OR - Notes PDF

RNS INSTITUTE OF TECHNOLOGY CHANNASANDRA, BANGALORE - 61 OPERATIONS RESEARCH NOTES FOR 6™ SEMESTER INFORMATION SCIENCE SUBJECT CODE: 06CS661 PREPARED BY DIVYAK ARNOSISO16 6% Semester Information Science and Engineering 1rn09is016@gmail.com In Association With AMBIKA S KARANTH ARNOBISO03 8 Semester Information Science and Engineering 4rn08is003@gmail.com REFERENCES: Manjunath Aradhya ‘SD Sharma Text Book Lieberman Text Book Class Notes (Manoj Kumar} CONTENTS: UNIT 1, UNIT 2, UNIT 3, UNIT 6, UNIT7 | rctes have been arcuted on slick, nobody canbe el esponsine anything wrong ors improper infrmation or naulet information provided initIMPORTANT TIPS TOPIC-WISE BREAK UP: (a) theory on OR fundamentals (b) problem on formulation of a LPP. (c) problem on graphical method (a) theory on simplex method (including definitions) (b) problem on simplex method in tabular form {c) problem on trial and error method ._ (a) theory/problem on BIG-M method (b) theory/problem on Two-Phase Mettiod. Revised simplex methad-1 (not covered in the notes) Revised simplex method ~2 (not covered in the notes] (a) problem on transportation models (b) problem on assignment models /. (a) problem on saddle point, dominance rule {(b) graphical method of solving 2xn or mx2 games/ decision analysis [trees] ‘Metaheurstis [not covered in the notes} PLEASE NOTE: These are the probable type of questions which we can assume to appear in the VTU examination after the analysis of previous year papers. These may be subjected to changes als. UNITS 1, 2, 3, 6, 7 are the best and easy ones to prepare and attend in the examination. ‘Theoretical aspects are almost covered in this notes based on previous paper analysis and we hope itis sufficient in the examination point of view. Problems such as formulation of LP model require wide subject knowledge since each and every problem is of different variety and there can be no generalisation. Hence, practice more varieties. Problems based on simplex methods ate generally easy to solve once we master the method. But, there are certain conditions such as Degeneracy, tie breaking etc, which have to be taken special care. Problems on Transportation: important techniques such as VAM, NWC are given more priority whereas RUSSELs method, Travelling salesman, MOD! method are dealt with only one example problem since they are not very frequently asked in VTU exam. By analysing an example problem, you must be capable enough to solve similar problems. This applies to assignment problem also. [you can solve anything if you know the steps properly] Problems on Game theory & Decision analysis: construction of decision trees once again requires good skill and control over the subject since it involves wide varieties of problems. CLASS ROOM ASSIGNMENTS are attached almost at the end of the notes. PLEASE REFER them since it involves different varieties of problems which are not solved in the notes. Finally, this NOTES gives a fair idea about the subject and the various methods/techniques. But, since it is a mathematical subject, VTU questions may not be restricted to only the ones solved here..!!! THE MORE YOU PRACTICE, THE BETTER YOU UNDERSTAND. ALL THE BEST oe ccensamners eaeDarga ! UNIT 1 Twtropuction To Linear Proarammine i | Oprnadions Reranch abbreviatid 0s OR may be desnibed as a Actentific approach to dectevon naaksinge Theis nine subject come inte extstomce dlurtng the second wooild war. OR & Aopined os om exprrimustal gelence dvetid te obsernring , wndertamdingn predicting, He behavior of purpertful own rrvohine, segpstems I i Moim Phaser of OR ae procedure for am OR Mioday, Lnsrolves the following 6 aajos phases Phase 1: Fevmulation ; “thie plaose maine the problem cbe be fommadalid de dy fom of am appraprinle madel . this Inchudir $ladsng, objectine fumckion, constrains 0% sestrichions, Arter selaBronships gouthle altrwats corer of action, Baws Uoanhla, fer makbngn ( wncentretlable Phase 2: Constuctin of a Mattemotical Model this prose tk concenmed ofilhe seferwmolion ef ther prshtem, Tw aan appraprialt foram colder is wteful, tw oomalysis,phose 3: Dectwabton of solutions prem modthemabital model This plrose I duvotid fo Se Computaliow of rere value of the decieton vontables colic wmorturts, Levpottbownks to orine ab the optimal solutton of the problem. Pi ee ey cece ar Tha counpleted mvodel ts Rertid por omens f ong. “the priwciple of judging tee volldlly rp a mobil ls Keotadnar ov wok tk predicts Be relative Apecli of the Live airmen courier of ation writ suffietent fpecmatip te poral & tromnd dectilen", A geod model td. be mppticoble for a longer Lame andl Gres uapdatis, thy madi from that te tine by hg” AS picount tha posts poem amd fotine spectpication Hf Hin proba Catal idbving Contrel ever the Solutiontotaal ovr the Aclution With gowe degree of coldifaction. This phase alia actoblichs a syttomadic: procedure for dutzating chamgee amd contialbingy she albwakin. Phase: Tmplumentoction Tha, ineplennerctasion of the conbrallid. glulton davoles,, te thamalotion of the model's Aavlls Ante : Instructions Te te dmperton In OK fe ensure tha Bre solution Us aceuralelip traralakid, tabs aon any procedure te saclthy, tanll, bro the solucbion . Adsoolager sf Speration Reconeh bs Ophimnaun ae a predactlin Foncrs Linear Pro Hechanlgper dncliclle bout seamarpr tame mast Ypelivey, sompley, bts produchioe fut by mare tity dean tnd dtig a. Tepe ally ion plelicne The computalton table gives a clear + He teithin they beste autiickions amd thy Splitter Of comrpouma behaviour of the elements Lnwolwed fin the protelom. “Te éhieck an profitabtiity duut te chamate tom the production pattern twonld be clearly, Amdteaded im tre solution 3. Preparation of Failiane q OR kedirrdqntr snbrtitate a meant fer Puproving, — ha rorcledge aan shill af cpowmngy managers. [flDivadvantager of OR Practical Seattle : oe he LP problem fe a diptude tee ‘commerpratt. Gist subamaible Hamenke of they agtiom cond he 4 spud, (tke uorwaane relabionsti) 5 Disko helen managers oud OR — oR problonn Aap a tmothemadiclan 0 MOLE on tole aks aot he amore Bb tne bwaimets- problems, Sonitlary, igen mot woolerstand re complx weikumg |. Ot wbdsle constlesa u OR- neem Notion of Operations Rescanch ; OR twelves “esearch ov operations 2 That, operations prrconch la applied 6 probtemn that conte how 6 | conduct and cohdinale Akt operations tolltsin am otgewltalion, | he valve of teakion is fumatertal , ond, Ln fork, - On ok been oppliod exkenatetly, dw tut diverse armas os F manufactianing., thanspertation, conatiuchiens, tele communications finamsioh planning, roll, tose fi mame just o Jove. Therefore, tre hrsadth o& application dt warrolly utde OR see om ath that aremblor Be To a Considerable extent, the aetenttfe method lu used, to warkigate the protlem of concern. OR Luvolver,The dempatk of 8 ation Ruscanch Qprations Research how had am bnyrestine iepack ow 4 ‘ the opirtency. of manmmaroure oAgautactions anowmd Aru wold. Tn the prceem, OR hos made a of the Leonenaier> of varteut Combis. Thue are now a fers dager omumbir commititis on the Trtormachiowol Fedirotion of OR bocieties (IF ORS) , wlth earch coi oS a national OR octal. wily Aawing, Th deals with the opbionniyabion Caxeimigption or imiudmtypsbion) of, av fumchion af vortables, Kunausn 4 objective fumekion , sabfeck to a set of neon te Requirements fer a. LEP . 7 - A LPP cam ba optioniged, Uy the follovoinge conditions pac gabipied! y mane unk be a Coll defined objective guacbion cartel Com be opetintyed amd tome be axpaerned as a Ldatass 3) Thre musk be am albomals comme f action 4) The decisteon vartables should be Lnber-solatid oma trom - negobine .1) Solution : A eh of variables §%)%y)Ky.-- Xn f ds called w golition i ib sabiegis thy constiodnle a) Frosthle Selubien: A sek of vartables £%),%, Xs---—- Xn} de colled a OAL wen -Vve- 3) Boric solutlon: A sslubion chtatued by setting ‘nr’ vortobles, Corewuornsy man vartobles) te yore amd. csolining, for pamnratankwcyr m variable, & colled wv basic solubion - These m vostobley are boric vartobler and mm vartobles, ane nen-basic vorlabte, Total we af basic solultong = (mind Cy 4) Borie Fearthle folution (ars) A basic beludion de cold a baste feartlle solution Y okt baste vatables 2 20. 9) Nen - degenerate BFS Lt ws ao BES ty which oll m vartables, Qne ve amd the Atmatuing, ‘ni vartobls ore zero Lod 6) Degenerate BFS Lt %& a BPS iw whith one 0% mere Of the m boarte variables wre tool to gene, Fi.=D) Opbinnnall BFS A BFS % called fhe optional OFS if 1b opbiwigte the objectne fom 8) Umbownrded Soludtion Ty toy volute of Hib objective: fumibion, com be dncrtased, o Aecruared Imclefinttelyy um bev solubton & Called om nbowaded Agludven. 9) Feasible Regier : hh 4 ton Rw whitey oll constealnlf ond manemigatinily conditions hold good. (0) Corer Port Feortble (cpr) solution _ De he a fenatbe salabin al devin’ I om oor Opbinads abt IDE fa the techs of chtntiingy he hast seal wader be gwen conditions Da 2 Th Lo the maltematicol tedulyne te Atormine dae opldunuum se of tre Amited owvallable arrources. Te a deletion mnakitng Lesbutane verdes die i gine Consliintnlé wader ty crndititen tok the i Aalationthip sawengn the variables Lvorelved Le MatarMatacmattcac Formuration Of A LPP Step 1: Define thy lector vartable ee Step ae Conitinck the ebperbioe fonction tohtch haste be aPiedged ae a Umtar tanation Bawelitng. deciston = vartablere m1 Step 3: Expsint ween, conaticton, ee ceca, wy Inercltige decision vartables, Step 42 Ei VTUPianeta womeforctiner prediacin B model TE, T omd IE of a certotm pracduck The following. Hable, give tru dake naw mofertale A and B, Raw Reqprirement per Unit maateriol ae wf me | Atallabiuey A 2 3 & 4000 8 4 Bs 1 6.000 Min Dem 200 200 150 - Paotit /Untt | 20 20 £0 | Fermaulale, Selution 2 thts problem os a LPP Lit tw taal ve. Gin profie [oe awd of aed T profit fe wth of model T profit fore wrk af model © prep lap wit of red profit [ont protit fp wk of Totol Proftkh = P= 30% + olOy, + 50%, . Objekt fowrction vs Mox Pe 30x + 204 + 503. Given that aacdely, L,T and W rtgytre 2 aamnd & unilf of raw matinials ft wn f rocdel OF medel IE (sune-ToLy 2004] Leb wa sata ne of wotlé of owedl T be ‘x Lax Be tial va. of antl of maodel Tbe %y! of wel ef model I be “yt ks Ro 30 as ‘ we is as as 30x Ree Lo to ¢ Ke. 50 #503. , ws befal Aaw-malaial A Aepdement = he +34 455, VTUP! iit |oveetlabttile