ic&tion o' the v&rio s types o' syste% %odels.

%&(&' System Con$e t

<he $ord system is co%%only sed in its bro&d %e&ning in & v&riety o' engineering &nd nonengineering 'ields. In si%ple $ords, & syste% is o'ten re'erred to &s & set o' ele%ents or oper&tions th&t &re org&ni.ed &nd logic&lly rel&ted to$&rd the &tt&in%ent o' & cert&in go&l or obDective. <o &tt&in the intended go&l or to serve the desired ' nction, it is necess&ry 'or the syste% to receive & set o' inp ts, to process the% correctly, &nd to prod ce the re8 ired o tco%es. <o s st&in s ch 'lo$, so%e control is re8 ired to govern syste% beh&vior. Given s ch & de'inition, $e c&n &n&ly.e &ny syste% 4S5 b&sed on the &rchitect re sho$n in >ig re 7.).

Figure %&'# 1e'inition o' the syste% concept. As sho$n in the 'ig re, e&ch syste% c&n be de'ined pri%&rily in ter%s o' & set o' inp ts 4 I 5 &nd & process 4P5 th&t tr&ns'or%s &nd ch&nges the ch&r&cteristics o' syste% inp ts to prod ce & speci'ic set o' o tp ts 4O5. <he process is per'or%ed thro gh & sets o' syste% ele%ents 4 5 governed by & set o' rel&tionships 4R5. An over&ll go&l 4!5 is o'ten de'ined to represent the p rpose &nd obDective o' the syste%. <o s st&in & 'l&$less 'lo$ &nd correct ' nction&lity o' I"P" O, so%e type o' control 4#5 is essenti&lly &pplied to syste% inp ts, process, &nd o tp ts. <h s, b ilding & syste% re8 ires pri%&rily si( b&sic ele%ents9 ). Speci'ying the set o' inp ts 4I 5 &nd the speci'ic&tions re8 ired to prod ce speci'ied o tco%es 4O5

2. 3. 7. 6. 6.

;isting syste% ele%ents S = 4 ), 2, 3, . . . , n5 &nd de'ining the ch&r&cteristics &nd the individ &l role o' e&ch ele%ent An&ly.ing &nd nderst&nding the logic&l rel&tionships 4 R), R2, R3, $ $ $ , Rm5 &%ong the set o' syste% ele%ents de'ined Speci'ying the set o' o tco%es 4O5 th&t sho ld be prod ced &nd their speci'ic&tions in order to re&ch the go&l 4!5 &nd obDective speci'ied Speci'ying the syste% controls 4#5 &nd their r les in %onitoring &nd &dD sting I"P"O 'lo$ to %eet inp ts 4I 5, process 4P5, &nd o tp t 4O5 speci'ic&tions 1e'ining the go&l 4!5 or the over&ll syste% obDective &nd rel&ting syste% str ct re 4inp ts, ele%ents, rel&tionships, controls, &nd o tp ts5 to go&l &tt&in%ent

@&sed on this nderst&nding o' the syste% concept, it beco%es essenti&l th&t &ny &rr&nge%ent o' things or obDects to be c&lled & syste% h&ve & de'ined set o' ele%ents. S ch ele%ents sho ld h&ve so%e logic&l rel&tionships &%ong the%, &nd there sho ld be so%e type o' go&l, obDective, or se' l o tco%e 'ro% the inter&ction o' the syste% ele%ents de'ined. <r&ns'or%ing syste% inp ts into desired o tp ts is o'ten per'or%ed thro gh the syste% reso rces. Correct processing is o'ten s pported by controls &nd inventory syste%s to &ss re 8 &lity &nd %&int&in ste&dy per'or%&nce. -nderst&nding the syste% concept is o r g&te$&y to the l&rgest &nd $idest s bDect o' syste% &n&lysis &nd design, $hich represents & "ey s"ill in si% l&tion %odeling. Co%%on e(&%ples o' syste% concept incl de & cl&ssroo%, & co%p ter syste%, &nd & pl&nt. We no$ &pply the de'inition o' syste% to & cl&ssroo%. <his le&ds to the 'ollo$ing &n&lyses9 ). St dents o' cert&in 8 &li'ic&tions o' &ge, &c&de%ic level, %&Dor, &nd so on, &re speci'ied &s the set o' inp ts 4I 5. -pon gr&d &tion, st dents &re &lso de'ined &s syste% o tco%e 4O5. 2. <he over&ll go&l 4!5 is set to provide & good/8 &lity ed c&tion to st dents &ttending cl&sses. 3. <o &chieve !, st dents &re s bDect to v&rio s ele%ents o' the ed c&tion&l process 4 P5 in the cl&ssroo%, $hich involves &ttend&nce, p&rticip&tion in cl&ss &ctivities, s b%itting &ssign%ents, p&ssing e(&%s, &nd so on, in order to co%plete the cl&ss $ith cert&in 8 &li'ic&tions &nd s"ills. 7. <he set o' syste% ele%ents is de'ined &s 'ollo$s9 S = Ht&bles, ch&irs, st dents, instr ctor, boo"s, $hitebo&rdI <he ele%ents de'ined in S &re logic&lly rel&ted thro gh & set o' rel&tionships 4R5. >or e(&%ple, ch&irs &re loc&ted &ro nd the t&bles, t&bles &re sit to '&ce the instr ctor, st dents sit on ch&irs, boo"s &re p t on t&bles in 'ront o' st dents, the instr ctor st&nds in 'ront o' st dents &nd ses the $hitebo&rd to e(pl&in the concepts, &nd so on. 6. Inter&ction o' ele%ents &s e(pl&ined in ite% 6 le&ds to the &tt&in%ent o' the go&l 4 !5 o' le&rning or ed c&ting st dents on & cert&in s bDect. *. >in&lly, the instr ctor &pplies reg l&tions &nd r les o' cond ct in the cl&ssroo%, &ttend&nce &nd gr&ding policies, &nd so on, &s & sort o' control 4 #5 on the ed c&tion process. 6. >in&lly, it is $orth %entioning th&t the ter% system covers prod cts &nd processes. A prod ct syste% co ld be &n & to%obile, & cell l&r phone, & co%p ter, or & c&lc l&tor. Any o' these prod cts involves the de'ined co%ponents o' the syste% in ter%s o' inp ts, o tp ts, ele%ents, rel&tionships, controls, &nd go&l. <ry to &n&ly.e 'ro% this perspective &ll the e(&%ples described. Jn the other h&nd, & process syste% c&n be & %&n '&ct ring process, &n &sse%bly line, & po$er pl&nt, & b siness process s ch &s b&n"ing oper&tions, & logistic syste%, or &n ed c&tion&l syste%. Si%il&rly, &ny o' these processes involves the syste% co%ponents de'ined in ter%s o' inp ts, o tp ts, ele%ents, rel&tionships, controls, &nd go&l. <ry to &n&ly.e &ll the e(&%ples 'ro% this perspective, too.

%&(&( )odeling Con$e t

<he $ord modeling re'ers to the process o' representing & syste% 4& prod ct or & process5 $ith & %odel th&t is e&sier to nderst&nd th&n the &ct &l %odel &nd less e(pensive to b ild. <he syste% represent&tion in the %odel i%plies t&"ing into &cco nt the co%ponents o' the syste%, &s disc ssed in Section 7.2.). <his incl des representing syste% ele%ents, rel&tionships, go&l, inp ts, controls, &nd o tp ts. 2odeling & syste% there'ore h&s t$o prere8 isites9 ). -nderst&nding the str ct re o' the &ct &l 4re&l/$orld5 syste% &nd the ' nction&lity &nd ch&r&cteristics o' e&ch syste% co%ponent &nd rel&tionship. It is i%per&tive to be '&%ili&r $ith & syste% be'ore &tte%pting to %odel it &nd to nderst&nd its p rpose &nd ' nction&lity be'ore &tte%pting to est&blish & se' l represent&tion o' its beh&vior. >or e(&%ple, the %odeler needs to be '&%ili&r $ith the prod ction syste% o' b ilding vehicles be'ore &tte%pting to %odel & vehicle body shop or & vehicle &sse%bly oper&tion. Si%il&rly, the %odeler needs to be '&%ili&r $ith v&rio s types o' b&n" tr&ns&ctions be'ore &tte%pting to %odel b&n"ing oper&tions. 2. @eing '&%ili&r $ith di''erent %odeling &nd syste% represent&tion techni8 es &nd %ethods. <his s"ill is essenti&l to choose the &ppropri&te %odeling techni8 e 'or representing the nderlying re&l/$orld syste%. 1 e to b dget&ry &nd ti%e constr&ints, the %odel selected sho ld be pr&ctic&lly &nd econo%ic&lly 'e&sible &s $ell &s bene'ici&l in %eeting the lti%&te go&l o' %odeling. As $e disc ss in Section 7.2.3, sever&l %odel types c&n be sed to cre&te & syste% %odel, &nd selection o' the %ost 'e&sible %odeling %ethod is & decision b&sed on econo%y, &tt&in&bility, &nd se' lness. <he "ey 8 estion to be &ns$ered by the %odeler is ho$ to %odel & syste% o' interest. <he &ns$er is & co%bin&tion o' &rt &nd science th&t is sed 'or &bstr&cting & re&l/$orld syste% into &n ed c&tion&l %odel. <he %odel sho ld be cle&r, co%prehensive, &nd &cc r&te so th&t $e c&n rely on its represent&tion in nderst&nding syste% ' nction&lity, &n&ly.ing its v&rio s post res, &nd predicting its ' t re beh&vior. >ro% this perspective, &nd &s sho$n in >ig re 7.2, system modeling is the process o' tr&ns'erring the &ct &l syste% into & %odel th&t c&n be sed to repl&ce the &ct &l one 'or syste% &n&lysis &nd syste% i%prove%ent. <he obDects o' & re&l/$orld syste% &re repl&ced by obDects o' represent&tion &nd sy%bols o' indic&tion, incl ding the set o' syste% ele%ents 4 5, ele%ent rel&tionships 4R5, syste% inp ts 4I 5, controls 4#5, &nd syste% o tp ts 4O5. <he &ct &l syste% ?C0CICCCJ is %i%ic"ed thoro ghly in the syste% %odel, le&ding to & represent&tion th&t c&pt res the ch&r&cteristics o' the re&l/$orld process. J' co rse, $e sho ld &l$&ys "eep in %ind th&t the syste% %odel is D st &n &ppro(i%&tion o' the &ct &l syste%. !o$ever, this &ppro(i%&tion sho ld not overloo" the "ey syste% ch&r&cteristics.

Figure %&(# Process o' syste% %odeling. <he science o' syste% %odeling ste%s 'ro% le&rning %odeling %ethodologies &nd h&ving 'l ency in &n&lytic&l s"ills &nd logic&l thin"ing. <he &rt &spect o' %odeling involves represent&tion, gr&phic&l &nd &bstr&ction s"ills o' obDects, rel&tionships, &nd str ct re. It is si%ply being &ble to i%&gine syste% ' nction&lity, oper&tion, &nd ele%ent inter&ction &nd possessing the c&p&bility o' %&pping rel&tionships &nd representing beh&vior. >or e(&%ple, t&"e & vie$ o' &n &sse%bly line,

nderst&nd its ' nction&lity &nd str ct re, &nd try to represent it in & process di&gr&%. 1oes the di&gr&% e(pl&in everything yo need 'or the s&"e o' &n&ly.ing the &sse%bly lineK Wh&t else sho ld be 8 &nti'ied &nd %e&s redK

%&(&* +y es of )odels
As %entioned in Section 7.2.2, sever&l %odeling %ethods c&n be sed to develop & syste% %odel. <he &n&lyst decides on the %odeling &ppro&ch by choosing & cert&in type o' %odel to represent the &ct &l syste%. <he &n&lyst choice $ill be b&sed on sever&l criteri&, s ch &s syste% co%ple(ity, the obDectives o' syste% %odeling, &nd the cost o' %odeling. As sho$n in >ig re 7.3, $e c&n cl&ssi'y the v&rio s types o' %odels into 'o r %&Dor c&tegories9 physic&l %odels, gr&phic&l %odels, %&the%&tic&l %odels, &nd co%p ter %odels. >ollo$ing is & s %%&ry o' these types o' %odels.

Figure %&*# <ypes o' syste% %odels. Physical Models Physic&l %odels &re t&ngible prototypes o' &ct &l prod cts or processes. Prototypes c&n se & ) 9 ) sc&le or &ny other 'e&sible sc&le o' choice. S ch %odels provide & close/to/re&lity direct represent&tion o' the &ct &l syste% &nd c&n be sed to de%onstr&te the syste%Es str ct re, the role o' e&ch syste% ele%ent, &nd the &ct &l ' nction&lity o' the syste% o' interest in & physic&l %&nner. <hey help designers &chieve & deeper nderst&nding o' syste% str ct re &nd det&ils &nd to try o t v&rio s con'ig r&tions o' design ele%ents be'ore the prod ct is &ct &lly b ilt &nd the process deployed. <hey &re sed pri%&rily in engineering o' l&rge/sc&le proDects to e(&%ine & li%ited set o' beh&vior&l ch&r&cteristics in the syste%. >or e(&%ple, & stre&% &nd d&% %odel is sed to si% l&te the r&te o' o t'lo$ 'ro% the d&% nder & v&riety o' circ %st&nces. Prototypes b ilt 'or vehicles in ne$ c&r progr&%s or 'or bridges, b ildings, &nd other &rchitect r&l designs &re &ddition&l e(&%ples o' physic&l %odels. Physic&l %odels s ch &s c&r prototypes &re b ilt 'ro% cl&y or $ood. <hree/di%ension&l prototype printing %&chines &re &lso &v&il&ble to develop prototypes sing di''erent %&teri&ls. L&rio s techni8 es o' r&pid prototyping &nd reverse engineering &re &lso sed to develop prod ct &nd process prototypes. >ig re 7.7 sho$s e(&%ples o' prod ct prototypes. Physic&l %odels c&n &lso be oper&tion&l %odels, s ch &s 'light si% l&tors &nd re&l/ti%e si% l&tors o' che%ic&l oper&tions. Another 'or% o' physic&l %odel c&n be 'o nd in ;ego/type %&chine &nd conveyor str ct res, &nd pl&nts or re&ctor %odels. <he bene'it o' physic&l %odels is the direct &nd e&sy/to/ nderst&nd t&ngible represent&tion o' the nderlying syste%. !o$ever, there &re sever&l li%it&tions to physic&l %odels. <hese li%it&tions incl de the cost o' physic&l %odeling, $hich co ld be enor%o s in so%e c&ses. #ot &ll syste%s c&n be %odeled $ith physic&l prototypes. So%e syste%s &re too co%ple( to be prototyped. Jther physic&l %odels %ight be ti%e cons %ing &nd re8 ire s perior cr&'ting &nd develop%ent s"ills to be b ilt. >or e(&%ple, thin" o' b ilding & physic&l %odel 'or & c&rEs intern&l co%b stion engine or &n &sse%bly line o' person&l co%p ters. Wh&t types o' cost, ti%e, co%ple(ity, &nd s"ill $o ld be involved in %odeling s ch syste%sK

Figure %&%# ?(&%ples o' prod ct prototypes. 4>ro% http9::$$$.centerpointdr&'ting.co%.5 Graphical Models Gr&phic&l %odels &re &bstr&ctions o' &ct &l prod cts or processes sing gr&phic&l tools. <his type o' %odeling st&rts $ith p&per &nd pencil s"etches, progresses to engineering dr&$ings, &nd ends $ith pict res &nd %ovies o' the syste% o' interest. Co%%on syste% gr&phic&l represent&tions incl de & syste% l&yo t, 'lo$ di&gr&%s, bloc" di&gr&%s, net$or" di&gr&%s, process %&ps, &nd oper&tions ch&rts. Since gr&phic&l represent&tions &re st&tic %odels, three/di%ension&l &ni%&tions &nd clip videos &re o'ten sed to ill str&te the oper&tion o' & dyn&%ic syste% or the &sse%bly process o' & prod ct. <his &dds to the di''ic lty o' &bstr&cting & co%ple( syste% gr&phic&lly, especi&lly $hen the nderlying prod ct &nd process &re still in the design ph&se. S ch gr&phic&l tools o'ten oversi%pli'y the re&lity o' the syste% in bloc"s &nd &rro$s &nd do not provide technic&l &nd ' nction&lity det&ils o' the process. <he &bsence o' dyn&%ic ' nction&lity in gr&phic&l represent&tions %&"es it di''ic lt to try o t $h&t/i' scen&rios &nd to e(pl&in ho$ the syste% responds to v&rio s ch&nges in %odel p&r&%eters &nd oper&ting conditions. S ch li%it&tions h&ve %&de %ost gr&phic&l tools p&rt o' or & prere8 isite to other %odeling techni8 es, s ch &s physic&l, %&the%&tic&l, &nd co%p ter %odels. >ig re 7.6 presents &n e(&%ple o' & gr&phic&l %odel 4the oper&tions ch&rt 'or & c&n opener5.

Figure %&,# Process gr&phic&l %odel. Mathematical Models 2&the%&tic&l %odeling is the process o' representing syste% beh&vior $ith 'or% l&s or %&the%&tic&l e8 &tions. S ch %odels &re sy%bolic represent&tions o' syste%s ' nction&lity, decision 4control5 v&ri&bles, response, &nd constr&ints. <hey &ssist in 'or% l&ting the syste% design proble% in & 'or% th&t is solv&ble sing gr&phic&l &nd c&lc l s/b&sed %ethods. 2&the%&tic&l %odels se %&the%&tic&l e8 &tions, prob&bilistic %odels, &nd st&tistic&l %ethods to represent the ' nd&%ent&l rel&tionships &%ong syste% co%ponents. <he e8 &tions c&n be derived in & n %ber o' $&ys. 2&ny o' the% co%e 'ro% e(tensive scienti'ic st dies th&t

h&ve 'or% l&ted & %&the%&tic&l rel&tionship &nd then tested it &g&inst re&l d&t&. 1esign 'or% l&s 'or stressCstr&in &n&lyses &nd %&the%&tic&l progr&%%ing %odels s ch &s line&r &nd go&l progr&%%ing &re e(&%ples o' %&the%&tic&l %odels. >ig re 7.6 sho$s & %&the%&tic&l %odel b ilt sing the 2A<;A@ so't$&re tool. So%e %&the%&tic&l %odels &re b&sed on e%piric&l %odels s ch &s regression %odels &nd tr&ns'er ' nctions. 2&the%&tic&l %odels &re sed e(tensively no$&d&ys to &n&ly.e $e&ther events, e&rth8 &"e ris"s, &nd pop l&tion dyn&%ics. <ypic&lly, & %&the%&tic&l 'or% l& is & closed/'or% rel&tionship bet$een & dependent v&ri&ble 4 %5 &nd one or %ore independent v&ri&bles 4& 5 in the 'or% % = f4& 5. S ch & 'or% l& c&n be line&r or nonline&r $ith & cert&in p&r&bolic order &s &n &ppro(i%&tion. <he dependent v&ri&ble is o'ten selected to %e&s re & "ey ch&r&cteristic o' the syste%, s ch &s the speed o' & vehicle, the yield o' & process, the pro'it o' &n org&ni.&tion, the de%&nd o' & prod ct, &nd so on. Independent v&ri&bles o' the 'or% l& represent the "ey or critic&l p&r&%eters on $hich syste% response 4the "ey ch&r&cteristic5 depends. ?(&%ples incl de the n %ber o' oper&tors in & pl&nt, s&les prices o' prod cts, oper&ting e(penses, %&chine cycle ti%es, the n %ber o' drive/thro gh $indo$s in & '&st/'ood rest& r&nt, &nd so on. ?(pressing the syste% response &s & ' nction o' syste% p&r&%eters &lso provides &n obDective ' nction 'or syste% e(peri%ent&l design, i%prove%ent st dies, &nd opti%i.&tion.A typic&l e(&%ple is the se o' line&r progr&%%ing %odels 'or &pplic&tions o' c&pit&l b dgeting, prod ction pl&nning, reso rces &lloc&tion, &nd '&cility loc&tion. Jther e(&%ples incl de 8 e ing %odels, 2&r"ov ch&ins, line&r &nd nonline&r regression %odels, bre&"/even &n&lysis, 'orec&sting %odels, &nd econo%ic order 8 &ntity %odels. -n'ort n&tely, not &ll syste% responses c&n be %odeled sing %&the%&tic&l 'or% l&s. <he co%ple(ity o' %ost re&l/ $orld syste%s precl des the &pplic&tion o' s ch %odels. !ence, & set o' si%pli'ic&tion &ss %ptions % st o'ten &cco%p&ny the &pplic&tion o' %&the%&tic&l %odels 'or the 'or% l&s to hold. S ch &ss %ptions o'ten le&d to i%pr&ctic&l res lts th&t li%it the ch&nce o' i%ple%enting or even, in so%e c&ses, considering s ch res lts. >or e(&%ple, thin" o' developing & 'or% l& th&t co%p tes & prod ction syste% thro ghp t given p&r&%eters s ch &s %&chine cycle ti%es, speeds o' convey&nce syste%s, n %ber o' &sse%bly oper&tors, si.es o' syste% b ''ers, &nd pl&nt oper&ting p&ttern. Wh&t type o' %&the%&tic&l %odel $o ld yo se to &ppro(i%&te s ch & responseK !o$ represent&tive $ill the %&the%&tic&l %odel beK C&n yo se the thro ghp t n %bers to pl&n sched le deliveries to c sto%ersK

Figure %&-# <$o/di%ension&l %&the%&tic&l %odel. Computer Models Co%p ter %odels &re n %eric&l, gr&phic&l, &nd logic&l represent&tion o' & syste% 4& prod ct or & process5 th&t tili.es the c&p&bility o' & co%p ter in '&st co%p t&tions, l&rge c&p&city, consistency, &ni%&tion, &nd &cc r&cy. Co%p ter si% l&tion %odels, $hich represent the %iddle$&re o' %odeling, &re virt &l represent&tions o' re&l/$orld prod cts &nd processes on the co%p ter. Co%p ter si% l&tions o' prod cts &nd processes &re developed sing di''erent &pplic&tion progr&%s &nd so't$&re tools. >or e(&%ple, & co%p ter progr&% c&n be sed to per'or% 'inite ele%ent &n&lysis to &n&ly.e stresses &nd str&ins 'or & cert&in prod ct design, &s sho$n in >ig re 7.*. Si%il&rly, sever&l %&the%&tic&l %odels th&t represent co%ple( %&the%&tic&l oper&tions, control syste%s, 'l id %ech&nics, co%p ter &lgorith%s, &nd others c&n

be b ilt, &ni%&ted, &nd &n&ly.ed $ith co%p ter tools. So't$&re tools &re &lso &v&il&ble to develop st&tic &nd dyn&%ic &ni%&tions o' %&ny ind stri&l processes.

Figure %&.# >inite ele%ent &n&lysis co%p ter %odel. Acc r&te &nd $ell/b ilt co%p ter %odels co%pens&te 'or the li%it&tions o' the other types o' %odels. <hey &re b ilt sing so't$&re tools, $hich is e&sier, '&ster, &nd che&per th&n b ilding physic&l %odels. In &ddition, the 'le(ibility o' co%p ter %odels &llo$ 'or 8 ic" ch&nges, e&sy testing o' $h&t/i's, &nd 8 ic" ev&l &tion o' syste% per'or%&nce 'or e(peri%ent&l design &nd opti%i.&tion st dies. Co%p ter %odels &lso provide the bene'its o' gr&phic&l %odels $ith %odern &ni%&tion &nd logic&l present&tion tools. Indeed, %ost co%p ter %odels &re b ilt sing gr&phic&l %odeling tools. Co%p&red to %&the%&tic&l %odels, co%p ter %odels &re % ch %ore re&listic &nd e''icient. <hey tili.e co%p ter c&p&bilities 'or %ore &cc r&te &ppro(i%&tions, they r n co%plic&ted co%p t&tions in little ti%e, &nd they c&n %e&s re syste% per'or%&nce $itho t the need 'or & closed/'or% de'inition o' the syste% obDective ' nction. S ch c&p&bilities %&de co%p ter %odels the %ost co%%on %odeling techni8 es in tod&yEs prod ction &nd b siness &pplic&tions. ;i%it&tions o' co%p ter %odels incl de co%p ter li%it&tions, &pplic&tion so't$&re li%it&tions, the li%it&tions o' syste% co%ple(ity, &nd li%it&tions o' &n&lysts. 1iscrete event si% l&tion 41?S5 4disc ssed in det&il in Ch&pter 65 is co%p ter si% l&tion th&t %i%ics the oper&tion o' re&l/$orld processes &s they evolve over ti%e. <his type o' co%p ter %odel is the %&Dor 'oc s o' this boo", $here si% l&tion is sed &s & te%pl&te 'or &pplying le&n si(/sig%& &nd design 'or si(/sig%& %ethods. 1?S co%p ter %odeling &ssists in c&pt ring the dyn&%ics &nd logics o' syste% processes &nd esti%&ting the syste%Es long/ter% per'or%&nce nder stoch&stic conditions. 2oreover, 1?S %odels &llo$ the ser to &s" $h&t/i' 8 estions &bo t & syste%, present ch&nges th&t &re %&de in the physic&l conditions, &nd r n the syste% 4o'ten 'or long periods5 to si% l&te the i%p&ct o' s ch ch&nges. <he %odel res lts &re then co%p&red to g&in insight into the beh&vior o' the syste%. >or e(&%ple, & 1?S pl&nt %odel c&n be sed to esti%&te the &sse%bly line thro ghp t by r nning the %odel dyn&%ic&lly &nd tr&c"ing its thro ghp t ho r by ho r or shi't by shi't. <he %odel c&n be &lso sed to &ssess % ltiple prod ction scen&rios b&sed on & long/ter% thro ghp t &ver&ge. As sho$n in >ig re 7.+, si% l&tion so't$&re tools provide & 'le(ible environ%ent o' %odeling &nd &n&lyses th&t %&"es 1?S inco%p&r&ble to %&the%&tic&l &nd physic&l %odels.

Figure %&/# 1?S environ%ent in the Si% l+ so't$&re p&c"&ge.

%&* Simulation )odeling

As sho$n e&rlier, modeling is the &rt &nd science o' c&pt ring the ' nction&lity &nd the relev&nt ch&r&cteristics o' re&l/$orld syste%s &nd presenting s ch syste%s in & 'or% th&t '&cilit&tes syste% "no$ledge, &n&lyses, i%prove%ent, &nd opti%i.&tion. Physic&l, gr&phic&l, %&the%&tic&l, &nd co%p ter %odels &re the %&Dor types o' %odels developed in di''erent engineering &pplic&tions. <his section 'oc ses on de'ining the simulation concept, developing the t&(ono%y o' v&rio s types o' si% l&tion %odels, &nd e(pl&ining the role o' si% l&tion in pl&nning, designing, &nd i%proving the per'or%&nce o' %odern b siness &nd prod ction oper&tions.

%&*&' Simulation Defined

Simulation is $idely sed ter% in re'erence to co%p ter si% l&tion %odels th&t represent & prod ct or & process. S ch %odels &re b ilt b&sed on both %&the%&tic&l &nd logic&l rel&tionships i%peded $ithin the syste% str ct re. >or e(&%ple, 'inite ele%ent &n&lysis is the %&the%&tic&l b&sis 'or & c&%sh&'t prod ct si% l&tion %odel, &nd the oper&tion o' & '&ctory is the logic&l design b&sis 'or & pl&nt process %odel. System simulation or simulation modeling, in gener&l, is the %i%ic"ing &ctivity o' the oper&tion o' & re&l syste%, in & co%p ter, $ith & 'oc s on process 'lo$, logic, &nd dyn&%ics. <here'ore, si% l&tion %e&ns %&"ing & si%pli'ied represent&tion o' &n origin&l. J st &s & %odel &ircr&'t c&pt res the %&ny o' the i%port&nt physic&l 'e&t res o' & re&l &ircr&'t, so & si% l&tion %odel c&pt res the i%port&nt oper&tion&l 'e&t res o' & re&l syste%. ?(&%ples o' syste% si% l&tion &re the d&y/to/d&y oper&tion o' & b&n", the v&l e o' & stoc" port'olio over & ti%e period, the r nning o' &n &sse%bly line in & '&ctory, &nd the st&'' &ssign%ent o' & hospit&l. <here'ore, inste&d o' &tte%pting to b ild e(tensive %&the%&tic&l %odels by e(perts, si% l&tion so't$&re tools h&s %&de it possible to %odel &nd &n&ly.e the oper&tion o' & re&l syste% by engineers &nd %&n&gers, not only progr&%%ers. <his &llo$s engineers to collect pertinent

in'or%&tion &bo t the beh&vior o' the syste% by e(ec ting & co%p teri.ed %odel inste&d o' observing the re&l one. In concl sion, the pri%&ry re8 ire%ents 'or si% l&tion &re & si% l&tion &n&lyst 4e.g., & si(/sig%& core te&% %e%ber or the te&% itsel'5 & co%p ter, &nd & si% l&tion l&ng &ge or so't$&re tool. <he &n&lyst is the "ey pl&yer in cond cting the si% l&tion st dy. !e or she is responsible 'or nderst&nding the re&l/$orld syste% &nd &n&ly.ing its ele%ents, logic, inp ts, o tp ts, &nd go&ls, in the %&nner presented in Section 7.2.). <he &n&lyst then oper&tes the co%p ter syste% &nd ses the so't$&re tool to b ild, v&lid&te, &nd veri'y the si% l&tion %odel. >in&lly, the &n&lyst &n&ly.es the res lts obt&ined 'ro% r nning the si% l&tion %odel &nd cond cts e(peri%ent&l design $ith the %odel in order to dr&$ concl sions on %odel beh&vior &nd deter%ine the best p&r&%eter settings. <he co%p ter syste% provides the h&rd$&re &nd so't$&re tools re8 ired to oper&te &nd r n the si% l&tion. <he si% l&tion so't$&re or l&ng &ge provides the pl&t'or% &nd environ%ent th&t '&cilit&tes %odel b ilding, testing, deb gging, &nd r nning. <he si% l&tion &n&lyst tili.es the si% l&tion so't$&re on & c&p&ble co%p ter syste% to develop & syste% si% l&tion %odel th&t c&n be sed &s & pr&ctic&l 4close/to/re&lity5 represent&tion o' the &ct &l syste%. >ig re 7., presents the si% l&tion concept.

Figure %&0# Si% l&tion concept.

%&*&( Simulation +axonomy

@&sed on the intern&l represent&tion sche%e selected, si% l&tion %odels c&n be discrete, continuous, or com'ined. 1iscrete event si% l&tion %odels, $hich &re the 'oc s o' this boo", &re the %ost co%%on &%ong si% l&tion types. 1?S %odels &re b&sed on & discrete intern&l represent&tion o' %odel v&ri&bles 4v&ri&bles th&t ch&nge their st&te &t discrete points in ti%e5. 1?S %ech&nics &re disc ssed in det&il in Ch&pter 6. In gener&l, discrete si% l&tion %odels 'oc s on %odeling discrete v&ri&bles th&t &re presented by r&ndo% or prob&bilistic %odels, $here the st&te o' the syste% ch&nges in discrete points in ti%e. A discrete v&ri&ble c&n be the n %ber o' c sto%ers in & b&n", prod cts &nd co%ponents in &n &sse%bly process, or c&rs in & drive/thro gh rest& r&nt. Contin o s si% l&tion %odels, on the other h&nd, 'oc s on contin o s v&ri&bles, r&ndo% or prob&bilistic, $here the st&te o' the syste% ch&nges contin o sly. ?(&%ples o' contin o s v&ri&bles incl de the level o' $&ter behind & d&%, che%ic&l processes in & petrole % re'inery, &nd the 'lo$ o' 'l ids in distrib tion pipes. Co%bined si% l&tion %odels incl de both discrete &nd contin o s ele%ents. >or e(&%ple, sep&r&te 4discrete5 'l id cont&iners &rrive to & che%ic&l process $here 'l ids &re po red into & reservoir to be processed in & contin o s %&nner. >ig re 7.)0 sho$s &n &ni%&tion o' & three/ di%ension&l discrete:contin o s si% l&tion o' pipe c&sting developed $ith & @roo"s A to%&tion A to2od si% l&tion p&c"&ge.