Riky op Acus—materfal & 4000. bet ee ee Gawew that modeler EE amd W requine €,2 and Tuwite of rou modertod B seupectively - => ketal raw motoriol B Aryudrement = 44 Iy + Te Ruud fatal avallabiliy wate of raw maliital B - 6000 PH hx tly + Tz 2 6000 Giron Min demamd dot model Fu 200 = x 2200 tiw dumand for model I ts 200 —> y 2 200 Mav demand yor meelel IL 1D = xy 16m Tae tad op of sel af onal Z, mndel Dal medal I Como be negative Fe. 70,4 FO, ZO LP medel : Max P = 30% + 20y + 50 STC ant 34 +52 = 4000 An + 24 +72 <= 6000 kR 2200 + 2200 L x 1d where 270, yx, 270 f iATV Congpoang, far te dectide om thy mumber of Atinch, ond Lo-dach TV els +e be produced ab one. of foveies Morket acrvearth éndiealis that atmos 40 Qtedme TV neti ond ortmresé 10 dentine TV 0b can be fold pe med, The v ne of reetk owas awellable Ja 800 bow/ month. A Btinck Ty Atqutros go workheran and a fo-dach TV Atqpines 10 wotk-howrs Cath Qtebreh TY sold Producer, a profit of £1208 amd each dovinch TV produces, a $80. A wohrolercaler t purdcrase oll ty TV aehi Moduced Jf treme donot exceed the maxtwuwn tnddcatid ban market Actearch , Fotmntote tris problem at a LP meet. [bec 2004] Solubion : deb thy teal ruber of, Btn THs be txt Lit thy total number of do-theh The be ‘y Girem L om of Atlich TV producer a profit af 4120 =x wi of Atetneh, TW produces a profit ok t20x Given 1 wok f Bo-beah Tr Prodlucer a paaget oy £90 —_ gy etl of dontich Ty reducer a prapcte f 80y Total Pacfit' = [20x + 80. & — Wofarbive Fluachion s Max P= eC Given Hank mac cabes “fL Attheth TV As 40 =F 2% eho Given that Mave Salem of Qorinsh TV te tg <> yelo Given ab ene tach TY Atacces AD wth heourge. => BF inch TM Atgushes 20m weikheuns, Grwem tak ont Roth TV Argues 10 werk hours, - a ip Qo-tnth TY Pugpone [Oy rorikhenne Bhf. Total beh how rows segptned = Don + (Oye Rut total nerkhewinr available 2 S00 ae lon + \0y & 500 max salefnenth = gotlo = 50 total w, of TY we = xty Glen? puhelraler Well all ty ele t tke total dettnet — Lecned thy namebeauu - + x+y £50 pie neti Haak ob be of AThach Tvs manujadirrd cannot be = wegabie te MO - foe nets Hak dotl ne, of Qorinch TVs manufactind tnrot be wegatine bey 70. LP Model Maximige P= I20x + 80y STC n 240 y elo 20x + Oy <= 500 K+ yp 450 tobine nwo YO VTUPlanetSA former haw to plomt tae btndin op tues P ond 8 ba & hamd of 4000 samt area. Each tee P rrquines a5 qm- Each teen G requires 4.0 tam of hard. “The awnual watt arguement for tach bie P ue 30 unig ~ ound for tach tree 8 de 1S umils. Atiwsst 3000 willy waken ts arellable tinnwolly» Te ts alee estimated dak the rable & dre total ns of Kus & te tte ns ¥ thas P shovld nob be Lose than 6/19 gud yb Q atguio ISu- vwike, of ronlie :fe Total water Aequdemenl = 30% + 15g 2000 wil of watt are avatlable Buk atest —> Bon + [54> cede) Given fob He ahowlh mot he lee otal tees -P ane 5 a e 4 Ghven otal ies -8, grew nok he more tran [1 total (Kee -P ° =e 42 17 we s 7 we of Kies P cannot he negabive fe. xzo Total wy of Kit B cannet be nigable Pe y70 LP Model Maximize P= [9% + yp STC Bon + Sy = aoa fee thy ted £75 al 19 * 20 a 8“A compos maneatactoais diner products Pr, P, and Pe. The profits ane 2d, 20 amnrd 40. Company has Kee machines ny od Ma. Prvce Aine Dy welvwtis _ 26h anche on couk predict om thy total Adme owrablonititig on cath marhiny mre ate fon the foUloretng table 7 r Machine} ites 1 waht at Total Time i se ‘i ee 7 , a 7 5 2000 | 2 - . Z + S000 Connyienye ramst mrasnurpachare atleast 100 Pi ad atleash Ao0 Ps und SOPs buck nok wre than 150% & . SE Up LP model fox tau qld pram de Aelee ycioaglan maBed: Solution t ee ror] Le He total we of wwlli of P, be x pee tee fatal me of woh of Pr be gy tek te vel ~s of rls of Ps be Gen profit /uik of Pr = 30 — Profit fe wll of P-dox | Quinn profit /utk of Pe = 20 > Pasftk fy eS of Pr ~20y, i agioeen profit Jade rfp Fo =O —* Profil ’g- wwnlt of Po = hoe Total Profik > Box + deg + 402 i Bbgective Funncltens Moxkeataye P= dont don + 40x | Gime PrifurPs agua 4.3 and & mlwalls | prccemscrg on Me ' otal whe. ay mtoulls ry procuncng on My = AK + Spt FyBuk Mm, te available for a mak rp olooe minutia > 4X+ By +52 ES dovo Ginew Py) Pp ond Py reqs §, 2 and g mtwulis of procuting om Ms. Total ‘mioulig, of frocming> am Ma = ont dy + he Buk Mr 4s owallalle fot man for 5000 mits =? ant dy + 4x = 5000 De Be ale talk Hak tempera rawntectiine otleest: 100 9, a 200, « "SDP ot te than 150 Pg Sete ra SER ie. x 2100 J x £150 y 7 200 LZ 7 So The Latal number of prrdect Pr Per Ps tmonupactiine Conmotk be wragacive he. R70, y 70, RTO. LP medel Maximize 4 * Box + Loy. +4On STC yxt By +52 42000 ane + 24 +z £5000 % Z 100 x 2150 fon 100 2x 2/50 N Vv a ° K 70,4 70,270 VTUPlanetyh farmer home 100 ace farm . He call sell oth Comatees, Atti & radicher ha tom naise. The prbee he tom obtain js E100 per kg of tomatoe, FOS a baad for Lettie amd F 2-00 pr ken por radichs othe anevage yield per aan & 20.00kap of domacow , 3000 heads of bethace and [000 kan of Aadtohes Fertiligeu 4s avotlable at F050 per ka, amd ce anaes Atgpabatd por acne ws 100kq tach gor Kowodetn and Metiace aud SDkgr for radehus. Labour Reged yr sourtng, culttvating. anc Aacwerttngs per acre Ag 5 man- tomaters and Aadishus and b-man fos Lkkuces A total of 400 mam~davye pf Aabonr are ovatlable at £ 20.0 per raan-daw« Fowles Unts problem Ax A Abner programmerg, model te maxtnige ter farmer's total prope: [pecemeer zou] . Setutien hoe Knew? Profit = TEncome ~ Expendtinre Objective Runclin + Trcemer Ginen tent paler g Ihgo of tomatoes = | Given bent paler ef doookg of bemakees = dooo a moun’ of mencgn former gains =Lo00% patee-/3000 wailk of pope. = 3000 X 0.15 = 8450 | Gann ftratt pate pee ome untk of fn = pS => Inveme from poomphains = 22504, iat Given that pater per one ky of Madvihy = 2 Price i000 ky~ of aadish = 2000 => Inve pram radish, = Lo00x at Total income = Qo00x + 22504, + A00Dg, Expemditare : Roto of thy, of fertilizers = 0.60 Giver that [dokg~ of fertiligerr e Aequibed Joe tomatoes acre DF total wok op fatiltyrs por domedeofre> 100K 0.50 250 WY Guten Hat [Olly oy fertilize 2s Aequted fou ppunepletns rere —> ketal cork of Lertllizers beam pumphins /ace = 100 X 0.50 = 50 Fuso, Gdmen rat Sdkg of fertiligens Hs aryutred fo racttibes fare GL wit of fertiligen gee radtihsslacie = SDK = aS 2. Teel expenditare beards fod lagers = 50x 1Sbg SH 7 Given. trak : 5 mon @ FQ ds required, fr tomateu Jarre. 6 mondays @F do te Aegpcred, foe purepbehs fore 5 man dage @ a0 & Aepudred for taddtes/acre fir lave, fabews cost f tomatete =100 a th pepe = 120 ; tr radishes = 100 “. Total enpenrdttirce taeverds Labont = 100% 4 [204 + 100% + Pashte = Treome — Experditine / t io Profit = @ooox + aesoy + 2000 x) —C Bx t Dy + 252) — Croom + (204, + [00 z/) Opjettive Fanthin P = [850x +adsoy + 18 152 Total labew manda, Maudud = Sat by 4 5y, But total na, ef chage aveilasle = 400 => Sx + by + ox = 400 Total howd ewyalloble = loo aug = ut yt 2 eloo De 3a obvious that fetal number of acre of fame Camwok be megane . “70, Gr0, x20 LP Model Monee | a eocne 1. ACSos [8752 7 STC Kt 4 +z €100 Sx + by +52 & 400 My, kh oODivya” © Old feng Com be bough ot ER ence andl gouma. ents, ak FS each, The old hems dou B egos per trek amd the yourap ones tow S tage per toeek » bach egg bing nth 30 pace A how (yom or Ud) torte F 1 per rocek to pd» You howe only, TSO te spend fer baying omg. Hers mowngs of, zach Kind abould gow buy. be gtee a profit of more tram F 6 per week eee meee comets Wewte mere thaw do hus. Foerwldie es problem a4 Om LPP and sole appl beally [DECEMBER 201i] Solution FORMULATION. heb x, be te rumber of old Dens and ty he the uLmban of gong beng te be bough Cinee oth Ineurs coo che torent: Late J egge por week, oma tee young tot Tay See por wal the fatal numb § eng obboiued per week wll be = 3%, +52, the tome of each Lage being 30 pai, the taka, godin wll be = Re. 0.30(3x) +522) total openditine fr pordeng (4%) rend ab tw Mali of fut com wlll be = Ri. t (%1+%5) thus» total profi x Looned por week will be xe total qain- total oxpendittvne, or ae 0630 (3%, + 5x5) = Oxy 4%) A KH = 050%, — OOK Sine lh Momus com be lornught at Re.2 gach and, young ents a& Reus oath ond teene ane erly REouroilable fon purchortngn tnens , the cenatratale Ja + ! xy + x, = 80 Abo, since Uh 2s nak powitle to howe mere thaw doling of a ti, %) +X, 220, Mio, since the profit ts pusbitcted te be mete thaw Re. 6 Unde wean drake the profil fumcbin x te te be maxinbal Thus, thre ie we ned to add ome mite constioted) be. 0-5x, —O-1n, 26 Again, ob is nok prnible te fuachos ragaline qraabity of bans, tosrefere my SO, % = 0. Lp Medel Maximize Zz 05x, - 091% Ste Lx, + $x, 80 ky er, €20 Kr 20 NOTE % SOLVE qrapricaLty ALSO fy VTUPlanetA Urnepome producs shoe tapes of Hats. Gach hat of the fast type saquinus Fiatie as muh Labour Hime as the second tigpe «Tp all the hats ane ef He second, ape only, the Lonep tay comm predate a telat 4 500 hati o day. “The awonkek Lowes docily gales af the fot and cecond tippe te tsb ond 250 hots, Atsuwning hak Ane profits por bok are Ree 8 ie ype A oh BS fer Fipe 8 formamloté tie problem an a Autor { model in order te deteomine the unabsr of hati te be produced of eacktae 40 08 te SOLUTION tek do vounpaag produce x, holt of type ® ovod x4, hate lige © sath: day: Se tte profit p after alley heater prada Ja gta be tho Uneor fumebions P= Bx, t 9x, Cobfective function’) Conte thy tomy Cou te ok the meth 500 Drake liv ov day amd A dpe of hats equine Bien az mach Hime 06 trol of tape 8, produchion ubsichion ts given og atx, +tx, = 500t, whan + ke Hew Labour tue par uni of second Hipe he An, +L, €500 wt, Bhar are Dinitadtens ov. the Aalew of hols, B So Ry, SIGO , %, £250 Lp model Maximize, Ps Sx, + 5%, sre 2x, 4%, £500 %, S150 uy = 250 x, x, 20 |7 A wmamufackiner of patra mediiwet as te a production plow fet medicimts A and @. hue ase Suefptetowe Z tovt owvollable fe make obo,000 bottles of medizines A ond 4v,000 bottles of medicine By but due are only, 5000 Sethe © pele ae te wodletines Com be pled. Frrrtian nt token Aree ours te prepare enough motatal pit, Wooo bottles of medderme A ound one her £6 gp, enough amakertol 46 42L. 000 bottles medtctar en dane Gb hours awalleble for Ute operadfon. “The pratt ee pr teotlle for medicine & omd Bet pen bottle for Meine B ee protium os a LPP. SOLUTION: Suppose thy mannfactire poducrs ty aud ry t f pottles of mrodieinis A oud &, . Cbece SL taki thace hours % prepare [000 bottles ef medicine A, the time siquinad. to fill x, thousand bottles warding A will be 3x, hows Sb | thy Game t nagpiud Be pepare Ky thousand bottles of medicine | B wilh be my hours. Therefore, tial, me soqhed 6 propore a, aroused leotties fy rudiclne Rand x bottles of mmedsctre & UL be Bx, tx, dows Nows- pine try total Monae awalloble for tits operation: os bb Rows « 2x, +x, = be cae aga bottles puratlabt for filling, medious A omd Bx, +t S45.Thaxe ase sufptetinds Lngrddinls gvailelte to make Ao trovrnand bettlles f medictre bk ond 40-Hromrand betes of medictre B henw 280 ond x, 2 ho N umber bottles being non-regphive, %%0, K-70 As Be vale of Be por bolle dea kitpe & madiiefnn and Bet gow bellle for Kine & aididine, We total prfite ow hy Mronsond bethles “of onedcetine A and x2 thersand bottles of mudictre B will becomes P= $X1000 x, + 1X1000%, & P= BOOOXn, + 1000 Xz FORMULATION: Max P = $000x; + 1000x, STC Bx, +%, S66 Ry +X, £45 : % 220 i %, #40 i ond Ke 2 O : +A fay compar wnamnfactions tse Hae ef delle, o baste version eA ees delux verccen-delt B. Eoth doth of ape B Kokey tivte as Longe ee prodace ax one of, tape Ny ond Ke pany. sould trowe Bim 4 make a martnuun of S000 A Lastre Ae gutecd Dt ge 2 the Su pplign of plow . sup ietend ee 1sv0 dole ye dos, ‘he ee worst Reg tres a fo dass. f which thie are ool. G00 pos low avaitadle- Ty tke Crop makin a prepit PL Re 8.00 amd Ry. 5.00 per doll, pregpecttnaly dott A amd 8 thon Irene nrug. of ARDY, dn pea Reed bee tae a des fk anaubice otalSOLUTION : Lak x, od %, be tho niumbuv ef dels. produced por - dawg of Hyper fk omd B, reppectinely - hate Aw doth A sequise Crt tw deAL B sequin Ab bx, Se the deta hime te mranvfastinn x oud ey belle strovlauot oxcerd ovat ha. Therefore, tux, + 26x, £20006. Lp model e Mariani se P= 3x, + Sap STC Hp + 2x, E2000 (Lime constraint) Hy +X, = 1900 (plastic constraint) Hy & 00 (dress constraint) and KO, MFO . A fom com Produce, Ahree, of Leth, say th, Band C« Thaee kinds of wool are Arqied for i, Gay. t sed, green and plus wool. One wut of length of tape let. needs 2 malies af nec eel and 3 wnaties of blac wool 5 one unit lew / dott nerde 3 makers of rod vovol, & nobies ft Op mabe th Seni ere uct of a chth neds. Smelter of quan wool amd A mulkis of “blue wv. “Tha foam Seo banleg a stock of Smalies of red wovel, (Ormolkis af gpesm vorcl amd 1S melkes of Bue tool. DE de arrumed Yak BRE Bareme ebtatned jrom ore wall: demat of pe A clove ts Ri. 300 bf tae B cloth te Rs.5.00 med of type C cloth % s+ 4.00 Detomuiw, hen Be fre should nse Mb araileble nrodintal 40 Aato moximige the Lncome few i pg. *p d cleth, aeSonuTion Quality of LL “Type _of Clete Total quavdl heel Aa) Bos) clas) | er Rad 2 3 oO fs ¢ Green ° B _ 5s fo Blu 3 2 Lrcows, PO wi - = ef uloty | As 300 | &.5.00 Ri. 4-00 AYE marta ond ae te Grantily preduced, a SE pant AB Le Sine L ntbhes 4, ned ecol ane fp each ameter of eee oan mie 4 ain gery debe gow produucd , 90 Ax, mols of red spol will be rcguired, Jor cloth A. Sievilonly, hot B requires 3x, mubies of nud wsoel ound alot dotnet regio ned rerols Tens, total: gyranrtity of red wool ie | Ax, + 3%, 40x, (red wort) L Sicnas t Soca Om) 2, HSH, Copenn wowol) i Dx, + At hr, Chlue worl) i Biace mete rare tran & malts of Aad, tomeler of gasew i oud (5 malig vf bine meet one avilable, tke vartobles, Ly yty amd x, mab sakccfirn bre following avskciclions 2 ax, +3x, 28 i AX, + SX3 slo Bayt x, +x, SI Asp 90, % 70, FO Goa total Barone of tin fivivehed cloth ts glnen by: (max) Pe 3x, + 5% + 4X, fF VTUPian® DIET PROBLEM ; Aa tiers B a Ore of the Laterecting. protons rey ey sonade of valarwed disk, Diebtctane Hill wt thak & bole tae a ae oa Soe saat fd iL 4, vitowins Oe oe Brad sreuld be Secommended from a Lange mumber of alltrmokive Fourie of trees maatstonts ae Yen tol work of fred Aabeh gh mmabnteniein Sagytrtimnerds halouned Aik te ter Lowest. the wedreal epee gud dteltetans bill Wn that tt ene _ aur adulb 26 comme ptleast 154 probivs, 854 of fale, and B00 of Canbalrgdratis Bails, Thee i Sag tanle 9% the food Hems Cusbtolr are 7 avelbate tr the weorkel) awualysis, anal tir Acspoctive costs, [eee econ “| - mute roves | FATS | Crewouvpentes | (Re) | it _ Bo (eo 35.0 1:00 a | _te0 1560 oT on 3 16,0 HO To h.d0 : 4 rr . = 6 [as | “willie Datey | ~ L. Reauigemencts | TS 95 SovuTion del Aap Ra XQ eer Cr of dood kipes t he used por days bo o disk amd tre total 1 anak ety beg, wibminmun roquiremunds, Ae hpec & He quimiming fetal coe © 4 pptThy objeckive fuckin tau becomes Zeit By + hts + Akg + 15% + IX6 bince 8) 16, 16,45 amd 25 quis of protemns oxe overitabfe, fem 100qm ust af cad tape of food respectively , tital eins anollable prem %),%s) Korky te ond x wl of wou food nespectivel coll be Bry + 1B + Lbs + hy HSE ASU gyre dacty « Buk pabvtnu, dots acgudvennsudk of prebeins as olescribel ds 1S qnas « Heww, the probaly neqpirements Constraint is Gey + 1B + long Fy + SX5t 25%, % TS Cpnctota') Stauton, fale ood: Conbelugdne tis seqpinemunck oendtaints ane obbatiud acrpeckively os below : | JS, + ISH, + Wy 20K, + ex, + 0% 285 (Fok) BGK + Ox, + Ty + ATHy + LOX 25% 7300 (Gabshadactis) L Heys PO, Xp 70, Xa FO, My HO, MFO, XO, ' ro nwcaanifa Eines oF bigots bs cowndortug for Tapes ob aye packs contatuing tone types of bireutts + range ercam (06), thevelate cam (CC) andl wafers (W). Market research Conducted | aacewtly be acess the pepe of the Consumers reves. ee | i ASSORTMENTS CONTENTS TT gece PRICE tT _ | pee eg CRs) A les ih af OC, not move tran 20% van aiyet | Q | Nok (ese than 20/+ of aC, not more than 40% #F Coamy aly of W L -_ & . ae lex than So% oc, not more than Of. ee, sory qty of Wl No vecksée-tonefor te bath, Lee masndachaing, cagetly, a cette ane gat | Bdoutts vowel 3 oc Ce w Plowck eoyoathy (xg /doy) 3 doo 200 150 Mawfedtioting Cost (Kj: 8 4 Feamualols ov Linsat programuntngn medal be find gurednde rtd mornlaretges tee propth antemingr that there arn ae marker nextrictions. SOLUTION ¢ eb sce dusteton tables %if (ih OOD: feb2o be depineb ns: - (i) Fer Be gift pak, ars ar > As duvet te quoudily bo hy af 06, 66 amd 10 tip bona (fer Hee ate pee B, Xer7 Xar> ea donetd HL gravely aw ka af 06, CO and © bipe of bescunte (ii) Pew few oie Ur donot Hee parang Bn kg a Oly CO ound 19 bape of bikuils | O) RA Bw gif pak Ds Hp Xor9%os Beret IRL quantity guia bg of OC, Ce aancl Wo Cape of bist Neus kv giaetn datos coum ba pa iw Lew fernn af Bem LPP ase Maxkratge P= dolxart tart Xn) + 25 (xe Kop Xes) frre Lert Hes) + 120 x51 Hy, 3) = & (Ry Fert Xey +21) =F (art Ror tXurt Xo) =T (reg t%ext Lea + Xa)eo Mogttwize Pe \rxay + Wart Dagyt Mei tlo%er + 18x03 | i ‘ FUE, TNBLeg T (SXegk py + 3x2 + 5S %pg Sudsjedt be cowiowls Cue PoukA + ta lar <= pero (ar + at Kaa) v 0-0 (ar + Kar 7 Kas) Gif fou Ay 7 O20 Xe + Ker t Xe) ign & P40 (ey + Xap + Koa) Gath pas er Hey 7 50 (toy +Kea # Xen) Ney = Oelo (Hey + Xr + Rea) Kay + Ke, + Ly + %, < 00 Rny + Xap +R + Xp2 H200 a rr—s—sés—SE=“‘ EE xy 70 (fer beaye, op amd fat,2, 3) A complete wk of ao Crtabn product consists of for wa i ¢ -hramd tree veils ef component B, Ture : a (hand B) ore mamufackured piem Taw diferent | Sow malidh qf tehich Lov ui ond 200 wb, doapockively, i ane ovntlaie . Thrse clepartiments ove, engaged By the P Producttory tere Wott each chepar torent wateg a olf ferent method for main fathering the compenerdé, | ee fottowtoge table gers Hhereaw motertal Wes per production Gur ound tae Aemlbing- kk beet:cenespensnk The objective & fe deberminr Ue armnber x} prodnchion suns” foe each departments tortel extll moxhwdye HL total wimbor sf comapenenrts vil of tee Jona pret. ae Tape pe awit Guckpuck per saan (unis) f : te matertol Ik TB | Cowpewat A_| Corpora B 3 6 4 2 4 8 5 8 = 2 2 1 1 3 SOLUTION ¢ / heb Sinko. % bo bo ruvnbsr of yroduchion ands fo By Aspeatowals (2,2 suxpeckinely.« 6x, + OXa + 1% Coomponente A) AX, PBR 3% (cormpenunt: ®) od sanpertcly, gore Mp tay + Uke + 2% = 100 5x, + BLT Te, = 200 Cbnin tie vbpfectiny frelon Ay fo maxing thw toral maunber ef urls of Hay final preduck vamd each Auch ww paqpires 4 omils of- component A ond 3 wil ef Comnpomsnls B, Ae marten rarenbor of vuatls of sea Lonol product Comme exceed the sirallr value of 64 B+ 1%) ond (4a * 8x, 4 3x5) x i| cot for. Unus becomes s : Moagtyy Zo ne 5 (ex, £5x,4 Ms), 4, (beet Sx.s2a Binrre His chfeckive fomultors Ls ovet Minton, 0 canctabie ba oo cow be ued to peduie kro above model, Suppose» wdin | (eer + te TH) L(t BaF 3r)) =v Theattes 1 (pee Sept Ths) z¥ ond 4 5 (4x1 + 8%&2 + 3x) ev din fasts, ak tant one of Are five Hnsgpallls ual gow golution ecause tie ruumber | old 04 ow cla t 4é ' 2 te, yb moxtrettad., oan upper nck ts xpedijed by tke Smalls of the taph hand odcles op Oboe tie Lranpalitin,, This Andicates thd: thw ocr frequalitis ae eplvalert te the Rathod, eqnotion defining Vs Meus, Rho above problem tm be put subs se oll Lage Lscenn programmtnge fen + Maximize Z=V Gulgeck te ematiounls Gxt 5K, + 1K - 4v 20 ! Ax + BR, + 3%y -3V- 70 TH th Gt2%, 200 5x, +8. + Tx, £200 Xr kay MV BOde Comnpants ros three oferaltonal departments (eccowtng, | pacing ) este Capatlty to produce tase Attjeent tips of cloths rarely, gulliings, abinkings and yoeleine welding ee profit of Bud , Rad and Bs. 3 per wd, expect ‘ Cre meter subttng Aeaives 3 raimntis in Say» Qo ovdnulia Ae fy Say OAL | mbna, tw Thing, BY aeslen of shistng “aque A minutiy Ow pecoutnge Tint Sw pfrotesings sued 8 ivwulis bo pack? reticle ewe wetiv deere wages sndnutes Aw Lach dept Gn a week, tatol yuu dome of each cleparbenerte 2 60, ho gud BO hours of ap telcuts oe or padetag, Avil wrntate Ae LPP. alepartrnends narpecint Point max (Total Profit) Pe ax, + 4x, +3%s aes ty Constraints Bu ther + 3% = 3600 Dept Xe 3%, = 2400 H+ 3x, + 3%, <= 4800 ard %, %0,%, 70, %3 FO gsocuTton Resources PRODUCT Total Avotlobettt Constraint going — | Suaring (WoeLen Cinta) Weaving Dept 3 4 3 gorse | . [Processtag Dept 2 \ 3 40 x60 [Yacking Dept v a 3 ~ 8 X60 tp model VTUPlanet,_"A bom Days castings castings of Pomd 8 Hype af park and sells thw ue tote pote efter. ean, pening. and potishing, The purchasing. cok dow cantlngs ane Rood mak Ring coche fr parks Pond d and telling coil ore RB ond Ri. 10 Biel, The pow hous f emachones bk fo abi ng wok pbc, fon test) product gion weet Copacoy Pew Hew — eee on Pp _ _ & Machining. _ 30 ___ 50 Boring 20 _ 4S Po Lealsengn 45 30 The aunnbng cons 4 machtviing. bathg ond pettshing. are Li 30 Pee BRT aid, Res QT per howr sapoctively FoRMULATE the LPP te find orb the product mtx to anaxiontge Hae prog. SOLUTION | Casting Opesation my Xe } | Machining 20/30 = 100 30/50 = 0.60 Bertng. 