Figure %&'1# 1iscrete:contin o s A to2od si% l&tion %odel. 2&the%&tic&lly, $e c&n represent & %odel st&te sing & r&ndo% v&ri&ble &, $here &(t) is & st&te v&ri&ble th&t ch&nges over ti%e t. >or e(&%ple, *(t) is & contin o s st&te v&ri&ble th&t represents the del&y o' the ith c sto%er in & b&n". Another e(&%ple o' & discrete/st&te v&ri&ble is the n %ber o' c sto%ers M+4t5N in & 8 e e &t ti%e t in &n 2:2:) 8 e ing syste%. > rther%ore, %odels &re either deterministic or stochastic depending $hether they %odel r&ndo%ness &nd ncert&inty in & process or not. A stoch&stic process is & prob&bilistic %odel o' & syste% th&t evolves r&ndo%ly in ti%e &nd sp&ce. >or%&lly, & stoch&stic process is & collection o' r&ndo% v&ri&bles M&4t5N de'ined on & co%%on s&%ple 4prob&bility5 sp&ce. ?(&%ples o' stoch&stic %odels oper&ting $ith r&ndo% v&ri&bles incl de inter&rriv&l ti%es o' c sto%ers &rriving &t & b&n" &nd service or processing ti%es o' c sto%ers re8 ests or tr&ns&ctions, v&ri&ble cycle ti%es, &nd %&chinesE ti%e to '&il re &nd ti%e to rep&ir p&r&%eters. 1eter%inistic %odels, on the other h&nd, involve no r&ndo% or prob&bilistic v&ri&bles in its processes. ?(&%ples incl de %odeling 'i(ed/cycle/ti%e oper&tions, s ch the c&se o' & to%&ted syste%s &nd &rriv&ls $ith preset &ppoint%ents to & doctor or & l&$yerEs o''ice. <he %&Dority o' re&l/$orld oper&tions &re prob&bilistic. !ence %ost si% l&tion st dies involve r&ndo% gener&tion &nd s&%pling 'ro% cert&in prob&bility distrib tions to %odel "ey syste% v&ri&bles. L&ri&bility in %odel inp ts le&d to v&ri&bility in %odel o tp ts. As sho$n in >ig re 7.)), there'ore, & deter%inistic %odel $ill gener&te stoch&stic response 4%5 $hen %odel inp ts 4&), &2, &nd &35 &re stoch&stic.

Figure %&''# 1eter%inistic %odel $ith stoch&stic response. >in&lly, &nd b&sed on the n&t re o' %odel evolve%ent $ith ti%e, %odels c&n be static or dynamic. In st&tic %odels the syste% st&te does not ch&nge over ti%e. >or e(&%ple, $hen & cert&in reso rce is &l$&ys &v&il&ble in & %&n '&ct ring process, the st&te o' s ch & reso rce is 'i(ed $ith ti%e. ?very ti%e this reso rce is re8 ired or needed, there $ill be no e(pected ch&nge in its st&t s. 2onte C&rlo si% l&tion %odels &re ti%e/independent 4st&tic5 %odels th&t de&l $ith & syste% o' 'i(ed st&tes. In s ch spre&dsheet/li"e %odels, cert&in v&ri&ble v&l es &re ch&nged by r&ndo% gener&tion, &nd & cert&in %e&s re or %ore &re ev&l &ted &s to s ch ch&nges $itho t considering the ti%ing &nd the dyn&%ics o' s ch ch&nges. 2ost oper&tion&l %odels &re, ho$ever, dyn&%ic. Syste% st&te v&ri&bles o'ten ch&nge $ith ti%e, &nd the inter&ctions th&t res lt 'ro%

s ch dyn&%ic ch&nges do &''ect the syste% beh&vior. !ence, in 1?S si% l&tion, the ti%e di%ension is live. As sho$n in >ig re 7.)2, & %odel c&n involve both st&tic &nd dyn&%ic responses. <his &g&in %&"es 1?S %odels the %ost co%%on &%ong si% l&tion types since they %odel v&ri&bles th&t ch&nge their st&te &t discrete points in ti%e.

Figure %&'(# St&tic &nd dyn&%ic %odel responses. 1yn&%ic si% l&tion %odels &re ' rther divided into ter%in&ting &nd nonter%in&ting %odels b&sed on r n ti%e. <er%in&ting %odels &re stopped by & cert&in natural event, s ch &s & doctor $ho sees & li%ited n %ber o' p&tients per d&y, & $or"shop th&t 'inishes &ll 8 &ntity in & cert&in order, or & b&n" th&t oper&tes 'or + ho rs & d&y. In s ch syste%s, the length o' & si% l&tion r n depends on the n %ber o' ite%s processed, or on re&ching & speci'ied ti%e period, $ith the %odel ter%in&ting pon co%pletion. 2odels th&t r n 'or & speci'ied period o'ten incl de initi&l conditions s ch &s cle&ning the syste%, 'illing b&n" A<2s, resetting co%p ters in & govern%ent o''ice, or e%ptying b ''ers in &n &sse%bly line. #onter%in&ting %odels, on the other h&nd, c&n r n contin o sly since the i%p&ct o' initi&li.&tion is negligible over the long r n &nd no n&t r&l event h&s been speci'ied to stop the%. 2ost pl&nts r n on & contin o s %ode, $here prod ction lines st&rt every shi't $itho t e%ptying the syste%. !ence, the r n ti%e 'or s ch %odels is o'ten deter%ined st&tistic&lly to obt&in & ste&dy/st&te response, &s $e disc ss l&ter. >ig re 7.)3 presents the t&(ono%y o' di''erent si% l&tion types $ith highlighted &ttrib tes o' 1?S. As noted in the 'ig re, 1?S %odels &re digit&l 4discrete5, stoch&stic, &nd dyn&%ic co%p ter %odels o' ter%in&ting or nonter%in&ting 4ste&dy/st&te5 response. S ch three ch&r&cteristics o'ten rese%ble the &ct &l beh&vior o' %&ny re&l/$orld syste%s &nd tr&ns&ction&l processes. >or e(&%ple, in %&n '&ct ring syste%s, the 'lo$ o' r&$ %&teri&l, &sse%bly co%ponents, &nd prod cts c&n be %odeled &s discrete, dyn&%ic, &nd stoch&stic processes. Si%il&rly, %&ny service '&cilities o'ten de&l $ith discrete entities th&t r n dyn&%ic&lly in & stoch&stic %&nner.

Figure %&'*# Si% l&tion t&(ono%y.

%&% +he Role Of Simulation

A'ter nderst&nding the v&rio s %e&nings &nd &spects o' the ter% simulation modeling, it is necess&ry to cl&ri'y the role th&t si% l&tion pl&ys in %odern ind stri&l &nd b siness 'ir%s. In this section $e cl&ri'y the role o' si% l&tion by 'irst D sti'ying the se o' si% l&tion both technically &nd economically &nd then presenting the spectr % o' si% l&tion &pplic&tions to v&rio s ind stries in the %&n '&ct ring &nd service sectors. It is &lso $orth %entioning th&t sing si% l&tion in ind stri&l &nd b siness &pplic&tion is the %ost co%%on b t not the only 'ield in $hich si% l&tion is tili.edO it is &lso sed 'or ed c&tion&l &nd le&rning p rposes, tr&ining, virt &l re&lity &pplic&tions, %ovies &nd &ni%&tion prod ction, &nd cri%in&l D stice, &%ong others.

%&%&' Simulation 2ustified

<he 8 estion o' $hy &nd $hen to si% l&te is typic&l o' those th&t cross the %inds o' pr&ctitioners, engineers, &nd %&n&gers. We si%ply si% l&te bec& se o' si% l&tion c&p&bilities th&t &re ni8 e &nd po$er' l in syste% represent&tion, per'or%&nce esti%&tion, &nd i%prove%ent. Si% l&tion is o'ten the &n&lystsE re' ge $hen other sol tion tools, s ch &s %&the%&tic&l %odels, '&il or beco%e e(tre%ely di''ic lt to &ppro(i%&te the sol tion to & cert&in proble%. 2ost re&l/$orld processes in prod ction &nd b siness syste%s &re co%ple(, stoch&stic, &nd highly nonline&r &nd dyn&%ic, $hich %&"es it &l%ost i%possible to present the% sing physic&l or %&the%&tic&l %odels. Atte%pts to se &n&lytic&l %odels in &ppro&ching re&l syste%s s &lly re8 ire %&ny &ppro(i%&tions &nd si%pli'ying &ss %ptions. <his o'ten yields sol tions th&t &re ns it&ble 'or proble%s o' re&l/$orld &pplic&tions. <here'ore, &n&lysts o'ten se si% l&tion $henever they %eet co%ple( proble%s th&t c&nnot be solved by other %e&ns, s ch &s %&the%&tic&l &nd c&lc l s/b&sed %ethods. <he per'or%&nce o' & re&l syste% is & co%plic&ted ' nction o' the design p&r&%eters, &nd &n &n&lytic&l closed/'or% e(pression o' the obDective ' nction or the constr&ints %&y not e(ist. A si% l&tion %odel c&n there'ore be sed to repl&ce the %&the%&tic&l 'or% l&tion o' the nderlying syste%. With the &id o' the %odel &nd r&ther th&n considering every possible v&ri&tion o' circ %st&nces 'or the co%ple( proble%, & s&%ple o' possible e(ec tion p&ths is t&"en &nd st died.

In short, si% l&tion is o'ten tili.ed $hen the beh&vior o' & syste% is co%ple(, stoch&stic 4r&ther th&n deter%inistic5, &nd dyn&%ic 4r&ther th&n st&tic5. An&lytic&l %ethods, s ch &s 8 e ing syste%s, inventory %odels, &nd 2&r"ovi&n %odels, $hich &re co%%only sed to &n&ly.e prod ction &nd b siness syste%s, o'ten '&il to provide st&tistics on syste% per'or%&nce $hen re&l/$orld conditions intensi'y to over$hel% &nd e(ceed the syste%/&ppro(i%&ting &ss %ptions. ?(&%ples incl de entities $hose &rriv&l &t & pl&nt or b&n" is not & Poisson process, &nd the 'lo$ o' entities is b&sed on co%ple( decision r les nder stoch&stic v&ri&bility $ithin &v&il&bility o' syste% reso rces. *ecision support is &nother co%%on D sti'ic&tion o' si% l&tion st dies. Jbvio sly, engineers &nd %&n&gers $&nt to %&"e the best decisions possible, especi&lly $hen enco ntering critic&l st&ges o' design, e(p&nsion, or i%prove%ent proDects $here the re&l syste% h&s not yet been b ilt. @y c&re' lly &n&ly.ing the hypothetic&l syste% $ith si% l&tion, designers c&n &void proble%s $ith the re&l syste% $hen it is b ilt. Si% l&tion st dies &t this st&ge %&y reve&l ins r%o nt&ble proble%s th&t co ld res lt in proDect c&ncell&tion, &nd s&ve cost, e''ort, &nd ti%e. S ch s&vings &re obt&ined since it is &l$&ys che&per &nd s&'er to le&rn 'ro% %ist&"es %&de $ith & si% l&ted syste% 4& co%p ter %odel5 th&n to %&"e the% 'or re&l. Si% l&tion c&n red ce cost, red ce ris", &nd i%prove &n&lystsE nderst&nding o' the syste% nder st dy. <he econo%ic D sti'ic&tion o' si% l&tion o'ten pl&ys & role in selecting si% l&tion &s & sol tion tool in design &nd i%prove%ent st dies. Altho gh si% l&tion st dies %ight be costly &nd ti%e cons %ing in so%e c&ses, the bene'its &nd s&vings obt&ined 'ro% s ch st dies o'ten recover the si% l&tion cost &nd &void % ch ' rther costs. Si% l&tion costs &re typic&lly the initi&l si% l&tion so't$&re &nd co%p ter cost, ye&rly %&inten&nce &nd pgr&de cost, tr&ining cost, engineering ti%e cost, &nd other costs9 'or tr&veling, prep&ring present&tions $ith % lti%edi& tools, &nd so on. S ch costs &re o'ten recovered $ith the 'irst t$o or three s ccess' l si% l&tion proDects. > rther, the cost &nd ti%e o' si% l&tion st dies &re o'ten red ced by &n&lyst e(perience &nd beco%e %in sc le co%p&red to the long/ter% s&vings 'ro% incre&sing prod ctivity &nd e''iciency.

%&%&( Simulation A

li$ations

A better &ns$er to the 8 estion F$hy si% l&teKA c&n be re&ched by e(ploring the $ide spectr % o' si% l&tion &pplic&tions to &ll &spects o' science &nd technology. <his spectr % st&rts by sing si% l&tion in b&sic sciences to esti%&te the &re& nder & c rve, ev&l &ting % ltiple integr&ls, &nd st dying p&rticle di'' sion, &nd contin er by tili.ing si% l&tion in pr&ctic&l sit &tions &nd designing 8 e ing syste%s, co%% nic&tion net$or"s, econo%ic 'orec&sting, bio%edic&l syste%s, &nd $&r str&tegies &nd t&ctics. <od&y, si% l&tion is being sed 'or & $ide r&nge o' &pplic&tions in both %&n '&ct ring &nd b siness oper&tions. As & po$er' l tool, si% l&tion %odels o' %&n '&ct ring syste%s &re sed9 <o deter%ine the thro ghp t c&p&bility o' & %&n '&ct ring cell or &sse%bly line <o deter%ine the n %ber o' oper&tors in & l&bor/intensive &sse%bly process <o deter%ine the n %ber o' & to%&ted g ided vehicles in & co%ple( %&teri&l/h&ndling syste% <o deter%ine the n %ber o' c&rriers in &n electri'ied %onor&il syste% <o deter%ine the n %ber o' stor&ge &nd retriev&l %&chines in & co%ple( & to%&ted stor&ge &nd retriev&l syste% <o deter%ine the best ordering policies 'or &n inventory control syste% <o v&lid&te the prod ction pl&n in %&teri&l re8 ire%ent pl&nning <o deter%ine the opti%&l b ''er si.es 'or $or"/in/progress prod cts <o pl&n the c&p&city o' s b&sse%blies 'eeding & prod ction %&inline >or b siness oper&tions, si% l&tion %odels &re &lso being sed 'or & $ide r&nge o' &pplic&tions9 <o deter%ine the n %ber o' b&n" tellers, $hich res lts in red cing c sto%er $&iting ti%e by & cert&in percent&ge

<o design distrib tion &nd tr&nsport&tion net$or"s to i%prove the per'or%&nce o' logistic &nd vending syste%s <o &n&ly.e & co%p&nyEs 'in&nci&l syste% <o design the oper&ting policies in & '&st/'ood rest& r&nt to red ce c sto%er ti%e/in/ syste% &nd incre&se c sto%er s&tis'&ction <o ev&l &te h&rd$&re &nd so't$&re re8 ire%ents 'or & co%p ter net$or" <o design the oper&ting policies in &n e%ergency roo% to red ce p&tient $&iting ti%e &nd sched le the $or"ing p&ttern o' the %edic&l st&'' <o &ssess the i%p&ct o' govern%ent reg l&tions on di''erent p blic services &t both the % nicip&l &nd n&tion&l levels <o test the 'e&sibility o' di''erent prod ct develop%ent processes &nd to ev&l &te their i%p&ct on co%p&nyEs b dget &nd co%petitive str&tegy <o design co%% nic&tion syste%s &nd d&t& tr&ns'er protocols

<&ble 7.) lists e(&%ples o' si% l&tion &pplic&tions in the %&n '&ct ring &nd service sectors. +a3le %&'# Exam les of Simulation A li$ations Jpen t&ble &s spre&dsheet )anufa$turing "ndustries A to%otive Aerosp&ce Pl&stics ind stry P&per %ills !o se &ppli&nces > rnit re %&n '&ct ring Che%ic&l ind stry Clothing &nd te(tile P&c"&ging Stor&ge &nd retriev&l Ser4i$e "ndustries @&n"ing ind stry !e&lth syste%s !otel oper&tions Co%% nic&tion services Co%p ter net$or"s <r&nsport&tion syste%s ;ogistics &nd s pply ch&in 0est& r&nts &nd '&st 'ood Post&l services Airport oper&tions

<o re&ch the go&ls o' the si% l&tion st dy, cert&in ele%ents o' e&ch si% l&ted syste% o'ten beco%e the 'oc s o' & si% l&tion %odel. 2odeling &nd tr&c"ing s ch ele%ents provide &ttrib tes &nd st&tistics necess&ry to design, i%prove, &nd opti%i.e the nderlying syste% per'or%&nce. <&ble 7.2 presents e(&%ples o' si% l&ted syste%s $ith %odel %&in ele%ents in e&ch syste%. +a3le %&(# Exam les of Simulated Systems Jpen t&ble &s spre&dsheet Simulated System 2&n '&ct ring syste% ?%ergency 0oo% @&n" 0et&il store Exam les of )odel Elements 5E-R-"-C-O6 P&rts, %&chines, oper&tors, conveyors, stor&ge P&tients, beds, doctors, n rses, $&iting roo% C sto%ers, b&n" tellers, A<2S, lo&n o''icers Shoppers, chec"o t c&sh registers, c sto%er service

+a3le %&(# Exam les of Simulated Systems Jpen t&ble &s spre&dsheet Simulated System Co%p ter net$or" >ree$&y syste% >&st/'ood rest& r&nt @order crossing point Cl&ss registr&tion o''ice S pply ch&in:logistics Exam les of )odel Elements 5E-R-"-C-O6 Server, client PCS, &d%inistr&tor, d&t& protocol C&rs, tr&''ic lights, ro&d eg%ents, interch&nges Servers, c sto%ers, c&rs, drive/thro gh $indo$s C&rs, c sto%s &gents, booths ,i%%igr&tion o''icers St dents, co rses, registr&tion st&tions, helpers S ppliers &nd vendors tr&nsport&tion syste%, clients

%&%&* Simulation !re$autions

;i"e &ny other engineering tool, si% l&tion h&s li%it&tions. S ch li%it&tions sho ld be de&lt $ith &s & %otiv&tion &nd sho ld not disco r&ge &n&lysts &nd decision %&"ers. Pno$ing li%it&tions o' the tool in h&nd sho ld e%ph&si.e sing it $isely &nd %otiv&te the ser to develop cre&tive %ethods &nd est&blish the correct &ss %ptions th&t bene'it 'ro% the po$er' l si% l&tion c&p&bilities &nd precl de si% l&tion li%it&tions 'ro% being & d&%ping '&ctor. !o$ever, cert&in prec& tions sho ld be considered in sing si% l&tion to &void the potenti&l pit'&lls o' si% l&tion. ?(&%ples o' iss es th&t $e sho ld p&y &ttention to $hen considering si% l&tion incl de the 'ollo$ing9 ). <he si% l&tion &n&lyst or decision %&"er sho ld be &ble to &ns$er the 8 estion o' $hen not to si% l&te. A lot o' si% l&tion st dies &re considered to be design over"ill $hen cond cted 'or solving proble%s o' rel&tive si%plicity. S ch proble%s c&n be solved sing engineering &n&lysis, co%%on sense, or %&the%&tic&l %odels. !ence, the only bene'its 'ro% &ppro&ching si%ple syste%s $ith si% l&tion is being &ble to pr&ctic %odeling &nd to provide &n &ni%&tion o' the t&rgeted process. 2. <he cost &nd ti%e o' si% l&tion sho ld be considered &nd pl&nned $ell. 2&ny si% l&tion st dies &re nderesti%&ted in ter%s o' ti%e &nd cost. So%e decision %&"ers thin" o' si% l&tion st dy &s %odel/b ilding ti%e &nd cost. Altho gh %odel b ilding is & critic&l ph&se o' & si% l&tion st dy, it o'ten cons %es less ti%e &nd cost th&n does e(peri%ent&l design or d&t& collection. 3. <he s"ill &nd "no$ledge o' the si% l&tion &n&lyst. @eing &n engineer is &l%ost essenti&l 'or si% l&tion pr&ctitioners bec& se o' the type o' &n&lytic&l, st&tistic&l, &nd syste% &n&lyses s"ills re8 ired 'or cond cting si% l&tion st dies. 7. ?(pect&tions 'ro% the si% l&tion st dy sho ld be re&listic &nd not overesti%&ted. A lot o' pro'ession&ls thin" o' si% l&tion &s & Fcryst&l b&llA thro gh $hich they c&n predict &nd opti%i.e syste% beh&vior. It sho ld be cle&r to the &n&lyst th&t si% l&tion %odels the%selves &re not syste% opti%i.ers. While it sho ld be &sserted th&t si% l&tion is D st & tool &nd &n e(peri%ent&l pl&t'or%, it sho ld &lso be e%ph&si.ed th&t co%bining si% l&tion $ith &ppropri&te st&tistic&l &n&lyses, e(peri%ent&l design, &nd e''icient se&rch engine c&n le&d to inv&l &ble syste% in'or%&tion th&t bene'its pl&nning, design, &nd opti%i.&tion. 6. <he res lts obt&ined 'ro% si% l&tion %odels &re &s good &s the %odel d&t& inp ts, &ss %ptions, &nd logic&l design. <he co%%only sed phr&se gar'age,in,gar'age,out 4GIGJ5 is very &pplic&ble to si% l&tion st dies. !ence, speci&l &ttention sho ld be p&id to d&t& inp ts selection, 'iltering, &nd &ss %ptions. 6. <he &n&lyst sho ld p&y &ttention to the level o' det&il incorpor&ted in the %odel. 1epending on the obDectives o' the si% l&tion st dy &nd the in'or%&tion &v&il&ble, the &n&lyst sho ld decide on the &%o nt o' det&il incorpor&ted into the si% l&tion %odel. So%e st dy obDectives c&n be re&ched $ith %&cro/level %odeling, $here&s others re8 ire

*.

+.

,.

)0.

)).

)2. )3.

%icro/level %odeling. <here is no need 'or the &n&lyst to e(h& st his or her %odeling s"ills trying to incorpor&te det&ils th&t &re irrelev&nt to si% l&tion obDectives. Inste&d, the %odel sho ld be 'oc sed on providing the %e&ns o' syste% &n&lysis th&t yields res lts directly relev&nt to st dy obDectives. 2odel v&lid&tion &nd veri'ic&tion is not & trivi&l t&s". As disc ssed l&ter, %odel v&lid&tion 'oc ses on %&"ing s re th&t & %odel beh&ves &s re8 ired by the %odel/designed logic &nd th&t its response re'lects the d&t& sed into the %odel. 2odel veri'ic&tion, on the other h&nd, 'oc ses on %&"ing s re th&t the %odel beh&vior rese%bles the intended beh&vior o' the &ct &l si% l&ted syste%. @oth pr&ctices deter%ine the degree o' %odel reli&bility &nd re8 ire the &n&lyst to be '&%ili&r $ith the s"ills o' %odel testing &nd the str ct re &nd ' nction&lity o' the &ct &l syste%. <he res lts o' si% l&tion c&n e&sily be %isinterpreted. !ence, the &n&lyst sho ld concentr&te e''orts on collecting reli&ble res lts 'ro% the %odel thro gh proper settings o' r n controls 4$&r%/ p period, r n length, &nd n %ber o' replic&tions5 &nd on sing the proper st&tistic&l &n&lyses to dr&$ %e&ning' l &nd &cc r&te concl sions 'ro% the %odel. <ypic&l %ist&"es in interpreting si% l&tion res lts incl de relying on & short r n ti%e 4not & ste&dy/st&te response5, incl ding in the res lts bi&ses c& sed by initi&l %odel conditions, sing the res lts o' one si% l&tion replic&tion, &nd relying on the response %e&n $hile ignoring the v&ri&bility enco%p&ssed into response v&l es. Si% l&tion inp ts &nd o tp ts sho ld be co%% nic&ted cle&rly &nd correctly to &ll p&rties o' & si% l&tion st dy. Syste% speci&lists s ch &s process engineers &nd syste% %&n&gers need to be &$&re o' the d&t& sed &nd the %odel logic in order to veri'y the %odel &nd incre&se its re&listic represent&tion. Si%il&rly, the res lts o' the si% l&tion %odel sho ld be co%% nic&ted to get 'eedb&c" 'ro% p&rties on the relev&ncy &nd &cc r&cy o' res lts. <he &n&lyst sho ld &void sing incorrect %e&s res o' per'or%&nce $hen b ilding &nd &n&ly.ing %odel res lts. 2odel per'or%&nce %e&s res sho ld be progr&%%ed correctly into the %odel &nd sho ld be represented by st&tistics collected 'ro% the %odel. S ch %e&s res sho ld &lso represent the type o' in'or%&tion essenti&l to the &n&lyst &nd decision %&"er to dr&$ concl sions &nd in'erences &bo t %odel beh&vior. <he &n&lyst sho ld &void the %is se o' %odel &ni%&tion. In '&ct, &ni%&tion is &n i%port&nt si% l&tion c&p&bility th&t provides engineers &nd decision %&"ers $ith & gre&t tool 'or syste% vis &li.&tion &nd response observ&nce. !ence, it is tr e th&t F& pict re is $orth & tho s&nd $ords.A S ch & tool is &lso se' l 'or %odel deb gging, v&lid&tion &nd veri'ic&tion, &nd present&tion to e(ec tives &nd c sto%ers. !o$ever, & lot o' people %is se %odel &ni%&tion &nd rely on their observ&tion to dr&$ concl sions &s to %odel long/ter% beh&vior. Given th&t si% l&tion %odels &re stoch&stic &nd dyn&%ic in n&t re, it sho ld be cle&r to the &n&lyst th&t & %odelEs st&t s &t & cert&in ti%e does not necess&rily re'lect its long/ter% beh&vior. Inste&d, %odel st&tistics &re & better represent&tion o' %odel response. <he &n&lyst needs to get the s pport o' pper %&n&ge%ent &nd decision %&"ers to %&"e & si% l&tion st dy 'r it' l &nd s ccess' l. >in&lly, the &n&lyst sho ld select the &ppropri&te si% l&tion so't$&re tools th&t 'it the &n&lystEs "no$ledge &nd e(pertise &nd th&t &re c&p&ble o' %odeling the nderlying syste% &nd providing the si% l&tion res lts re8 ired. <he criteri& 'or selecting the proper si% l&tion so't$&re tools &re &v&il&ble in the liter&t re &nd &re not the 'oc s o' this boo". It sho ld be "no$n, ho$ever, th&t si% l&tion p&c"&ges v&ry in their c&p&bilities &nd incl siveness o' di''erent %odeling syste%s &nd techni8 es, s ch &s conveyor syste%s, po$er &nd 'ree syste%s, & to%&ted g ided vehicle syste%s, "ine%&tics, & to%&ted stor&ge &nd retriev&l syste%s, h %&n %odeling c&p&bilities, st&tistic&l tools, opti%i.&tion %ethods, &ni%&tion, &nd so on.

%&, Simulation Soft7are

Si% l&tion so't$&re tools &re in co%%on se &%ong si% l&tion &nd si(/sig%& pr&ctitioners in v&rio s types o' &pplic&tions &nd 'or di''erent p rposes. <he %&Dor bene'it th&t si% l&tion p&c"&ges provide to si(/sig%& pr&ctitioners is the &bility to %odel dyn&%ic syste%s nder close/ to re&l/$orld conditions. Processes o' & tr&ns&ction&l n&t re in both %&n '&ct ring &nd services o'ten beh&ve in & co%ple(, dyn&%ic, &nd stoch&stic %&nner. S ch c&p&bility there'ore &llo$s &n&lysts to better e(press ne$ concepts &nd designs 'or tr&ns&ction&l processes, to %e&s re process per'or%&nce $ith ti%e/b&sed %etrics, to cond ct st&tistic&l &n&lyses, &nd to i%prove or opti%i.e ne$ &nd c rrent syste%s. Indeed, %&ny si% l&tion so't$&re tools $ere initi&lly b ilt by pr&ctitioners to %odel p&rtic l&r re&l/ $orld syste%s, s ch &s %&teri&l h&ndling syste%s, he&lth syste%s, &nd &sse%bly lines. <hey $ere b ilt 'ro% entities &nd processes th&t %i%ic the obDects &nd &ctivities in the re&l syste%. <o ' rther %eet the needs o' si(/sig%& pr&ctitioners &nd engineers in gener&l, si% l&tion vendors developed &nd integr&ted %&ny %od les into their so't$&re prod cts to si%pli'y %odel b ilding, '&cilit&te %odel c sto%i.&tion, &llo$ 'or the cre&tion o' i%pressive &ni%&tion, &nd en&ble &n&lysts to cond ct st&tistic&l &n&lyses &nd r n opti%i.&tion se&rches. So%e o' these %od les incl de si(/ sig%& de'initions &nd c&lc l&tors &nd provide lin"s to spre&dsheet &nd st&tistic&l so't$&re to r n e(peri%ent&l designs &nd si(/sig%& &n&lyses directly. >ig re 7.)7 sho$s &n e(&%ple o' & 'le(ible si% l&tion environ%ent developed $ith the WI<#?SS si% l&tion p&c"&ge. WI<#?SS incl des & %od le 'or si(/sig%& c&lc l&tion, & %od le 'or opti%i.&tion, &nd & direct lin" to the 2I#I<A@ st&tistic&l p&c"&ge.

Figure %&'%# >le(ible si% l&tion so't$&re 'or %odeling &nd &n&lyses. <he spectr % o' si% l&tion so't$&re c rrently &v&il&ble r&nges 'ro% proced r&l l&ng &ges &nd si% l&tion l&ng &ges to speci&l si% l&tors. <he v&st &%o nt o' si% l&tion so't$&re &v&il&ble c&n be over$hel%ing 'or ne$ sers. Co%%on so't$&re p&c"&ges incl de A to2od,WI<#?SS, Si% l+, &nd Aren&. In Ch&pter )) $e provide & co%prehensive list o' si% l&tion so't$&re tools &nd vendors &nd provide ' rther det&ils on si% l&tion so't$&re 'e&t res, &pplic&tions, &nd co%p&risons.

%&- Summary
@&sic concepts o' si% l&tion %odeling incl de the syste%, %odel, &nd si% l&tion concepts. Syste%s incl de inp ts, entities, rel&tionships, controls, &nd o tp ts. Syste% ele%ents sho ld be t ned to$&rd &tt&ining &n over&ll syste% go&l. A syste% %odel is & represent&tion or &n &ppro(i%&tion o' & re&l/$orld syste%. 2odels c&n be physic&l, gr&phic&l, %&the%&tic&l, &nd co%p ter %odels. <he go&l o' %odeling is to provide & tool 'or syste% &n&lysis, design, &nd

i%prove%ent th&t is che&per, '&ster, &nd e&sier to nderst&nd. Si% l&tion is the &rt &nd science o' %i%ic"ing the oper&tion o' & re&l/$orld syste% on the co%p ter. It is &i%ed &t c&pt ring the co%ple(, dyn&%ic, &nd stoch&stic ch&r&cteristics o' & re&l/$orld process, $here other types o' %odels '&ll short. @&sed on the type o' st&te v&ri&bles, co%p ter si% l&tion %odels c&n be discrete, contin o s, or co%bined. <hey c&n &lso be deter%inistic or stoch&stic, b&sed on r&ndo%ness %odeling. >in&lly, they c&n be st&tic or dyn&%ic, b&sed on the ch&nges o' syste% st&te. <he 'oc s o' this boo" is on discrete event si% l&tion %odels, $hich represent syste%sE co%ple( &nd stoch&stic beh&vior dyn&%ic&lly in ter%s o' discrete st&te v&ri&bles. 1?S is D sti'ied by its gro$ing c&p&bility, so't$&re tools, &nd $ide spectr % o' re&l/$orld &pplic&tions. 1?S si% l&tion is the b&c"bone o' si% l&tion/b&sed le&n si(/sig%& &nd design 'or le&n si(/sig%&.