22-5/30 = 015 225/45 = 0-50 Polishing 22.5/ 45 = 0.50 22-5/30 = 015 Purdrase 3 4 Totol cost 5.257 : 6.85 [ Sole price 8 __ Jo Prchtk TE, : 41S Lak oy sey be the ruember of Pond type pants to be | produced per our respe nanpertinvela. Fes Ha profil poate of, xy and | ay, we coluntabe He total cost de each ond deo tale pace vl that park fr]em iOm tne moakint, type of P pants consumes 0 of they | aureiloble hour, a type B park Conrmes Youth of cin howk 2 d4o ARADHYA TUTORIALS :; 9972731111, 9972851111, 9845642144,4 NZVv 5 t 4 | i Sy [on ° + 6 : Pivot = 23 e = ogy = 51 : i rev = ¥ : : NZV : ' bs Ral 7 = 40 4 PR A : t Pot = 2 ft . ' avemcrm | rev i CR > Rifas wR > Ry Yaa NeW” Ge xe ey St Sal tp SG o \ ea x a Lrrr—S—r— 7 QO _z=130; 0 0 S % A alt hs! tadPeatetr ose men ove, Kmae = 130 and ik | Atos of 4 x= 40 Gnd y= 60 id i ARADHYA TUTORIALS : 9972751111, 9972851111, 9845642444, 9901Bxry 292 ¢ Qe + by = Dy | xy 0 } 4 | Solute » Tinted ver cobler Sond $ ey ths 22 sie forte [S ie : a T Appl, Re‘ Re eee 7 R Rieke ee a + 3 { |_| -¥s 7 b © | Ya | [Sind Fase" | | 0 o_| 4s ' Ae oll the" Yndtratou era spline poecsdust plete e's Zone 32 pb wh j | 2) Ualng Afnablee anedhed tiatolye b= Bet hga’ “ste 2eey Sto | i Qx+5y = igo I i ae a ! 1 4 l +t ARADHYA TUTORIALS ; 9972731111, 9972851111, 9845642144, 9901942144 ' FYSoludin + Bothy Si = ho L Qn $Sy + Se = [80 | aS 4 ‘ > 2 VeeeigivemeCes (a tect tat fs, Sa Rodntes St Oo 40 (2 ' t ° dof, = ho Se ow 2 6 © 1 | 18/52 36 4 cE eon eo b, cep fa Zongx, = I TS ot x = Sf ye . J ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642144, notte. “cd SE fy 3 ke . th : St One 2), ° t o Ys Oo «— i fF S Oo mbm 0 o \ 4 | ke I i eae Vea Onseas Ya 1 i} 00 eee ers) + ‘0 fa it ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642 144, 9901942144 VTUPIP, t Le 4» Ri/2 picotal olemndaake 2 Appl Bi! => Rifa . H OGV FS Ru > Ri. Sky i Dey te x gi > R.- YeRy Nzy Ce | wy SSS: LS — Se ve End 5 i 2 = 32 at 5) Maxtinats Pz box mee po STC 4x +2y "2.0 4 (Ont 2y 2°60: ‘| 3 Fey yp ZO ARADHYA TUTORIALS : 9972731111, 9972851111, 9848642144, 9901942144"Ib 7 1 i t | 1 “ 4 iy a + { t ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642144, 9901942144 i \ | VTUPianet,1) Maxtrsaliye. subj + Buy + Be, STC mut + 3x eu Quit 3m + Sey 22 Jvetue ¢ [Optional sith | | t Med Hy hy Spbra + Sey=h No. - [ - I i [Ba,+ Say 7 2 T | AA Sky = No. 4 No | at + Fug =7 lk 3 ert 284 25] nol 6 | Yes Quit 3t22 7] _ L | o) All the thse one baste Jepsttle 8) ¢) Solution 3) fs epbinal | | i { | 1 \ ARADHYA TUTORIALS: soraabiai, 9972851111, 9845642144, 990192744 UI Piai STC Gx, + 4%, + les + %q = too i tad + Beg pb ey = TS | ' oh) Maxinige, Za tiet Fey + loca lore | lett type ba A nksasinonst toad tive 09 uation [—__setotel ne if baat sobutioas total ie. doss Hoetile = 40, 6 [ Smee at ¢ * & im bake [Baste OF t ate,| Yarebles_| ental [Canotions Solubon ” - : | {1 {4s Hed — i | Xy tr=0] eal “No - - | L tye Her i | 2 My =10a4 33 i : « | 2 yO te Gane Kye (00 =o 2 32d ky wet Byes | %=25 | Yes oo No. r 7 i + i | % Kyo i r | ran t : 4 | te ee | deer deg too] y= | _L ty =O Be “Pat 4eegets | e275 Mex 200 No | 20 4 i + } - } 8 = oO | 1 Sey thy 100 | ay = $33 LL L Xs 20 Ky ay thy eIS | y= Yes I I zo | . ny 2 583} HK 24 Say tdeaet00 | gy = 18.054) Xye0 x Ray TBH, y Ave Yes l zy chide) ued ARADHYA TUTORIALS ; 9972731111, 9972851111, 9845642144, 9901942144 givi i i Value OF E Nev ev & SAN Be oe Dobie sf] | i 3 t 1 | 6) ese) | uy |S ete, 100 | my oaieoel | . myeo | asta es] ee Be] ve, | 200 a i Rae i a! . 1 1 3 too | Rut eet 2 i} ae Beye Xe = Nh Res [i i 320 | Ji i ARADHYA TUTORIALS : 9972731111, 9972851112, 9845642144, 99019111 TUPlamet i ¢2x + bx, + axe + My = 3 Gx hts + ea the, =o ih rebuction i j Givin _Aaysdieon hod Ded nabiocs Le 4 unk Sx (e-?) 2 2 unknonns ty 2er0 atiq titre ‘edotel ne of awic: aalnbions!s tC = v V L Scene bee ey 8 Xy. “sh tr + bu, =e ( rd 3 20 Xe bur + 2x3 b3 Feo [a Akt eed iu § \ t | - + 4 20 % Dey Fey 3 I X20 Kay fey +6 %y 22 { Seem ed) Ly Wy + ax373 jy 20 x, bry rhs od ARADHYA TUTORIALS : 9972751111, 9972851111, 9845642144, 9901942144 WAV TUPlanet_Oe : seamnamnanen ze Le a ma - Mobhad : : @_ Mac Ps fxavdy + 6% ceerees eae se 4xwon 2 470 Dube “Gig! = 450 Solution, Sa ‘ fe Pe wal ho 23 t 2 0) 220 i “040 OL CS) ° } 0 ANY, c ae) _ 5.0 o o B57 j Peo “4 -3 4 98 9 T Pst = 3 Aggy 8a" op Bag ogy Se Rg’ > R= 2e,! eens 1, ARADHYA TUTORIALS : 9972731111, 9972881111, 9845642144,a Nzv Ce peeneet SiS, S. Bottos om % O10 1 + % ob so/y Zz 6 4% te ot o 4% 9% 3; 0 $30 2 5 0 5B oO 1 880%— y-g P= 440 = © ° 2 0 = i Pivot = 3 Apply. Ri > Bia _i Ri > Ry 5h" Js all tndica tors ose agen -regabine Sisaloe pple Comgjlebe + Prax = 3200 gt X= O : = 390 = lo + 7 shat t Bz i yt dz <7 : po ey 2 HI WGn+3y 482 < lo § Siimplex methed sll obs sey—<—<— ' Hila Hhnwngheuts by 1, 7 2, the temitiaink ate =yr22 ¢ =O & hy = | ex £34 + 82 = 10 ee ee SoclKodute. shat yintables § = =yt22 Tie uy oe Sb = lh Whee + Bie Re ies TAA Cognore "fo4 Crgnore) A eso: Bods Ye V5 4 a Ys Wr (eanore Pivot = S/3 Apple 2: —» Ws Ry dgv = 5 Ri —> Rp + 2% Ry = re = x : Ri > Bt % Ry! ra V. ARADHYA TUTORIALS ; 9972731111, 9972851111, 9845642144, 9901942144\ | —— - t ZV fe Xe XZ ye % SSS: i oc “Ws 1 0s Ys 0 Ve ‘ ba 0 2% 0 0 18 os 1 We i 3 sH/s 1 5 ky 0/5? i pt. 6 0 la 4% 0H I Phra = 13 i > = = ok wo gl yu 25e zs ot sgt, yle lez <0 5 rae) Le Viwhke tral beck caw | 430 Li 2 I ‘ 460 3 0 2. 420 I 4 0 4 cs Poti 3 2 5 i LP mada { i Max Ps 92 + 2y + Sz l I STC Kr dy eZ 1g belo ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642144, 9901942144 VTUPlanet,r — i 4 §- ue t i yi 2 |) a i TT ARADIA TUTORIALS : 9972731111, 9972851111, 9845682144, 9901942 44UNITS SIMPLEX METHOD-2 Port Opbiolity ty Avaliysis Post optimality amalsin “Ae amalipts dae: Of aw optimal sabution ie dbtained for A itil woicrn of he omnedal - comstifis 0 wong. tmafory amd very. Znnpartoact: Ale: da words of, OR stad ies - The pack that pork optimality amalyris 2 mony Ppt 2s pay, ie oh ype’ UF opting tet Steps dew fx poskceplimatty cosets bee LP f Sete anal, Aha mightiest oe ae tome aatttveds on abenglon meturd. = TASK | __ PURPOSE, - | TECHNIQUE ] inedel Delagging | Pak owwnt L weahiness Bos medal | Reeptinaiya tion | | i Medal Volidadiors [Retaocpective Test e| Demenstiabe volt t of Kinal modal Final wranagabl make apprprtate dtvisten division of Shadow decdstens on Hew Aganinptiorel poms beeen pieces | allecattons aukatites index oud | Lnnporbaud actin! etl | | Evaluate ebm eben wouckal atimalis thet) Sen stent | of model j mag apfect eptimal, mn arabyets L parameters | fut dirk» | Evaluate trade | . | Parametece | che between | Delowending bet ade | Ancor ’ rege ming \ medi paramediasReoptimization t t : a e < 4 = for one version of LP medal, we pegpartly musk solve Jim gos tae, selection of a Abighdly defpermt vewien of ba meds . Ome approche Bs recopply sani retina dyer mau more afftolent approach la do Re-eplrmigt. ay the forol solution dk feastble, we ap; ual sé aerate ae ere Binraglex rtrd» The bige advantage of this ae-opbinigahn Lieluddgyer over arradvinge fume the 4enokch ds thak tun, vite L solution fev the revtied model probably, ts going: Ie a the prioy optimal dolton, Hnom 16 ow initial BFS. : Retvospective Test A sep tioncttin appracch te Tesling te model called a retrespetive tet. This tik Bavelver ting the De reconnect the past amd tine Aatirmiat frees eel thy wed and He Aenatting solution would he perypmds Hy dag had leon wsed. Shadow Pisses pp vcftin com be Inltipartid 4 allocating Meson 46 eabinthe. Dr mong tases, the may be dont: Labitide haw the arnwounk that will be made availableonformmattion, ore Ae evonombe conlribublon of the Avvorrees | te Hen measure of performance (2) fpr ty worent eae wef Ther srbaphen mat bved cae a nr ren nas Semsitintly Analysis ae a rap oot my he a ppadicting the tyck of change conbiel vantoblor on tay enrall syrtinn cfpckiventis, The usepulner of a omedel te tintid by cdutermining Rew swell Eb pradktcts Bau eds of, there changes. Sud an te called The vor validity of Ae Foluctiiow ony gi te Aly nl Wttrenk applying te solution tte the mesulls obtatned bolum WK Ue used. Ponowmetric Hear Pregra Stas Sonate Ba pe pre dalirnviimed condiinuone Vortobions of Vortows Co-cfiivionts on the optimal solution, De is extimmon of analysts oud. cri ok fine thy vortewe tite atts Laake bona opto, one aftr © Be ether, Os thy Co-eftietends ¢ ed eeeeelaee a ovis The catffielonls homage a6 a Miments punchion fo eagle, parannlir » Dame ty manne paramebiceI ue ry ee an & Luplementodion oe | “et vf ete cony eo I a com us ohh wedeww pole ag) be oni’ of 0. Avphiabicatid Aefooare package joc Py he ptt sp ce te di os Bait fe “vento ‘uty or tpt gla i te aad Me oo matriy bom Comal, ae pantie Atmnblax mame that lg ma tl ne ie ca eek ore i SO A a a cal ee tis whale Computing, cand hud a x lige PP, 2 Pte Mek owe : a emaie wilh be made Luilially | an ie CS GSI re a eee oe Cen pet sain oP of pt examine This nthe procun 45 getty fe be carried ov dvternctinvelyy , FP res thew! Available selene options fer finean Pregaamming. | 1. Knock amd i prumion solver for _ | wimg LP models on a sprradshect format + Gpreadshec — j solver are becermaing. Hoes Lovsiogly, popular, fr Mncon produced boy, Prawntlinn Austins for, Mieadelt Excel anol Othe Apreadsherk packages. a mpL / CPLEX fe efppedenctly povmulating and solrt, donge Uinta programing ae reer pagpaded te be a paabioclanty prntrful package for solving, maasine probtonns is CPLEX o predusck of IL0G, CPLEX offi user the Adwplex meted ard 2 vastonk te seve these massive problems, w LINgo amd Ae solver (shored with Linpo) for om alkerwatione wing of eppietenty formtationg and gotrb Aang Mrrear Pregromning- Medel, LUDO Estiont pos LINEAR , TAITERACTIVE AND Discxere optimize] Se a Acfkoane bobich adres LP mvolel Sw a sEalahtforcoar$ ad +‘ | - ‘ UNIT 3 ! i | t L ARTIEICiAT VARIABLE ‘TecuMaues | \ i LPP. in yohtcd. Conable of the “type = andl = sane presents Surplus vantabls md ontigictal vautahler ase Pubirealusced 4 “Tinto - Puase Meru ty j———-hlerkuiag: Rule, Passe -T “ ‘ i uicen LPP is consuatid ced. Thus, tie lnk _staaidond dey Shep 3 2 4 tote le tp cack ae AV awd a cot tp | uit basil, Te" J i _| tind te Lud Hoe elubion dp Prose 2D i “ee . i t s t (7) Maximum 272 Vf atleast ent jAV appease bn | ‘ s, : 1 i optima bass ab aps i ARADHYA TUTORIALS : 9972751111, 9972851111, 9845642144, 9901942144 . : Fi VTUPianet52> Net deasitl - 7 SP. We Anbieduce tm AY, Ay and a Pig Be-ye7z7-5, +a, =3 I= tye te — Set) = 2M 23 Fayed Phase- I design a Cov Hl te Avs Oy ond ap and 2 rest 0 ARADHYA TUTORIALS : 9972751111, 9972851111, 9845647144, 990198744 TUPlanet; mmet |eo o © .co ~ — NZV Ce Xe % 4 2 Sp ; Roses | a 7 3: @ - 3 OY ai Al 1 t a =) 2 pot th ot 2 Li Arc; PMs -F eet 4 ° t Pivetal element = 3 Ogy = a Tey = % Be og wa AV we shadl d By fe the neck stmt NZV (a Xm x. Qo \ Ay -l t Zeg Ps Pivot = 4/3 ie = Oe Tey = 2 Ce, % x oS } ~% Dd z o 4% o B 4 gy pF -0 oO. 0 0 ARADHYA TUTORIALS wun PP + 9972751111, 9972851111, i | Pie Loe net stat PP* 20 ond y Op Linncce boss 2 TL Ay ha Appears Gr iw Ab Lg. SBAs6r{s44, opo1942144 | t | i iDidopdinte 0 wae ay Pee ys Sy ot Rly — Se th, oh ARADHYA TUTORIALS : 9972751111,, 9972851111, 9845642144, soo1sypaayT i Ewit Phasead : Aerio a cork -| ts AY a, aud) cot 6 ty all \ then vasa i t NZV (a Xe my 5 a, Rad oO { Vite, mac Tt ee & 12 p Feed Vo contains wnite of Vilanin B/oan ~The a vi 0 | ase. tno iwi £ WO wisi nespesdivdh,. Eland ste dntuvnn, ob prodist# wvnulate, the prolom asa LP noacol A solve tt Aliclting LPP hy eon phase mottend E | i I —Giein abi ie pant in taboos showset ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642144, 9901942144 i . 