Cha ter ,# Dis$rete E4ent Simulation

,&' "ntrodu$tion
As disc ssed in Ch&pter 7, co%p ter si% l&tion, is & techni8 e th&t i%it&tes the oper&tion o' & re&l/$orld syste% &s it evolves over time. 4See Section 7.2.) 'or the de'inition o' system sed in this boo".5 It is considered &n e(peri%ent&l &nd &pplied %ethodology th&t see"s to describe the beh&vior o' syste%s, constr ct hypotheses th&t &cco nt 'or observed beh&vior, &nd predict ' t re beh&vior. A si% l&tion is there'ore the e(ec tion o' & %odel, represented by & co%p ter progr&% th&t gives in'or%&tion &bo t the syste% being investig&ted. Process si% l&tion, o'ten sed to %odel prod ction &nd b siness processes in both the %&n '&ct ring &nd service sectors, is re'erred to &s discrete e-ent simulation 41?S5. 1iscrete event syste%s &re dyn&%ic syste%s th&t evolve in ti%e by the occ rrence o' events &t possibly irreg l&r ti%e interv&ls. Since this rese%bles the n&t re o' the %&Dority o' re&l/$orld prod ction &nd b siness syste%s, 1?S %odels &re sed $idely in re&l/$orld &pplic&tions. ?(&%ples incl de tr&''ic syste%s, %&n '&ct ring syste%s, co%p ter/co%% nic&tions syste%s, c&ll centers, b&n" oper&tions, hospit&ls, rest& r&nts, prod ction lines, &nd 'lo$ net$or"s. 2ost o' these syste%s c&n be %odeled in ter%s o' discrete events $hose occ rrence c& ses the syste% to ch&nge 'ro% one st&te to &nother in & stoch&stic %&nner. <he simulation modeling approach 'or &n&ly.ing & prod ction or b siness syste% is opposed to the &n&lytic&l &ppro&ch, $here the %ethod o' &n&ly.ing the syste% is p rely theoretic&l or &n &ppro(i%&tion o' & theoretic&l %odel. As & %ore re&listic &ltern&tive to the &n&lytic&l &ppro&ch, the si% l&tion &ppro&ch gives %ore 'le(ibility to the si(/sig%& te&% &nd convenience to the decision %&"er by tili.ing cert&in e''ective %ech&nics. ?vents &ctiv&tion, r&ndo% n %ber gener&tion, &nd ti%e &dv&nce%ent &re the %&in 1?S %ech&nics. ?vents, $hich &re &ctiv&ted &t cert&in points in ti%e, &''ect the over&ll st&te o' the syste% 4syste% v&ri&bles, reso rces, entity 'lo$, etc.5. <he points in ti%e &t $hich &n event is &ctiv&ted &re s &lly r&ndo%i.ed sing & r&ndo% n %ber gener&tion techni8 e th&t is b ilt in the %odel so't$&re tool so th&t no inp t 'ro% o tside the syste% is re8 ired. Si% l&tion events e(ist & tono%o sly in & discrete %&nner. <i%e 4& si% l&tion cloc"5 is there'ore &dv&nced to & ne(t sched led event since nothing $ill h&ppen bet$een the e(ec tion o' &ny t$o consec tive events. In &ddition to po$er' l 1?S %ech&nics, '&st co%p t&tions on tod&yEs high/speed processors, &long $ith gro$ing gr&phics c&p&bility, h&ve contrib ted gre&tly to the e''ectiveness &nd &ni%&tion or vis &li.&tion c&p&bility in 1?S %odels. <his ch&pter provides & deeper nderst&nding o' the 1?S process, co%ponents, &nd %ech&nis%s. ?(&%ples o' %&n &l si% l&tions &re sed to cl&ri'y 1?S ' nction&lity.

,&( System )odeling 8ith DES

@y tili.ing co%p ter c&p&bilities in logic&l progr&%%ing, r&ndo% gener&tion, '&st co%p t&tions, &nd &ni%&tion, 1?S %odeling is c&p&ble o' c&pt ring the ch&r&cteristics o' & re&l/$orld process &nd esti%&ting syste% per'or%&nce %e&s res &t di''erent settings o' its design p&r&%eters. <his is p&rtic l&rly i%port&nt in si(/sig%& 1A2IC &nd 1>SS &pplic&tions, $here syste% per'or%&nce is opti%i.ed &nd i%proved. <o %e&s re s ch per'or%&nce, 1?S i%it&tes the stoch&stic &nd co%ple( oper&tion o' & re&l/$orld syste% &s it evolves over ti%e &nd see"s to describe &nd predict the syste%Es &ct &l beh&vior. <he 1?S &ppro&ch h&s n %ero s po$er' l c&p&bilities th&t c&n provide the si(/sig%& te&% $ith in'or%&tion ele%ents th&t 8 &nti'y the syste% ch&r&cteristics, s ch &s design or decision p&r&%eters &nd per'or%&nce %e&s res 4C<Bs:C<Ss5. S ch in'or%&tion %&"es & 1?S %odel &n e''ective decision s pport syste%. >or e(&%ple, 1?S c&n be sed to esti%&te the n %ber o' reso rces needed in & syste% 4e.g., ho$ %&ny oper&tors, %&chines, tr c"s5. <he %odel c&n &lso

be sed to &rr&nge &nd b&l&nce reso rces to &void bottlenec"s, cross/tr&''ic, b&c"tr&c"ing, &nd e(cessive $&iting ti%es or inventories. <o$&rd this end, syste% %odeling $ith 1?S incl des %i%ic"ing the str ct re, l&yo t, d&t&, logic, &nd st&tistics o' the re&l/$orld syste% &nd representing the% in & 1?S %odel. An &bstr&ction o' syste% %odeling $ith 1?S is ill str&ted in >ig re 6.). Abstr&cting the re&l/$orld syste% in & 1?S %odel c&n be &ppro&ched by precise nderst&nding &nd speci'ic&tion o' the det&ils o' the 'ive syste% %odeling ele%ents sho$n in the 'ig re.

Figure ,&'# ?le%ents o' syste% %odeling $ith 1?S.

,&(&' System Stru$ture

A syste% 1?S %odel is e(pected to incl de the str ct re o' the &ct &l syste% being si% l&ted. <his str ct re is b&sic&lly the set o' syste% ele%ents in ter%s o' physic&l co%ponents, pieces o' e8 ip%ent, reso rces, %&teri&ls, 'lo$ lines, &nd in'r&str ct re. 1epending on the n&t re o' the syste% being %odeled &nd the obDective o' the si% l&tion st dy, the si(/sig%& te&% o'ten decides on $h&t ele%ents to incl de in the si% l&tion %odel. ?le%ents o' & %&n '&ct ring syste% &re di''erent 'ro% the ele%ents o' & b siness syste%. Where&s %&n '&ct ring syste%s &re %odeled sing %&chines, l&bor, $or"pieces, conveyors, &nd so on, b siness syste%s &re %odeled sing h %&n st&'', c sto%ers, in'or%&tion 'lo$, service oper&tions, &nd tr&ns&ctions. 2odeling s ch ele%ents thoro ghly is $h&t %&"es & %odel re&listic &nd represent&tive. !o$ever, the level o' det&ils &nd speci'ic&tions o' %odel str ct r&l ele%ents depends pri%&rily on the obDective &nd p rpose 'or b ilding the %odel. >or e(&%ple, the det&ils o' %&chine co%ponents &nd "ine%&tics %&y not be help' l in esti%&ting its tili.&tion &nd e''ectiveness. !ence, b&sic gr&phic&l represent&tions o' str ct r&l ele%ents &re o'ten sed in 1?S %odels 'or &ni%&tion p rposes. Jn the other h&nd, physic&l ch&r&cteristics s ch &s di%ensions, sp&ce, dist&nces, &nd p&th$&ys th&t o'ten &''ect 'lo$ ro tes, cycle ti%es, &nd c&p&city sho ld be p&rt o' the %odel str ct re. >or & pl&nt or & %&n '&ct ring syste%, <&ble 6.) sho$s e(&%ples o' str ct r&l ele%ents &nd their i%p&ct on %odel per'or%&nce '&ctors. +a3le ,&'# Exam les of Stru$tural Elements in DES Jpen t&ble &s spre&dsheet Stru$tural Element )odeled Conveyor length -nit lo&d di%ensions @ ''er si.e ;ength o' &isles &nd $&l"$&ys Si.e o' & to%&ted g ided vehicle AGL ;ength o' %onor&il )odel !erforman$e Fa$tor Affe$ted Conveyor c&p&city # %ber o' nits st&c"ed @ ''er c&p&city W&l"ing dist&nce &nd ti%e # %ber o' AGL c&rriers C&rrier tr&veling ti%e

+a3le ,&'# Exam les of Stru$tural Elements in DES Jpen t&ble &s spre&dsheet Stru$tural Element )odeled 1og sp&cing o' po$er &nd 'ree syste% 1i%ensions o' stor&ge nits )odel !erforman$e Fa$tor Affe$ted Po$er &nd 'ree thro ghp t Stor&ge &nd retriev&l ti%e

,&(&( System Layout

A syste% l&yo t is si%ply the con'ig r&tion pl&n 'or & syste%Es str ct r&l ele%ents. <he l&yo t speci'ies $here to loc&te pieces o' e8 ip%ents, &isles, rep&ir nits, %&teri&l/h&ndling syste%s, stor&ge nits, lo&ding &nd nlo&ding doc"s, &nd so on. Si%il&r to syste% str ct re, pl&cing &nd si.ing %odel ele%ents &ccording to the l&yo t speci'ied res lts in & %ore represent&tive 1?S %odel. Stic"ing to l&yo t speci'ic&tions helps c&pt re the 'lo$ p&th o' %&teri&l or entities $ithin the syste%. !ence, 'lo$ di&gr&%s &re o'ten developed sing syste% l&yo ts. When designing ne$ syste%s or e(p&nding e(isting ones, the l&yo t o'ten pl&ys &n i%port&nt role in &ssessing design &ltern&tives. ?(&%ples o' 8 estions &ns$ered in & syste% l&yo t incl de9 Wh&t is the syste% str ct reK Wh&t does the syste% loo" li"eK !o$ close is the p&c"&ging dep&rt%ent to the shipping doc"K !o$ l&rge is the stor&ge &re&K >&cility pl&nning is the topic nder $hich the l&yo t o' & pl&nt or & '&cility is designed. 1ep&rt%ent &re&s &nd &ctivity/rel&tionship ch&rts &re o'ten sed to provide & design 'or & '&cility l&yo t. ;oc&tions o' dep&rt%ents, dist&nces bet$een the%, &nd interdep&rt%ent&l 'lo$ need to be c&pt red in the 1?S %odel to provide &cc r&te syste% represent&tion.

,&(&* System Data

0e&l/$orld syste%s o'ten involve & tre%endo s &%o nt o' d&t& $hile ' nctioning. 1&t& collection syste%s 4%&n &l or & to%&tic5 &re o'ten sed to collect critic&l d&t& 'or v&rio s p rposes, s ch &s %onitoring o' oper&tions, process control, &nd gener&ting %&n&ge%ent reports. 1?S %odels &re d&t&/ drivenO hence, pertinent syste% d&t& sho ld be collected &nd sed in the %odel. J' co rse, not &ll re&l syste% d&t& &re necess&ry to b ild &nd r n the 1?S %odel. 1eciding on $h&t type o' d&t& is necess&ry 'or 1?S is highly dependent on the %odel str ct re &nd the go&l o' si% l&tion. Gener&lly spe&"ing, &ll syste% ele%ents &re de'ined sing syste% d&t&. ?(&%ples incl de p&r&%eter settings o' %&chines, %&teri&l/h&ndling syste%s, &nd stor&ge syste%s. S ch p&r&%eters incl de inter&rriv&l ti%es, cycle ti%es, tr&ns'er ti%es, conveying speed, ti%e to '&il re 42<@>5, ti%e to rep&ir 42<<05, prod ct %i(, de'ect &nd scr&p r&tes, &nd others. 2odel per'or%&nce &nd res lts &re highly dependent on the 8 &lity &nd &cc r&cy o' s ch d&t&, b&sed on the co%%only sed ter% gar'age,in,gar'age,out. <h s, developing & 1?S %odel re8 ires the si(/sig%& te&% to de'ine precisely the d&t& ele%ents needed, the %ethod o' collecting s ch d&t&, &nd ho$ the d&t& $ill be represented &nd sed in the 1?S %odel. Ch&pter 7 h&s shed light on the "ey topic o' inp t %odeling &nd st&tistic&l distrib tions sed in 1?S %odels. <&ble 6.2 presents the d&t& th&t need to be collected 'or %odeling v&rio s si% l&tion ele%ents. +a3le ,&(# Data Colle$ted for 9arious DES Elements Jpen t&ble &s spre&dsheet Element )odeled 2&chine Conveyor !ertinent Simulation Data 4;o&d, cycle, nlo&d5 ti%e, 2<@>,2<<0 Speed, c&p&city, type, &cc % l&tion

+a3le ,&(# Data Colle$ted for 9arious DES Elements Jpen t&ble &s spre&dsheet Element )odeled Jper&tor @ ''er A to%&ted g ided vehicle Po$er/&nd/'ree syste% P&rt:lo&d !ertinent Simulation Data W&l" speed, $or" se8 ence, $&l" p&th C&p&city, discipline, inp t:o tp t r les Speed, &cceler&tion:deceler&tion, ro te Speed, dog sp&cing, ch&in length Attrib tes o' si.e, color, 'lo$, %i(

,&(&% System Logi$

Syste% logic co%prises the r les &nd proced res th&t govern the beh&vior &nd inter&ction o' v&rio s ele%ents in & si% l&tion %odel. It de'ines the rel&tionships &%ong %odel ele%ents &nd ho$ entities 'lo$ $ithin & syste%. <he progr&%%ing c&p&bility o' si% l&tion l&ng &ges is o'ten tili.ed to i%ple%ent the syste% logic designed into the 1?S %odel developed. Si%il&rly, re&l/ $orld syste%s o'ten involve & set o' si%ple or co%ple( logic&l designs th&t control syste% per'or%&nce &nd direct its beh&vior. Abstr&cting relev&nt logic into & 1?S %odel is & critic&l %odeling t&s". It is prob&bly the %ost di''ic lt %odeling ch&llenge th&t '&ces si% l&tion si(/sig%& te&%s, especi&lly $ith li%it&tions o' so%e si% l&tion so't$&re p&c"&ges. It is $orth re%e%bering, ho$ever, th&t co%ple( logic is & %&in driver behind the co%%on&lity o' si% l&tion %odeling in prod ction &nd b siness &pplic&tions. In & typic&l si% l&tion %odel, it is o'ten the c&se th&t sever&l decision points e(ist $ithin the %odel oper&tions, s ch &s splitting &nd %erging points. At these decision points, cert&in sched ling r les, ro ting sche%es, &nd oper&tion&l se8 ences %&y need to be b ilt into the 1?S %odel to re'lect the &ct &l per'or%&nce o' the nderlying syste%. <&ble 6.3 provides e(&%ples o' s ch logic&l designs. +a3le ,&*# Exam les of )odel Logi$al Designs Jpen t&ble &s spre&dsheet )odel A$ti4ity P&rts &rriving &t lo&ding doc" 0e8 esting co%ponents Prod cing &n order 2&teri&l h&ndling St&tistic&l process control 2&chining & p&rt >or"li't 'loor oper&tion Stor&ge &nd retriev&l syste% 4AS:0S5 Logi$al Design Sorting &nd inspection sche%e 2odel %i( r les 2&chine sched ling r les C&rrier ro ting r les 1ecision r les Se8 ence o' oper&tion 1riversE disp&tching r les AS:0S vehicle %ove%ent r les

,&(&, System Statisti$s

Syste% st&tistics &re %e&ns o' collecting r n/ti%e in'or%&tion &nd d&t& 'ro% & syste% d ring r n ti%e &nd &ggreg&ting the% &t the end o' si% l&tion r n ti%e. 1 ring r n ti%e, s ch st&tistics &re necess&ry to control the oper&tion &nd 'lo$ o' syste% &ctivities &nd ele%ents. At si% l&tion end, these st&tistics &re collected to s %%&ri.e syste% per'or%&nce &t v&rio s syste% design &nd p&r&%eter settings. In & syste% 1?S %odel, there'ore, st&tistics &re collected &nd &cc % l&ted to provide & s %%&ry o' res lts &t the end o' r n ti%e. S ch st&tistics &re sed to %odel re&l/ti%e %onitoring g& ges &nd cloc"s in & re&l/ $orld syste%. @ec& se o' %odel 'le(ibility, ho$ever, so%e st&tistics th&t &re sed in the %odel %&y not &ct &lly be in the re&l/$orld syste%. <his is bec& se st&tistics do not &''ect %odel per'or%&nce. <here'ore, $e c&n de'ine st&tistics th&t &re necess&ry to syste% oper&tion &nd other st&tistics th&t %&y provide se' l in'or%&tion d ring r n ti%e &nd s %%&ri.e the res lts &t the end o' r n ti%e. Appendi( A covers st&tistics collected 'ro% si% l&tion &nd o tp t &n&lyses per'or%ed sing s ch st&tistics. <&ble 6.7 provides e(&%ples o' s ch st&tistics. +a3le ,&%# Exam les of model Statisti$s Jpen t&ble &s spre&dsheet )odel Statisti$ Jobs prod ced per ho r Percent o' %&chine b sy ti%e # %ber o' nits in syste% <i%e nits spend in syste% # %ber o' de'ectives # %ber o' %&chine '&il res # %ber o' nits on & conveyor # %ber o' nits on & b ''er 9alue )easured Syste% thro ghp t 2&chine tili.&tion Wor"/in/progress level 2&n '&ct ring le&d ti%e Process 8 &lity 2&inten&nce pl&n Conveyor tili.&tion @ ''er tili.&tion

,&* Elements Of Dis$rete E4ent Simulation

A 1?S %odel is b ilt sing & set o' %odel co%ponents 4b ilding bloc"s5 &nd is e(ec ted b&sed on & set o' 1?S %ech&nis%s. <he n&t re &nd ' nction&lity o' these co%ponents &nd %ech&nis%s %&y v&ry 'ro% one si% l&tion p&c"&ge to &nother. !o$ever, b&sic co%ponents o' 1?S %odels &re o'ten de'ined sing & co%%on set o' 1?S ele%ents. <hese ele%ents $ere de'ined in Ch&pter 7 in ter%s o' syste% inp ts, process 4ele%ents &nd rel&tionships5, o tp ts, &nd controls. !ence, 1?S ele%ents incl de syste% entities 4 .5, o'ten ch&r&cteri.ed by & set o' &ttrib tes 4/T5. <he syste% st&te 4S5 is described by st&te v&ri&bles 40Rs5. <he syste% st&te ch&nges pon the occ rrence o' & set o' events 4 5. 1 ring oper&tion, di''erent syste% &ctivities 4 /5 or t&s"s &re per'or%ed by syste% reso rces 4R5. >in&lly, the %odel involves del&ys 4*5 th&t t&"e pl&ce in the syste% &nd the logic 415 th&t governs the syste% oper&tion. <hese ele%ents &re sho$n in >ig re 6.2. In this section $e disc ss these ele%ents &nd in the ne(t section disc ss "ey %ech&nis%s sed to e(ec te 1?S %odels.

Figure ,&(# 1?S syste% ele%ents.

,&*&' System Entities

?ntities &re ite%s th&t enter & syste% &s inp ts, &re processed thro gh the syste% reso rces &nd &ctivities, &nd dep&rt the syste% &s o tp ts. As 1?S dyn&%ic obDects, entities &re %odel/ tr&ce&ble ele%ents th&t &re o'ten o' interest to syste% designers, %&n&gers, &nd si(/sig%& oper&tives s ch &s green belts, bl&c" belts, &nd %&ster bl&c" belts. ?(&%ples incl de p&rts or prod cts in %&n '&ct ring, c sto%ers in b&n"ing, c&lls in & c sto%er/service center, p&tients in he&lth syste%s, letters &nd p&c"&ges in post&l services, doc %ents in &n o''ice, ins r&nce policies in &n ins r&nce co%p&ny, d&t& in &n in'or%&tion syste%, &nd so on. In 1?S, entities &re ch&r&cteri.ed by &ttrib tes s ch &s price, type, cl&ss, color, sh&pe, I1 n %ber, origin, destin&tion, priority, d e d&te, &nd so on. Speci'ic v&l es o' s ch &ttrib tes &re tied to entities &nd c&n di''er 'ro% one entity to &nother. !ence, &ttrib tes c&n be considered &s local v&ri&bles th&t &re tied to individ &l entities. <hese &ttrib tes c&n be sed &t di''erent loc&tions &nd inst&nces $ithin the 1?S %odel to %&"e v&rio s decisions 'or directing the 'lo$ o' entities, &ssigning the% to stor&ge loc&tions, &ctiv&ting reso rces, &nd so on. Also, the type o' entity is the b&sis 'or cl&ssi'ying 1?S syste%s bet$een discrete &nd contin o s. 1iscrete entities &re %odeled $ith discrete syste%s. Contin o s entities s ch &s 'lo$ o' b l" %&teri&ls, 'l ids, &nd g&ses &re %odeled $ith contin o s syste%s.As %entioned e&rlier, o r 'oc s in this boo" is on discrete entities since they represent the %&Dority o' tr&ns&ction/b&sed syste%s in both %&n '&ct ring &nd service ind stries.

,&*&( System State

<he syste% st&te is & description o' syste% st&t s &t &ny point in ti%e d ring & si% l&tion. It describes the condition o' e&ch syste% co%ponent in the 1?S %odel. As sho$n in >ig re 6.2, these ele%ents incl de syste% inp ts 4I5, process 4P5, o tp ts 4O5, &nd controls 4#5. !ence, &t &ny point in ti%e, the syste% st&te de'ines the st&te o' syste% inp ts 4i.e., types, &%o nt, %i(, speci'ic&tions, &rriv&l process, so rce, &ttrib tes5. It &lso de'ines the syste% process in ter%s o' type o' &ctivities 4/5, n %ber o' &ctive reso rces 4R5, n %ber o' nits in the 'lo$, tili.&tion o' reso rces, ti%e in st&te, del&y ti%e, ti%e in syste%, &nd so on. Si%il&rly, & syste% st&te de'ines the st&te o' %odel o tp ts 4i.e., types, &%o nt, %i(, speci'ic&tions, dep&rt re process, destin&tion, &ttrib tes, v&ri&bility5. <&ble 6.6 incl des e(&%ples o' syste% st&tes th&t typic&lly t&"e pl&ce in di''erent syste% ele%ents. +a3le ,&,# exam les of states of different model Elements Jpen t&ble &s spre&dsheet )odel Element 2&chine System State @ sy, do$n, idle, bloc"ed

+a3le ,&,# exam les of states of different model Elements Jpen t&ble &s spre&dsheet )odel Element @ ''er Conveyor ;&bor Lehicle System State ?%pty, ' ll, h&l'/' ll ?%pty, ' ll, bro"en, bloc"ed, st&rved Idle b sy, $&l"ing 2oving, stopped, bro"en, p&r"ing

1?S records ch&nges in & syste% st&te &s it evolves over ti%e &t & discrete point in ti%e to provide & represent&tion o' syste% beh&vior &nd to collect st&tistics &nd per'or%&nce %e&s res th&t &re essenti&l to syste% design &nd &n&lysis. Syste% st&te v&ri&bles &re sed to 8 &nti'y the description o' & syste% st&te.

,&*&* State 9aria3les

1?S %odels incl de & collection o' v&ri&bles th&t describe the syste% st&te &t &ny speci'ic point in ti%e. S ch v&ri&bles cont&in the in'or%&tion needed to describe & %odel co%ponent st&t s &nd %e&s re its per'or%&nce. ?(&%ples incl de the n %ber o' nits in the syste%, the percent&ge o' e&ch reso rce st&t s 4i.e., Q idle, Q b sy, Q bro"en, Q bloc"ed5, &nd the n %ber o' b sy or idle oper&tors. In &ddition to the over&ll syste%/level v&ri&bles, %odel v&ri&bles incl de inp t '&ctors 4 &), &2, . . . , &n5. <hey represent & set o' independent v&ri&bles $hose ch&nged v&l es &''ect syste% beh&vior, &nd di''erent settings o' s ch '&ctors o'ten le&d to di''erent sets o' %odel o tco%es. ?(&%ples incl de inter/ &rriv&l ti%es, n %ber o' oper&tors &nd reso rces, service ti%e, cycle ti%e, c&p&city o' b ''ers, &nd speeds o' conveyors. 2odel v&ri&bles &lso incl de syste% response 4o tp t5 v&ri&bles 4%), %2, . . . , %n5, $hich represent dependent %e&s res o' syste% per'or%&nce. ?(&%ples incl de syste% thro ghp t, &ver&ge tili.&tion o' reso rces, &nd %&n '&ct ring le&d ti%e. >ig re 6.3 is & sche%&tic represent&tion o' st&te v&ri&bles.

Figure ,&*# Syste% st&te v&ri&bles.

,&*&% System E4ents

An event is &n inst&nt&neo s occ rrence th&t ch&nges the syste% st&te. As disc ssed e&rlier, the syste% st&te is the description o' syste% st&t s &t &ny ti%e, $hich is de'ined by & set o' st&te v&ri&bles. <he event is the "ey ele%ent in 1?S %odels since the %odels &re ch&r&cteri.ed by being event/driven. -pd&ting the syste% st&te, collecting syste% st&tistics, &nd &dv&ncing the si% l&tion cloc" t&"e pl&ce &t event occ rrence. <he set o' events 4 ), 2, 3, . . . , n5 th&t occ r &t cert&in corresponding ti%es 4T), T2, T3, . . . , <n5 &re stored chronologic&lly in &n event list 4disc ssed l&ter in 1?S %ech&nis%s5. Syste% st&tes 4S), S2, S3, . . . , Sn5 &re ch&nged b&sed on

the events &nd their i%plic&tionsO hence, %odel beh&vior is re'erred to &s event/driven. >ig re 6.7 sho$s ho$ & syste% st&te is pd&ted &t event occ rrences.

Figure ,&%# ?vent/driven syste% st&te. ?vent occ rrence in & 1?S %odel c&n be the &rriv&l o' &n entity, the st&rt o' & co''ee bre&", the end o' & shi't, the '&il re o' & reso rce, & ch&nge in b&tch si.e, the st&rt o' & ne$ prod ction sched le, the dep&rt re o' &n entity, &nd so on. Since 1?S %odel ele%ents &re interrel&ted &nd 1?S environ%ents &re dyn&%ic, the occ rrence o' s ch events o'ten le&ds & series o' ch&nges to the syste% st&te. >or e(&%ple, the &rriv&l o' & c sto%er &t & b&n" incre&ses the $&iting line i' the b&n" teller is b sy, or ch&nges the teller st&t s 'ro% idle to b sy i' & teller is &v&il&ble. Si%il&rly, $hen the c sto%er dep&rts the b&n", the server is ch&nged b&c" to idle, &nother c sto%er is re8 ested 'ro% the $&iting line, &nd the n %ber o' c sto%ers served is incre&sed. !ence, the st&te v&ri&bles &''ected &re pd&ted, relev&nt st&tistics &re &cc % l&ted, &nd the si% l&tion cloc" is &dv&nced to the ne(t event in the event list.

,&*&, System A$ti4ities

An &ctivity is & t&s" per'or%ed in & %odel th&t h&s & speci'ied ti%e d r&tion. Activities &re deter%ined si%ply by est&blishing & se8 ence 'or the oper&tions needed to process &n entity. ?(&%ples o' s ch oper&tions incl de receiving, directing, tr&ns'erring, cle&ning, %&chining, p&c"&ging, &nd shipping. S ch &ctivities either process the entity 4v&l e/&dded5 directly, s ch &s c tting &nd serving, or indirectly, s ch &s %&teri&l h&ndling, inspection, &nd stor&ge &ctivities 4non/v&l e/&dded5. A process %&p o'ten incl des & se8 ence o' &ll &ctivities or oper&tions re8 ired to process &n entity, &long $ith speci'ic&tions &nd the cl&ssi'ic&tion o' di''erent types o' oper&tions. A process ch&rt o'ten cl&ssi'ies oper&tions into process 4& circle5, tr&nsport 4&n &rro$5, stor&ge 4& tri&ngle5, del&y 415, &nd inspection 4& s8 &re5, &s sho$n in <&ble 6.6. Si%il&rly, &ctivities in & 1?S %odel c&n be cl&ssi'ied to provide & better nderst&nding o' the process 'lo$ &nd & deeper "no$ledge o' process oper&tions. +a3le ,&-# A$ti4ities in a !ro$ess O erations Chart Jpen t&ble &s spre&dsheet Sym3ol !ro$ess )eaning Processing oper&tion Exam le 1rilling oper&tion

<r&nsporting oper&tion

>or"li't tr&ns'ers & nit lo&d to stor&ge

Stor&ge oper&tion

>inished goods &re stored be'ore shipping Wor"/in/process nits &re del&yed be'ore 'in&l &sse%bly

1el&y oper&tion

+a3le ,&-# A$ti4ities in a !ro$ess O erations Chart Jpen t&ble &s spre&dsheet Sym3ol !ro$ess )eaning Inspection oper&tion Exam le Jne o' every )00 nits is inspected &t & 8 &lity control st&tion

Syste% &ctivities c&n be &lso cl&ssi'ied &s v&l e/&dded or non/v&l e/&dded9 ). 0alue,added 4LA5 acti-ities. L&l e is &dded thro gh &ctivities th&t tr&ns'or% the ch&r&cteristics o' &n entity 'ro% one 'or% to &nother.L&l e is de'ined by the c sto%er. @y %&"ing ch&nges to entities, s ch &ctivities incre&se their v&l es. !ence, the price o' & ton o' rein'orce%ent steel is % ch higher th&n the price o' & ton o' steel billet. Steel billets &re he&ted, 'or%ed, 8 enched, cooled, &nd c t be'ore beco%ing rein'orce%ent steel. S ch oper&tions gr&d &lly incre&se the v&l e o' steel entities. 2. .on,-alue,added 4#LA5 acti-ities. 2&ny oper&tions per'or%ed on entities %&y not &dd v&l e to the% b t they &re still needed to co%plete the process. ?(&%ples incl de tr&nsporting &nd tr&ns'erring oper&tions, storing &nd del&ying %&teri&ls, &nd 8 &lity control inspections. <he re&der is enco r&ged to revisit Ch&pter 2 'or the le&n si(/sig%& perspectives &bo t #LAs. <he ti%e d r&tion 'or syste% &ctivities is speci'ied in three $&ys, depending on the n&t re o' the &ctivity9 ). Fi2ed time duration. <i%e &ssigned to &n &ctivity h&s & 'i(ed v&l e &nd possesses no v&ri&bility. <ypic&l e(&%ples incl de 'i(ed cycle ti%es in & to%&tic oper&tions s ch &s C#C %&chining, pl&nned stopp&ge ti%es, &nd ti%ed inde(ing oper&tions. 2. Pro'a'ilistic time duration. <he ti%e &ssigned in this c&se incorpor&tes r&ndo%ness &nd v&ri&bilityO hence, the &ctivity ti%e ch&nges 'ro% one entity to &nother. S&%pling 'ro% theoretic&l or e%piric&l st&tistic&l distrib tions is o'ten sed to represent &ctivity ti%e d r&tion. 3. Formula,'ased time duration. In this c&se the &ctivity ti%e is c&lc l&ted sing &n e(pression o' cert&in syste% v&ri&bles. >or e(&%ple, lo&ding ti%e is deter%ined b&sed on the n %ber o' p&rts lo&ded, $&l"ing ti%e is deter%ined &s & ' nction o' lo&d $eight, &nd %&chine cycle ti%e is & ' nction o' c tting p&r&%eters.

,&*&- System Resour$es

A syste% reso rce represents the tool or %e&n by $hich %odel &ctivities &re c&rried o t. ?(&%ples incl de pieces o' e8 ip%ent, oper&tors, personnel, physici&ns, rep&irpersons, %&chines, speci&li.ed tools, &nd other %e&ns th&t '&cilit&te the processing o' entities. 0eso rce &lloc&tion &nd sched ling sche%es &re o'ten i%ple%ented to provide best t&s" &ssign%ents to reso rces. <&ble 6.* incl des e(&%ples o' syste% reso rces in %&n '&ct ring &nd service &pplic&tions. +a3le ,&.# Exam les of System Resour$es Jpen t&ble &s spre&dsheet )anufa$turing System 2&chines &nd %&chine centers Jper&tors &nd gener&l l&bor Inspectors &nd 8 &lity controllers 0ep&irpersons &nd %&inten&nce cre$s Physici&ns @&n" tellers !ospit&l beds 1rivers Ser4i$e System

+a3le ,&.# Exam les of System Resour$es Jpen t&ble &s spre&dsheet )anufa$turing System Asse%bly st&tions &nd tools Ser4i$e System 1rive/thro gh $indo$s

Pey '&ctors th&t &''ect the per'or%&nce o' reso rces incl de c&p&city, speed, &nd reli&bility. C&p&city &''ects reso rce tili.&tion, $hich %e&s res the percent&ge o' reso rce se. Since reso rces cons %e c&pit&l, syste% designers &nd %&n&gers pre'er to incre&se the tili.&tion o' reso rces thro gh better sched ling &nd reso rce &lloc&tion. 0eso rce speed deter%ines the prod ctivity o' the reso rce, o'ten %e&s red &s thro ghp t or yield. Si%il&rly, reso rce thro ghp t is & l&rger/the/better %e&s re. ?li%in&ting $&ste in ti%e &nd ine''iciencies in oper&tion incre&ses thro ghp t. >in&lly, reso rce reli&bility deter%ines the percent&ge o' reso rce pti%e 4&v&il&bility5. It is &l$&ys re8 ired to incre&se the pti%e percent&ge o' reso rces thro gh better %&inten&nce &nd $or"lo&d b&l&ncing. <&ble 6.+ s %%&ri.es the three reso rce '&ctors9 +a3le ,&/# Fa$tors Affe$ting System Resour$es Jpen t&ble &s spre&dsheet Resour$e Fa$tor C&p&city Speed 0eli&bility ). !erforman$e )easure -tili.&tion <hro ghp t -pti%e :nit )etri$ Percent b sy -nits prod ced per ho r 4-P!5 2e&n ti%e v&l e bet$een Exam le 22 is +6Q tili.ed 22 prod ces 60 -P! 22 h&s & 2<@> '&il res o' 200 ho rs

Resource capacity9 & '&ctor th&t is o'ten %e&s red by ho$ %&ny entities &re &ble to &ccess or enter & reso rce. 2ost reso rces, s ch &s %&chines, b&n" tellers, &nd physici&ns, o'ten tre&t one entity &t & ti%e. So%e, ho$ever, per'or% b&tch processing o' identic&l or di''erent entities, s ch &s 'illing %&chines, &sse%bly st&tions, testing centers, &nd gro p/b&sed &ctivities. C&p&city li%it&tions o'ten &''ect the tili.&tion o' reso rces, $here syste% designers o'ten strive to stri"e & b&l&nce bet$een %e&s res. <here'ore,

2.

Resource speed9 & '&ctor th&t deter%ines the thro ghp t or yield o' & reso rce. >&st reso rces o'ten r n $ith & short processing ti%e &nd res lt in processing %ore entities in the ti%e nit. ?(&%ples incl de the speed o' & tr&ns'er line, the cycle ti%e o' & %&chining center, the service ti%e &t & b&n" teller, &nd the di&gnostic ti%e &t & clinic. -nits prod ced per ho r 4-P!5 or per shi't 4-PS5 &re co%%on thro ghp t %e&s res in %&n '&ct ring syste%s. Cycle ti%e is o'ten the ter% th&t is sed to indic&te the speed o' %&chines &nd reso rces in gener&l. 2&chine thro ghp t is there'ore, the reciproc&l o' cycle ti%e. <h&t is,

>or e(&%ple, & %&chine $ith & cycle ti%e o' 60 seconds prod ces ):60 nit per second or ) nit per %in te. <&"t ti%e is o'ten sed to indic&te the speed o' & to%&tic prod ction lines. 3. Resource relia'ility3 & reso rce '&ctor th&t deter%ines the reso rce pti%e or &v&il&bility. It is o'ten %e&s red in ter%s o' %e&n ti%e bet$een '&il res 42<@>5, $here & reso rce '&il re is e(pected to occ r &t 2<@> ti%e nits. 0ep&ir ti%e, &'ter & '&il re occ rs, is %e&s red by %e&n ti%e to rep&ir 42<<05. <he reso rce pti%e percent&ge is deter%ined by dividing 2<@> by &v&il&ble ti%e 4both 2<@> &nd 2<<059

,&*&. System Delay

A syste% del&y is &n &ctivity th&t t&"es pl&ce $ithin the syste% b t does not h&ve & speci'ied ti%e d r&tion. <he d r&tion is deter%ined d ring r n ti%e b&sed on dyn&%ic &nd logic&l inter&ctions &%ong syste% ele%ents. ?(&%ples incl de c sto%er $&iting ti%e, del&ys in nit se8 encing, &nd del&ys c& sed by logic&l design. 2e&s ring del&ys by &cc % l&ting the ti%e o' del&y occ rrences is the principle &dv&nt&ge o' 1?S %odels. St&tistics i%port&nt to decision %&"ers c&n be collected sing this c&p&bility9 'or e(&%ple, the &ver&ge $&iting ti%e 'or b&n" c sto%ersO %&n '&ct ring le&d ti%eO &nd ti%e sp&n 'ro% the st&rt to the end o' p&rt %&n '&ct ring, incl ding del&ys.

,&*&/ System Logi$

Syste% logic controls the per'or%&nce o' &ctivities &nd dict&tes the ti%e, loc&tion, &nd %ethod o' their e(ec tion. 1?S %odels r n by e(ec ting logic&l designs s ch &s r les 'or reso rces &lloc&tion, p&rts se8 encing, 'lo$ ro ting, t&s" prioriti.&tion, &nd $or" sched ling. 0o te sheets, $or" sched les, prod ction pl&ns, &nd $or" instr ctions &re e(&%ples o' %ethods sed to i%ple%ent the syste% logic&l design. <&ble 6., s %%&ri.es e(&%ples o' syste% logic. +a3le ,&0# Exam les of System Logi$ Jpen t&ble &s spre&dsheet Logi$al Design 2&teri&l 'lo$ Wor"'orce sched ling Prod ct/%i( sched ling <&s" prioriti.ing St&nd&rd oper&tions "m lementation )ethod 0o te sheet Wor" sched le Prod ction pl&n Shop 'loor controls Wor" instr ctions Exam le 2&chining se8 ence # rsing shi'ts in hospit&ls Prod ction lot si.e ?(pedited processing Prod ct &sse%bly

Si% l&tion p&c"&ges o'ten provide b ilt/in ro ting &nd disp&tching r les 'or si% l&tion entities. <hey &lso provide &n ?nglish/li"e synt&( to progr&% the %odel logic. Si% l&tion p&c"&ges o'ten %&"e se o' ind stri&l &nd b siness ter%inology &s synt&(9 'or e(&%ple, process, p&rt, l&bor, %&chine, conveyor, c&rrier, &nd so on. !o$ever, si% l&tion so't$&re tools o'ten v&ry in the c&p&bility o' their logic&l progr&%%ing pl&t'or%. While so%e tools provide open editors 'or $riting co%plic&ted logic in & C/li"e or L@/li"e progr&%%ing str ct re, others provide li%ited progr&%%ing c&p&bility 'or %&"ing the %odel b ilding e&sier &nd '&ster. ;i"e &ny other progr&%%ing l&ng &ge, $riting the logic in si% l&tion re8 ires developing & 'lo$ch&rt o' the logic&l design &nd tr&nsl&ting the logic&l design into code sing the so't$&re synt&(. Since si% l&tion logic is typic&lly developed to %i%ic re&l/$orld oper&tion&l logic, &

thoro gh &nd pr&ctic&l nderst&nding o' the nderlying prod ction or b siness process is &n essenti&l ingredient 'or developing represent&tive si% l&tion logic.

,&% DES )e$hanisms

As disc ssed e&rlier, 1?S %odels &re dyn&%ic, event/driven, discrete in ti%e, stoch&stic 4r&ndo%i.ed &nd prob&bilistic5, &nd co%p ter/&ni%&ted. S ch ch&r&cteristics &re est&blished in the 1?S %odel b&sed on cert&in po$er' l %ech&nis%s. 2&in 1?S %ech&nis%s incl de the cre&tion &nd pd&ting o' &n events list, the ti%e/&dv&nce%ent %ech&nis%, the c&p&bility o' s&%pling 'ro% prob&bility distrib tions $ith r&ndo% n %ber gener&tion, the c&p&bility o' &cc % l&ting st&tistics over r n ti%e, &nd the po$er o' t$o/ or three/di%ension&l dyn&%ic gr&phic&l represent&tion $ith &ni%&tion %ech&nis%. <hese %ech&nis%s represent the b&c"bone o' 1?S ' nction&lity in si% l&tion so't$&re tools since they &re critic&l to cre&ting oper&tion&l %odels. Altho gh si% l&tion so't$&re tools v&ry in %ethods &nd &lgorith%s to i%ple%ent these %ech&nis%s, the 1?S ' nction&lity o'ten incl des the cre&tion o' &n event list, &dv&ncing the si% l&tion cloc" 4event/driven &dv&nce%ent5, pd&ting the event list, pd&ting pertinent st&tistics, &nd chec"ing 'or ter%in&tion. >ig re 6.6 is sho$s & 'lo$ch&rt o' 1?S ' nction&lity.

Figure ,&,# 1?S ' nction&lity. ?&ch 1?S %ech&nis% involves %&ny theoretic&l det&ils, &nd in this section $e disc ss the "ey ch&r&cteristics &nd %ethods sed in these %ech&nis%s.

,&%&' Dis$rete E4ent )e$hanism

?vents in & 1?S &re discrete since they t&"e pl&ce &t discrete points in ti%e. 1?S ' nction&lity is, there'ore, b&sed on cre&ting &n event list &nd e(ec ting the events o' the list chronologic&lly. S ch ' nction&lity is b&sed on & discrete event &nd cloc"/&dv&nce%ent %ech&nis%. <he discrete event %ech&nis% is the %ost distinctive 'e&t re o' 1?S si% l&tion %odels. As disc ssed e&rlier, &n e-ent is de'ined &s &n occ rrence, & sit &tion, or & condition th&t res lts in ch&nging the state o' the syste% %odel inst&nt&neo sly. <he st&te o' & syste% is represented by & collection o' st&te v&ri&bles th&t describe the syste% st&t s. Arriv&l &nd dep&rt re o' entities s ch &s p&rts &nd c sto%ers &re events th&t ch&nge the st&te o' & prod ction syste% or & service st&tion. In'or%&tion &bo t event types &nd their sched led occ rrence ti%es &re stored in &n event list 4?;5. An event list co%prises & set o' events 4?5 &nd eventCti%e 4<5 co%bin&tion ). <h&t c&n be e(pressed &s 'ollo$s9 ?; = M4 ), T)5, 4 2, T25, 4 3, T35,. . . , 4
n

, Tn5N

At e&ch event occ rrence, t$o &ctions t&"e pl&ce in the %odel9 <he ?; is pd&ted in ter%s o' content &nd ti%e, &nd the st&tistics collected &re pd&ted. -pon the occ rrence o' &n event, the event e(ec ted $ill be re%oved 'ro% the list &nd the ne(t %ost i%%inent event $ill top the list. Jther events %&y enter or le&ve the list &ccordingly. -pd&ting &n event list incl des t$o %&in oper&tions 4>ig re 6.659

Figure ,&-# ?vent list oper&tions. ). -ent remo-al. -pon the occ rrence o' &n event, the list %&y be pd&ted by re%oving one or %ore events 'ro% the list, res lting in ch&nging the chronologic&l order o' event e(ec tion. So%e events %&y be deleted 'ro% the list be'ore their e(ec tion. An event %&y be deleted 'or %&ny re&sons9 'or e(&%ple, $hen the occ rrence o' one event precl des the occ rrence o' &nother. ?vent re%ov&l c&n t&"e pl&ce in t$o loc&tions in &n eventlist9 &. From the top of the 1. When &n event is re%oved 'ro% the top o' &n ?;, it is done or e(ec ted. >or e(&%ple, event ) in >ig re 6.6 $&s processedO hence the event is re%oved 'ro% the top o' ?;) $hen pd&ting the list to beco%e ?;2. b. From any other location 4ithin the 1. When &n event is re%oved 'ro% &ny other loc&tion in the ?; other then the top o' the list, it is c&nceled. >or e(&%ple, event 7 in >ig re 6.6 $&s c&nceledO hence the event is re%oved 'ro% ?;) &nd did not sho$ p in ?;2. 2. -ent addition. -pon the occ rrence o' &n event, the list %&y be pd&ted by &dding one or %ore events to the list, res lting in ch&nging the chronologic&l order o' event e(ec tion. ?vent &ddition c&n t&"e pl&ce in t$o loc&tions in &n ?;9 &. To the top of the 1$ When &n event is &dded to the top o' the ?;, it is considered &s the %ost i%%inent event to t&"e pl&ce. >or e(&%ple, event * in >ig re 6.6 $&s &dded to the top o' ?;2 &nd $ill be processed 'irst. ?* $&s not & %e%ber o' ?;) $hen pd&ting the list to beco%e ?;2. b. To any other location 4ithin the 1$ When &n event is &dded to &ny loc&tion in &n ?; other then the top o' the list, it is D st being &dded &s & ' t re event. >or e(&%ple, event 6 in >ig re 6.6 $&s &dded to ?;2 &nd $ill be processed right &'ter 2. <h s, $ith the discrete event %ech&nis%, $e vie$ & syste% &s progressing thro gh ti%e 'ro% one event to &nother r&ther th&n ch&nging contin o sly. As %entioned e&rlier, %&ny re&l/$orld

syste%s 4e.g., pl&nts, b&n"s, hospit&ls5 oper&te in & discrete %&nner 4tr&ns&ction/ or event/ b&sed5. Since these syste%s o'ten involve $&iting lines, & si%ple discrete event %ech&nis% is o'ten %odeled sing 8 e ing %odels. 1?S %odels c&n be vie$ed &s 8 e ing net$or"s th&t %&y or %&y not be &%en&ble to 8 e ing theory. 1iscrete events occ r in the si% l&tion %odel si%il&r to the $&y they occ r in the 8 e ing syste% 4&rriv&l &nd dep&rt re o' events5. >or inst&nce, & c sto%er Doins & 8 e e &t & discrete inst&nt o' ti%e, &nd &t & l&ter discrete inst&nt the c sto%er le&ves the b&n". <i%e is &cc % l&ted pon the occ rrence o' events &t discrete points in ti%e. <he c sto%er is either in or o t o' the 8 e e. <his vie$ res lts in & co%p t&tion&lly e''icient $&y o' representing ti%e. <he &bility to represent ti%e in &n event/driven %&nner is the gre&test strength o' 1?S since it c&pt res the dyn&%ic beh&vior o' re&l/$orld prod ction &nd b siness syste%s.

,&%&( +ime-Ad4an$ement )e$hanism

<he ti%e &ssoci&ted $ith the discrete events 4T), T2, T3, . . . , Tn5 in & si% l&tion %odel is %&int&ined sing & simulation cloc5 ti%e v&ri&ble. <his v&ri&ble is pd&ted thro gh the ne2t,e-ent time,ad-ancement mechanism, $hich &dv&nces the si% l&tion cloc" to the ti%e o' the %ost i%%inent event in the event list. >or e(&%ple, >ig re 6.* sho$s &n event list o' )0 events 4 ), 2, 3, . . . , )05 ordered chronologic&lly &s T), T2, T3, . . . , T)0. In 1?S, the ti%e periods bet$een events 4e.g., ) &nd 25 &re s"ipped $hen e(ec ting the %odel, res lting in & compressed si% l&tion ti%e. !ence, the tot&l ti%e re8 ired to process the )0 events in the 1?S co%p ter %odel is % ch shorter th&n the &ct &l cloc" ti%e.

Figure ,&.# <i%e &dv&nce%ent &nd co%pression in 1?S. <he ti%e/&dv&nce%ent %ech&nis% si%ply st&tes th&t i' no event is occ rring in & ti%e sp&n, there is no need to observe the %odel &nd pd&te st&tistics, so the ti%e is si%ply s"ipped to the ne(t event sched led in the ?;. <his event/driven ti%e/&dv&nce%ent process contin es ntil so%e speci'ied stopping condition is s&tis'ied. >or e(&%ple, i' & b&n" teller st&rts serving & c sto%er in & b&n", there is no need to tr&c" every second d ring the service ti%e. Inste&d, the cloc" is &dv&nced to the ne(t event, $hich %&y be the end o' the service ti%e i' no other event is e(pected to t&"e pl&ce d ring the service ti%e.

,&%&* Random Sam ling )e$hanism

<he stoch&stic n&t re o' si% l&tion %odels is est&blished by %e&ns o' s&%pling 'ro% prob&bility distrib tions sing r&ndo% gener&tion techni8 es. Stochastic, $hich %e&ns r&ndo%, is incorpor&ted into the %odel sing %&the%&tic&l %ethods to gener&te stre&%s o' pse dor&ndo% n %bers. A r&ndo% n %ber gener&tion 40#G5 %ech&nis% &llo$s 'or c&pt ring the r&ndo% v&ri&tions in re&l/$orld syste%s in 1?S %odels. <he &bility to &llo$ 'or r&ndo%ness is one o' the gre&t strengths o' 1?S.

>ro% & pr&ctic&l point o' vie$, r&ndo% n %bers &re b&sic ingredients in the si% l&tion %odels o' %ost re&l/$orld syste%s. 2&chine cycle ti%es, c sto%er service ti%es, e8 ip%ent '&il res, &nd 8 &lity control tests &re e(&%ples o' r&ndo%ness ele%ents in 1?S %odels. 0&ndo% %odel d&t& inp ts o'ten le&d r&ndo% %odel o tp ts. <h s, si% l&tion o tco%es, representing re&l/$orld syste%s, &re o'ten stoch&stic in n&t re &nd le&d to inconsistent per'or%&nce levels. <he prob&bilistic n&t re o' the ti%e period th&t sep&r&tes s ccessive events $hen e(ec ting & 1?S %odel represents the %erit o' r&ndo%ness in si% l&tion %odels. ?vents 4e.g., c sto%er &rriv&l, %&chine '&il re, prod ct dep&rt re5 c&n occ r &t &ny point in ti%e, res lting in & stoch&stic si% l&tion environ%ent. <o gener&te s ccessive r&ndo% ti%e s&%ples 4 T = T), T2, . . .5 bet$een events, sampling 'ro% prob&bility distrib tions is sed. !ence, %ost si% l&tion l&ng &ges, &s $ell &s generic progr&%%ing l&ng &ges, incl de &n 0#G engine to gener&te event ti%es &nd other r&ndo% v&ri&bles %odeled. >or e(&%ple, inter&rriv&l ti%es o' cert&in entities to & syste% &re gener&ted r&ndo%ly by s&%pling 'ro% & contin o s e(ponenti&l distrib tion $hile the n %ber o' entities &rriving per nit ti%e is r&ndo%ly gener&ted 'ro% & discrete Poisson distrib tion. 2odeling r&ndo%ness in discrete event si% l&tion re8 ires 0#G &nd & %ethod 'or selecting d&t& points r&ndo%ly 'ro% st&tistic&l distrib tions. A line&r congr enti&l gener&tor 4;CG5 is sed 'or 0#G, &nd the inverse %ethod is sed 'or s&%pling 'ro% st&tistic&l distrib tions. RNG with LCG <he p rpose o' &n 0#G %&chine in 1?S is to prod ce & 'lo$ o' n %bers 'ro% & contin o s ni'or% distrib tion bet$een 0 &nd ). S ch n %bers represent r&ndo% &nd independent observ&tions o' & cert&in r&ndo% '&ctor th&t is ni'or%ly distrib ted bec& se o' the e8 &l prob&bility o' the ni'or% r&ndo% v&ri&ble. <hese r&ndo% n %bers c&n &lso be sed to s&%ple 'ro% prob&bility distrib tions other th&n & ni'or% distrib tion. In &ddition to the ;CG %ethod, co%%on 0#G techni8 es incl de co%bined line&r congr enti&l gener&tors. ;CG is, ho$ever, the %ost co%%on &rith%etic oper&tion 'or gener&ting ni'or% 40,)5 r&ndo% n %bers. <he r&ndo%ness o' s&%pling is &l%ost g &r&nteed sing & congruential method. ?&ch r&ndo% n %ber gener&ted sing this %ethod $ill be & deci%&l n %ber bet$een 0 &nd ). Given p&r&%eters 60, a, c, &nd m, & pse dor&ndo% n %ber Ri c&n be gener&ted 'ro% the 'ollo$ing 'or% l&9

$here 6i v&l es &re integers bet$een 0 &nd m / ), b t v&l es o' the r&ndo% n %bers gener&ted th&t &re sed in si% l&tion &re deci%&l n %bers bet$een 0 &nd ).<he oper&tion %od4 m5 %e&ns to divide 4aRi/) S c5 by m &nd ret rn the re%&inder o' the division &s the ne(t 6i$ <he initi&l v&l e 60 is re'erred to &s the seed o' the r&ndo% gener&tor. <he v&l es o' a, c, &nd m const&nts &re selected properly to prod ce & good 'lo$ o' r&ndo% n %bers. >or e(&%ple, i' m T 20, a T )0, c T 6, &nd 60 T 60, the rec rsion gener&ting o' Ri v&l es is Ri T 4)0Ri/) S 65 %od4205, $hich res lts in the 'ollo$ing v&l es9 60 T 60 6) T H)04605 S 6I %od4205 T 606 %od4205 T 6 &nd R) T 6:20 T 0.26 62 T H)04605 S 6I %od4205 T 66 %od4205 T )6 &nd R2 T )6:20 T 0.*6 . . . An 0#G %ech&nis% gener&tes n %bers th&t &re not tr ly r&ndo% bec& se they c&n be gener&ted in &dv&nce &nd bec& se o' the rec rsive n&t re o' ;CG, $here & si%il&r se8 ence o'

r&ndo% n %bers c&n be repe&ted &g&in &nd &g&in. >or this re&son, 0#G %ethods &re o'ten c&lled pseudo,R.!. <&bles o' re&dy/to/ se r&ndo% n %bers &re &v&il&ble, especi&lly 'or 2onte C&rlo si% l&tions, $here r&ndo% s&%pling is sed to esti%&te cert&in e(peri%ent&l o tp t. Also, %ost si% l&tion so't$&re p&c"&ges h&ve the c&p&bility o' & to%&tic r&ndo% n %ber gener&tion. Appendi( @ presents & t&ble o' r&ndo% n %bers. ampling with the !n"erse Method -sing &n 0#G $ith ni'or%40,)5 v&l es, s ccessive r&ndo% s&%ples c&n be gener&ted 'ro% prob&bility distrib tions sing three %ethods9 ). In-erse method3 sed 'or s&%pling 'ro% e(ponenti&l, ni'or%, Weib ll, &nd tri&ng l&r prob&bility distrib tions &s $ell &s e%piric&l distrib tions. 2. #on-olution method3 sed 'or s&%pling 'ro% ?rl&ng, Poisson, nor%&l, &nd bino%i&l distrib tions. 3. /cceptance"re7ection techni8ue3 sed 'or s&%pling 'ro% bet& &nd g&%%& distrib tions. <he inverse %ethod involves prob&bility distrib tions $ith & closed/'or% c % l&tive density ' nction 4C1>5, s ch &s the e(ponenti&l &nd ni'or% distrib tions. <he %ethod consists o' t$o steps9 r&ndo% n %ber gener&tion o' R &nd co%p ting the r&ndo% v&ri&ble 425 'ro% the C1> th&t corresponds to R. Since the R v&l es &re &ll bet$een 0 &nd ), the C1> v&l es, or f425 = P4y 25 &re &lso bet$een 0 &nd ), $here 0 f425 ). <he v&l e o' 2 is co%p ted by deter%ining the inverse o' C1> &t R or 2 = F)4R5. <he prob&bility density ' nction 4P1>5 o' the e(ponenti&l distrib tion is f(2) T e/2 $here 2 9 0 <o deter%ine & r&ndo% s&%ple 425 'ro% f425, $e 'irst deter%ine the C1> o' the e(ponenti&l distrib tion by integr&ting f425 'ro% 0 to 2, $hich res lts in f(2) T ) / e/2 $here 2 9 0 <he r&ndo% s&%ple is then deter%ined &s 2 = F)4R5, $here R is gener&ted $ith &n 0#G. <his res lts in the 'ollo$ing9

>or e(&%ple, 'ive c sto%ers 4 = 65 enter & grocery store e&ch ho r, &nd 'or R = 0.*6, the ti%e period ntil the ne(t c sto%er &rriv&l is t = 4):65 ln 4) 0.*65 = 0.2** ho r or )6.62 %in tes. <he convol tion %ethod e(presses the r&ndo% s&%ple &s & st&tistic&l s % o' other e&sy/to/ s&%ple r&ndo% v&ri&bles. ?(&%ples incl de the prob&bility distrib tions o' & s % o' t$o or %ore independent r&ndo% v&ri&bles, s ch &s Poisson &nd ?rl&ng, $hose s&%ple c&n be obt&ined 'ro% the e(ponenti&l distrib tion s&%ple. >or e(&%ple, the m/?rl&ng r&ndo% v&ri&ble is de'ined &s the st&tistic&l s % o' m independent &nd identic&lly distrib ted e(ponenti&l r&ndo% v&ri&bles. !ence, the m/?rl&ng s&%ple is co%p ted &s

S&%pling 'ro% & nor%&l distrib tion c&n &lso be &chieved $ith convol tion %ethods, since &s the centr&l li%it theore% st&tes, the s % o' n independent &nd identic&lly distrib ted r&ndo% v&ri&bles beco%es &sy%ptotic&lly nor%&l &s n beco%es s ''iciently l&rge.

<he &ccept&nceCreDection %ethod is developed 'or distrib tions $ith co%ple( P1>s, $here the co%ple( P1> f425 is repl&ced by & %ore &n&lytic&lly %&n&ge&ble pro(y P1> h425. 0&ndo% s&%ples &re then t&"en 'ro% h425.

,&%&% Statisti$al A$$umulation )e$hanism

St&tistics co%prise & set o' per'or%&nce %e&s res &nd %onitoring v&ri&bles th&t &re de'ined $ithin the 1?S %odel to 8 &nti'y its per'or%&nce &nd to collect observ&tions &bo t its beh&vior. <hose st&tistics &re de'ined in the %odel 'or v&rio s re&sons9 ). :onitoring simulation progress$ L&l es o' the set o' %odel st&tistics c&n be de'ined to be p&rt o' %odel &ni%&tion. S ch v&l es &re pd&ted discretely thro gh %odel r n ti%e. <he si(/sig%& te&% c&n observe the progress o' si% l&tion by %onitoring the ch&nges th&t occ r to the v&l es o' the st&tistics de'ined. >or e(&%ple, the si(/sig%& te&% c&n observe contin o sly the ch&nges th&t occ r to the n %ber o' nits or c sto%ers in the syste%. 2. #onducting scheduled simulation re-ie4s$ In %&ny si% l&tion st dies it is o'ten necess&ry to revie$ the syste% st&te &t cert&in points in si% l&tion r n ti%e. <hose co ld be the end o' one $ee" o' oper&ting, the co%pletion o' & cert&in order, or &t & %odel/%i( ch&nge. With %odel st&tistics de'ined, the si(/sig%& te&% c&n h&lt si% l&tion &t &ny point in ti%e re&ched by the si% l&tion cloc", &nd revie$ si% l&tion progress by chec"ing the v&l es o' s ch st&tistics. >or e(&%ple, the si(/sig%& te&% c&n stop the %odel &t the end o' the d&y shi't &t & '&st/'ood rest& r&nt &nd revie$ the n %ber o' c sto%ers $ho $&ited %ore th&n )0 %in tes &t the rest& r&nt drive/thro gh $indo$. 3. Summari;ing model performance$ At the end o' the si% l&tion r n ti%e, the %odel gener&tes & report th&t incl des &ver&ges 4point esti%&tes5 o' the st&tistics de'ined, &long $ith v&ri&bility %e&s res s ch &s v&ri&nce, r&nge, st&nd&rd devi&tion, &nd con'idence interv&ls. <hose st&tistics c&n be sed &s & s %%&ry o' the over&ll syste% per'or%&nce. >or e(&%ple, the &ver&ge thro ghp t 'or & prod ction line is 'o nd to be 66.6 nits per ho r. Collecting st&tistics 'ro% & 1?S %odel d ring r n ti%e &nd &t the end o' si% l&tion is &chieved sing & st&tistic&l &cc % l&tion %ech&nis%. <hro gh this %ech&nis%, the %odel "eeps tr&c" o' the st&tistics de'ined, &cc % l&tes their v&l es, &nd provides &ver&ges o' their per'or%&nce. <his %ech&nis% is per'or%ed &t the occ rrence o' %odel events. As sho$n in >ig re 6.6, %odel st&tistics &re pd&ted &'ter the e(ec tion o' e&ch event &s $ell &s &t the end o' si% l&tion. >or e(&%ple, $hen de'ining c sto%er $&iting ti%e &s & st&tistic in & 1?S %odel o' & b&n", & $&iting/ti%e st&tistic&l &cc % l&tor pd&tes the over&ll c sto%er $&iting ti%e pon the &rriv&l or dep&rt re o' c sto%ers. As & si% l&tion r n ends, the tot&l &cc % l&ted $&iting ti%e is divided by the n %ber o' c sto%ers 4processed in the b&n" d ring r n ti%e5 to obt&in the &ver&ge $&iting ti%e 'or b&n" c sto%ers. In %&n '&ct ring, & pl&nt st&tistic c&n be the &ver&ge n %ber o' nits prod ced per ho r. Jther st&tistics in & b&n" &nd & pl&nt e(&%ple &re sho$n in <&ble 6.)0. +a3le ,&'1# Exam les of Bank and !lant Statisti$s Jpen t&ble &s spre&dsheet Statisti$s in a Bank Exam le Aver&ge c sto%er $&iting ti%e Aver&ge ti%e spent in & syste% Percent o' ti%e & b&n" teller is idle # %ber o' c sto%ers served per d&y 2&(i% % length o' the $&iting line Statisti$s in a !lant Exam le Aver&ge thro ghp t per ho r # %ber o' de'ectives per shi't Aver&ge %&n '&ct ring le&d ti%e Percent&ge o' %&chine tili.&tion Aver&ge n %ber o' nits in b ''er

,&%&, Animation )e$hanism

Ani%&tion in syste% si% l&tion is & se' l tool to both the si% l&tion %odeler &nd the decision %&"er. 2ost gr&phic&lly b&sed si% l&tion so't$&re p&c"&ges h&ve de'& lt &ni%&tion ele%ents 'or syste% reso rces, b ''ers, oper&tions, l&bor, &nd so on. <his is 8 ite se' l 'or %odel deb gging, v&lid&tion, &nd veri'ic&tion. <his type o' &ni%&tion co%es $ith little or no &ddition&l e''ort &nd gives the %odeler &ddition&l insight into ho$ the %odel $or"s &nd ho$ to test di''erent scen&rios. > rther, the %ore re&listic the &ni%&tion, the %ore se' l the %odel beco%es to the decision %&"er in testing scen&rios &nd i%ple%enting sol tions. Si% l&tion %odels c&n r n &nd prod ce res lts $itho t &ni%&tion. <he %odel logic is e(ec ted, %odel st&tistics &re &cc % l&ted, &nd & s %%&ry report is prod ced &t the end o' the %odel r n ti%e. Si% l&tion/b&sed si(/ sig%& te&%s sed to rely on the res lts o' the %odel progr&% to v&lid&te the %odel &nd correct errors in logic progr&%%ing &nd %odel p&r&%eters. Jnly the si% l&tion %odeler $ill be &ble to veri'y %odel correctness &nd se' lness. 1ecision %&"ers revie$ the %odel res lts &nd s ggest $h&t/i' scen&rios &nd e(peri%ents to be per'or%ed. #o$&d&ys, ho$ever, $ith &ni%&tion being & %ech&nis% o' %ost si% l&tion p&c"&ges, decision %&"ers c&n $&tch the %odel r n 4in & t$o/ or three/di%ension&l gr&phic&l represent&tion5, tr&c" %odel entities, observe the i%p&ct o' ch&nges, &nd test %odel beh&vior &t &ny selected point in the process 'lo$. S ch c&p&bility h&s helped %odelers sell ne$ concepts, co%p&re &ltern&tives, &nd opti%i.e per'or%&nce. >in&lly, &ni%&ted %odel c&n be sed 'or tr&ining p rposes &nd &s & de%onstr&tion tool in v&rio s types o' ind stries. Along $ith &ni%&tion c&p&bility, 1?S %odels &re o'ten co%bined $ith good %odel %&n&ge%ent tools. -ser/'riendly gr&phic&l ser inter'&ces 'or gr&phics editing $ith libr&ries o' gr&phic&l tools h&ve %&de %odel &ni%&tion &n e&sy &nd enDoy&ble process. So%e o' those tools h&ve been developed co%bined $ith & si% l&tion d&t&b&se to store %odels, d&t&, res lts, &nd &ni%&tions. 2od les o' both gr&phic&l &nd p&r&%etric represent&tion 'or b&sic %odel co%ponents s ch &s p&rts, %&chines, reso rces, l&bor, co nters, 8 e es, conveyors, &nd %&ny others &re &v&il&ble in %ost si% l&tion p&c"&ges. >ig re 6.+ sho$s &n e(&%ple o' three/di%ension&l si% l&tion &ni%&tion b ilt $ith A to2od so't$&re.