7 fy VTUPlanet i nokFood x & Fd Y 4 ee wgatre =) xu 20 _——_Fend ¥ cewtioue fo wil [ah Vit 8 oleae eee => total VIR APS be poy Buk a nais Of 100 aadis of VIA Ove reguived z bu + Sy Ip food exfhisiny “Twos yl vit BO food Y Centar 2 unids of Vit 6S tal VHB = Teete of (20 wats of Vib Bo ausised > [xr l2y = 120 " 0 Mia. Z_= fle fp 20g) | gre | 6 tay em Ino x lay bea a 2,4 ZO + Mar 2" = “fay ote i a Bay orianet ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642144, 99019! _ 7Luce sunplae vortabler & fence 2 Ay bu tf Trt tay Ss bar 2120 Titioclu. 1a Sj, #41 = top Phase ET: Myign. a tock il fy bathe Abd a rose 0 ta all Pahl 8 Nay G = a = a -l xy Lae Oavc av Fey + y Pivot = [2 2 os 0 i es Oto tp ol ti 0 me ops. 2 opts mop atin et I “Ys 2 = a er A : 0ay =o Apelig. Rim SBR, od Tey = % Ryl => Ra Yin Ky" ! AS gy fh ow BY, ae ata Nala fed cols i LBL! pen A, fa te next 2tmploy tobl t ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642144, 9901942144 GV TUPlanet,3 a : Nzv Co Xp x Pee eh Sy 4 x x o Is 1 Deseo nee - yo. he ot ee _ x “ef ut 0 o © 0 0 _ 1 T =: 1 ds max 28 /P0 fk as, OM fn AP ra bres nen = ny Give, gheue st + do complete, whralio note that Av, dewet Appear Ln itte, tlie + ee eee egg eee a nent L T Le oes ab ee > Der aon), “ User two cnlie wwliod in “nuiuimisn Z2o.4e + OS, | i SIC O.3" + Ov OSx poe Ovb 2 4+ Bol) i ad Sols we why =~ (maid) a Tidete way Fs Oly OSI 7 ARADHYA TUTORIALS : 9972751111, 9972851111, 9845642144, seo1siaegsy TUPI janetSlick variable | 0.3% Oly +S, 5 2.6% 7 2 tend _benctrnint) Ut of ye te eitodsace ann Ay, ar wx OSX tOSy ta, £ We mete tat did cometrulat fof ciggel 2 2 hte nchodice o Surplus lartakte $5 2 2 ay at Doky tbiey Sata, = | SOLVE LatEe t f ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642144, 9901942144 | WV TuPlanet!i Chars Peal, Melted (Bip Method) be tee OF naaximizatian ts b - hi NOTE: Sf ov given conshatat ix of egy tubieclced len tte Lot trond dide 2 Rout fe ste NOTE + Whee ow ve _hevena Gy, we shall invere e 40. colsman, Cehntipon did He Ov. in the Sabergueni —.? aniplue Zeb L ARADMYATUTORIALS : 9972751111, 9972851111, 9645042144, 9901942IAW TUPlancei 7 er Wisp. Perrallic, method abu ondae. Zs dury Y Z STC 3x + y = fn+3y 2 Xt yh v ey Solukin Ag the first Cowsliacne i Kinteod Ah Av a Zur yt a = 3 24 ZO at Hy tye “= Penny —M a, — May i o 2 7 = i Nzv s Baya Radi t ay L ° ° : t © 3,2) +— ay 3 = o oo! by tS wa Sit 2 0 t oo hel Zj-G “om 0 o, 0 - it Pivot = Bpplg By ~9 R/3 hav 244 : Riis Ran thy’ | Tey «x Ro Qa eR! | f t AS tae ny fratt > 5 . i ARADHYA TUTORIALS 9972731111, 9972851111, 9845642144, 9901942144 VTUPlanet_= Si Sy Ax o. o© 2 =o i o_, ol o Me Ry! > Ra X39 Rl 9 R ~ Kea! Rg mo Rin S/R ener 0. he 0 Vs J f 0 Ls ARADHYA TUTORIALS: 9972851111, 9845642144, 9901942144Babbitt, tube 3 dy $$. =3 ' te ie vbVie OAe introduced, the hen) pOF i fax " 2=2e + ye Ma, — Mar 1 i Le \ 2 2 &, - 28 ty 2% Bol —> Re- 3a is Bs . le € Neve Xe 2 yg 5 BS, hats Ss han fit fea: Sy OES. Oo -t i gL) 5 + Ea 63 i i oo =! Tenors [tnd azole oO 2% +b D0 6 = i _t Piveok = ogv = Ieve $. ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642144, 9901942144 givtpl Ral Ru/s Bu” ER ake Ra’ Rs Re! SS i Nzv Ca Xe el Sy Ss fitiee i S 2 Me o sh 1 Ho Ss of o “Wo at 6 ~ 1% 0 Yao Ind. 22s © Oo v9 2 0 Dh fodbtcates one nem: nigabit, tfutles proadure. b caingl Znoax = 42 Mand W oteuss at 228 and ned. (select 2 cabin. 08" piverbal alsns) a ho bee BAD been are N2v Ce xd A o 8 oO iS | Sp oO 24 o i a eS ' — b Zet o mM 6 - eka! : ma i Rb Ry - 3a. i Ra’ > Rr -28 : — mm iy s Se Sy ate ' -| 90 | o @ 2+ a ee ee ee \ \ iu ° cal TE ANoKE f 20. 6 (0 be i a Ae Ril fi _ Ral 9 Ry HEIR)” Bl Ba +t RY ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642144, o9019#21¥4l'U Planet. Ll} = 7 oe @ © rT: NZV Ce OX % pee) eee Rothees gf puma ae ae eer eae ea yt Ss om FX 0-43 yo tls i or (eae ibe ee) emma / ope One On ee es iz if Bnd Z240 —“'% 1G 43 0 0 a if +) : i nays S> LCL i Lev ze pe Ry > Rye Ye'RA! o Ra! <> Ra ~ 2/2 R) , 3 i Nz) a a i St Women ti ae “Ys 45 0 “ u 4 5 Of) Ys 5 0} ! a ~ | STC R+y = . i t i we Bxtytz 210 Ruy, RO \ Petr c Ube Metiect te aii tule 2 SH + Bip el é oe Tt SHC pm eyy ele i SN \ 2X +2y) +10 f 7 ; 4 \ v Su ty! = (0 i a 1 a My 20 i ¢ 7 F ARADHYA TUTORIALS : 9972731111, 9972851111; 9845642144, 9901942144 | } fy VTUPlal* GAME THEORY {- DECISION ANAL YS): aleh € : d ho2° 3 ----- 9 ' ‘TDs ee = [y = | Th Waly, i Voy [Vix } Vos i Fala Wal Yad | 1d vel val vel: ARADHYA TUTORIALS : 9972731111, 9972851111, 9845642144, 9901: Wry Planet’UNIT 6 Davega A transportation problem ds to thomapert a adage haqutns emeily at I fam? ovis 1 dipprunts durtinodions puch that Be OE eanagedatin mins. Sealine ay | Cat) Gon Ga oe an (Cau) Coe Cag oe t+ je Om * This -tabie nepal o * pand rn destinations Kaansperdodter nw proble me(ma need ret be equal ton) Ty Zap = Shy, Hew ut caléed ov a balanced docoosperdatton pacbloes + cthemedce If ds called ag oon unbalanced Hamspertation problem. babar cs te fe the ict Of aeetlga 4 uk in ticnsperts is Le o uw 1 yo Spe wp “g Cenuneerdlity pee th righ be jh desbiatien . Be VTUPlanetFeasthle, Solution of Tramiportation Probleac | A sek of posttine vallecations velbcl semoves depietencien ts tolled m4 a frautbler saluction Baxte Feorihle Solution (OFS) & posthle solubien ds called as 0 Caaie Feovthle solution My fetal mambr of allocabions =m+n—| ree one hess thaw Ae Lelal mumbir of nowt anol columane of tke thanspettabbn table, the tetal allocations exactly tenet, thaw tke solution ts alo called as mon-deportrali baie ferdbie selubin. Zp the total allocations ts bos how men-l, Ban Hk called as algenerats baste feortle tobedion, - Opkimal Solution . A feordle solution Cet wecunconily barte frovibe) A cable Bu optimal satution y We oreo Metter = To 1. Nonth West Commer Rule (nwe) 2. Matrix ~ Mintiwa Method 3. Vogel's Approximation Method (vara) 4 rr )RtH-Wesz Cogne ce (NWC Wow hing Rules Step lt Tduntty xhe Nerth- hist cones of te table will be the col (69). (tin tre pst sep , the Ninic the Moralt xy = mm (4,561) (ose 12 Dy Oy Sb)» thew past sent qh completa . coeds Sp b, a, , thew firse Column gels completed. Cove 32 TE A, = br > Hom Hue he a tie ond a and Shep d: hank fom the seus santh weal Cherie obi. Poses ee! polubton fer the fallowing- bramepordelion spool bap Nwe methed. | Di Ds Ds Dy Supply 6, {14} a0[s0] 0} 4 02} ay] apt : {22} 29h go] 4 Cs [40] & . _ Demand 5 8 7 €: @ Cotte ane wriltin ab the botlonn lf comer @ Mlecotions orc walltin of the conlke of thy box Supply 2 144416 = 34 Demand = Ste +1414 = 34Dies Sup i o sht? 422 a i Axo i eb 6 3 le Y xo o: i 4 iz ste |, i. 4 to !4 Iyer o Demand gg” 7 2 % a a Stepls NING B® CHI). Xe emiw (1,6) = 5+ Cy ~tompletéx Shep&: NIC ke (2) 6 y= malin (2,8) = 2. Re ccomnplelis Sep 31 NINC Le (88) 6 hay e lw (106) = 6 - Co ~ Comnplelis Step 4 Nile 28 (8,3) 2 yg = we (3:1) = 3. Recompletis Steps: NWIC BB). tase MGW) = Gh - Co-completer Step 6: Mlocote 14 + tee cell (3,4). fue note Bak fated sablocafiors =6 = 344 —} f. The tolution bk parte fearthle . i re xia) (2x20) + (6x30) + (3x49 + 4X10) + (la x29 = Ms 1015 @ Finck am Ubtial soliton by NUIC mtbied Di Pa Supply To CC os lede os al rae 40 Os, EF TET! Go Demand 65 42 43 hee Supply = Jor3ot50 2160 Durand = 65442443 = 150 Ds supply = demond , we TP Ls bolemend Eqvzue.Di De_—ODs Supply ~ | 0 65], os |, : gt 07 er : o |, | 3» t hos i Of fat? a] Se a, Demand . cs Step | NIG 38 (by). sy emetinl 70,65) 26S » G Compl Step 2> NIC Be (Hd). hap een (5) G2) HS. Re commplebis step3s NING ke (BD. Rrpemiw (20, a1) = Bo. Re ~covnplels : Step es NINE as (3.2) tay 2 rive (150) =F. Cy complet Shep 5: Miocate 42 40 cell (3,3) Total ws, of ollecedions =F = 3ta-) pe tae solution Le tuifial baste feasible Gort = (65x5) + (ox) + (30x0 POixt + GEXT = 810 _ gun Jabltal foledton Ing Nie matdhurd o [elala er) 2? f % | 3 Lr r ol siala by 4 i Creal | lay 1h ; Demo 7 4 18 Setutron Supply. 2 S+e+1tl = 3G Doman = THI t1E = 34 > tr be blond - Bee. Auapply = dunn BUY TUPianeteB Sey Or i 6 te wo OF eee ese | «fo 0 | 3 | 4 Tee % pd ol dye CDevvased. us = ° Big #, ° ’o Step | nie 6 (1) my een (87) 2S Rives Sepa NINE he (GND. Ky een (B,2) = 2. Cy ~tomeppletis Sep: Ke fa (20). Xyremrn (64) = b. Rpm complebis Stepgt Nae 4 (20). Kyewin (1, = 2. Crrvomplelia seeps: Mae As (3,5). Ksyz mce (4,18) & « Rar eonapltis ie te coll (4,9 Total ws of allocatans b= 4+3-] ce ghe solution Lthtal beste feastlle Cook 2 EXD) + (2K3) +(x) +(3x 4) + (EXD) + 4K) Srp b> Alves = [or ® fard aw Anitlal dst fensthle, tobsction by Nie DD Ps De Supply. S&lelali{s} Sle | al2iq) ¢ *s | 47/3) 6 42 3 Dword § to 1 Ge Sotation © g elke be +3 = 23 Duronoynal = Etloristh = Be he snepply + domond =9 TP a unbalanced Danny nos = 35-23 12 yaara De Supply Or fp 8 iy * re 7 oe Bie Os |. __|h - we eee HE eo Denno Wi 15 Ho Step ls we (U1). ry erm (thy 6) = 6, Crncomepleis : Step ds Nae % (U2) ayy aml (6510) = B. By -tomplolie z e Step 3) NC (2d)- Kare Wem (62) = 2 Ry -completie i Es (QD) + Kgp = hw GI) =h. 0 completa Sep s: NC Le (3,3). 