Figure ,&/# <hree/di%ension&l si% l&tion &ni%&tion.

,&, )anual Simulation Exam le

1iscrete event si% l&tion %odels &re driven by the %ech&nis%s o' discrete event 4tr&ns&ction/ b&sed5, ti%e &dv&nce%ent, r&ndo% s&%pling, st&tistic&l &cc % l&tion, &nd &ni%&tion, &s disc ssed in Section 6.7. S ch %ech&nics c&n be better nderstood by &n&ly.ing & 8 e ing syste%. B e ing syste%s &re closely rel&ted to si% l&tion %odels. B e ing %odels &re si%ple &nd c&n be enco ntered &n&lytic&lly to co%p&re the res lts to si% l&tion %odel res lts. Jther

&n&lytic&l %odels th&t c&n be &n&ly.ed $ith si% l&tion incl de inventory %odels, 'in&nci&l %odels, &nd reli&bility %odels. Since $e h&ve closed/'or% 'or% l&s to &n&ly.e si%ple 8 e ing %odels, the need to &ppro&ch s ch %odels $ith si% l&tion is o'ten 8 estion&ble. Si% l&tion is o'ten sed in s ch c&ses to cl&ri'y the &pplic&tion o' 1?S %ech&nis%s. Also, $hen &ss %ptions th&t &re re8 ired 'or developing &n&lytic&l 8 e ing syste%s do not &pply, si% l&tion is sed to &n&ly.e the syste%s. ?vents &rriv&l &nd dep&rt re, deter%ining e&ch eventEs ti%e o' occ rrence, &nd pd&ting the event list &re e(&%ples o' those %ech&nics. @&sed on the %ech&nics o' %odeling discrete events, the %odeling st&te v&ri&bles &nd st&tistics collected &re deter%ined &t di''erent inst&nts $hen e(ec ting the %odel. <o cl&ri'y the %ech&nics th&t t&"e pl&ce $ithin & 1?S %odel, &n e(&%ple o' & discrete event si% l&tion %odel th&t represents & si%ple single/server 8 e ing %odel 42:2:)5 is presented. In this %odel, c&rs &re &ss %ed to &rrive &t & single/b&y oil/ch&nge service st&tion b&sed on &n e(ponenti&lly distrib ted inter&rriv&l ti%e 4t5 $ith & %e&n o' 20 %in tes. <$o types o' oil ch&nges t&"e pl&ce in the st&tion b&y9 & reg l&r oil ch&nge th&t t&"es &bo t )6 %in tes on &ver&ge, &nd ' ll service, $hich t&"es &bo t 20 %in tes on &ver&ge. <he st&tion history sho$s th&t only 20Q o' c sto%ers &s" 'or & ' ll/service oil ch&nge. A si% l&tion %odel is b ilt to %i%ic oper&tion o' the service st&tion. 2odel &ss %ptions incl de the 'ollo$ing9 <here &re no c&rs in the oil/ch&nge st&tion initi&lly 4the 8 e e is e%pty &nd the oil/ch&nge b&y is e%pty5. <he 'irst c&r &rrives &t the beginning o' si% l&tion 4cloc" ti%e T = 05. C&rs inter&rriv&l ti%es &re distrib ted e(ponenti&lly 4this &ss %ption is essenti&l in &n&ly.ing the %odel &s & 8 e ing syste%5, $here&s the oil/ ch&nge ti%e is distrib ted discretely, $ith & )6/%in te service ti%e h&ving & prob&bility o' +0Q &nd & 20/%in te service ti%e h&ving & prob&bility o' 20Q. <he %ove ti%e 'ro% the 8 e e to the oil/ch&nge b&y is negligible. C&rs &re p lled 'ro% the 8 e e b&sed on 'irst/in/'irst/o t 4>I>J5 discipline. #o '&il res &re e(pected to occ r &t the oil/ch&nge b&y. <he logic o' the si% l&tion %odel is per'or%ed b&sed on the 'ollo$ing %ech&nics9 In & discrete event %ech&nis%, entities 4c&rs5 &rrive &nd dep&rt the service st&tion &t cert&in points in ti%e. ?&ch &rriv&l or dep&rt re is &n event, &nd e&ch event is stored chronologic&lly in &n event list &ccording to the 'ollo$ing 'or% l&s9 ne(t &rriv&l ti%e ne(t dep&rt re ti%e = = c rrent si% l&tion cloc" ti%e + gener&ted inter&rriv&l ti%e c rrent si% l&tion cloc" ti%e + gener&ted service ti%e

>or e&ch c&r &rriv&l, cert&in logic is e(ec ted b&sed on the discrete event &nd ti%e/&dv&nce%ent %ech&nis%s. I' & b&y is idle 4e%pty5, the c&r &rriving &t ti%e T enters the b&y 'or service. An 0#G is sed to s&%ple r&ndo%ly & service ti%e s &nd sched le the c&r dep&rt re to be &t ti%e t = T + s. In c&se the b&y is b sy, the c&r enters the st&tion 8 e e &nd $&its 'or the b&y to be e%pty b&sed on & >I>J sched le. Jnce the c&r 'inishes service, &n inter&rriv&l ti%e 4 a5 is s&%pled r&ndo%ly sing the 0#G, &nd & ne$ &rriv&l is sched led &t ti%e t = T + a. St&tistics &re &cc % l&ted &nd collected &nd si% l&tion contin es si%il&rly 'or &nother &rriv&l. <he e(ec tion o' event &rriv&l is sho$n in & 'lo$ ch&rt &s depicted in >ig re 6.,.

Figure ,&0# ?(ec tion o' & c&r &rriv&lCdep&rt re process. C&rs inter&rriv&l ti%e is gener&ted r&ndo%ly 'ro% & contin o s e(ponenti&l distrib tion 4%e&n = 20 %in tes5, $hile the service ti%e is gener&ted r&ndo%ly 'ro% & discrete distrib tion 4)6 %in tes $ith prob&bility +0Q &nd 20 %in tes $ith prob&bility 20Q5. S&%pling 'ro% the e(ponenti&l distrib tion $ith & C1> o' f4t5 = ) et, t > 0, res lts in t = 4):5 ln4) R5, $here R = f4t5.L&l es o' R sed to obt&in s ccessive r&ndo% ti%e s&%ples 4t5 &re selected 'ro% & ni'or% 40,)5 distrib tion sing the ;CG %ethod. Si%il&rly, R is gener&ted 'or s&%pling 'ro% the discrete distrib tion to deter%ine the service ti%e 4s5, $here s = )6 %in tes i' 0 R 0.+ &nd s = 20 %in tes i' 0.+ < R ).0. <here'ore, $hen st&rting the si% l&tion 4T = 0, the ti%e &t the si% l&tion cloc"5, $e &ss %e th&t the 'irst c&r 4c sto%er )5 &rrives &t T = 0. <he &rriv&l o' c sto%er ) is the event th&t ch&nges the st&te o' the service st&tion 'ro% idle to b sy.<he service ti%e &ssigned 'or the 'irst c sto%er 4c sto%er )5 sing R = 0.66*7 is s = )6.00 %in tes. <h s, c sto%er ) dep&rt re ti%e is sched led &s 40 + )6.00 = )6.00 %in tes5. While c sto%er )Es c&r is receiving & reg l&r oil ch&nge, &nother c&r 4c sto%er 25 %&y be driven into the service st&tion. <he &rriv&l ti%e o' c sto%er 2 is deter%ined sing R = 0.+623 &s t = 4205 ln4) 0.+6235 = 3+.27 %in tes. !ence, the &rriv&l ti%e 'or c sto%er 2 is sched led &s 40 + 3+.27 = 3+.27 %in tes5. @&sed on this %ech&nic, the 'irst event list 4?;5 is 'or%ed $ith t$o sched led events on the si% l&tion cloc"9 4 ), t)59 dep&rt re o' c sto%er )O the dep&rt re ti%e is T = )6.00 %in tes. 4 2, t259 &rriv&l o' c sto%er 2O the &rriv&l ti%e is T = 3+.27 %in tes. App&rently, c sto%er ) $ill dep&rt the service st&tion be'ore c sto%er 2 &rrives. ) $ill be re%oved 'ro% top o' the ?; &nd the list $ill be red ced to 2 &lone. <he si% l&tion cloc" $ill be

&dv&nced to the ne(t event 4 25 ti%e 4T = )6.00 %in tes5 in order to e(ec te 2. <he dep&rt re o' c sto%er ) is the event th&t ch&nges the st&te o' the service st&tion 'ro% b sy to idle. <he service ti%e &ssigned to c sto%er 2 sing R = 0.++6* is s = 20.00 %in tes. !ence, the dep&rt re ti%e o' c sto%er 2 is sched led &s 43+.27 + 20.00 = 6+.27 %in tes5. A'ter the &rriv&l o' c sto%er 2, &nother c&r 4c sto%er 35 is sched led to &rrive &t t = 4205 ln4) 0.26635 = 6.,2 %in tes sing R = 0.2663. <he &rriv&l o' c sto%er 3 is sched led &s 3+.27 + 6.,2 = 77.)6 %in tes. @&sed on this %ech&nic, the event list is pd&ted $ith the t$o sched led events 4 39 dep&rt re o' c sto%er 2 &nd 7 = &rriv&l o' c sto%er 35. <he t$o events &re sched led on the si% l&tion cloc" &s 'ollo$s9 39 T = 6+.27 %in tes 79 T = 77.)6 %in tes App&rently, c sto%er 3 $ill &rrive &t the service st&tion be'ore the dep&rt re o' c sto%er 2, res lting in & $&iting ti%e o' 46+.27 77.)6 = )7.0+ %in tes5. Si% l&tion r n ti%e contin es sing the %ech&nis% described ntil & cert&in ter%in&ting condition is %et &t & cert&in n %ber o' ti%e nits 4e.g., three prod ction shi'ts, ) ye&r o' service, 2000 nits prod ced, &nd so on5. A si% l&tion t&ble th&t s %%&ri.es the dyn&%ics o' the 'irst 20 c&rs &rriving &t the service st&tion is sho$n in <&ble 6.)), $here the 'ollo$ing sy%bols &re sed9 +a3le ,&''# Simulation +a3le for the First (1 Cars Jpen t&ble &s spre&dsheet A ) 2 3 7 6 6 * + , )0 )) )2 )3 )7 )6 B 0.00 3+.27 6.,2 3).70 )3.26 2+.)2 )7.36 )6.22 2).+* )3.,+ 36.67 ,.,6 )0.67 23.67 )).*0 C 0.00 3+.27 77.)6 *6.66 ++.+) ))6.,3 )3).2+ )76.60 )6+.3* )+2.36 2)+.+, 22+.+7 23,.3+ 263.02 2*7.*2 D )6.00 20.00 20.00 )6.00 20.00 )6.00 )6.00 )6.00 20.00 )6.00 )6.00 )6.00 )6.00 20.00 )6.00 E 0 3+.27 6+.27 *+.27 ,3.27 ))6.,3 )3).,3 )76.,3 )6+.3* )++.3* 2)+.+, 233.+, 27+.+, 263.+, 2+3.+, F 0.00 0.00 )7.0+ 2.6+ 7.73 0.00 0.66 0.73 0.00 6.02 0.00 6.06 ,.6) 0.+* ,.)* ; )6.00 6+.27 *+.27 ,3.27 ))3.27 )3).,3 )76.,3 )6).,3 )++.3* 203.3* 233.+, 27+.+, 263.+, 2+3.+, 2,+.+, H )6.00 20.00 37.0+ )*.6+ 27.73 )6.00 )6.66 )6.73 20.00 2).02 )6.00 20.06 27.6) 20.+* 27.)* " 0.00 23.27 0.00 0.00 0.00 3.6, 0.00 0.00 6.77 0.00 )6.62 0.00 0.00 0.00 0.00

+a3le ,&''# Simulation +a3le for the First (1 Cars Jpen t&ble &s spre&dsheet A )6 )* )+ ), 20 Sum B )6.,0 2+.*0 26.66 22.76 )2.60 <=>$>2 C 2,0.62 3),.32 377.,* 36*.72 3*,.,2 D )6.00 )6.00 )6.00 20.00 )6.00 <<0$00 E 2,+.+, 3),.32 377.,* 36*.72 3+*.72 F +.2* 0.00 0.00 0.00 *.60 6?$66 ; 3)3.+, 337.32 36,.,* 3+*.72 702.72 H 23.2* )6.00 )6.00 20.00 22.60 <>?$66 " 0.00 6.73 )0.66 *.76 0.00 =2$@2

A9 c sto%er n %ber 4),2, . . . , 205 @9 r&ndo%ly gener&ted inter&rriv&l ti%e in %in tes 'or e&ch c sto%er C9 &rriv&l ti%e o' e&ch c sto%er 19 r&ndo%ly gener&ted service ti%e in %in tes 'or e&ch c sto%er ?9 ti%e &t $hich service 4oil ch&nge5 begins >9 ti%e th&t e&ch c sto%er $&its in the 8 e e $&iting 'or service to begin G9 ti%e &t $hich service 4oil ch&nge5 ends !9 tot&l ti%e & c sto%er spends in the syste% 4$&iting pl s service5 I9 idle ti%e 'or the st&tion b&y

Sever&l st&tistics c&n be c&lc l&ted %&n &lly 'ro% d&t& in the si% l&tion t&ble. <hese st&tistics &re collected by &cc % l&ting st&te v&ri&bles &nd by &ver&ging the &cc % l&ted v&l es over the r n ti%e. In co%p ter si% l&tion, %odel st&te v&ri&bles &re tr&c"ed dyn&%ic&lly to &cc % l&te ti%es &nd occ r/ rences o' events over r n ti%e &nd to provide &ver&ges o' st&tistics o' interest &t the end o' the si% l&tion r n ti%e. <he 'ollo$ing c&lc l&tions sho$ e(&%ples o' "ey st&tistics th&t &re o' interest to the &n&lyst in the service st&tion e(&%ple9 Aver&ge c sto%er $&iting ti%e = 4tot&l $&iting ti%e in 8 e e5:4n %ber o' c sto%ers5 = 6+.66:20 = 3.73 %in tes. Aver&ge c&r service ti%e = 4tot&l service ti%e5:4n %ber o' c sto%ers5 = 330.00:20 = )6.60 %in tes. Aver&ge inter&rriv&l ti%e = 4s % o' inter&rriv&l ti%es5:4n %ber o' c sto%ers5 ) = 3*,.,2:420 )5 = ),.,, %in tes. Aver&ge $&iting ti%e 'or $&iting c sto%ers = 4tot&l $&iting ti%e5:4n %ber o' c sto%ers $ho $&it5 = 6+.66:)2 = 6.*2 %in tes. Aver&ge ti%e/in/syste% = 4tot&l ti%e in syste%5:4n %ber o' c sto%ers5 = 3,+.66:20 = ),.,3 %in tes. <his c&n &lso be c&lc l&ted by s %%ing the &ver&ge $&iting ti%e &nd &ver&ge service ti%e 4i.e., 3.73 + )6.6 = ),.,35. <he prob&bility th&t & c sto%er h&s to $&it = 4n %ber o' c sto%ers $ho $&it5:4n %ber o' c sto%ers5 = )2:20 = 60Q. <he tili.&tion o' the st&tion b&y = H).00 4tot&l idle ti%e:tot&l r n ti%e5I = )00Q H).00 4*2.72:702.725I = +2.00Q.

,&- Com uter DES Exam le

In this section $e present & 1?S co%p ter si% l&tion %odel 'or & s%&ll he&lth c&re clinic. Si% l&tion is tili.ed to b ild & clinic %odel th&t c&n be sed to &ssess oper&tion&l &ltern&tives 'or %&(i%i.ing the n %ber o' p&tients served $hile %ini%i.ing the p&tient ti%e/in/syste% 4clinic5. <hro gh this si%ple e(&%ple $e describe ho$ 1?S %ech&nics &re tili.ed to represent clinic

ele%ents &nd p&tient 'lo$ &nd to gener&te si% l&tion d&t& 'ro% relev&nt s&%pling distrib tions. We brie'ly disc ss %odel &ss %ptions, %odel constr ction, &nd %odel v&lid&tion &nd veri'ic&tion. >in&lly, & %odel &n&lysis s&%ple $ill be presented to de%onstr&te ho$ 1?S provides &ns$ers to "ey 8 estions on the beh&vior o' the clinic syste%. <he clinic tre&ts both rgent &nd &c te c&re p&tients. 2ost p&tients re8 ire &c te c&re. 2ost rgent c&re p&tients &re &ssessed by clinic doctors &nd sent to the hospit&l. <he clinic &d%its t$o types o' p&tients9 ne$ p&tients 4i.e., $&l"/ ins5 &nd ret rn p&tients 4i.e., p&tients $ith &ppoint%ents5. A p&tient &rrives &t the clinicEs $&iting roo% &nd $&its &t the registr&tion co nter. P&tients then $&it 'or the %edic&l &ssist&nt to &d%it the% to the clinic. <he %edic&l &ssist&nt c&lls p&tients, t&"es their blood press re, chec"s their te%per&t re, &nd &s" the% relev&nt 8 estions be'ore they enter the e(&%in&tion roo%. <he t$o physici&ns in the e(&%in&tion roo%s di&gnose p&tients, rele&se the%, or direct the% to the l&b or the ph&r%&cy. I' the p&tient is sent to the l&b he or she ret rns $ith the l&b res lts to the e(&%in&tion roo% to see the s&%e physici&n th&t he or she s&$ e&rlier. P&tients $ho h&ve been rele&sed le&ve the clinic. >ig re 6.)0 sho$s p&tient 'lo$ in the clinic. Potenti&l p&tient $&iting points &re incl ded on the %&p.

Figure ,&'1# Process %&p o' p&tient 'lo$ in the clinic e(&%ple. -nli"e %&n &l si% l&tion, the clinic logic &nd 'lo$ sho$n in >ig re 6.)) is not e&sy to %odel %&n &lly 4&n&lytic&lly5. A net$or" o' 8 e es c&n be sed to &ppro(i%&te the sol tion $ith %&ny &ss %ptions. !ence, co%p ter 1?S $ith A to2od so't$&re is sed to %odel the clinic, esti%&te its per'or%&nce %etrics, &nd select the best oper&ting str&tegy.

Figure ,&''# A to2od si% l&tion %odel o' the clinic e(&%ple. <he st dent version o' the A to2od so't$&re p&c"&ge is sed to si% l&te the clinic e(&%ple. ?ven $ith the li%ited n %ber o' si% l&tion entities in the st dent version, A to2od provides & 'le(ible %odel b ilding &nd logic progr&%%ing pl&t'or% th&t %&"es it e&sy to develop three/ di%ension&l str ct res, incorpor&te co%ple( logic, &nd insert process d&t&. It &lso esti%&tes %&ny si% l&tion st&tistics &nd provides &n e(cellent %od le to &n&ly.e si% l&tion o tp ts $ith A toSt&t. <he 'ollo$ing &ss %ptions &re sed in b ilding the clinic si% l&tion %odel9 <he p&tientsE inter&rriv&l ti%e is distrib ted e(ponenti&lly $ith & %e&n o' )6 %in tes. Arriving p&tients register &nd $&it to be c&lled. <he clinic $&iting roo% h&s & %&(i% % c&p&city o' 60 se&ts. <he clinic oper&tion p&ttern is + ho rs per d&y 'ro% +900 &.%. to 6900 p.%. $ith & )/ho r l nch, $hich is set initi&lly )2900 to )900 p.%. 1 e to the n&t re o' the clinic $or", ho$ever, st&'' %e%bers t&"e their l nch bre&" b&sed on their &v&il&ble idle ti%e. <$enty/'ive percent o' p&tients &re re8 ired to t&"e & l&b blood test, $hich t&"es 26 %in tes on &ver&ge. P&tients $ho t&"e their l&b test % st ret rn to the physici&n $ho re8 ested the test. P&tients ret rning 'ro% the l&b test h&ve & di''erent processing ti%es by physici&ns. Si(ty percent o' p&tients &re re8 ired to resched le &ppoint%ents $ith & clinic physici&n 'or ' rther tre&t%ent. >i'ty percent o' p&tients &re sent to the ph&r%&cy to get prescription dr gs. >i'teen percent o' p&tients &re tre&ted by & clinic physici&n &nd rele&sed 'ro% the clinic in their 'irst visit. <he %edic&l &ssist&nt per'or%s three ' nctions9 t&"ing blood press re, chec"ing te%per&t re, &nd &s"ing p&tients 8 estions &bo t their he&lth. Clinic st&'' &re distrib ted &s sho$n in <&ble 6.)2. +a3le ,&'(# Distri3ution of Clini$ Staff Jpen t&ble &s spre&dsheet Resour$e Co nter 0egistr&tion 2edic&l c&re ?(&%in&tion <est l&b Ph&r%&cy ) cler" 2 cler"s 2 &ssist&nts 2 physici&ns ) technici&n ) ph&r%&cist Staff

<o %odel the stoch&stic n&t re o' the clinic oper&tion, the si% l&tion d&t& th&t h&ve been collected &re 'itted to st&nd&rd distrib tions. 2odel inp ts &re gener&ted r&ndo%ly 'ro% A to2odEs b ilt/in st&tistic&l distrib tions. In A to2od, distrib tions &re s&%pled sing &n inverse %ethod $ith ;CG 0#G. > rther det&ils on %odeling si% l&tion inp ts &re presented in Ch&pter 6. <&ble 6.)3 s %%&ri.es the s&%pling distrib tions sed in the clinic %odel. +a3le ,&'*# Sam ling Distri3utions :sed in the Clini$ Exam le Jpen t&ble &s spre&dsheet !ro$ess Arriv&l Co nter 0egistr&tion 2edic&l &ssist&nt !atient +y e All p&tients J st entering clinic <o be rele&sed 'ro% clinic #e$ &nd $&l"/ins 0et rning <&"ing blood press re Chec"ing te%per&t re B estionn&ire All p&tients 0et rning 'ro% l&b test Per'or%ing blood test >irst ti%e Second ti%e Distri3ution ?(ponenti&l 4)6 %in5 -ni'or% 4).2, 0.65 %in ?(ponenti&l 47 %in5 -ni'or% 43.6, ).65 -ni'or% 42, 0.65 #or%&l 460, 65 sec #or%&l 460, 65 sec <ri&ng l&r 42, 7, )05 %in <ri&ng l&r 4)0, )6, 705 <ri&ng l&r 47, 6, )05 %in #or%&l 426, 35 %in <ri&ng l&r 4).6, 7, +5 %in <ri&ng l&r 4), 3, 65 %in

?(&%in&tion ;&b test Ph&r%&cy

Since A to2od is lo&d/driven, clinic %odel b ilding $ith A to2od begins by de'ining & process 'or cre&ting p&tients 4lo&ds5. An initi&li.&tion process c&n be sed to cre&te & d %%y lo&d &nd clone to v&rio s clinic processes. P&tients &re directed thro gh clinic oper&tions b&sed on &ssigning lo&d &ttrib tes. <he t$o %&in lo&d types de'ined &re 'or $&l"/ins &nd sched led 4ne$ or ret rning5 p&tients. B e es &nd reso rces &re then sed to constr ct the clinic %odel. ?&ch 8 e e is de'ined in ter%s o' c&p&city &nd loc&tion, &nd e&ch reso rce is de'ined $ith & cert&in distrib tion cycle ti%e. Synt&( is $ritten to direct p&tient 'lo$ &nd to control inter&ctions &%ong clinic processes. <he %odel intended 'lo$ logic is sed to veri'y the %odel, &nd the per'or%&nce d&t& e(pected or observed &re sed to v&lid&te the %odel. >ig re 6.)) sho$s & si%ple &ni%&tion o' the A to2od clinic si% l&tion %odel. Jnce v&lid&ted &nd veri'ied, the %odel c&n be sed to collect clinic st&tistics & to%&tic&lly. -nli"e %&n &l si% l&tion, ho$ever, co%p ter si% l&tion cre&tes & si% l&tion events list, tr&c"s events, pd&tes the event list, &nd develops &n intern&l si% l&tion t&ble to &cc % l&te si% l&tion st&tistics. We obt&in the res lts by r nning the %odel 'or 70 ho rs. <he 'ollo$ing st&tistics &re gener&ted 'ro% the %odel9 &ver&ge ti%e th&t p&tients spend in the clinic = &ver&ge service 4tot&l tre&t%ent5 ti%e = &ver&ge $&iting ti%e = n %ber o' p&tients tre&ted d&ily = +7.62 %in tes 66.) %in tes 2+.62 %in tes 2* p&tients

We 'o nd th&t the p&tient e(&%in&tion process contrib ted the %&Dor del&y in the clinic. <he e(&%in&tion process t&"es on &ver&ge 72.,0 %in tes, $ith & $&iting ti%e o' &bo t 20.,* %in tes &nd & tre&t%ent ti%e o' &bo t 2).,3 %in tes. Also, p&tients co%pl&in th&t physici&ns &re

press red to co%plete tre&t%ent 8 ic"ly. Still, ho$ever, d e to ro ting p&tients, the t$o clinic e(&%in&tion physici&ns &re tili.ed ,3.6Q &nd +,.6Q. <his is & rel&tively high tili.&tion v&l e. <he obDective is to se the %odel to red ce $&iting ti%e &nd incre&se service ti%e $itho t ' rther incre&sing the tili.&tion o' e(&%in&tion reso rces &nd $itho t red cing the n %ber o' p&tients tre&ted d&ily. It is there'ore s ggested &s &n &ltern&tive process con'ig r&tion to %ini%i.e cost th&t one o' the registr&tion cler"s be re%oved &nd the e(&%in&tion physici&ns be incre&sed by one. <he &ltern&tive clinic process str ct re incl des ch&nging clinic reso rces &s sho$n in <&ble 6.)7. +a3le ,&'%# Alternati4e Clini$ !ro$ess Stru$ture Jpen t&ble &s spre&dsheet Resour$e Co nter 0egistr&tion 2edic&l c&re ?(&%in&tion <est l&b Ph&r%&cy ) cler" ) cler" 2 &ssist&nts 3 physici&ns ) technici&n ) ph&r%&cist Staff

@y r nning &nother si% l&tion $ith the ne$ clinic str ct re, the 'ollo$ing st&tistics &re gener&ted 'ro% the %odel9 &ver&ge ti%e th&t p&tients spend in the clinic = &ver&ge service 4tot&l tre&t%ent5 ti%e = &ver&ge $&iting ti%e = n %ber o' p&tients tre&ted d&ily = *7.30 %in tes 60.)2 %in tes )7.)+ %in tes 32 p&tients

<he pd&ted st&tistics sho$ & red ction in the &ver&ge ti%e spent in the clinic, $ith &n incre&se in the tre&t%ent ti%e. <he red ction is &chieved pri%&rily in $&iting ti%e. ?(&%in&tion physici&ns spend & longer ti%e $ith p&tients.<he tot&l n %ber o' p&tients tre&ted d&ily h&s incre&sed to &n &ver&ge o' 32. Aver&ge tili.&tion o' the three physici&n reso rces is &lso red ced to &n &ver&ge o' *6Q 4*7.+Q, *6.6Q, &nd *3.7Q5. <he pd&ted process con'ig r&tion is &lso v&lid&ted $ith s bDect %&tter e(perts &nd reco%%ended 'or i%ple%ent&tion. I%prove%ent $&s &lso D sti'ied $ith & costCbene'it &n&lysis.

,&. Summary
<he 1?S techni8 es presented in this ch&pter &re %ost c&p&ble &nd po$er' l tools 'or b siness process si% l&tion o' tr&ns&ction&l n&t re $ithin 1>SS &nd si(/sig%& proDects. A 1?S provides %odeling o' entity 'lo$s $ith c&p&bilities th&t &llo$ the si% l&tion %odeler to see ho$ 'lo$ obDects &re ro ted thro gh the syste%. 1?S $itnesses gro$ing c&p&bility, so't$&re tools, &nd & $ide spectr % o' re&l/$orld &pplic&tions. Sever&l consider&tions sho ld prep&re the te&% 'or pl&nning & tr&ns&ction/ b&sed process si% l&tion st dy &nd 'or &ddressing the s it&bility o' their choice o' &dopted si% l&tion p&c"&ge. >irst, & cle&r cl&ssi'ic&tion o' the proDect b siness processes is needed &s & prod ct develop%ent &nd prod ction process 'or proble% solving, incl ding distrib tion processes or service/b&sed processes. Process cycle ti%es &nd reso rce re8 ire%ents &re s &lly h&ndled by si% l&tion,

prod cing %ore &cc r&te res lts. 2&ny 1?S si% l&tion p&c"&ges &re c&p&ble o' representing highly v&ri&ble steps $ith tightly co pled reso rces. 2odeling shi'ts, do$nti%e, overti%e, &nd le&rning c rves $ith % ltiple replic&tions 'or reso rces &re so%e o' the i%port&nt consider&tions in b ilding & v&lid si% l&tion %odel. In the conte(t o' si(/sig%&, discrete event si% l&tion is 8 ite s it&ble 'or prod ction process proble% solving, $here o tp ts &re prod ced in & b&tch or contin o s/'lo$ %ode in rel&tively high vol %es. <&s"s s ch &s &sse%bly, dis&sse%bly, set p, inspection, &nd re$or" &re typic&l steps in prod ction processes $ith 8 e ing r les &nd do$nti%e %odeling. S ch processes &re s &lly %odeled to obt&in ste&dy/st&te beh&vior p&st the $&r%/ p period. In %odeling distrib tion processes in & prod ction environ%ent, it is i%port&nt to de'ine &ttrib tes 'or 'lo$ entities in order to "eep tr&c" o' ni8 e ch&r&cteristics s ch &s v&l e/&dds, cost, &nd dist&nce tr&veled. 1 e to the tr&nsient n&t re o' distrib tion processes, the si% l&tion %odel $&r%/ p period is s &lly longer th&n other prod ction processes. In & service/b&sed ind stry, processes present & %&Dor &re& 'or e%ploy%ent o' si% l&tion st dies. <hese processes &re typic&lly ch&r&cteri.ed by & non/ v&l e/&dded ti%e th&t e(ceeds the v&l e/&dded ti%e 4processing5. <he si% l&tion o' service processes represents & ch&llenge bec& se the entities &nd processing reso rces &re both typic&lly h %&n. 2odeling h %&n beh&vior is co%ple( &nd npredict&ble. >or e(&%ple, c sto%ers c&lling & service center %&y hold or h&ng p. 2odeling &d&pt&tion is re8 ired to %odel s ch sit &tions. Also, processing ti%es &re highly v&ri&ble &nd c sto%er &rriv&ls &re r&ndo% &nd &dd to the %odel co%ple(ity. 2odel &cc r&cy de%&nds represent&tive prob&bility distrib tions. Jther$ise, & ste&dy st&te %&y not be re&ched.

Cha ter -# +he Simulation !ro$ess

-&' "ntrodu$tion
<he set o' techni8 es, steps, &nd logic 'ollo$ed $hen cond cting & si% l&tion st dy is re'erred to in the conte(t o' this boo" &s & simulation process. <he det&ils o' s ch & process o'ten depend on the n&t re o' the si(/sig%& proDect, the proDect obDectives, the si% l&tion so't$&re sed, &nd even on the $&y the te&% h&ndles si% l&tion %odeling.Altho gh the t&ctics o' s ch & process o'ten v&ries 'ro% one &pplic&tion to &nother &nd 'ro% one si% l&tion proDect to &nother, the over&ll str ct re o' the si% l&tion process is co%%on. It is necess&ry to 'ollo$ & syste%&tic %ethod $hen per'or%ing the si% l&tion process. <his ch&pter is 'oc sed on &n&ly.ing the v&rio s &spects o' the si% l&tion process, the process 'ollo$ed by & si(/sig%& te&% c&rrying o t & co%plete si% l&tion st dy. Si% l&tion proDects &re cond cted pri%&rily 'or one o' three p rposes9 syste% design, proble% solving, &nd contin o s i%prove%ent. <he three p rposes c&n, ho$ever, be &chieved sing & generic syste%&tic &ppro&ch. <his &ppro&ch st&rts $ith & cle&r de'inition o' the design ch&llenge, the proble%, or the i%prove%ent opport nity. Altern&tives o' syste% design, proble% solving, &nd i%prove%ent opport nity &re then developed.<hese &ltern&tives &re ev&l &ted sing si% l&tion/ b&sed per'or%&nce criteri& chosen by the te&%. <he &ltern&tive $ith the best esti%&ted per'or%&nce is 'in&lly selected &nd i%ple%ented. Si% l&tion is sed &s &n e''ective tool 'or di&gnosing the syste% &nd de'ining proble%s, ch&llenges, &nd opport nities. It is &lso sed 'or developing &nd ev&l &ting &ltern&tives &s $ell &s 'or &ssessing the per'or%&nce o' e&ch &ltern&tive. !ence, & co%plete si% l&tion st dy o'ten incl des proble% de'initionO setting si% l&tion obDectivesO developing & concept &l %odelO speci'ying %odel &ss %ptionsO collecting pertinent %odel d&t&O b ilding, veri'ying, &nd v&lid&ting the si% l&tion %odelO &n&ly.ing %odel o tp tsO &nd doc %enting the proDect 'indings.

-&( Categories Of Simulation Studies

@&sed on the princip&l go&l o' the proDect, si% l&tion st dies c&n be c&tegori.ed into three cl&sses9 syste% design, proble% solving, &nd contin o s i%prove%ent. <hese c&tegories &re sho$n in >ig re 6.). <he t&ctics &nd techni8 es sed in & si% l&tion st dy o'ten v&ry depending on these three c&tegories. >in&lly, it is $orth %entioning th&t si% l&tion is the environ%ent o' si% l&tion/b&sed si(/sig%&, le&n si(/sig%&, &nd design 'or si(/sig%& %ethods. >ollo$ing is & disc ssion o' the three c&tegories o' si% l&tion st dies.

Figure -&'# C&tegories o' si% l&tion st dies.

-&(&' Simulation-Based System Design

2&ny si% l&tion st dies &re sed to &ssist the design process. <his is o'ten &ligned $ith & typic&l obDective o' & 1>SS proDect. !o$ever, the 'oc s here is on & event/driven or tr&ns&ction/b&sed process &nd service design r&ther th&n the prod ct design 4see, e.g., =&ng &nd ?l/!&i", 20035. Process co%p&tibility st dies c&n bene'it gre&tly 'ro% si% l&tion/b&sed design. Applic&tions o' si% l&tion proDects incl de & $ide spectr % o' proDects, s ch &s the develop%ent o' ne$ '&cilities,

%&Dor e(p&nsions o' & %&n '&ct ring syste%, & ne$ clinic, & ne$ b&n", & ne$ vehicle progr&%, &nd & ne$ tr&nsport&tion net$or". 1esigning & ne$ syste% or process $ith or $itho t si% l&tion o'ten involves testing ne$ concepts, b ilding ne$ str ct res or l&yo ts, &nd developing ne$ logic. <ypic&l obDectives o' design si% l&tion st dies incl de developing & ne$ syste% design, &ssessing &nd v&lid&ting & proposed design, &nd co%p&ring t$o or %ore design &ltern&tives. We c&n concl de th&t so%e synergy &nd &lign%ent c&n be lever&ged by dr&$ing on the si%il&rities bet$een syste% design &nd the si% l&tion process. <his is e(plored ' rther in l&ter ch&pters o' this boo". <he 'le(ibility o' si% l&tion %odeling provides &n e(cellent pl&t'or% 'or developing &nd testing design &ltern&tives &s $ell &s 'or &n&ly.ing & v&riety o' process sche%es.<he rel&tively lo$ cost &nd high e''iciency o' si% l&tion o'ten pro%otes the se o' si% l&tion st dies, solely on s ch %erits. S&ving &chieved 'ro% si% l&tion/b&sed design st dies o'ten co%e 'ro% testing design 'e&sibility, enh&ncing the syste% str ct re, &nd reco%%ending process ch&nges p 'ront 4see Section +.75. 1esign ch&nges h&ve lo$ cost &nd reso rce i%plic&tions $hen c&rried o t &t the concept st&ge 4see Ch&pter + 'or & theoretic&l conte(t &nd Ch&pter )0 'or &n &pplic&tion c&se st dy5. >ig re 6.2 sho$s the proDect cost rel&tion to the st&ge o' the proDect. As sho$n in the 'ig re, the cost o' engineering ch&nges incre&ses e(ponenti&lly $ith ti%e in &ny proDect. ?(pensive ch&nges o'ten t&"e pl&ce &t the constr ction &nd oper&tion st&ges o' the proDect, $here the proDect is s bDected to costly del&ys, oper&tion interr ptions, &nd di''ic lt physic&l ch&nges.

Figure -&(# Cost o' %&"ing ch&nges &t v&rio s proDect st&ges. As sho$n in >ig re 6.3, the design o' s ch syste%s incl des syste% str ct r&l design, logic&l design, &nd syste% p&r&%etric design.

Figure -&*# Syste% design $ith si% l&tion. ). Syste% str ct r&l design incl des developing & syste% str ct re in ter%s o' ele%ents, l&yo t, &nd represent&tion. In & pl&nt e(&%ple, this %&y incl de selecting %&chines, pieces o' e8 ip%ent, &nd %&teri&l/h&ndling syste%s, &nd developing &n &ppropri&te l&yo t o' the syste% str ct re. <he l&yo t $o ld incl de dep&rt%ent&l sp&cing, $or"sp&ce design, &isles, %&teri&l 'lo$ di&gr&%, &nd so on. An &ctivity rel&tionship di&gr&% is o'ten sed to g ide &nd str ct re the syste% l&yo t. >in&lly, represent&tion is %&de sing gr&phic&l tools &nd %ethods to present the %odel str ct re de'ined. 2. Syste% logic&l design incl des developing the logic th&t governs the beh&vior o' the nderlying syste%. <he syste% ' nction&lity &nd the $&y thro gh $hich syste% ele%ents $or" on syste% entities &re de'ined thro gh the syste% logic&l design. ?(&%ples o' logic&l design incl de 'lo$ logic, se8 ence o' oper&tion, priority sche%es, %&"ing decisions, sched ling r les, &nd so on. 3. P&r&%etric design incl des setting p&r&%eters 4control '&ctors5 to v&l es or levels th&t le&d to 'est per'or%&nce. <hese p&r&%eters &re de'ined &s settings 'or %odel str ct r&l

ele%ents. ?(&%ples incl de %&chine cycle ti%es &nd reli&bility d&t&, %&teri&l h&ndling &nd tr&ns'er speed, b ''er c&p&city, &nd so on. 1i''erent levels o' s ch p&r&%eters &re tested sing the si% l&tion %odel thro gh e(peri%ent&l design &nd opti%i.&tion se&rch to &rrive &t the p&r&%eter settings th&t le&d to best syste% per'or%&nce. >or e(&%ple, di''erent p&r&%eters in & %&n '&ct ring process 4e.g., n %ber o' oper&tors, %&chine cycle ti%es, conveyor speeds, b ''er c&p&cities5 &re set so th&t the thro ghp t o' & pl&nt is %&(i%i.ed. Syste% str ct r&l, logic&l, &nd p&r&%etric design &''ects syste% per'or%&nce 4%odel response5 directly in ter%s o' syste% thro ghp t, le&d ti%e, inventory levels, &nd so on. In designing syste%s, s ch responsesH)I &re sed to g ide the design process. >or e(&%ple, one l&yo t %&y res lt in longer %&teri&l 'lo$ &nd conse8 ently, longer %&n '&ct ring le&d ti%e. 1el&ys cre&ted bec& se o' i%proper logic&l design &lso &''ect le&d ti%e &nd thro ghp t. >in&lly, & longer cycle ti%e or & high percent&ge o' do$nti%e &t & cert&in $or"st&tion %&y cre&t & syste% bottlenec" th&t c& ses del&ys &nd res lts in thro ghp t. <he design 'or si(/sig%& 41>SS5 %ethod is &pplied to enh&nce the syste% design so th&t best per'or%&nce c&n be &chieved.

-&(&( !ro3lem-Sol4ing Simulation

Prod ction &nd b siness syste%s o'ten '&ce ch&llenges th&t &''ect their oper&tion &nd per'or%&nce.<he i%p&ct o' s ch ch&llenges v&ries 'ro% red ced e''iciency or 're8 ent del&ys &nd '&il res to %&Dor sh tdo$ns &nd c&t&strophes. !ence, & big portion o' the $or" o' prod ction %&n&gers, syste% engineers, &nd oper&tions %&n&gers is o'ten 'oc sed on %onitoring the syste% per'or%&nce, t&c"ling oper&tion&l proble%s, &nd &tte%pting to prevent ' t re proble%s. In &ddition, %&ny o' these proble%s %&y be concl ded 'ro% c sto%er co%pl&ins, %&r"et 'eedb&c", &nd &ct &l s&les n %bers. 1i&gnosing s ch proble%s &nd providing the proper sol tion is not &l$&ys & trivi&l t&s". 1i&gnostic &n&lysis, testing $h&t/i' scen&rios, &nd v&lid&ting the 'e&sibility o' proposed 'i(es &re typic&lly re8 ired. L&rio s engineering &n&lyses th&t &re b&sed on &n&lytic&l %odeling or engineering sense &re o'ten tili.ed to solve proble%s. ?(&%ples incl de inventory %odels, 8 e ing %odels, reli&bility %odels, '&il re %ode &nd e''ect &n&lysis, root/c& se &n&lysis, '& lt/ tree &n&lysis, &nd so on. !o$ever, &s disc ssed in Ch&pter 7, the &ss %ptions &nd &n&lytic&l li%it&tions o' s ch tools li%it their &pplic&tion 'or solving proble%s o' consider&ble co%ple(ity. 1?S c&p&bilities provide & 'le(ible proble%/solving pl&t'or%. !ence, &s sho$n in >ig re 6.7, si% l&tion %odeling is tili.ed to per'or% the %&in proble%/solving techni8 es o' di&gnosing syste%s, testing oper&tion&l sche%es, &nd v&lid&ting prescribed sol tions. <hese techni8 es 'it the obDective o' si(/sig%& 12AIC proDects.

Figure -&%# Proble% solving $ith si% l&tion. As sho$n in >ig re 6.7, solving proble%s $ith si% l&tion incl des proble% di&gnostic &nd de'inition, testing sol tion sche%es, &nd v&lid&ting proble% sol tion. ). Proble% di&gnostic $ith si% l&tion is not &l$&ys &n e&sy t&s". It depends he&vily on the credibility o' the %odel b ilt. <he %odel h&s to re'lect the c rrent syste% st&te 4incl ding the i%p&ct o' the proble%5. In so%e c&ses, the proble% %&y be 'o nd $hile the %odel is being b ilt. <his is bec& se, &s disc ssed in Ch&pter 7, syste% str ct re, l&yo t, d&t&, &nd logic &re e(plored thoro ghly $hen b ilding the %odel. Jnce the %odel is b ilt, observing the %odel per'or%&nce d ring r n ti%e %&y &lso le&d to .oo%/in on the proble% or proble% &re&. <he 8 &lity o' &ni%&tion &nd gr&phic&l represent&tion o'ten pl&y & %&Dor

role in proble% di&gnostic by $&tching the %odel. In other c&ses it is necess&ry to per'or% so%e tests, inserting so%e logic, or st dying the si% l&tion report in order to di&gnose the so rce o' the proble%. 2. A'ter di&gnosing the proble%, si% l&tion c&n be sed to test & v&riety o' proble% sol tion sche%es. S ch sche%es &re o'ten tested sing $h&t/i' &n&lysis, $here the i%p&ct o' e&ch sche%e on %odel per'or%&nce is tested &nd &n&ly.ed. <he %ost 'e&sible sche%e th&t le&ds to & co%prehensive 'i( o' the proble% 4$itho t &dding &nother proble%5 is selected &s & sol tion to the proble% de'ined. 2odel 'le(ibility &nd e''iciency &llo$ 'or testing l&rge n %ber o' sol tion sche%es &nd r nning v&rio s $h&t/i' scen&rios. 3. >in&lly, the vi&bility &nd rob stness o' the sol tion sche%e proposed 'or the nderlying proble% is tested. Sensitivity &n&lysis c&n be sed to test sol tion rob stness to v&rio s so rces o' v&ri&tions. #ot &ll sol tion sche%es &re &ble to ' nction properly $ith other syste% ele%ents nder syste% constr&ints &nd $or" conditions. S%&ll v&ri&tions in s ch conditions %&y le&d to dr&stic ch&nges in syste% per'or%&nce. !ence, si% l&tion is sed to test sol tion sensitivity &nd rob stness be'ore considering the sol tion sche%e to be 'in&l.

-&(&* Continuous "m ro4ement Simulation

It is o'ten &sserted th&t the s ccess o' prod ction &nd b siness syste%s in s st&ining & cert&in level o' per'or%&nce depends on e''ort in est&blishing &nd i%ple%enting pl&ns 'or contin o s i%prove%ent. Co%p&nies do not &l$&ys $&it ntil & proble% &rises to t&"e correction &nd i%prove%ent &ctions. 2&n&gers &nd engineers o'ten believe th&t there is &l$&ys & $indo$ 'or i%prove%ent in the $&y th&t co%p&nies prod ce prod cts or provide services.<hro gh this $indo$, syste% %&n&gers &nd pl&nners o'ten 'oresee opport nities 'or %&"ing the syste% better &nd %ore prep&red to '&ce ' t re ch&llenges. As disc ssed in Ch&pter 2, this %&tches the philosophy o' le&n %&n '&ct ring &nd le&n si(/sig%& 4;SS5 st dies. <he J&p&nese ter% Aai;en is sed $idely to re'er to contin o s i%prove%ent %ethods &nd techni8 es th&t co%p&nies 'ollo$ to boost e''iciency &nd s st&in high per'or%&nce. !ence, contin o s i%prove%ent e''ort is o'ten 'oc sed on de'ining i%prove%ent opport nities, proposing i%prove%ent pl&ns, &nd testing the v&lidity o' the i%prove%ent &ctions proposed.A lot o' those &ctions &rise 'ro% developing ' t re oper&tion&l pl&ns 'or e(p&nsion &nd b siness gro$th. Jpport nities o' i%prove%ent c&n be de'ined by thin"ing o' better $&ys to do things, br&instor%ing on cost &nd $&ste red ction %e&s res, &nd bench%&r"ing co%petition. Si% l&tion c&n be tili.ed to se&rch 'or s ch opport nities &nd 'or testing di''erent i%prove%ent pl&ns. So%e o' those pl&ns %&y not be 'e&sibleO hence, the si% l&tion c&n test pl&n 'e&sibility &nd &pplic&bility. >in&lly, the %odel c&n be sed to v&lid&te the i%prove%ent pl&n selected &nd to chec" its rob stness &nd vi&bility. >ig re 6.6 sho$s the tili.&tion o' si% l&tion in &iding the contin o s i%prove%ent e''ort.

Figure -&,# Contin o s i%prove%ent $ith si% l&tion. As sho$n in the 'ig re, contin o s i%prove%ent $ith si% l&tion incl des de'ining opport nities, testing i%prove%ent pl&ns, &nd v&lid&ting the pl&n selected. ). 1e'ining &n i%prove%ent opport nity c&n be &ppro&ched in di''erent $&ys. <hro$ing ide&s 'or i%prove%ent into ro tinely sched led %eetings, holding br&instor%ing sessions, &nd bench%&r"ing co%petition &re e(&%ples o' techni8 es &dopted to gener&te ide&s &nd opport nities 'or i%prove%ent. !&ving & represent&tive si% l&tion %odel th&t c&n be $&tched &nd &n&ly.ed by & contin o s i%prove%ent te&% c&n be bene'ici&l in this reg&rd.

Ch&nges s ggested &nd ide&s r&ised by te&% %e%bers c&n be tested &nd ev&l &ted 8 ic"ly &t virt &lly no cost. 2. A'ter & cert&in i%prove%ent opport nity is de'ined, it o'ten necess&ry to provide & pl&n &nd %ech&nis% 'or &chieving the de'ined i%prove%ent. 1i''erent %ethods &nd techni8 es c&n be sed to i%prove per'or%&nce, c t cost, &nd red ce $&ste. ;e&n %&n '&ct ring techni8 es &re co%%only sed by engineers to &chieve i%prove%ent pl&ns. In le&n %&n '&ct ring, the 'oc s is on cost &nd $&ste red ction, red cing %&n '&ct ring le&d ti%e &nd co%p&ny response ti%e, &nd incre&sing e8 ip%ent &nd l&bor e''ectiveness. Si% l&tion c&n &id in &pplying &nd testing the i%p&ct o' the i%prove%ent pl&ns proposed. 3. >in&lly, the vi&bility &nd rob stness o' the i%prove%ent &ction or pl&n proposed is tested. Sensitivity &n&lysis c&n be sed to test the pl&nEs rob stness to v&rio s so rces o' v&ri&tions. -sing the si% l&tion %odel, i%prove%ent &ctions &re i%ple%ented &nd per'or%&nce is ev&l &ted to 8 &nti'y the i%prove%ent. <he sensitivity o' the pl&n to & v&riety o' oper&ting conditions is then tested to %&"e s re th&t there no neg&tive i%p&ct $ill occ r on other syste% ' nctions &nd th&t the i%prove%ent level c&n be s st&ined.
H)I

We &lso sed critic&l/to/8 &lity ch&r&cteristics 4C<Bs5 &s &n &ltern&tive si(/sig%& ter%inology.

-&* Systemati$ Simulation A

roa$h

<he &ppro&ch 'ollo$ed 'or &pplying e&ch o' the three c&tegories o' si% l&tion st dies disc ssed in Section 6.2 h&s & speci'ic n&t re &nd re8 ire%ents. >or the three c&tegories, ho$ever, $e c&n 'ollo$ & generic &nd syste%&tic &ppro&ch 'or &pplying & si% l&tion st dy e''ectively. <his &ppro&ch consists o' co%%on st&ges 'or per'or%ing the si% l&tion st dy, &s sho$n in >ig re 6.6.

Figure -&-# Syste%&tic si% l&tion &ppro&ch.

<he &ppro&ch sho$n in the 'ig re is & typic&l engineering %ethodology 'or the three c&tegories o' si% l&tion st dies 4i.e., syste% design, proble% solving, &nd syste% i%prove%ent5 th&t p ts the si% l&tion process into the conte(t o' engineering sol tion %ethods. ?ngineers &nd si% l&tion &n&lysts o'ten &dopt &nd se s ch &n &ppro&ch i%plicitly in re&l/$orld si% l&tion st dies $itho t str ct ring it into st&ges &nd steps. <he &ppro&ch is &n engineering %ethodology th&t consists o' 'ive iter&tive st&ges, &s sho$n in >ig re 6.69 Identi'y the si% l&tion proble%, develop sol tion &ltern&tives, ev&l &te sol tion &ltern&tives, select the best sol tion &ltern&tive, &nd i%ple%ent the sol tion selected.

-&*&' "dentifi$ation of the !ro3lem or O

ortunity

Any si% l&tion st dy sho ld st&rt by de'ining the proble% to be solved &nd &n&ly.ed thro gh si% l&tion. Syste% design, per'or%&nce, &nd i%prove%ent ch&llenge &re three types o' si% l&tion proble%s. In contin o s i%prove%ent st dies, the proble% is vie$ed &s &n i%prove%ent opport nity i' there is no speci'ic proble% to be solved. <he si% l&tion proble% is de'ined in ter%s o' st dy scope, st dy obDectives, &nd %odel &ss %ptions. ). 1e'ining the scope o' & si% l&tion st dy st&rts by describing the syste% str ct re, logic, &nd ' nction&lity. A process %&p or 'lo$ch&rt c&n be sed to de'ine the process str ct re &nd 'lo$. <he st dy scope &lso incl des & cle&r description o' the syste% proble% or opport nity. 1escribing the ch&llenges, li%it&tions, &nd iss es $ith the c rrent st&te &nd e(pl&ining $hy s ch iss es t&"e pl&ce provide the si% l&tion &n&lyst $ith & s&tis'&ctory de'inition o' the si% l&tion proble%. I' possible, the proble% c&n be str ct red %&the%&tic&lly or by sing & sche%&tic di&gr&% to %&"e it e&sy 'or the &n&lyst to nderst&nd v&rio s &spects o' the proble%. 2. @&sed on the proble% scope, the si% l&tion st dy obDectives &re set &nd e(pressed in the %ost concise $&y possible. <ogether $ith the over&ll obDective o' the st dy, $hich is o'ten &i%ed &t solving the proble% th&t h&s been de'ined, the st dy s bobDectives provide the criteri& &nd %ech&nis% 'or solving the proble%. >or e(&%ple, the over&ll obDective %ight be to incre&se c sto%er s&tis'&ction, &nd the speci'ic obDectives %ight be to red ce both c sto%er $&iting ti%e &nd processing ti%e. It is &l$&ys better to de'ine 8 &nti'i&ble %etrics th&t %e&s re the obDectives th&t h&ve been de'ined. 3. 2odel &ss %ptions &re de'ined to n&rro$ do$n the scope o' the proble% &nd to sh&rpen the st dy obDectives. 2odel &ss %ptions &lso help in directing %odel &n&lysis &nd in selecting proper %ethods &nd techni8 es to solve the proble% de'ined. 2ost %odel &ss %ptions c&n be rel&ted to the s rro nding syste% conditions, the constr&ints on syste% ' nction&lity, &nd the &v&il&bility o' %onet&ry &nd technic&l reso rces. >or e(&%ple, the pl&nt oper&ting p&ttern %ight be & restriction on its d&ily prod ctivity. 2odel &ss %ptions &re &lso 'oc sed on the so rce &nd type o' d&t& sed in the %odel, s ch &s reli&bility d&t&, processing ti%es, &nd decision/%&"ing r les. St&tistic&l %ethods &nd distrib tions &re &l$&ys tili.ed in this reg&rd.

-&*&( De4elo ment of Solution and "m ro4ement Alternati4es

Jnce & proble% or &n opport nity h&s been identi'ied, sol tion &ltern&tives c&n be e(plored &nd developed. S ch &ltern&tives &re e(plored by &n&ly.ing the proble% v&ri&bles, obDectives, &nd constr&ints &nd &tte%pting to de'ine the p&r&%eter set p, the str ct r&l ch&nges, or the logic&l design necess&ry to %eet the proble% obDectives $itho t viol&ting its physic&l &nd logic&l constr&ints. Sol tion &ltern&tives &re then developed by str ct ring the set o' %ethods &nd &ctions $ithin e&ch &ltern&tive e(plored. It is i%port&nt to p&y &ttention to the $&y in $hich $e present the set o' sol tion &ltern&tives th&t h&s been developed. <he &n&lyst sho ld be &ble to pinpoint the di''erences &%ong sol tion &ltern&tives. <&bles, gr&phs, &nd s %%&ry sheets c&n be sed to present &nd co%p&re the sol tion &ltern&tives developed.

).

1e'ining the set o' sol tion &ltern&tives is &ppro&ched by e(ploring the sol tion do%&in &nd listing the potenti&l sol tion %ethods. In &ddition to & thoro gh nderst&nding &nd solid e(perience in the nderlying syste%, & co%bin&tion o' cre&tivity &nd engineering s"ills is re8 ired to gener&te & set o' sol tion &ltern&tives. @r&instor%ing sessions &nd ide&/gener&ting techni8 es c&n be sed in this reg&rd in & %&nner si%il&r to th&t sed 'or gener&ted i%prove%ent ide&s. With si% l&tion, ho$ever, sol tion ide&s c&n be gener&ted in & % ch e&sier $&y. @y providing & close/to/re&lity represent&tion o' the syste%, the %odel c&n be sed to observe the syste%Es dyn&%ic beh&vior, pro%ote ide& gener&tion, &nd predict the per'or%&nce 'or e&ch ide& gener&ted. 2. ?(ploring the sol tion do%&in o' the nderlying proble% o'ten le&ds to de'ining & set o' potenti&l sol tion ide&s. <hose ide&s &re tr&ns'or%ed into & set o' sol tion &ltern&tives by str ct ring e&ch ide& in ter%s o' & %ethod &nd & set o' &ctions &nd ch&nges. <he 'e&sibility o' e&ch ide& is chec"ed &nd v&lid&ted in the process o' developing the set o' sol tion &ltern&tives. ?&ch 'e&sible sol tion ide& is p t in ter%s o' & concise &nd speci'ic pl&n th&t is &i%ed &t %&"ing the ch&nge re8 ired to eli%in&te the proble% or to i%prove syste% per'or%&nce. Str ct ring &n ide& into & sol tion &ltern&tive re8 ires "no$ledge in engineering syste%s, represent&tion %ethods, &nd best pr&ctices.

-&*&* E4aluation of Solution Alternati4es

In this st&ge, the set o' sol tion &ltern&tives selected &re ev&l &ted b&sed on the obDectives de'ined, &nd sol tion &ltern&tives &re r&n"ed &ccordingly. ?v&l &ting sol tion &ltern&tives is essenti&l to co%p&re their per'or%&nce &nd %&"e it e&sier 'or decision %&"ers to select the best sol tion &ltern&tive. ?v&l &ting sol tion &ltern&tives incl des de'ining & set o' per'or%&nce criteri& &nd ev&l &ting e&ch sol tion &ltern&tive 'or the set o' criteri& de'ined. Si% l&tion pl&ys & pri%&ry role in per'or%&nce ev&l &tion nder co%ple(, dyn&%ic, &nd stoch&stic beh&vior o' re&l/$orld syste%s. ). 1e'ining & set o' per'or%&nce criteri& incl des providing the proper 8 &ntit&tive %etrics th&t %e&s re &n &spect o' the syste% per'or%&nce. <his incl des %onet&ry criteri& s ch &s cost, pro'it, &nd r&te o' ret rn, &s $ell &s technic&l &nd oper&tion&l criteri& s ch &s thro ghp t, e''ectiveness, &nd delivery speed. <his %&y &lso incl de so%e sociotechnic&l criteri& th&t %&y be di''ic lt to %e&s re, s ch &s &ppe&l, rep t&tion, &nd c sto%er s&tis'&ction. It is i%port&nt to select, &s % ch &s possible, & set o' %e&s r&ble criteri& th&t c&n be &ssessed sing the si% l&tion %odel &nd c&n be sed &s & b&sis to co%p&re sol tion &ltern&tives. In &ddition to the si% l&tion c&p&bility o' 'le(ible progr&%%ing, %odel co nters, t&llies, &nd st&tistics provide the &n&lyst $ith & v&riety o' techni8 es to %e&s re per'or%&nce criteri&. So%e o' those criteri&, s ch &s thro ghp t, le&d ti%e, &nd inventory level, c&n be t&"en &t & syste% level. Jthers, s ch &s tili.&tion, e''ectiveness, &nd reli&bility, c&n be t&"en &t the process or oper&tion level. >in&lly, so%e o' the %e&s res, s ch &s 8 &lity, cost, &nd n %ber o' nits processed, c&n be rel&ted to the syste% entity. 2. Jnce the set o' per'or%&nce %e&s res &re de'ined, & si% l&tion %odel c&n be sed to ev&l &te the set o' sol tion &ltern&tives. ?&ch si% l&tion/b&sed ev&l &tion provides & set o' per'or%&nce %e&s res. Sol tion &ltern&tives &re then co%p&red &nd the &ltern&tive $ith the best per'or%&nce is selected. <he si% l&tion %odel is set to 8 &nti'y co%p&rison criteri& 4& set o' per'or%&nce %e&s res5 $itho t the need 'or & closed/'or% de'inition o' e&ch %odel response. >or e(&%ple, in & %&n '&ct ring syste% %odel, syste% thro ghp t is & ' nction o' b ''er si.es, %&chine cycle ti%es &nd reli&bility, conveyor speeds, &nd so on. It %&y not be possible &n&lytic&lly to provide & closed/'or% de'inition o' the thro ghp t ' nction. !o$ever, r&ndo%ness o'ten co%bines the response sign&ls o' si% l&tion %odels. L&ri&bility enco%p&ssed in the si% l&tion %odel le&ds to v&ri&bility in the %odel o tco%e. !ence, the stoch&stic n&t re o' si% l&tion o tco%es sho ld be incl ded in &n &ltern&tive co%p&rison. 2 ltiple si% l&tion replic&tions &re o'ten sed to re'lect v&ri&bility in %odel o tco%es, &nd st&tistic&l %e&s res s ch &s o tp t %e&ns, v&ri&nces, &nd con'idence interv&ls &re co%p ted to represent s ch v&ri&bility.