5p TP & balanced. VTUPianetD De Supply 2000 * bie aoc 0s |, 1000 : Ho G000 Foe, 7 +500] to0o~ on, Demand. 200-3000 4090- sa00 Step Ts NWC te (11). xu = suelo (2000, 2008) 2 2000+ Ry ~comepletis Hepds NAC is (81). x44 = hebn (Gore, 1000) = 000. Cy rcomapletis Seep Be NING 85 (R12) Ky 2 Mur (5000, 3008) = 3000» Cy completes Step 4: Nale £4 (2,3. Xyq = beta (2000, woos) = 2000, Ry ~conaplelis NIC 2 (39) + Hoy <7 (T000, 2008) = 2000. Cyr complies Sheps Mlocokt S000 t (3/4) frat we of ollecations = 6 = ate! tock = (\3x2000) + (141006) H(i4xB008) + (i2x2009) + (15 x2008) + (lax 50 = bA%000 Steps! © Find am anibal goludtin of Har following, by Nie Di Pe 2, d. Stipply, i ate le [os Cae : | te . Oey aleve etrouleremal: ro : O | ig | a to [ig 30 i Laie | 3 fe | go | Demand 25 35 195 20 Sobrction. Supply = 50¢10+30 +50 = 200 5 teo = 18s Demand = ast aerlo hk apply pdemand, =y TPE unbalanced Woe Subredlce 0 olermung, column nibh demand = 15 © fp VTUPlanetSupply | : p BO 356 i de ere i 30 6 SO we — fe (iN. Xy em (50,28) = 25. C, ~ commplelis = Steps: Moiale 15 to Xu - total wg of allecations as Ne te (H2) . tien (25,35) = 15. Rytompletes Nie U (22) Term (10,10) = 10. Co-tampletia Stepps NWE i (253). t= (60,105) =60. Ry -tompletis Sep Si Nw ts (3,3). Lag = mer (30,45) = 30. Ry-completis Steps NAC & (4,3. Xyg =m ( 50,15) = 15 C5 ~tornplalis Sep RWC & Choe) = Raye MY (S80) = 20, Cy coneplele —_——— ate-l lok = Qsxy) + (osxe) + (itxte) + (eoxte) +0oxse) + (raxts)+(2oxe) +(15x0) * Leg @ Faad on initial tolution of be golinsing ag NC Diarra eee ela fiels le i} ve | ig fio [8 | iz} ite Oye |r lal} Semand 90 geo 300 300 600 rs Supyly 400 500 Joo Supply =(00 + 500 +100 =[400 Demand = 500 + 00 + 3004300 +600 = Loo _ a eee eof 4g i | i := 200 -1600 > $00 Dwrmy Sone 5 = De re a . Ds Supply. we? dig yt | ee os fig [poor UL gee Oo fo (|p 300 | [30 ca 109 p0%ic0-, mg |p i [ mo, Demand Se €00 302 0 e050 uch roe NWO te (Li). xu = min Goo, 500) = 400 » Ry ~ compli NWO £6 (oy) ty =m (S00, 100) = 100. Cy ~complefis Nwe 26 (2,2). ore ™ (ty00, 00) = 400, Ry ~Completis step NIC te C3s2) + Xp = mon (T00,0) = 0. Cy completes Step NWO Bs (3,3) + gs = me (100, 300) = 300, Cy-complatis Sepa NINE ke (3,4). Kay =r (00 , 200) = 300» Cy —coraplelis Nwe te (35). Ras = MY (100,600) 2100 . Ry comp labia Step | Stepa: Step dé Shep 1: 3: Mlocat’ S00 t %e- Petol we a allocates =8 = G+5-] ce The salution tc curilkiot bane feasible. Cok = hoo Xe) + (too x18) + (400 x10) + BoonY) + (200x15)-4f00x1 2) = )3, 100 aLeast (osr Ewrey Merton (Maraix Miran Merton) Working Rube: Ship 1: Datounine the smallask covt in fable. det UE be cy. Allocate xi = i ate ok) Steph: Oy Ny = Oe, thon oveu-out 1% wow. Goto stp 3. @ 4 ny = bj » thew ener ort J column. gre Hep 3 () Dp ti aie by, Hum crook 1 yon or jt column, but wet beth Step3 : Raped a to ee ug Whur minkmum crt ts net wiigue, make ow aabibiary. chetee omneng, thy whntuna PROBLEMS O Seber bry Mabute - Mawma, Method DP Da Dy Supply & | 14 | 30] 9} wf 7 S [10] 30} go} bo} 4 Cs | 40] 8 | tol 20] 12 Pemowd F 8 7 1h Solution Supply = 749418 > By Demand = 5441414 = 34 = TP ls balanced Pp ty By Sp Or | ph go re 2 ti Ty Or |ap 3b a w-_| 47° oo Le i | +7 Iie fe Dowand 5 fo | i i LStep le: Janis cot Js 8. Agpz min (18,8) =8 - Cy ~ Comnplelis Stipa: dansk cost B10. ay = ran (44,14) = 7 8, completes Sep 3? dansk cost us 20. Xgy 2 nC 10,7) 27 Corvenng lelis Step 4s Jak cok 40. Thine Be a din betiwww (4,3) & (3.1). We whol seleet (2,3) a ESM - comple « StepS> kanak torte 40 aD hay = pub (3,5) = 3, Ry~ Completa Stepe : Mlocake t wolk (29 = Total ve of pllecations = 3+k -1 = b, Sethe solution i Enllial baste feasible. beak 2 Ax10) + (2x10) + (1848) + (sxeo + (exe) + (1020) = ole cement @ Pind om daittal tolubton for the follewlnge TP by matbih-mbnna Di De Ps De Sapplap a i | 4] 40 fs Ts pel 50 | ol¢le [5s] 4] 2» i emanh 20 go 30 | ee Supply = 30490 420 = (00 | Demand = 20-+40 +30 +10 = 100 } => Tp % bolonced Pa Ps Dy Supyly ; _ WN TUPlanet |Step bi Laask cok 1, We droll geleer CL). + gy ~completis thy = ww (30,20) = 20, heost Cott te |, Ayg Zratin (10, 30) 210» Ry -complelis Laas’ cote BB Le yy = tn (57110) = 10. Cyrcompletis 1 done Cork & 2, Xy3 + mh (20/0) = 20. Ca vompltia ep 51 Raask cok LQ. Kgp erate (2E0) = 20. Re rcompllis Step b+ Mlocaté 20 te (2,2). Total ne of ablocabions = 6 = ara -l tae eoliction ke Anitel borte feovtite Corte =(Q0K1) (OXI E20x) + C20xe) + (tort) +Goxe) = {80 Q@ Vstngy moatbue-mbndma rattid, find oun Labial roludion of TP Dib De Supply {Sees Ct [6 | 7 50 Deming 65 Gr 43 i } Supply = 10430450 =!150 Dumoard = 65442 +43 = 150 =e TP 6 balanced De i ' i I iSay Vt leat test Ae kyy mmm (30, 42) 230, Ry completes Step ds heme co Be Ry Hw ( 10,85) = 6S « Cy ~completis Step ds Kame cok fe Te Kye mtw(5,12) = 56 Aco completes step 4 dane vere AT, Rap = en (H,1) > 7 Cy -conaplatiz sey o * Mlocot? €3 +v (3,3) Total we of allocations =F = B+8—] The noltion he PBltol bate foriele Coe = G5x5) + (6x1) + (20x4) +C1x7) +(48x1) = i i Bo Solve the following TP by, matrix rattan wretied. Dy De P32 Supply =p Tp i balanced. Sepplep F xo wo 7 Ao Oy .7 D —— =p") | Were No Ty. lew Step 12 byork bok Le I gy = Bate (ls 1) = Te Cy-Complain Geep Ls faact cost BL. age mabe (8,18) = @. Ky compel seep 3° Least cok & Q. Xy32 hur (0,1) =7+ Ru completes Stip |? henek cost fh he Hag erm (1/9) = 1, Rabo Pe LOPlanet eeStep 6: bene we Mlowte 2 to (1,2). tl ve of allocable 6 =ut3-l oe solution's Suitiol baste feastble desk = (2x1) (3X4) + COX) HET a) HORN + (182) &VTUPlanet _ Jone Seay DestinationSovvtion By Monirien ‘Drstrieowen Mertwop Ortimar Working Rules Steph: Find an Inthiol baate frowtble solution (prefrably wing WAM Ledintqus). Agrigh vartables U1 stg stty---- peut » VieVa Vg eee columns Blep2: Conner the nos o& the Column wth maxtmium auambor of allocations. Th there be ote, break ib aabitinatly. A uw ved. Paad te olher Voluss ou andy balag, te donate [EGET NoTe: uz and ye Mola ane cabewtated Reh the help f He Step 3: Fink uit yy fa the cells pettrerly allocobtons t (cthese orden are wattben ak the bolton, Right corny): L Step 43 Find dg = i — (uty) fur the ells Lure there one | we allocations . (These diy valu one wrltten Ob the i Hop Right commun). F 2 o belin les Step Sz Ie applin the ‘ellentmge Peinolttg e (0) ay ost dj 70, then de cusnent Goludion ke te u colition Way stleat ene diy ot, than he colatton fe moth splinal, To find tee optimal solabion, gx 6 sips. Step 6: Seleck the wht chit the Least magaleire dij. (Te i token anbitiarily ). ubltecaté an Winkie 4 ee, ang. | 2 & shh elt. Constinsk 0 Loop that starlé : end, At he ell. Daaw she Loop fw ducky Pee le Comers: othe lap to te els Ragollocatfors ( A can have @ maximum ef & allocated call bw 0 meus ov column). The omeunh of 6 te added ond tnbtracted from the i corner ln of. the Lovp im Goth prong thal He demand | and supply, values Auman | Step Bt Artuan to step > and repeat he Procedune until optimal golidion tu obtained be all die ane man - née 7“pel |: aes em cell ee | el BREED. ote rand #70 Shao Woe 4 [aE Beate Feactble colubion =Tolal cock =% = = (19x5) + (30%2) + C30xb) + (40x8)+ (Tox) + (20x14)por —— i (Rind the BF colubien for trey abou probln sing Reo. minimiin method bl te, | BP Colubon = Tolal cost» 2Cax7)+(20K8)4 C408!) + Ca0x6) + C10%¢) +€20X7) nt o. BF &) Z=1O : \ : { Procedure + : (> Search ofr the mfnimum cock fin the es Pal nec whee ancl attomte to them| accor a i Aepencling _ and demand mertpened : & i etter of thy ov clemand i | | (> ty, q | : cdo gA the nest minimum eee | Oy ratte and continue tre Attocahon prog a Linki att supptios ond clemande ane Feros | ke Oo: Cie, tf the Auppy “| the poorke 88 becomes zero, HH & as \ee clemenaling thot vow. Lemé Uae demanal of pastteutins beemec reve ent ie as Lee alerniaabing Haak prsteouton tabhre). i] i Ti, Mettop: CoLuMN MINIMDM _ meTHob: “Final We BF Spleibion for the fottorsengi abe problem weeeng ee mentmuns:: metheel - I : / i tw ¢. BF éclubien =Telal eset = Z=Ciaxe)+(exe)+(4ox7) + | Ctax2)+ 6ox2.) + C20x10) t | ', 8 (Z= tH y Proceduss Licelumnetee andaltecale |: () Select the minimum test, bo trem. accorcing hy Aapercting en the of hh % Aemand preeteledd in Pre problem @ dt te “op hy [aseriand becomes xean offer Akocahen , “a ag as elernenaling that eteeetan Bove feclurms veepeckrocl « @® Atlocabion & made unkl att cupplee [Aemancle | become: LerneTe MetHop: MATRIX MINIMUM METHOD: ee oe ae 7.) bo ¥ ' “BP eolubion =Tobal cock = 2? Ciox#)+ (tox2) tLH0%7) + 'Cyox3) + (axe )+ O0xT)4 oe |Z. ec Sl4 ~~ Rrecedueret \\ (© earch Jor the sobabe mabrse wtlh the east mentmiim cock ang attecate PE | Aspendeng onthe cemanel and OY OT wes Ac anc rohan the bupply [Aare i becomes zero after ‘ ‘cancel Oe Lecular ‘row /eobemn veepechsely thus no move a erticcabone cchoute be made te these cancetled sows eoburmls. (3) Continue the sume above 2 sleps unl the demanel /eobpeg becernes zeros.I j ¥. merci! Moga's APPROXIMATION Merton: tor the’ obewe 5 [i Ponty | fel [A [ae | wa. > co]Peay | 1 Z= Cexé) + (14xs) + C20xI6) W Heaben: 3 Mffen aieinaking 2” 7 2 | ee [ey | W Therahen: cebinien “z= Coxa) + Clans) + C20%Ws) + C10X2) ew ho q Ee al |, BF Soluben = Total eost-ZX oo Ze Cane) + Cian) + Or (nad honda (bona) | | | | (® Prope 2 ‘oe ia BF Sole fe, Slep | 4 check Supply > Denraned _ Alene ery sDimand +410 2. BF eelybien = Ke (3x3) +Crx2)+ (ani )+(3xi)+(an1) 5x2) & [z=48]et tener Ze (22) + C2K0) i fy VTUPlanet,_Baste = Nartoble = need aproate emane|Asstanmenk robleme: t eee j —=| (@ PROBLEM 3: ‘4 Eelne by Ve sa method te find the i Bouts 4 Roath etl Beluhen focthe betas pte - att3h Or ls fsZe Cae axa + | Cann) xC3x3) “E Heohins andar the column DI emeacaelac _ raat he bh Fl +. =) a LF it , te a Ze ‘Cixey+ C442) + (AXA) + (air “al W theratin eluntnabing -Hu|Row ‘Sirus ia ge pao [ea Pe | ide! | i oe ph 21, [eer - Zl t rot) jaemand | Ajo |The | a savesemfene es T oo * LARD) + (44)- eee “Faia cennntonesenna i 1 i nOXI ~ 1 j j | - : 1 “_Metob: NertH voksr CORNER METHOD: oo BF ¢ jn oe LaR8) ¥C1im1) + CHRD IK) + (2K2) + (8x2)+ CLexe) glep i Sup py -Dernanel HSa——eer EL EeEEEEEEEESEEEIE OD pean TL Texeher oe ceo os [epi [Pe 3 Is 16 | Wt pt OE Co ail pf |. pp fie et [ee Or i Ze Caxie)+ Gxt); Spa’ a Henhts A r z | a2 demand Why \3 | | Ot = | we CARIE)E CRACKS)| ‘ TD Hewdsog: After cleminabing (the row “62/ hoe gee bi [oe RY peta [ele Ae | . ws | - Son = CARIR) + Cant) Fim) -+(2xa)+ (sxe) | : | 62° td Peete ot |i | *h, | BP dolubton= Z~ CaxIH)+ C(lex3)t (XS) + xD+C4 X12)«Ruesei's Approximation MeTtHop + PRopiem | Daman 89% I Ve Va Vo Ys | = Colummnustee max) hate 3s 14 2023 M ee 3 Ta ta 20 23 ~ = : Rann BDREONR t | ': Au = Cy -T-V = 16-22-M = ~5-m Aln> Co- -W > Ie- 22-11 25 L A> Cat NG 713-22 Me x@ Mia = 4B ~ = 22-22-93>-23 ‘ i Ais> Gs ti -Ve> I-22 M>—-5°M i xo | Ao) = Coi- TV = Wy - 1AM = | Ao? Co2-- Oa Vp = 14-14 14 oa C2 - Ua Va=13-19-m Avy > Con - Ua Va = 1-19-23 > 23 xo Aos> Gis- tia- Ve = 15-14-M> pM. dat = Ca)- Us- We am = a Aeo= C32- Us-Wo>14-M- 19 = ss >Cag- Us, Ng-20-M-M =202m Aa > Cay - U3 - Nae 93-M~93 = ain} Ae6> Cag - Ua-Newm- M- Mam | Ray = Ciy Ug -W - M-MemM eM sa ys. Cia -Ta~Ve~|O- M142 M49 4 Aus > CHa -U-Va 2]M.