-&*&% Sele$tion of the Best Alternati4e

In this st&ge, the best sol tion or i%prove%ent &ltern&tive is selected b&sed on the si% l&tion ev&l &tion in the preceding step. <he best sol tion in ter%s o' v&l es o' per'or%&nce %e&s res is selected by co%p&ring its over&ll per'or%&nce to the rest o' the sol tion &ltern&tives. Co%p&ring the per'or%&nce o' sol tion &ltern&tives b&sed on % ltiple obDectives, ho$ever, is not &n e&sy t&s". Selecting the best sol tion &ltern&tive is & decision/%&"ing process th&t is b&sed on &n over&rching &ssess%ent o' sol tion &ltern&tives. I' the co%p&rison is %&de b&sed on & single obDective, the &ltern&tive $ith best per'or%&nce is selected. When % ltiple per'or%&nce %e&s res &re sed, % lticriteri& decision %&"ing 42C125 is re8 ired to &ssess sol tion &ltern&tives b&sed on % ltiple decision criteri&. I' one &ltern&tive do%in&tes other &ltern&tives 'or &ll per'or%&nce criteri&, the nondo%in&ted sol tion is selected. Jther$ise, 2C12 techni8 es s ch &s go&l progr&%%ing &nd &n&lytic&l hier&rchy process &re sed to s pport the % lticriteri& decision. S ch %ethods &re b&sed on both s bDective &nd obDective D dg%ents o' & decision %&"er or gro p o' e(perts in $eighting decision criteri& &nd r&n"ing sol tion &ltern&tives. An over&ll tility ' nction Ho'ten re'erred to &s & % lti&ttrib te tility ' nction 42A->5I is developed by co%bining criteri& $eights &nd per'or%&nce ev&l &tion. 2A-> is sed to provide &n over&rching tility score to r&n" sol tion &ltern&tives. St&tistic&l co%p&r&tive &n&lysis &nd hypothesis testing &re &lso sed to co%p&re sol tion &ltern&tives. It is $orth %entioning, ho$ever, th&t other '&ctors th&t %&y not be ev&l &ted sing the si% l&tion %odel, s ch &s econo%ic&l '&ctors, s&'ety, &nd environ%ent&l concerns, &re o'ten incl ded in the selection process. <he sol tion selected is then reco%%ended 'or i%ple%ent&tion in the re&l/$orld syste% &s & design 'or & ne$ syste%, &s & sol tion to &n e(isting proble%, or &s &n i%prove%ent pl&n 'or &n e(isting syste%.

-&*&, "m lementation of the Alternati4e Sele$ted

>in&lly, the sol tion &ltern&tive selected is considered 'or i%ple%ent&tion. 1epending on the n&t re o' the proble%, i%ple%ent&tion prep&r&tions &re o'ten t&"en prior to the &ct &l constr ction&l ch&nges on the 'loor. ;i"e &ny other proDect, i%ple%enting the sol tion reco%%ended by the si% l&tion st dy is per'or%ed into ph&ses. ProDect %&n&ge%ent techni8 es &re o'ten sed to str ct re the ti%e 'r&%e 'or the e(ec tion pl&n &nd to &lloc&te the reso rces re8 ired. Altho gh so%e o' the p&r&%eters &nd logic reco%%ended by the si% l&tion %&y be &ltered d ring the inst&ll&tion st&ge, the %odel c&n still be sed &s & tool 'or g iding the i%ple%ent&tion process &t its v&rio s st&ges. 2ilestone %eetings 'or progress revie$s c&n bene'it 'ro% si% l&tion res lts &nd si% l&tion &ni%&tion o' the syste% str ct re.

-&% Ste s "n A Simulation Study

In this section $e present & proced re 'or cond cting si% l&tion st dies in ter%s o' & step/by/ step &ppro&ch 'or de'ining the si% l&tion proble%, b ilding the si% l&tion %odel, &nd cond cting si% l&tion e(peri%ents. <his proced re is & det&iled tr&nsl&tion o' the syste%&tic si% l&tion &ppro&ch presented in >ig re 6.6. >ig re 6.* is & 'lo$ch&rt o' the step/by/step si% l&tion proced re.

Figure -&.# <he si% l&tion proced re. <he si% l&tion syste%&tic &ppro&ch sho$n in >ig re 6.* represents the engineering 'r&%e$or" o' the si% l&tion st dy. <he steps %&y v&ry 'ro% one &n&lyst to &nother bec& se o' '&ctors s ch &s the n&t re o' the proble% &nd the si% l&tion so't$&re sed. !o$ever, the b ilding bloc"s o' the si% l&tion proced re &re typic&lly co%%on &%ong si% l&tion st dies. <he si% l&tion proced re, o'ten represented by & 'lo$ch&rt, consists o' the ele%ents &nd the logic&l se8 ence o' the si% l&tion st dy. It &lso incl des decision points thro gh $hich the concept &nd %odel &re chec"ed, v&lid&ted, &nd veri'ied. Iter&tive steps %&y be necess&ry to &dD st &nd %odi'y the %odel concept &nd logic. >in&lly, the proced re sho$s steps th&t c&n be e(ec ted in p&r&llel $ith other steps.

-&%&' !ro3lem Formulation

<he si% l&tion st dy sho ld st&rt $ith & concise de'inition &nd st&te%ent o' the nderlying proble%. <he proble% st&te%ent incl des & description o' the sit &tion or the syste% o' the st dy &nd the proble% th&t needs to be solved. >or% l&ting the proble% in ter%s o' &n over&ll go&l &nd & set o' constr&ints provides & better represent&tion o' the proble% st&te%ent. A thoro gh nderst&nding o' the ele%ents &nd str ct re o' the syste% nder st dy o'ten helps in developing the proble% st&te%ent.

>or% l&ting & design proble% incl des st&ting the over&ll design obDective &nd the constr&ints on the design process. >or e(&%ple, the go&l %ight be to design & %&teri&l/h&ndling syste% th&t is c&p&ble o' tr&ns'erring & cert&in ite% 'ro% point A to point @. <he constr&ints on the design process %&y incl de cert&in thro ghp t re8 ire%ents, b dget li%it&tions, nit lo&d c&p&city, inclin&tion &nd declin&tion re8 ire%ents, &nd so on. <h s, the design proble% is 'or% l&ted s ch th&t the go&l de'ined is %et $itho t viol&ting &ny o' the constr&ints. -nli"e line&r progr&%%ing, ho$ever, in si% l&tion there is no need to e(press the proble% go&l &nd constr&ints %&the%&tic&lly. @oth &re %odeled sing di''erent ele%ents &nd reso rces o' 1?S. Si%il&rly, 'or% l&ting & proble% in &n e(isting syste% incl des st&ting the over&ll proble%/solving obDective &nd the constr&ints on the sol tion proposed. >or e(&%ple, the proble% %ight be identi'ied &s & red ction in syste% thro ghp t by & cert&in percent&ge. !ence, the si% l&tion go&l is set to boost the syste% thro ghp t to re&ch & cert&in t&rget. <he constr&ints on the sol tion proposed %&y incl de the li%ited c&p&city o' $or"st&tions, conveyor speeds, n %ber o' oper&tors, prod ct %i(, &nd b dget li%it&tion, &%ong others. <h s, the proble% is 'or% l&ted s ch th&t the go&l de'ined is %et $itho t viol&ting &ny o' the constr&ints. >in&lly, the 'or% l&tion o' &n i%prove%ent proble% %&y incl de st&ting the over&ll i%prove%ent obDective in ter%s o' % ltiple &nd o'ten co%peting go&ls $hile %eeting process constr&ints. >or e(&%ple, the 'irst go&l %ight be to red ce the %&n '&ct ring le&d ti%e by & cert&in percent&ge in order to &pply le&n %&n '&ct ring principles, &nd the second go&l, to incl de %eeting cert&in thro ghp t re8 ire%ents. @oth types o' go&ls &re o'ten s bDect to & si%il&r set o' process constr&ints, s ch &s b dget li%it&tions, v&ri&tions in %&n '&ct ring oper&tions, &nd 'lo$ re8 ire%ents. <h s, the i%prove%ent proble% is 'or% l&ted s ch th&t the t$o go&ls &re %et $itho t viol&ting &ny o' the constr&ints th&t h&ve been de'ined.

-&%&( Setting Study O3<e$ti4es

@&sed on the proble% 'or% l&tion, & set o' obDectives c&n be set to the si% l&tion st dy. S ch obDectives represent the criteri& thro gh $hich the over&ll go&l o' the st dy is &chieved. St dy obDectives si%ply indic&te 8 estions th&t sho ld be &ns$ered by the si% l&tion st dy. ?(&%ples incl de deter%ining c rrent/st&te per'or%&nce, testing design &ltern&tives, st dying the i%p&ct o' speeding p the %&inline conveyor, &nd opti%i.ing the n %ber o' c&rriers in & %&teri&l/h&ndling syste%. Speci'ying st dy obDectives serves v&rio s p rposes. >irst, $e c&n decide i' si% l&tion is the best tool to se to solve the nderlying proble%. 1o $e h&ve eno gh d&t& to deter%ine %etrics 'or the obDectives th&t h&ve been de'inedK C&n $e se &n&lytic&l %ethods to &ns$er the 8 estions th&t h&ve been r&isedK Is the so't$&re tool c&p&ble o' presenting &nd &n&ly.ing st dy re8 ire%ents so th&t the st dy obDectives &re &chievedK <hese types o' 8 estions c&n be better &ns$ered by st&ting the obDectives o' the si% l&tion st dy n&%big o sly. In ter%s o' %odeling, %odel ele%ents &nd %odel logic &re selected &nd designed to provide &ppropri&te %e&s res o' per'or%&nce th&t 8 &nti'y the st dy obDectives. >or e(&%ple, to %eet st dy obDectives, &n &cc % l&ting conveyor is sed &s & %&teri&l/h&ndling syste%. Also, st&tistics th&t represent thro ghp t, le&d ti%e, del&ys, &nd n %ber o' c&rriers &re inserted into the %odel over ti%e to &cc % l&te the d&t& re8 ired &t the end o' the r n. Speci'ying st dy obDectives provides & cle&r vision 'or &n&ly.ing %odel o tp ts so th&t the 8 estions r&ised &re &ns$ered by the si% l&tion report.

-&%&* Con$e tual )odeling

1eveloping & concept &l %odel is the process thro gh $hich the %odeler &bstr&cts the str ct re, ' nction&lity, &nd essenti&l 'e&t res o' & re&l/$orld syste% into & str ct r&l &nd logic&l represent&tion th&t is tr&ns'er&ble into & si% l&tion %odel. <he %odel concept c&n be & si%ple or

& co%ple( gr&phic&l represent&tion, s ch &s & bloc" di&gr&%, & 'lo$ch&rt, or & process %&p th&t depicts "ey ch&r&cteristics o' the si% l&ted syste%, s ch &s inp ts, ele%ents, p&r&%eters, logic, 'lo$, &nd o tp ts. S ch & represent&tion sho ld event &lly be progr&%%&ble &nd tr&ns'er&ble into & si% l&tion %odel sing &v&il&ble si% l&tion so't$&re tools. <h s, & s ccess' l %odel concept is one th&t t&"es into consider&tion the %ethod o' tr&ns'erring e&ch &bstr&cted ch&r&cteristic, b ilding e&ch %odel ele%ent, &nd progr&%%ing the concept &l logic sing the so't$&re tool. <he &rt o' concept &l %odeling is & co%bin&tion o' syste% "no$ledge &nd %odel/b ilding s"ills. <he %odeler st&rts by & est&blishing & thoro gh nderst&nding o' the syste%, $hether it is & ne$ or &n e(isting syste%. <he %odeler st dies syste% inp ts, ele%ents, str ct re, logic, 'lo$, &nd o tp ts &nd &bstr&cts the over&ll str ct re &nd the interrel&tionships o' str ct re ele%ents into & concept &l %odel. <he %odel concept is presented considering the co%ponents &nd c&p&bilities o' the si% l&tion environ%ent. >or e(&%ple, developing & concept th&t incl des & po$er/&nd/'ree conveyor syste% sho ld 'ind o t i' there is & c&p&bility 'or %odeling s ch & syste% in the si% l&tion so't$&re tool being sed. Pey p&r&%eters o' syste% ele%ents &re &lso speci'ied &s & p&rt o' the %odel concept. >or e(&%ple, the concept o' sing & convey&nce syste% to tr&ns'er entities 'ro% point A to point @ sho ld incl de p&r&%eters o' conveyor type, speed, reli&bility, &nd c&p&city. S ch p&r&%eters g ide d&t& collection &nd ele%ent selection. >in&lly, the %odel concept is developed t&"ing into consider&tion the proble% &s it $&s 'or% l&ted &nd the obDectives o' the si% l&tion st dy. >ig re 6.+ sho$s the re8 ire%ents 'or concept &l %odeling.

Figure -&/# 1eveloping & %odel concept.

-&%&% Data Colle$tion

Si% l&tion %odels &re d&t&/driven co%p ter progr&%s th&t receive inp t d&t&, e(ec te the logic designed, &nd prod ce cert&in o tp ts. !ence, the d&t& collection step is & "ey co%ponent o' &ny si% l&tion st dy. Si% l&tion d&t& c&n, ho$ever, be collected in p&r&llel to b ilding & %odel sing the si% l&tion so't$&re. <his is reco%%ended since d&t& collection %&y be ti%e cons %ing in so%e c&ses, &nd b ilding the %odel str ct re &nd designing %odel logic c&n be independent o' the %odel d&t&. 1e'& lt p&r&%eters &nd generic d&t& c&n be sed initi&lly ntil the syste% d&t& &re collected. <he 8 &lity o' d&t& sed in the %odels drives the over&ll %odel 8 &lity &nd v&lidity &nd &''ects the &cc r&cy o' the %odel res lts &nd st&tistics collected. !ence, the ter% gar'age,in,gar'age,out 4GIGJ5 is o'ten co%%on in %odeling. >or & %odel to be represent&tive, it h&s to be driven by represent&tive d&t& in ter%s o' d&t& integrity, &cc r&cy, &nd co%prehensiveness. 1&t& ele%ents re8 ired 'or constr cting & si% l&tion %odel &re o'ten deter%ined b&sed on the %odel concept, %odel str ct re, &nd the n&t re &nd type o' %odel o tco%es &nd st&tistics to be collected. >or e(&%ple, typic&l d&t& collected 'or %odeling & b&n" oper&tion incl de c sto%er inter&rriv&l ti%es &nd the service ti%e distrib tion.

1epending on the n&t re o' the si% l&tion st dy, %odel d&t& &re collected by revie$ing historic&l d&t&, by observing &nd %onitoring syste% oper&tions, &nd by sing bench%&r" d&t& &ss %ptions. <he three types o' d&t& collected in %odel develop%ent &re sho$n in >ig re 6.,.

Figure -&0# Collecting %odel d&t&. !istoric&l d&t& &re o'ten sed $hen %odeling e(isting syste%s th&t h&ve been in oper&tion 'or & cert&in ti%e. ?(&%ples o' historic&l d&t& incl de &ct &l prod ction d&t&, %&inten&nce d&t&, inventory records, c sto%er 'eedb&c"s, &nd $ee"ly &nd %onthly reports on oper&tions per'or%&nce. St&tistic&l %ethods o' inp t %odeling 4disc ssed brie'ly in Appendi( A5, s ch &s descriptive st&tistics, re%oving o tliers, &nd 'itting d&t& distrib tions pl&y &n i%port&nt role in &n&ly.ing historic&l d&t&. In c&se historic&l d&t& &re not &v&il&ble 'or & s ''icient period o' ti%e, the &ct &l per'or%&nce o' syste% c&n be observed to collect pertinent d&t&. <i%e st dies sing & stop$&tch &re co%%on d&t& collection %ethods in this reg&rd. A ti%e st dy ses & st&nd&rd 'or% to collect d&t& &nd st&rts by %onitoring syste% beh&vior. Syste% %onitoring incl des $&tching the oper&tion o' & cert&in syste% ele%ent, nderst&nding its ' nction&lity, &nd deciding on the p&r&%eters to be collected &nd the ti%es o' collecting d&t&. <he d&t& collected sho ld be st&tistic&lly represent&tive &nd sho ld be distrib ted to cover the entire spectr % o' syste% oper&tion. Selecting the &ppropri&te s&%ple si.e &nd ti%es o' observ&tions &re t&"en is essenti&l to obt&in represent&tive d&t&. >in&lly, $hen no historic&l d&t& &re &v&il&ble &nd it is not per%issible to collect d&t& sing ti%e st dies, si% l&tion d&t& %&y be bench%&r"ed or &ss %ed. ?d c&ted g esses, bench%&r" d&t&, &nd theoretic&l st&tistic&l %odels c&n be sed to develop the d&t& re8 ired 'or si% l&tion. <his o'ten occ rs $hen %odeling ne$ syste%s &nd testing proposed design &ltern&tives. ?(perience in both %odeling &nd syste% &n&lysis o'ten e8 ips the si% l&tion &n&lyst $ith the "no$ledge &nd s"ill th&t is essenti&l to provide ed c&ted g esses o' & v&riety o' %odel d&t&. 1&t& 'ro% & si%il&r b siness or %&n '&ct ring process c&n be sed in the %odel, &t le&st &s & st&rting point. >in&lly, theoretic&l st&tistic&l distrib tions $ith esti%&ted p&r&%eters c&n be sed 'or %odel d&t&. All types o' &ss %ed d&t& &re s bDect to %odi'ic&tion &nd &lter&tion &s the %odel b ilding progresses &nd %ore insight is g&ined into %odel beh&vior. !nput Modeling <he p&rt o' si% l&tion concerned $ith selecting proper d&t& represent&tion 'or deter%inistic %odel p&r&%eters &nd stoch&stic %odel v&ri&bles is re'erred to &s input modeling. 1eter%inistic %odel d&t& &re typic&lly e&sier to predict &nd set since their v&l es $ill be 'i(ed in the %odel over the r n ti%e. ?(&%ples incl de const&nt de%&nd, sched led deliveries, 'i(ed cycle ti%es, 'i(ed b ''er c&p&city, &nd const&nt n %ber o' l&bor or %&chine reso rces. L&l es 'or these p&r&%eters &re 'i(ed initi&lly i' "no$n 'or cert&in or c&n be vie$ed &s decision or design v&ri&bles 4control '&ctors5 th&t &re set to their opti% % levels thro gh e(peri%ent&l design or opti%i.&tion %ethods. Stoch&stic %odel inp ts, on the other h&nd, c&n be vie$ed &s noise 4r&ndo%5 '&ctors th&t ch&nge over ti%e r&ndo%ly or &ccording to cert&in prob&bility distrib tions. Jver the si% l&tion r n period, v&l es 'or these v&ri&bles &re s&%pled 'ro% theoretic&l or e%piric&l prob&bility distrib tions. !ence, it is 'irst re8 ired to 'it & cert&in prob&bility distrib tion to the d&t& collected, insert into the %odel the distrib tion selected, &nd sing the s&%pling %ech&nis% disc ssed in Ch&pter 6, gener&te s&%ples o' the v&ri&ble over the r n ti%e. I' v&l es ch&nge r&ndo%ly, ni'or% r&ndo% n %bers c&n be sed to gener&te v&l es 'or the v&ri&ble. As disc ssed e&rlier, in

&ddition to the v&ri&bility o' &d&ption over the r n ti%e, di''erent r&ndo% n %ber stre&%s o'ten res lt in di''erent sets o' d&t& s&%pled, $hich le&ds to stoch&stic v&ri&bility in %odel o tco%es. !ence, % ltiple si% l&tion replic&tions &re o'ten sed to esti%&te e&ch %odel per'or%&nce %e&s re in ter%s o' & %e&n &nd & v&ri&nce. 2&ny st&tistic&l %ethods &re typic&lly sed 'or %odeling si% l&tion inp ts. !o$ever, the b&sic concept in these %ethods is si%ple. I' $e co%p te the %e&n o' & s&%ple o' 20 n %bers, the v&l e obt&ined $ill gener&lly be % ch di''erent 'ro% the pop l&tion %e&n 4i.e., the s&%pling distrib tion is not & good &ppro(i%&tion o' the d&t& distrib tion5. @ t i' $e s&%pled %&ny sets o' 20 n %bers over &nd over &g&in, co%p ted the %e&n 'or e&ch set, &nd constr cted & rel&tive 're8 ency distrib tion o' &ll sets, $e $o ld event &lly 'ind th&t the distrib tion o' %e&ns is & very good &ppro(i%&tion to the s&%pling distrib tion. <he s&%pling distrib tion o' the %e&n is & theoretic&l distrib tion th&t is &ppro&ched &s the n %ber o' s&%ples in the rel&tive 're8 ency distrib tion incre&ses. As the n %ber o' s&%ples &ppro&ches in'inity, the rel&tive 're8 ency distrib tion o' &ct &l d&t& &ppro&ches the s&%pling distrib tion in the si% l&tion %odel. <he s&%pling distrib tion o' the %e&n 'or & cert&in s&%ple si.e is D st &n e(&%pleO there is & di''erent s&%pling distrib tion 'or other s&%ple si.es.<here &re &lso & n %ber o' st&tistic&l tests designed to &ssess $hether & given distrib tion describes & d&t& set '&ith' lly. Altho gh inp t %odeling c&n be &ppro&ched di''erently, & generic proced re 'or inp t %odeling 4'itting d&t& collected to prob&bility distrib tions5 c&n be developed. <his proced re is &lso b ilt in &s & %od le in %&ny si% l&tion p&c"&ges. <he proced re incl des the 'ollo$ing steps9 Step B3 Plot the d&t&. o -se & histogr&% &nd s %%&ry st&tistics to deter%ine the gener&l ch&r&cteristics o' the nderlying distrib tion. Step 23 Select & '&%ily o' distrib tions. o -se the res lts o' step B to select & set o' Fre&son&bleA distrib tions. o >it e&ch distrib tion to the d&t& observed &nd esti%&te the distrib tion p&r&%eters. Step <3 Select the best distrib tion. o 1eter%ine $hich o' the 'itted distrib tions best represents the d&t& observed sing one or %ore &ppropri&te st&tistics. Step @3 Chec" the distrib tion 8 &lity. o 1eter%ine the distrib tion goodness o' 'it9 Chi/s8 &re test Pol%ogorovCS%irnov test AndersonC1&rling test 2&ny si% l&tion boo"s incl de & det&iled description o' co%%on prob&bility distrib tions in si% l&tion st dies. 42&ny o' these &re disc ssed in Appendi( A5, $hich incl des & s %%&ry o' p&r&%eters, &pplic&tions, &nd gr&phic&l represent&tions o' prob&bility distrib tions th&t &re co%%only sed in conD nction $ith si% l&tion. Si% l&tion boo"s &lso provide s ggestions 'or selecting prob&bility distrib tions 'or di''erent d&t& sit &tions. <&ble 6.) presents & s %%&ry o' so%e co%%only sed distrib tions in si% l&tion st dies o' v&rio s d&t& sit &tions in %&n '&ct ring &pplic&tions. +a3le -&'# Summary of Commonly :sed Distri3utions in Simulation Studies Jpen t&ble &s spre&dsheet Simulation 9aria3le 2&chine cycle ti%e Speed 'or conveyor,& to%&ted !ro3a3ility Distri3ution -ni'or% <ri&ng l&r -ni'or% =ey Distri3ution !arameters 4%in.,%&(.5 4%in.,%&(.,%ode5 4%in.,%&(.5

+a3le -&'# Summary of Commonly :sed Distri3utions in Simulation Studies Jpen t&ble &s spre&dsheet Simulation 9aria3le g ided vehicle 4AGL5, 'or" tr c",etc. <r&vel ti%e 'or AGL, 'or" tr c",etc. 1o$nti%e 4ti%e to rep&ir5 -pti%e 4ti%e bet$een '&il res5 Scr&p r&te Inter&rriv&l ti%e <ri&ng l&r ;ognor%&l <ri&ng l&r ?rl&ng ?(ponenti&l @ino%i&l ?rl&ng ?(ponenti&l ;ognor%&l @ino%i&l ?%piric&l discrete 4%in.,%&(.,%ode5 4%e&n,st&nd&rd devi&tion5 4%in.,%&(.,%ode5 4%e&n, A C 35 T ):2<@> Q scr&pped 4%e&n, A T 25 T ):2<@A 4%e&n,st&nd&rd devi&tion5 Q prod ct %i( Q A,Q @,etc. !ro3a3ility Distri3ution =ey Distri3ution !arameters

P&rt type &ssign%ent

-&%&, )odel Building

1&t& collection &nd %odel b ilding o'ten cons %e the %&Dority o' the ti%e re8 ired 'or co%pletion o' & si% l&tion proDect. <o red ce s ch ti%e, the %odeler sho ld st&rt b ilding the si% l&tion %odel $hile d&t& &re being collected. <he concept &l %odel c&n be sed to constr ct the co%p ter %odel sing &ss %ed d&t& ntil the d&t& collected beco%e &v&il&ble. <he overl&p bet$een %odel b ilding &nd d&t& collection does not &''ect the logic&l se8 ence o' the si% l&tion proced re. Constr cting %odel co%ponents, entity 'lo$, &nd logic depends %ostly on the %odel concept &nd is in %ost c&ses independent o' %odel d&t&. Jnce the %odel is re&dy, %odel inp t d&t& &nd p&r&%eter settings c&n be inserted into the %odel l&ter. Also, since & l&rge portion o' & si% l&tion st dy is o'ten spent in collecting %odel d&t&, b ilding the %odel si% lt&neo sly red ces signi'ic&ntly the over&ll d r&tion o' the si% l&tion st dy &nd provides %ore ti%e 'or %odel &n&lysis &nd e(peri%ent&tion. <here is no st&nd&rd proced re 'or b ilding & si% l&tion %odel. <he proced re is o'ten b&sed on the %odeler &ppro&ch &nd on the si% l&tion so't$&re tool sed. !o$ever, & generic proced re 'or b ilding & si% l&tion %odel e''ectively is e(pected to incl de "ey b&sic steps. <his proced re is sho$n in >ig re 6.)0.

Figure -&'1# 2odel/b ilding proced re. <he %odel/b ilding proced re st&rts by i%porting & co%p ter/&ided design 4CA15 'ile o' the syste% l&yo t. <he CA1 'ile is & t$o/ or three/di%ension&l l&yo t 'ile developed in &ny co%%erci&l dr&'ting &nd design so't$&re tool. Altho gh the l&yo t h&s no i%p&ct on %odel per'or%&nce, it is & se' l gr&phic&l represent&tion o' the syste% th&t provides re&listic syste% &ni%&tion, especi&lly & three/di%ension&l l&yo t. @ec& se o' the l&rge si.e o' three/di%ension&l 'iles, %ost si% l&tion %odels &re b ilt on & t$o/di%ension&l l&yo t, $hich provides loc&tions &nd sp&ces 'or v&rio s syste% ele%ents, s ch &s %&chines, pieces o' e8 ip%ent, conveyors, stor&ge, &nd lo&dingC nlo&ding dec"s. It &lso provides & physic&l re'erence to the entity 'lo$ by sho$ing physic&l syste% ele%ents s ch &s &isles, $&l"$&ys, $&lls, e%pty sp&ces, &nd o''ices. <h s, $hen r nning the %odel, the entity 'lo$ $ithin di''erent %odel ele%ents $ill %&tch the 'lo$ o' p&rts or c sto%ers in the re&l/$orld syste%. 1i''erent %odel ele%ents &nd co%ponents &re b ilt sing the set o' %od les or b ilding bloc"s de'ined in the si% l&tion so't$&re tool sed. ?(&%ples incl de reso rces or %&chines, 8 e es or b ''ers, conveyors, &nd lo&ds or p&rts. 2ost si% l&tion tools h&ve b ilt/in, re&dy/to/ se, &nd dr&g/&nd/drop %od les o' s ch co%ponents $ith so%e gr&phic&l represent&tion. Constr cting & %odel co%ponent incl des selecting the si% l&tion ele%ent th&t best represents the co%ponent &nd loc&ting the ele%ent on its &ct &l loc&tions in the syste%Es CA1 l&yo t. At this st&ge, %odel ele%ents &re not rel&ted to e&ch other. ?le%ent interrel&tionships, decision points, &nd entity 'lo$ &re de'ined by developing the %odel logic. 1epending on the si% l&tion tool sed, the %odel logic c&n be developed by de'ining in &nd o t ro ting r les &t e&ch si% l&tion ele%ent, the %odel logic c&n be $ritten &nd deb gged in & cert&in editing environ%ent, or the t$o %&y be co%bined. Altho gh %ost ne$ si% l&tion so't$&re

tools strive to red ce the progr&%%ing e''ort, $riting code is still necess&ry to i%ple%ent the %odel logic. Jnce the %odel logic is developed, $e st&rt r nning the %odel. !o$ever, the %odel per'or%&nce &nd res lts %&y not re'lect the beh&vior o' the syste% o' interest $itho t inserting represent&tive d&t& into the %odel &nd setting the p&r&%eters o' v&rio s syste% co%ponents. S ch d&t& sho ld be collected $hile the 'irst three steps in the %odel/b ilding proced re &re being e(ec ted. As the %odel d&t& beco%e &v&il&ble, $e st&rt inserting the d&t& in the %odel 'ollo$ing the instr ctions s ggested by the si% l&tion tool sed. Gener&lly, this is &n e&sy &nd 8 ic" step in the %odel/ b ilding proced re. <he l&st step in b ilding & si% l&tion %odel is to &dd the &ni%&tion &nd gr&phic&l represent&tions o' %odel co%ple%ents &nd the s rro nding environ%ent. <ypic&l e(&%ples incl de developing &ni%&tion th&t re'lects & pl&nt environ%ent, & b&n"ing syste%, & rest& r&nt, or others. Altho gh &ni%&tion does not re&lly &dd to the 8 &lity o' the si% l&tion res lts, it helps gre&tly in %odel v&lid&tion &nd veri'ic&tion. It is &lso & gre&t selling &nd present&tion tool, especi&lly i' developed $ith three/di%ension&l gr&phics. So%e si% l&tion environ%ents &llo$ 'or only t$o/ or 2U/ di%ension&l gr&phic&l represent&tion. In &ddition to providing & gr&phics editing environ%ent, si% l&tion so't$&re tools incl de libr&ries o' re&dy/to/ se gr&phics th&t s it di''erent syste%s si% l&ted 'or both %&n '&ct ring &nd service &pplic&tions.

-&%&- )odel 9erifi$ation

2odel veri'ic&tion is the 8 &lity control chec" th&t is &pplied to the si% l&tion %odel b ilt. ;i"e &ny other co%p ter progr&%, the si% l&tion %odel sho ld per'or% b&sed on the intended logic&l design sed in b ilding the %odel. Altho gh, %odel logic c&n be de'ined sing di''erent %ethods &nd c&n be i%ple%ented sing di''erent progr&%%ing techni8 es, e(ec tion o' the logic $hen r nning the %odel sho ld re'lect the initi&l design o' the progr&%%er or %odeler. 1i''erent %ethods &re sed 'or deb gging logic&l 4progr&%%ing5 errors &s $ell &s errors in inp tting d&t& &nd setting %odel p&r&%eters. Corrected potenti&l code &nd d&t& discrep&ncies sho ld &l$&ys be veri'ied by c&re' l observ&tion o' ch&nges in %odel beh&vior. <o veri'y & %odel, $e si%ply chec" $hether the %odel is doing $h&t it is s pposed to do. >or e(&%ple, does the %odel re&d the inp t d&t& properlyK 1oes the %odel send the right p&rt to the right pl&ceK 1oes the %odel i%ple%ent the prod ction sched le prescribedK 1o c sto%ers in the %odel 'ollo$ the 8 e ing discipline proposedK 1oes the %odel provide the right o tp tK And so on. Jther veri'ic&tion techni8 es incl de &pplying r les o' co%%on sense, $&tching the %odel &ni%&tion periodic&lly d ring r n ti%e, e(&%ining %odel o tp ts, &nd &s"ing &nother %odeler to revie$ the %odel &nd chec" its beh&vior. <he observ&tions %&de by other &n&lysts &re v&l &ble since the %odel b ilder $ill be %ore 'oc sed on the progr&%%ing det&ils &nd less 'oc sed on the i%plic&tion o' di''erent progr&%%ing ele%ents. When the %odel logic is co%ple(, %ore th&n one si% l&tion &n&lyst %&y h&ve to $or" on b ilding the %odel.

-&%&. )odel 9alidation

2odel v&lid&tion is the process o' chec"ing the &cc r&cy o' the %odel represent&tion to the re&l/ $orld syste% th&t h&s been si% l&ted. It is si%ply &bo t &ns$ering the 'ollo$ing 8 estion9 1oes the %odel beh&ve si%il&rly to the si% l&ted syste%K Since the %odel $ill be sed to repl&ce the &ct &l syste% in e(peri%ent&l design &nd per'or%&nce &n&lysis, c&n $e rely in its represent&tion o' the &ct &l syste%K Pno$ing th&t the %odel is only &n &ppro(i%&tion o' the re&l/$orld syste%, "ey ch&r&cteristics o' &ct &l syste% beh&vior sho ld be c&pt red in the %odel, especi&lly those rel&ted to co%p&ring &ltern&tives, dr&$ing in'erences, &nd %&"ing decisions. !ence, necess&ry ch&nges &nd c&libr&tions th&t &re %&de to the %odel to better represent the &ct &l syste% sho ld be ret rned

to the %odel concept. <he %odel concept represents the %odelerEs &bstr&ction o' the re&l/$orld syste% str ct re &nd logic. <h s, i' the %odel $ere not ' lly v&lid, the %odel concept needs to be enh&nced &nd then tr&nsl&ted into the si% l&tion %odel. Sever&l techni8 es &re s &lly 'ollo$ed by %odelers to chec" the v&lidity o' the %odel be'ore sing it 'or s ch p rposes. ?(&%ples incl de chec"ing the d&t& sed in the %odel &nd co%p&ring the% to the &ct &l syste% d&t&, v&lid&ting the %odel logic in ter%s o' 'lo$, se8 ence, ro ting, decisions, sched ling, &nd so on, vis/V/vis the re&l/$orld syste%, &nd %&tching the res lts o' the %odel st&tistics to those o' &ct &l syste% per'or%&nce %e&s res. Cross/v&lid&tion sing &ct &l syste% res lts &nd r nning cert&in $h&t/i' scen&rios c&n &lso be sed to chec" %odel v&lidity. >or e(&%ple, l&st ye&rEs thro ghp t d&t& c&n sed be to v&lid&te the thro ghp t n %ber prod ced by the %odel 'or the s&%e d r&tion &nd nder si%il&r conditions. We c&n &lso do ble the cycle ti%e o' & cert&in oper&tion &nd see i' the syste% thro ghp t prod ced is &''ected &ccordingly or i' the %&n '&ct ring le&d ti%e d&t& re'lect this incre&se in cycle ti%e.

-&%&/ )odel Analysis>(?

!&ving & veri'ied &nd v&lid&ted si% l&tion %odel provides &n&lysts $ith & gre&t opport nity since it provides & 'le(ible pl&t'or% on $hich to r n e(peri%ents &nd to &pply v&rio s types o' engineering &n&lyses e''ectively. With the l&test &dv&nces in co%p ter speed &nd c&p&city, even l&rge/sc&le si% l&tion %odels o' intensive gr&phics c&n be r n 'or sever&l replic&tions in & rel&tively short ti%e. !ence, it t&"es only & 'e$ %in tes to r n % ltiple si% l&tion replic&tions 'or long periods o' ti%e in %ost si% l&tion environ%ents. As sho$n in >ig re 6.)), %odel &n&lysis o'ten incl des st&tistic&l &n&lysis, e(peri%ent&l design, &nd opti%i.&tion se&rch. <he obDective o' s ch %ethods is to &n&ly.e the per'or%&nce o' the si% l&tion %odel, co%p&re the per'or%&nce o' proposed design &ltern&tives, &nd provide the %odel str ct re &nd p&r&%eter setting th&t $ill le&d to the best level o' per'or%&nce. St&tistic&l &n&lyses incl de representing per'or%&nce $ith descriptive st&tistics s ch &s %e&n, v&ri&nce, &nd con'idence interv&ls. It &lso incl des per'or%ing v&rio s st&tistic&l tests 'or hypothesis testing &nd co%p&r&tive &n&lysis. ?(peri%ent&l design $ith si% l&tion incl des cond cting & p&rti&l or ' ll '&ctori&l design o' e(peri%ents to provide the best settings to %odel control v&ri&bles. >in&lly, se&rch &lgorith%s incl de the &pplic&tion o' opti%i.&tion se&rch %ethods s ch &s e(h& stive se&rch, genetic &lgorith%, si% l&tion &nne&ling, &nd t&b se&rch to opti%i.e &n obDective ' nction.

Figure -&''# Contin o s i%prove%ent $ith si% l&tion. <he speed &nd 'le(ibility o' the si% l&tion %odel &lso '&cilit&te cond cting e(peri%ent&l design &nd r nning opti%i.&tion &lgorith%s. Also, the &v&il&bility o' di''erent %ethods &nd &n&lyses in co%%erci&l si% l&tion so't$&re tools '&cilit&tes the se o' o tp t &n&lyses. 2ost ' ll versions o' si% l&tion p&c"&ges &re e8 ipped $ith %od les 'or st&tistic&l &n&lyses &nd &dd/ins o' e(peri%ent&l design &nd opti%i.&tion %ethods.

-&%&0 Study Do$umentation

<he 'in&l step in & si% l&tion st dy is to doc %ent the st dy &nd report its res lts. Proper doc %ent&tion is cr ci&l to the s ccess o' & si% l&tion st dy. <he si% l&tion process o'ten incl des co%% nic&ting $ith %&ny sides, $riting co%ple( logic, enco ntering enor%o s &%o nts

o' d&t&, cond cting e(tensive e(peri%ent&tion, &nd going thro gh sever&l progress revie$s &nd %ilestones. <h s, $itho t proper doc %ent&tion, the &n&lyst loses tr&c" o' &nd control over the st dy &nd c&nnot deliver the re8 ired in'or%&tion or %eet the st dy e(pect&tions. <his o'ten res lts in &n in&cc r&te si% l&tion %odel $ith poor res lts, in&bility to D sti'y %odel beh&vior &nd e(pl&in %odel res lts, &nd loss o' othersE con'idence in st dy 'indings &nd reco%%end&tions. 1oc %enting the si% l&tion st dy is the develop%ent o' & st dy 'ile th&t incl des the det&ils o' e&ch si% l&tion step. Co%prehensive doc %ent&tion o' & si% l&tion st dy co%prises three %&in ele%ents9 det&iled doc %ent&tion o' the si% l&tion %odel, the develop%ent o' &n engineering si% l&tion report, &nd the present&tion o' si% l&tion res lts to c sto%ers &nd p&rtners o' the si% l&tion proDect. >ig re 6.)2 presents the three ele%ents o' doc %enting & si% l&tion st dy.

Figure -&'(# 1oc %enting & si% l&tion st dy. <he 'irst ele%ent in doc %enting & si% l&tion st dy incl des doc %ent&tion o' the si% l&tion %odel. <his incl des doc %enting both the concept %odel &nd the si% l&tion progr&%. <he doc %ent&tion o' the %odel concept incl des doc %enting the syste% process %&p, 'lo$ch&rts, bloc" di&gr&%s, &nd s"etches th&t e(pl&in the %odel concept, %odel ele%ents, %odeling &ppro&ch, &nd %odel logic&l design. Si% l&tion %odel doc %ent&tion incl des & description o' %odel str ct re, progr&% det&ils &nd 'lo$ch&rts, code &nd e(pl&n&tion o' ro tines sing $ithin/ code co%%ents &nd notes, &nd &n e(pl&n&tion o' the sched ling &nd se8 encing r les, ro ting r les, &nd the decision/%&"ing process 'ollo$ed &t decision points $ithin the %odel. 2odel reporting &lso incl des & st&te%ent o' %odel &ss %ptions &nd & description o' %odel inp ts &nd o tp ts. @y revie$ing s ch doc %ents, $e sho ld be &ble to nderst&nd ho$ the %odel re&ds inp ts, processes the logic, ses &v&il&ble si% l&tion ele%ents, &nd provides the re8 ired o tp ts. S ch doc %ent&tion '&cilit&tes %&"ing %odel ch&nges, e(pl&ining the d&t& sed in the %odel, deb gging code &nd logic, nderst&nding %odel beh&vior, &nd interpreting %odel res lts. <he %&Dor deliver&ble o' the si% l&tion st dy is the develop%ent o' & si% l&tion report. <his incl des & 'or%&l engineering report &nd s pple%ent&ry %&teri&ls o' the si% l&tion process. S ch & report constit tes the co%% nic&tion %edi % bet$een the si% l&tion &n&lyst &nd the other p&rties involved in the si% l&tion st dy. So%e o' those p&rties %&y not even see the %odel or t&l" to the si% l&tion &n&lystO hence, the report sho ld be co%prehensive, n&%big o s, &nd nderst&nd&ble by people o' v&rio s s"ills &nd b&c"gro nds. <he ele%ents o' & 'or%&l si% l&tion report incl de & description o' the nderlying syste%, 'or% l&tion o' the si% l&tion proble%, & st&te%ent o' st dy obDectives &nd &ss %ptions, so rces o' %odel d&t&, & description o' the %odel str ct re, &nd & s %%&ry o' si% l&tion res lts, 'indings, concl sions, &nd reco%%end&tions. S pple%ent&ry %&teri&ls incl de the d&t& inp ts sed, syste% &nd %odel gr&phs, s"etches &nd dr&$ings, det&ils o' e(peri%ent&l design &nd o tp t &n&lysis, &nd & printo t o' si% l&tion res lts. It is &lso co%%on &%ong si% l&tion engineers to provide & one/p&ge s %%&ry o' "ey in'or%&tion provided by the si% l&tion st dy. <his type o' e(ec tive revie$ provides concise &nd 'oc sed in'or%&tion 'or those decision %&"ers, %&n&gers, &nd 'ir% e(ec tives $ho %&y not be interested in the det&ils o' %odel inp ts, str ct re, &nd res lts &n&lysis. A si% l&tion report incl des the 'ollo$ing ele%ents9 ). <he Syste% @eing Si% l&ted &. @&c"gro nd b. Syste% description c. Syste% design 2. <he Si% l&tion Proble% &. Proble% 'or% l&tion

3.

7. 6. 6.

b. Proble% &ss %ptions c. St dy obDectives <he Si% l&tion 2odel &. 2odel str ct re b. 2odel inp ts c. 2odel &ss %ptions Si% l&tion 0es lts &. 0es lts s %%&ry b. 0es lts &n&lysis St dy Concl sion &. St dy 'inding b. St dy reco%%end&tions St dy S pple%ents &. 1r&$ings &nd gr&phs b. Inp t d&t& c. J tp t d&t& d. ?(peri%ent&l design e. Jthers

<he si% l&tion st dy is o'ten concl ded by presenting the si% l&tion res lts to c sto%ers, %&n&gers, &nd other p&rties to the si% l&tion proDect. S ch & present&tion o'ten incl des & s %%&ry o' the si% l&tion res lts 4s %%&ry o' st dy steps, res lts &nd 'indings, concl sions, &nd reco%%end&tions5. <he present&tion sho ld incl de r nning &ni%&tions, %ovies, or sn&pshots o' the si% l&tion %odel in & v&riety o' sit &tions. S ch &ni%&tions help e(pl&in the st dy res lts &nd &id the &n&lyst in selling the design proposed, co%p&ring design &ltern&tives, &nd sec ring %&n&ge%ent s pport thro gh proposed sol tions &nd pl&ns o' &ction. <he co%% nic&tion &nd present&tion s"ills o' the si% l&tion &n&lyst pl&y & %&Dor role in developing & s ccess' l present&tion &nd g&ining %&n&geri&l s pport to i%ple%ent the reco%%end&tions o' the si% l&tion proDect.
H2I

We provide ' rther insight on &n&ly.ing si% l&tion o tp ts in Ch&pter * &nd Appendi( A.

-&, Exam le# A lying +he Simulation !ro$ess +o A Hos ital Emergen$y Room
In this section $e &pply the si% l&tion process to &n e%ergency roo% 4?05 si% l&tion. <he %&in obDective here is to se si% l&tion to st dy the ?0 syste%, i%prove the ?0 oper&tion by red cing tot&l ti%e in the syste%, &nd predict the e''ect o' ch&nges on long/ter% per'or%&nce. ?0 si% l&tion st dy re8 ire%ents incl de the 'ollo$ing9 Identi'ying %e&s re%ent points necess&ry 'or syste% ev&l &tion Selecting &nd &pplying the &ppropri&te si% l&tion &ppro&ches &nd techni8 es to represent the syste% processes Identi'ying ?0 oper&tion&l $e&" points &nd bottlenec"s S ggesting i%prove%ents to the c rrent sit &tion Constr cting & represent&tion o' the ?0 syste% con'ig r&tion proposed Leri'ying the syste% proposed #R Process Description As depicted in >ig re 6.)3, the ?0 e(&%ple consists o' three st&tions9 the tri&ge st&tion, $here p&tients record their n&%es &nd 'ill o t the tri&ge 'or%O the reception st&tion, $here the st&tion &ttend&nt enters their d&t& into & co%p ter &nd cre&tes %edic&l 'iles 'or ne$ p&tientsO &nd the tre&t%ent roo%, 'or physici&n di&gnosis. I' & p&tient needs ' rther tre&t%ent, he or she $ill be sent to & speci&list doctor 4i.e., internist, c&rdiologist, gynecologist, ophth&l%ologist, etc.5 >or serio s %edic&l e%ergencies 4i.e., &ccidents, he&rt &tt&c"s,

p&rt ritions, etc.5, p&tients &re tre&ted directly by doctors $itho t p&ssing thro gh the tri&ge &nd d&t&/entry st&tions.

Figure -&'*# Concept &l %odel o' the ?0 e(&%ple. An ?0 syste% c&n be vie$ed &s & 8 e ing syste% $ith & single ch&nnel 4i.e., &d%ission, tre&t%ent, rele&se5. <he p&tient 'lo$ $ithin the ?0 is se8 enti&l &nd c&n be co%bined into & one/ stop service in order to &pply the 'or% l&s o' the single/server 8 e ing %odel. In pr&ctice, ho$ever, di''erent ?0 reso rces provide service &t di''erent loc&tions. <he ?0 is not & typic&l 'irst/in/'irst/served syste%O the ?0 oper&tes sing speci'ic r les to ro te p&tients. <he priority 'or & p&tient to enter the ?0 depends on his or her %edic&l c&se &nd the ris" involved in %&"ing the p&tient $&it. Altho gh %ore co%ple( 8 e ing %odels 4e.g., & net$or" o' 8 e es5 c&n be sed to &ppro(i%&te the syste%, si% l&tion $ill be & better &ltern&tive, given the st dy re8 ire%ents %entioned. <he syste% c&n still be described by the r&te o' p&tient &rriv&ls &nd the service %ech&nis%, $hich incl des the 'ollo$ing three st&ges9 ). P&ssing by tri&ge 2. ?ntering & p&tientEs rel&ted d&t& in co%p ter records 3. Physic&l tre&t%ent In the ?0 1?S %odel, t$o %&in events c& se &n inst&nt&neo s ch&nge in the st&te o' the ?0 syste%9 ). <he entry o' & p&tient into the syste% 4p&tient &rriv&l5 2. <he rele&se o' & p&tient pon co%pletion o' tre&t%ent 4p&tient dep&rt re5 @&sed on in'or%&tion obt&ined 'ro% s bDect %&tter e(perts, observ&tion, &nd historic&l d&t&, the 'ollo$ing logic governs the 'lo$ o' p&tients in the ?09 <he order o' tre&t%ent priority is high ris", %edi % ris", &nd nor%&l c&ses, respectively. ?0 p&tients &re cl&ssi'ied &s 'ollo$s9 o Ris5y case3 & prob&bility o' 70Q o :edium,ris5 case3 & prob&bility o' 30Q o .ormal case3 & prob&bility o' 30Q When & ris"y c&se &rrives &t the ?0, it does not p&ss thro gh the tri&ge or d&t&/entry st&gesO the person is tre&ted directly by & physici&n. When & %edi %/ris" c&se &rrives, the person p&sses thro gh tri&ge b t does not p&ss thro gh the d&t&/entry &re&. When & nor%&l c&se &rrives, the person % st p&ss thro gh &ll three st&ges9 tri&ge, d&t& entry, &nd tre&t%ent. Jne physici&n c&n tre&t p to t$o p&tients in the e%ergency roo%. P&tients le&ving the e%ergency roo% &re ro ted &s 'ollo$s9 o <hose $ho $ill le&ve the hospit&l9 & percent&ge o' 70Q o Those 4ho 4ill continue the treatment in other specialist units in the hospital3 & percent&ge o' 60Q

Data Collection and !nput Modeling A'ter getting per%ission 'ro% the hospit&l &d%inistr&tion, the proDect te&% st&rted their $or" by observing the sit &tion in the ?0 &nd collecting pertinent d&t&, incl ding co nting the 8 e e o' p&tients, %e&s ring the service ti%e &t v&rio s process st&ges &nd the inter/&rriv&l ti%e o' p&tients, tr&c"ing p&tient 'lo$, &nd so on. Also, historic&l d&t& 'ro% the ?0 $ere revie$ed &nd s %%&ri.ed. <he te&% $&s '&ced $ith & l&c" o' historic&l d&t& 4records o' p&tient &rriv&ls &nd service d r&tions5 essenti&l to esti%&ting the service &nd inter&rriv&l ti%es. <h s, it $&s decided to collect d&t& sing & stop$&tch. Pertinent d&t& 'or 6, s&%ples 4p&tients5 $ere collected. <&ble 6.2 s %%&ri.es the d&t& collected in ter%s o' p&tient &rriv&l ti%es &nd the ti%e 'or end o' service &t the three ?0 st&tions. -sing the d&t& collected, v&l es &re esti%&ted 'or e&ch p&tientEs inter&rriv&l &nd service ti%es &t the three ?0 st&tions. +a3le -&(# ER Data Colle$ted Jpen t&ble &s spre&dsheet !atie nt @um3 er +im e of Arri 4al at +ria ge 5'6 +930 ,900 ,920 ,960 )09) 6 )092 0 )092 2 )093 6 )097 0 )097 6 )097 6 ))90 0 )290 0 +ime of Arri4al at Re$e ti on 5(6 +ime of Startin g the +reatm ent +ime of !atient De art ure 5*6 +ime Bet7e en Arri4a ls 5min6 Ser4i $e +ime ' 5min6 Ser4i $e +ime ( 5min6 Ser4i$ e +ime * 5min6

) 2 3 7 6 6 * + , )0 )) )2 )3

+93) ,90) ,926 ,96) )09)6 )092) )0927 )0936 )097) )0976 )0976 ))90) )290)

+936 ,906 ,930 ,966 )09)+ )0930 )0930 )093+ )0972 )097* )097+ ))906 )2906

,966 ,9)0 ,976 )0906 )0930 ))9)0 ))930 ))900 )29)0 ))976 )2906 29)2 )29)0

0 30 20 30 26 6 2 )3 6 6 0 )6 60

) ) 6 ) ) ) 2 ) ) ) ) ) )

7 7 6 7 2 , 6 2 ) ) 2 7 7

30 6 )6 )0 )2 70 60 22 ++ 6+ ** )2* 6

+a3le -&(# ER Data Colle$ted Jpen t&ble &s spre&dsheet !atie nt @um3 er +im e of Arri 4al at +ria ge 5'6 )29) 0 )293 0 )293 0 )293 0 )293 6 )297 0 )296 0 )296 0 )296 0 )296 0 )392 0 )393 0 )393 0 )393 6 )790 0 )790 +ime of Arri4al at Re$e ti on 5(6 +ime of Startin g the +reatm ent +ime of !atient De art ure 5*6 +ime Bet7e en Arri4a ls 5min6 Ser4i $e +ime ' 5min6 Ser4i $e +ime ( 5min6 Ser4i$ e +ime * 5min6

)7 )6 )6 )* )+ ), 20 2) 22 23 27 26 26 2* 2+ 2,

)29)) )293) )293) )293) )2936 )297) )296) )296) )296) )296) )392) )393) )393) )3936 )790) )790)

)29)2 )2936 )2936 )293+ )2936 )2976 )2966 )296+ )2966 )2966 )3922 )3936 )3936 )7900 )79)0 )7906

)29)6 )293+ )2970 )2970 )2970 )2960 )3900 )39)0 )39)6 )3900 )3930 )7900 )3970 )7930 )79)2 )7920

)0 20 0 0 6 6 )0 0 0 0 30 )0 0 6 26 0

) ) ) ) ) ) ) ) ) ) ) ) ) ) ) )

) 7 6 * ) 6 6 + 6 6 2 6 6 26 )0 6

7 3 7 2 7 6 6 )2 ), 6 + 26 6 30 2 )6

+a3le -&(# ER Data Colle$ted Jpen t&ble &s spre&dsheet !atie nt @um3 er +im e of Arri 4al at +ria ge 5'6 0 30 3) 32 33 37 36 36 3* 3+ 3, 70 7) 72 73 77 76 76 7* )797 6 )693 0 +92, +96* ,906 ,90* ,970 ,960 ,966 )090 0 )090 0 )09) 0 )093 6 )093 6 )097 0 )096 6 ))90 0 ))9) )7976 )693) +933 +96, ,906 ,90+ ,97) ,96) ,966 )090) )090) )09)) )0936 )0936 )097) )0966 ))906 ))9)6 )797+ )6936 +973 ,900 ,936 ,9)6 )0900 )0920 )0930 )0970 )0960 )09)2 )0970 )097) )0976 )096+ ))9)0 ))920 )6930 )6900 +966 ,906 ,970 ,92) )0920 )0930 ))900 ))900 ))970 ))900 ))920 ))900 ))930 )29)0 )2930 )2900 76 36 0 )+ + 2 33 )0 6 6 0 )0 26 0 6 )6 6 )6 ) ) 7 2 ) ) ) ) ) ) ) ) ) ) ) ) 6 ) 3 6 )0 ) 30 + ), 2, 37 3, 7, ) 7 6 7 2 6 7 72 26 )0 6 6 7 20 )0 30 20 60 7+ 70 ), 76 *2 +0 70 +ime of Arri4al at Re$e ti on 5(6 +ime of Startin g the +reatm ent +ime of !atient De art ure 5*6 +ime Bet7e en Arri4a ls 5min6 Ser4i $e +ime ' 5min6 Ser4i $e +ime ( 5min6 Ser4i$ e +ime * 5min6

+a3le -&(# ER Data Colle$ted Jpen t&ble &s spre&dsheet !atie nt @um3 er +im e of Arri 4al at +ria ge 5'6 6 7+ 7, 60 6) 62 63 67 66 66 6* 6+ 6, ))93 0 ))97 0 ))97 0 ))97 6 ))97 6 )290 0 )29) 0 )393 0 )397 0 )790 0 )790 6 )692 0 ))93) ))97) ))97) ))976 ))976 )290) )29)) )393) )397) )790) )7906 792) ))936 ))976 ))960 ))97+ ))966 )29)0 )2920 )3937 )7900 )7920 )79)0 )6930 )2900 )3900 )3900 )29)0 )2920 )2970 )39)6 )7900 )79)6 )7930 )6900 )*9)0 )6 )0 0 6 0 )6 )0 +0 )0 20 6 )26 ) ) ) ) ) ) ) ) ) ) ) ) 7 7 , 2 , , , 3 ), ), 7 , 26 *6 *0 22 26 30 66 26 )6 )0 60 70 +ime of Arri4al at Re$e ti on 5(6 +ime of Startin g the +reatm ent +ime of !atient De art ure 5*6 +ime Bet7e en Arri4a ls 5min6 Ser4i $e +ime ' 5min6 Ser4i $e +ime ( 5min6 Ser4i$ e +ime * 5min6

<o 'it d&t& collected to their corresponding st&tistic&l distrib tions, the inp t %odeling %od le in &n A0?#A si% l&tion p&c"&ge $&s sed. <&ble 6.3 s %%&ri.es the res lts 'or inter&rriv&l ti%es. +a3le -&*# Fitting Data Colle$ted to Statisti$al Distri3utions >a? Jpen t&ble &s spre&dsheet Patients Interarri-al Time 1istrib tion s %%&ry Goodness o' 'it9

1istrib tion9e(ponenti&l ?(pression9 C0.00) S ?WPJ4)7.65 S8 &re error90.002,06 1&t& s %%&ry # %ber o' d&t& points T 6, 2in.d&t& v&l e T 0 2&(.d&t& v&l e T )26 S&%ple %e&n T )7.6 S&%ple std.dev. T 2).2

Chi/s8 &re test # %ber o' interv&ls T2 1egrees o' 'reedo% T 0 <est st&tistic T 0.)67 Corresponding p/v&l e X 0.006

Ser-ice Time at the Triage Station 1istrib tion s %%&ry 1istrib tion9lognor%&l ?(pression90.6 S ;JG#40.667,0.37*5 S8 &re error90.0)307* 1&t& s %%&ry # %ber o' d&t& points T 6, 2in.d&t& v&l e T ) 2&(.d&t& v&l e T 6 S&%ple %e&n T ).22 S&%ple std.dev. T 0.+32 Goodness o' 'it9 Chi/s8 &re test # %ber o' interv&ls T 2 1egrees o' 'reedo% T /) <est st&tistic T 6.72 Corresponding p/v&l e X 0.006

Ser-ice Time at the Registration Station 1istrib tion s %%&ry 1istrib tion9lognor%&l ?(pression90.6 S ;JG#4+.),,)2.*5 S8 &re error90.07766* 1&t& s %%&ry # %ber o' d&t& points T 6, 2in.d&t& v&l e T ) 2&(.d&t& v&l e T 7, S&%ple %e&n T +.76 S&%ple std.dev. T ,.,6 Goodness o' 'it9 Chi/s8 &re test # %ber o' interv&ls T * 1egrees o' 'reedo% T 7 <est st&tistic T )6., Corresponding p/v&l e X 0.006

Ser-ice Time at the Treatment Station 1istrib tion s %%&ry 1istrib tion9Weib ll ?(pression92 S W?I@4)6.+,0.675 S8 &re error90.00)6,2 1&t& s %%&ry # %ber o' d&t& points T 37 2in.d&t& v&l e T 2 2&(.d&t& v&l e T )2* S&%ple %e&n T 23.) S&%ple std.dev. T 2+.*
H&I

Goodness o' 'it9 Chi/s8 &re test # %ber o' interv&ls T 2 1egrees o' 'reedo% T C) <est st&tistic T 0.)22 Corresponding p/v&l e X 0.006

See Appendi( A 'or &ddition&l st&tistic&l s bst&nce.

Model $uilding% &erification% and &alidation -sing A to2od si% l&tion so't$&re, the b&se %odel is b ilt &s sho$n in the A to2od &ni%&tion sho$n in >ig re 6.)7. <he %odel logic $&s 'irst veri'ied &nd the %odel beh&vior $&s v&lid&ted sing the d&t& collected &nd observ&tions o' ?0 s bDect %&tter e(perts. A to2od %odel ele%ents sed incl de9

Figure -&'%# A to2od &ni%&tion o' the ?0 %odel. ntity3 p&tients /ttri'utes3 he&lth 4cl&ssi'ic&tion o' ?0 p&tients5 /cti-ity3 tre&t%ent 2ogenous e-ent3 &rriv&l o' p&tients ndogenous e-ent3 co%pletion o' tre&t%ent 4dep&rt re o' p&tients5 State -aria'le3 n %ber o' p&tients $&iting, n %ber o' b sy doctors, etc. Model Results <he %odel is set to r n )0 d&ys contin o sly 4270 ho rs, &ss %ing th&t the ?0 oper&tes three shi'ts & d&y5 &'ter & $&r%/ p period o' + ho rs. As sho$n 'ro% the 8 e e st&tistics in >ig re 6.)6, & tot&l o' )00) p&tients $ere tre&ted &t the ?0, o' $hich 3*3 p&tients $ere cl&ssi'ied &s serio s. A tot&l o' 62+ p&tients $ent thro gh the three ?0 st&tions, st&rting $ith tri&ge. <he &ver&ge ?0 p&tient spent &bo t 662 seconds $&iting, &bo t )22 seconds &t tri&ge, &bo t *,0 seconds &t the reception registr&tion, &nd &bo t 2**2 seconds in tre&t%ent. <he &ver&ge tot&l ti%e/in/syste% is &bo t 3677 seconds.

Figure -&',# S %%&ry o' A to2od 8 e e st&tistics. >ig re 6.)6 sho$s the A to2od s %%&ry o' reso rce st&tistics. <he %odel is set to se one reso rce 4teller5 'or reception, one 'or co%p ter d&t& entry, &nd & high n %ber o' reso rces 'or physici&ns since the n %ber o' physici&ns sed &t &ny point is b&sed on the n %ber o' p&tients in the ?0 roo%. -tili.&tion is & "ey reso rce st&tistic th&t c&n be sed in directing ?0 i%prove%ent &ctions.

Figure -&'-# S %%&ry o' A to2od reso rce st&tistics. @&sed on %odel observ&tions &nd %odel st&tistics, the 'ollo$ing %e&s res $ere t&"en to i%prove ?0 oper&tions. I%prove%ent e''orts &re 'oc sed on red cing p&tient $&iting ti%e be'ore service, registr&tion ti%e, &nd $&iting 'or tre&t%ent ti%e. <his $ill signi'ic&ntly red ce the over&ll ti%e spent in the syste%. Jther %e&s res &re &i%ed &t incre&sing the e''ectiveness o' ?0 oper&tions by opti%i.ing the n %ber o' physici&ns &nd incre&sing the tili.&tion o' reception &nd tri&ge reso rces. Jther %icro %e&s res c&n &lso t&"e pl&ce to i%prove $or" proced res, especi&lly 'or prep&ring p&tient 'iles &nd entering p&tient in'or%&tion. >ollo$ing is & s %%&ry o' ?0 i%prove%ent &ctions9 <he long ti%e spent in the reception st&ge $&s red ced 'irst since & long $&iting ti%e %&y &''ect p&tient he&lth neg&tively 4i.e., %&y convert it 'ro% nor%&l or %edi % to serio s5. <he reception &nd registr&tion ti%e is i%proved by developing & %ore e''ective $or" proced re 'or prep&ring p&tient 'iles &nd entering p&tient in'or%&tion. A ti%e &nd %otion st dy $&s reco%%ended 'or this p rpose. An integr&ted hospit&l d&t&b&se o'ten provides & s bst&nti&l ti%e s&ving in this reg&rd. <i%e spent $hile p&tients $&it 'or physici&ns &t the tre&t%ent st&tion $&s red ced by ch&nging the r le o' Fone doctor to t$o p&tientsA &nd &llo$ing the i%%edi&te &ssign%ent o' &nother &v&il&ble doctor to the second p&tient. <he 'in&l n %ber o' physici&n reso rces is set to three b&sed on the tili.&tion o' physici&n reso rces 4see >ig re 6.)65. <he tri&ge &nd registr&tion reso rces $ere co%bined, d e to lo$ tili.&tion o' the tri&ge reso rce. Ch&nges in the tot&l ti%e/in/syste% $ere observed thro gh & con'ir%&tion si% l&tion r n $ith 'ive replic&tions, &s sho$n in >ig re 6.)*. <he ne$ &ver&ge ti%e/in/syste% is 2*32 seconds. Co%p&red to &n &ver&ge ti%e/in/syste% o' 3677 seconds in the initi&l ?0 %odel, i%prove%ent &ctions res lted in &bo t & 26Q red ction in &ver&ge p&tient ti%e/in/syste%.

Figure -&'.# >ive replic&tions o' the ?0 ti%e/in/syste%.

-&- Summary

In this ch&pter $e presented & det&iled description o' the si% l&tion process, incl ding the det&ils o' si% l&tion proced res 'or proDect scoping, d&t& collecting, %odel b ilding, %odel &n&lyses, &nd %odel doc %ent&tion. <his process is l&ter integr&ted into the si(/sig%& proDect ch&rter in the 3S/ ;SS &nd 3S/1>SS &ppro&ches.