-M-M = -m : Ant = Cry “UVa | O°- M-93 =-20-M ne Nas» Ces Ta - Ve < O- M- Ms ome so7 New clentvate: We vor & afbe | a i to Aas, Since the athe hae besernes Kero - \ ForetThexehion 2+ Ay = 16722~14 =-35 ° Alo = 16-22-19 = 25 | Bia =13-22~'20 = 24 : 4 Ay = 22~22°'2. Ais = 1-4 -22 SF Any =p -}4~ 1 = — 24 Asa = 14 “I4- 19 = -a4y Ale Ass = 3-19-20. -26 © Aon =l4-1q-93--23 fos = 15-192 M > ~My ; a -23 6M Na) = 14-M- ta on Ao2=14-M~ 19 =—M : Ass = 20 -~M-20 = = | As =23-M 93 6—M Nos 2M “MM eM Now eleminats Hae sobuman D5) aftattofabing Dheralfon 3+ By = 16-22-17 = 25 A= e-2271T = 2S Re oa ae one Aty = 22-22-23 * 22 Aay> ty IV-D 24 | oo = eo 14 “19 = 24h } op 8719-20-26 Day 19 “14-837 2 Ags 14723714 2 929 Boe 4-22-19 = 23 Los = 29-23-29 7-23 Day -23-23- 22> 23News clumriale the vou’ Si5 Myer to Aj as the “enpy betermes Tercker 4: Day = t4 14-19 = 24 Rag < 14-14-14 = 3 ' ugg 2 13> 11-20 Pe Koy - 11-41-23 > -23 Subpuy sho allocation Ag #I4-23-M1 2-23 Aap - (9-23-19 2-28 Asa ~ jp - tan 924 q ay = 14 -19-23-}23 i Az)-}q-23-4 i Aso= iq—23-1 : Dons 83-23- 93>'-23 : News climtate the vow ‘so'after altorale, Supply > Demansl>204 Séree here both Demanel qd : and buemee xno affi attocahen rettheed || ew ‘62! & celum DI’ can be elemtnated. i i 2s BE Scbuibsn = Z= Loxev)+ Utxie)+ Liaxto)+(03x30)4 i i C1420) + C1RKOD+( 14x20) C23K20) : i 2540@ Ana the Entbta) Baste frocvble solutaon Copimeat Belubindfer the protien Vogel's imaben ‘wl, Lon ae [or _rretio| 7 [-— as Fy 14 3° so ro oo Po 1° 3o HO 60 Polubén: Vogts ApPRoximA~riIoN METHOD Bberddesh ts fuoa +| Ze Cexe)tl C1axs) Teoohong: Mts eliminating the ectumn 'vot! | cox Ze Cexe)+€ taxs)+Coprio) \- tlae nasa oa ea "PB awe get wo _ dy |” Sapp Peratty | | oi ' i op : | Ft i | BO , [thle doh | -—| “ : Sempron i a 20! l eee por | | ere : n= eee ; X= (axa)t ‘CiansD+ (20x10) + (rox2) chops Theahin 5: aanrats the vow Ei rsae pets O35 Sp pe (re [wel feel, | 9/2fo | [oer Yo "ear, | Ze (Oxe)+ ClAKE)H C20 mI0) + |CtORd) + LAO T) HEN) zferr] | SteprmTo creck che ate Vase 4Z te opkimal | Aoluber or 6 t Nen - pouabe =metnel iy : “Tt ig+H-! | \ etl +6 altocaben: rat tence 6 aticoalions a+ made. Hence ™o ty, He altocahin mate Powid nt fenpea Ve=20 Ve =o ShepBs Conatober: Cty = ULtNy ' ey a x2) 5 ( | Uy t¥4 z!0 } UrtV32HO if Ust Vy 80 Ug4+ Ve ee UgtVg 220 tm CAE Vig columm hae the i New make Men OC Va JPPatioeahens mods) max ne dee for ott the locations tr Lhe abe wile bitocatfon te net made - it spay ee i Ayn =Ca- Lt, -Vo.= 80-10 -CA2)= 22.. F aya: Cia U7 Ye D-H +20 = bo ‘dope Gy -Us-¥) > To-Go -4= 4 daa» Cro- Us-Va= 30 -Go-12=-18 Bares da) = Ce —-Ue-Vi ce bO-on Gert i Idsa- lo-20+20 2 fo \ do Abes! Er clay Constder tre nur bE gue The ENTERING BAgiC VARIABLE- i [aoa = td = esteaeey 7 Slee: de fine leaceng Baste Vartalll. Ne | | [Pj Z sca aa, ey 108% Aubbrack 2 tr the Loop + tdrtcrever Wer value boiernres’ Zex0, then a ===-«- Baste Varteble | _* ina tdo find Uy, Wo. Ua, V),Ve,Va Valu, Ae fe Ue; = , , Ege : Unt Vn* ca (Tha th the meus exquac! - Ut Va PHO i: a UgtV2>2 ' UstV4 =20 New, futl V4.0 ! " Uysto : | wird [Vd otetle Ds Aik. = B0-10 +12. = 22 r did= 50-1042 = 42 : 4e)-Jo— 42-4214 | doy. 60 -— 4Q-O = 18 i : da)=#0-2042 = BQ 2 > Aa3-10 -20 42-52 Kine, att aj fotki the final eolubrir tan be ctbared i Za(are)+ tions) + Gon) + (40K 1 Bx6)+ (20x12) | eof X= THS & Bre optimal £0 Bolaben.@ PROBLEM 2: “find the Cntlinl Baste teagtldd Soluben the forewing Novth- heck Corner - nnn ' | fs ys i [tha Droad ab | Vay |"Aa po | See . a! z: fone} Cena) (ax) + (4x3) + C7 xt CORI) 2. Baste feaatthe selation e102] iSep: de find optimal getubion \ Nen- a. emtned ~4+3-& =6 Caotesfetel) Conetalen Cpe Ue” Ly +¥; =2. Ustnvy = 3 ' sea 23 7 Ug+No 2H i ; UatVs9F! . Ue Va = 2 * Baste Yartetibes are | Ly, ls, Xo2 +%ax, Maa Hs Emr bacte Comeciles ard azume U2O - oMeS uy=-l V272 i gel a | Nab Ug ett a aig > 9- (Des 5 diss h-CN-6 = 1 dos+ 1-0-6 2 *G s Eatening Be Oi ene | fy VTUPlangt eta jape(-4)-2 =a 4dag> @nterng Baste Vastatle. Plea minimum verues |O 2. daz = lecerng Prete Variable. Ze (2s) + (2x2) + C3x2) sttsibe Coxe caxe) ww [eS] oe IE ctlerabion: lo fa vagal Ug Has Vis Van's vatleas- Up? 2 Usty = 2 ust V2 UstV3 2! Ust¥s 2 i Ug t Vo 22 Pub Un,‘ rue } ay+2-¢-1)-3 =o: oyzey-CI-1 HH | i | day >5-C1)-3- Be 2 > Enkering Baste Variath. -I-|-0 ' i 22 Ay) ~!entereng baste: warkelleso Mthtmwhn yelue=[6=B ] ?. lays Lemania beee vestable . Ze Cres)t C2x2) + Che) 4x7) + Crea) + C212) a [A=%6 | Nove aque cn Uy, Ua,Ua, Ug .Vj,Vs,V3 Verlies Up =o UstVs23 U24+Vs =} t { H34V2 2H ! | ag ayia ! 0 Ug t¥g eh t i Fob Us\20 t Vae3 Viel Use! Ly U4el Vj=0 bo Ajo-1-21)- (BS Aig = C1) CID = Aatj=37(0)- (3) 20 Ag] =S-(1)=C3) = leg. J C1)- 01) 5 lye 6 -CI)—-C3) 22 ence am. ole ane pesthie, the Selabin | Cbberined tow i opbimnes VTUPlangt ey(2) PREBIEM 3: (ASSIGNMENT) [__for fou Toe” eTlof af, 8 EDA Tae dt 1 + pode a] is} el] i 7 oor ind | 4/. 34 | | | | 4 a it | Le CA xe) + (242) + (1aK1) &i i —_L__ % 2C4xd) + (21x24 Clay1)+ C2axs) Temben 8: Etuninal Ye vow ‘Ss’ - a [Peni oe | ee (# \* Ze C4 ¥¢) + ‘(a92) + [12n1) + (23x85) + C13, Sy 1axe C29x2) + (lant) + (23x85) + 1axy) +(8x2) + CIbXL + [aii] Sate: Ae fot penal colubion CEEXE) +( SvxO) ae a OT eg | = t ’ Consoles: Ces uttvr i UpraVy =F i UstVg +8 IL Unt M5 +13 Cstv, 296 U3 4Ny obs 43t\q 50fechie. Arye Cop BV ay = 8-642) - 5D= das = 13~(C0- so! 5) = We-C0)- 462 O dap = 65- (0 -65°O days 5D- 0-50 = =0 Alyy 2 29-G38)-67 «0 dys? !2- C-38)-2-O eg 283 + Cin 65-0, Ke ltox | A> A- ES8)-6F'=O. ° Aw aye ae i : fF] VTUPlaneUj “58 Age Hd ly20 G>-3¢ Iga-42 Siepl “eget a Aemand eer 3 ele 5 2 80 &o m Len ee ethed CCelubén): = 1 Sppy | Patty 2 +t 2 60 4 jo * 1% Ge i oy Be 5 + 2 | <=Gxe) ¥ z Tt 3 2 9 “eo 7 e/o) 8% ao py Oe Ze (oxe8)+ (3x68) cephiy | Pnauty 60 a I i Dd a ecm arn ' i_ de AA =f A 3@ | i fol wel B 1@ | 0 i i ce SO | % abemand | hol | ; 4 pF ce Z= Caxed)+ C3xec) + (axéc)4 Cax2e) + (Sxe) i fo|K= 450 i Slepa: So foe! | Nien -clegenestite eimtoa-l ) a = BtSe! 2 : i : > Ne of ee | Be (oa, Sor Xa) Hes) Baste ‘ varcable fee Bence [Mace < Newetgenssete on Bose verre 4Hwo (D chee! A, eusch a tay thal TE choutd net fom acleced loop as rs) chewed! A, Much thak Bre evel ob Uieele paatecutar locahon 4 WatVo> 3 Ugt Ve25 dif - 8-3-4a=1) dig = 13-2) °6 | doo. -@-9-C-2) -2 | ag, = \W-5-C-€) = 12,PROBLEM 2: C2 HON 2 vB tat 4S AO 3 t 6 F2 5 io 3 4 & | 80 F3 3 Ss é 3) 2 a0 pa f. » | Ga mf Ae 4» OF 2@ | 10/s/o _|2 Gs t ! ~ po al 204 hoy, Top fy Sh ee A eed : J Crx3d)+ (10%2)+ (2x20) + Lions) + (6x5) 7 « CtSxs)+( sant (385) « c ° : 4 ¢ teal 2 i ; al wa, SEbR ' al Nen- ahequrenale F ment a, = 5thof 1 db a i i Prete Variable = Non -ctegenesatexe + Gonacoley UjeMyot Put! ctp=O Vee te tutV2 23 aes LatV222 uae-l bo ta =2 aod = Ug tNg 26 U3 22 Se lst Vg 23 — V4 . A U4 + Vg = 5 Ug 24 ate 3 1 Need G7g EN. dine} Om an Aig = 2-0-4, | As = 6-0-CI=1 ~ ON) Cy = 2 Aay= y- CNP ey i das=5-(4)-G = ly Alsi = 3-2-Chy --3) Clay Entering Rage. nastatele (A mi) Anas HOW) = ly Tp ford vatus of & VTUPlanft! ‘gntn = (0-30, 8-25, 8-5, 8-5) Ext ar +Flere conetcler dys 08 only the Lemuorg I Baste Vartable, rether tran alzy: because i cok of ag UatVe 23 , Ugeyy, 22 | Ug tVer3 |a % ASSIGNMENT PROBLEM CHONGARIAN M@THOD)? UW) PRopuem ©: i A a e ; [vo teo Be. | 2 & 10 Woe i (20 3g {to ko | Slept Pact the least element tr Seite lll ubbrow th wetth at thre elements ef Ehat pertteular Kaw . Matrix has bb be 2 square iA 6 ie rmabix to Acs ent 4 : Rreblem 1 | &e ao 7 0) 2,e Jo, 30 | 2 3.0. B40 | a Stepa: Hera the beak element tr each colummand Aublracktl yetth an the ~o af that facteceslas columy + A B a 1) ao 1; Lol i ' ' at 2] Tl. 30 = 3 {J] 20 ep 3: Cheek mevistecfeolumnietee th har ene Hero or nob. tf Gewfrolunm hat one Kei, mark the cuctgnment | Er colar [oo cteps: check [m= N a i hee none of anttgnmente a : Nene ef oor and cancet the ethes Kearse i 7 hevom abew - weep tel ae theo conr| _ By 1UP laneand plo | @ 3 a —rts—SS R i2s 4 3 6 14 le 1h. O Step 2 Eoturmistze + LA Be p Riel 14 4, 3 el@ 20 Je] 2 elias fol 3 sla bo we He. N=? Conclusion & ofamal Je Aastgnmentone: & [=e] POA aoc ROB oe gotubon: X= 84h HIT#IO yr'2) PROBLEM 2: i i i i toy 105 i at —£— fi {, \ glo o 30 30 46 rleao o 36 aS Ho | Slep2: Cekumriodeg, i Pp |20 35 @ Is @ —-@ 15 — 0 eh { , | ee eo) 80 | Os fol 20 oe. Be R |20 S35 ex ‘So as 1B le Sore Hove nN: Slepa: cheese the mintmiim aiumber tribe Matix ishee the Line clees not feccee. tex &. \min = 1SCH) PRoeu: Pla 22 ss Nf 94 alHs Ts 12s 63 48 i. | rela 28 4 a7 4S 33 : ‘ foe tepe roe ee] 49 21.2 "4 5 +t fae on 51 22 a5 18 : ‘ vfs am 52 3i 7 2 °F Skepl: “Resrze: . [A 8 ¢ op el Fk ' plo 13 #4 tor ale chef 4 2 4 3 Boe a5 4 2a ~ ct 4 52 Herve nN=N oy ophmat golubon Zo M4 HE+284274+2543 oof Zah(©) Prope: Colubén: : ele ee ee > there are S conbackere Shep: and 4. Reade. JsAlence clummy Rec for each tordoacber hae bo be addled tp make malex a Savark MATRIK: | YR go Re Re URS i ala" 4 14 Is. ° co] 7 rt \ ec] 4 is a i 0 ‘ cy | 1° 12 Ie 14 i cs | te Is ts) te Step 2: Cobinmivize oe ‘we RI Fe a ST g 0 aUNI]- ite : GAME CFHEORY AND ot SS ~~ Becision ANALYSIS eo - 4 B & weinner ° 0 ~ tain Game-} «DOMINANCE STRATEGY! » Progiem fe Player 8 | cay its) r| 3 Ss 4 T)4 c 3 IF Plager. * Lr j ple. 2 @ | Aclucier Since A is wena? ie \ O Plage A —obice fo wud imum | Player —> biker be fox Pini fae! because Raups A co renner (2 go mahi) me Slept: ate checking courvorien oe chedang eat erhane oi minimum Sena B72, ers, ye Vole’ ain the mastmum ain Hoe! BLS, Ary 476,973 B74» #72, 778 meena Lt, ans 7 Roya &é ia Play & oo @ 1X8, IK9 oh- i (D PRoBiem 2° ee Rea|eoee ot ie : WAV TUPlanbt% SADPLE PoInT MBTHop: | Plades B 1 a_b a 4 1]-3 2 6 ; { Ce 2\-s Places B 3 tlmxlo Ss 8 : ; ! Ve min mass 6 \ i i : woNo addte pork ! f tenes the vatue ef game tenet prechecbaul. HodadbtePRopiem 3: ote Hie ~ | PNapoos. ne ae en ha ae eee Ve¥ ot C gacdde peer -Gxfisen tin ne | | ) PRoeiemi: pe s Pale | 4 -2 h , \ ‘pew pay-off mabix “ cheatly Dl Solutasie | Conetzhone fr Asberminatle + Va 7 Ve = Nalue ~~Vemox. ms =e“2 | 2, Plaujen: |B. te bechnen « ged on sy Set Te sas i acm: uel sme — fo nly 220.mabo) | pProeiem 1° & ‘ Mate Pty Ate lin ir} -2 4734 a ! wg | Verret min = ababate eet W omtn maxed 7 Since WREV~ | com ebrake = ste Pl = trates { cheeses io a Shen let the pal et chectee 2 I fa > Te it ieee, ci-p> (3) gent page Bd [od ON, tt = '-3p4 4CI-p>+0 pea> POP [ I casap = -PPTIP | Mp3 = -hptl E P eek cml [P=], Jp Ee > Is i | f ager A chooses fy Brae of iD peo tes CTS times! i chrategy T- Va ames Sp od pew i hue. vs eg + (9x (4) if . afte 22 i Step «flee pers He | | ee pb the “pers | hat ie cheoces ' Be ab fees peeked four of P es reese) t Expesbed feuy ie | = perp exp expsted boy He qf By renee +4) yr 34347 -katl fa + rape)sYatus of rie ° Beep Coneluder: te Cts 5283 > Cis Yo) | L CYpoysd = CVs, Wis) i : 4 PloyecA ! 2 4 ~ | dee) Demtrance 6 Step 2° = 2, tote, 122-& i "ence to rh)? ek. 7 : i » efiminate “IT evo” i u sl urn Oy, 12710, ebirinale 5 col B>-4 10 S10 ,ekminate t* co cobungry VTUPIa B>6,HY2, ebinnak 374 column, $3 heel& Phy 8 ep &; ipa! —app + Ci-pt0 Sp 2-3p + mp sto-top — 2 pt 2m = Nye 410 gp oatow Ae ' I ey+ Eeplt-qd = Pat tocr-q) | i Sq - bt = 24 tlEtoq, | log- = -8qt10 Y teq = ly rad | 4) = % ' ™ oo Vetie ef ame = 64+ Chr) (rq) - out © (4) + alg : ' TF 3 ee eae ete 5] © Cemclusteny : oe (hy %r%8) = £07 %4,%) a ‘Yatfoo Yar JOD > (0, %. 01% oDEtininate tuis’2m ores Sex’ te! column - i anes by Bat 2bhso, “yore ee ee Player B OE ; 2 Player A , 4 © 4 | o 8 a ee waa Bowe 25 io, WOK | edecninale: I ve $ oe Bor HlOB Player'B [Row min ° 4 VoD a 4 2 (Ve max min 20. Ec ae up| © @ {@ | toe i Vie min maxes | | kines VEV t LO fer : | = page lap + Q-p) Co) ——— © ' of apc) @ Pe O«®: up 1p = opt OPO | Oe =@ a ' s \ | famieoa ae polo) a ame apt Of atne of ( Alp) Layer oe) Yame “5/, Der Hegee creeped oD ReENTTEITON Og, + Os & Oarnt@: 44+ tq) = = eres | Aq = 4 -8y \94, * aa |. Vatiee tee mr Cinq lo) « oefe Cenclueven > (ot; Xa, 28%) = (0.0, %, 7s) | Coes Yo yo f= (0,0, 4a.Vs) CHS PROBLEM dt Play Rows min { 6 alo) ‘ 5 a Plage te Ye rma mtn =I ) Opel | tala [Pr 8 2% Je rlie mare ol : VRV be adic port net fund | boo, dle 2! Deriirance method: ee diepe erie TO : atl Sonce Sil yl, Orsy2> Ber | 2 i a Zz i i 1 eliminate tre Theo" _?EE Propabelty tee a4 | a H i i ion x | eres ne tn os bounces E ee : ‘oxrls mabrd the potnk Brohen te then sel ef te oentd. RE coca) s Ont 80-9 Bx lee = F8% Gu-| > een [ya = a Fe %]Nowe chao Gx -(l-x) Ki) He te 4S 5 Ty | sang! 20 \ he | ' Ce | ‘— oe 8 | e oye ote y+ ely) i 7 fy «lyse | » Prone [aye ly = 2 } 4 | i Ple _ ge A (ja) | | co Vous of wee sy rp ecu) 40 it oNewas of (er “=8 ' | o. Conclusion | (4 to ta= (Vy? Hy) Cyr yey > C4, 10)iraprient. MeTHer | Nete(d Enaphteal methe we appt fe = be (2x2) mabe: i 7 Prtbabitly tea Gem) er Cnx2) ee te fe aa lye ge fer Caen) rnabPe, ribximum alive of te chosen « Peink B & tnbessediion then pon 2% - Lacs Oorl®. FY UPIFede Bis ‘ ant Sek ay Ie tl 2 oan tothe) = Nae (oes Bn+S Bx = Met2-2n | 4 Ban = AWxtD t i, Ils 3 { : Cs % | : B (ey 5 {2 sepa Ma Mgebrate met melbod: te is Bo +5 Clo) = Ie+ otis) Bae ee Eee = le 4 2-20 -QuteS = Axt2 Nx = 3 Ire) Si | i i came? og SCI“) = one +2(%) oe Natue 49SERENE more = “tor Jolasper Az By t ee = syt 2. p ay tlt Wg 4 $2039 “fe = aye ial o 2 Vatu of ame> 3: yenceg) t vatagy en Ch = 21 4 22, Tow condbueuon : E Ca, a2) = Chr eh) Cire p Co. Vi» en Fp Shep: Grophceal metal fey myn mabe: —ea tr” et BLIND) = Bn = Xt 2 -OE gut 20-2) out 3 4B 2 DHA Bn = | | i mle aSRR | wNatur ff game = Qt 3 Cl 2) ita i 30) = 3] ewe ipod a) tee. playpen Be fag a cry Byo scr) en “Hy = By FOG > = +h ete i te ame + Roap 7 re 2. Coneticvon + ( a, XS, 240) ~ (4: 23,0) Copotfe)= (7-3) anthe above calte, ehebee bred Lites at Ye point ct tnbertechlon aslthartl ancl the Vete ame | Remaine < een oe fie BL (D ter te bias bgp goth \ ~ 7 V6" 7 KF 1 Acnd B, cHlee Uae are bee miaxumure Cether Lum ean be eheocen Fe Cateutabens - ray faethe } By &| BS) PRepiem 8: nete A br 3 Cra) = - BK 41 1-7) dyn - 3420 =- Be 4 1-% fa-3 = -he +), Vin = 4 ae A en 2, Neste of game ohx- 30-2 24 (f)- 2 (“ws ee dy- altey) 744 )= 2h) ae wow : Con chueire : Gj 765) * CMM) Cantos = C4 te) 4: ch) Prowtem 4: Seine, i Ve pin max 216 it Atnee | V2EV , No Sactelts pent - ! Step 2: Borruniance method: 1 ape ave. jSyte- tra, (b>? 1200! plane | ova » ji la b € A a By { - Plas polit is) te alo sate. e So el minalei IW 7le, 18 >S 4 eliminate 2 columm ed (ee)t ! i ve i x Pott. id thaltabeichin of lew Ibe Oat ® se of bree Leino isn p20lIl-e) = 16x + 6Ctr) 20 - 2out im = [614+ 5-Sa ao- Bly — x = 5} | ‘ OG cee te | eae bx= "Yiee Vv. 2 of: re Nei 4 20(1-%) 2 ISTE 4 20x. le 1e ten eee ey ley) = 204 +6lsy) i Isy + “4 soy 45-54 i | Letiy = By +5 Gy = Fl ! 2h Party Ins | =} | J ie | Note of | > 1Stu i = 165+ Go ir > bs 6 3 £ 225430 Qe fe 16 = By + te limp) SU lee 16 “3 Cenctusion: (a Ma) = COs Ya) Grr por Jargd = CO; "her Or Fah tm f iwre her aaa tne nner eenetre nhentemeeeeecetaennene east ! |15.2 DECISION MAKING WITHOUT EXPERIMENTATION Before seeking a solution tothe first Goferbroke Co. problem, we will formilate a general framework for decision making. In general terms, the decision maker inust choose’ an, alternative from a set of possible decision alternatives. The set contains all the feasible alteriatives under consideration for how to proceed with the prab- Jem of concern. : . This choice of an alternative myst be made in the face of uncestainty, because the outtome will be af- fected by random factors that are obtside the control of the decision makes, These random factors deter- ‘mine what situation wil be found al the time that the decision alternative is executed. Bach of these possible situations is referred to as 2 posible state’ of nature. 5 For each'combination of a decision altemative and a state of nature, the decision maker knows. what the resting PRFBE would be. The payolEis-e-quantitalive measure of the valué to the decision maker of the consequences of the outcome: For ekample; the payotf frequently is represented by the net monetary gain (profit), although other measures-also:cambe used (as, described in'Sec, 15.6). If the consequences of the outcome’ do riot Bedoiné completely certain even whed the state of sfiture is gives then the payoff becomes an expected value (ia the statistical sense) of the measure of the Consequences. A payoff table’com- ‘monly is used to provide the payoff for each’ combination of an action and a state of nature, Ifyou previously studied gamje theory (Chap. 14), we should point yut an interesting analogy between Leste in Chap. 14. The decision this decision analysis framework and the two-person, zero-sum games d maker and nature can be viewed as the pv0 players of sucha garne, The alternatives and the possible states of nature can then be viewed és'the available strategies for these respective players, where each combination of strategies results in some payoff to player 1 {the decision maker). Fromthis viewpoint, the decision analysis framework can be summarized as follows: 1. The decision maker needs to choose one of the decision alternatives. 2. Nature then would choose one of the possible states of nature. 3, Bath combination of a decision altemative arid state of nature would result in 9 payoff, which is given as one of the entries in a payoff table. 4, This payoff table should be used th find-an optimal altertiative for tife decision maker according to an appropriate criterion. a Soon we will present three possibilities for this ci where the first ohe (the maximin payoff criterion) comes from game theory. However, this analogy to two-person, zero-sum galnes breaks down in one important respect. In game theory, bath players are assumed tobe rational and choosing their strategies to promote their own welfare. This description still fits the decision maker, but certainly not nature. By contrast, nature gow isa passive player that chooses its strategies (tates of natufe) in fone random fashion. This change means that the game theory criterion for how to choose an optimal strategy (alternative) will not appeal to many decision makers in the current context One additional element needs to bé added to the decision analysis framework. The decision maker generally will have some information that should be taken ihto account about the relative likeliboBl gf hd Bde? janet, sible states of nature. Such informétion can an usually ben translated to a tfobeblity distribution, acting! lari fiscalcoctmedeereonemtnineaet net ‘@ TABLE 15.2 Payoff table-for the decision analysis formulation of the Goferbroke Co. problem ne renner State of Nature 1. Bal fo oi 2. Sel th tnd— of an individual. The probabilities forthe respéive states of nature provided by the prior distribution are called prior probabilities. Formulation of the Prototype Examiple in This Framework AS indicated in Table 15.1, the Goferbroke Co. Has two, possible decision alternatives under consideration: drill for oil or sell the land, The possible states af nature are thatthe land. contains oil and that it does not, as designated in the column headings of mf L by oil and dry. Since the consulting geologist has @8- timaied that there is i chance in 4 of oil (and sb 3 chances in 4 of no oil), the prior probabilities of the io states of nature are 0.25,and 0.75, respectively. Therefore, with the payoff in units of thousands of dollars of profit, the so il can be, obtained directly from Table 15.1, as showin in Table 15.2 We will use this payoff tdble next to find the optimal altemative according to each of the three criteria described below, The Maximin Payoff Criterion If the decision- maker's problem were to be'viewed as.a game against nature, then game theory. would say | to choose the decision alternative according to the minimax criterion (as described in Sec. 14.2). From'the * viewpoint of player 1 (the decision'maker), this hierion is more aptly named the maximin payoff criterion, as summarized below. Maxiinin payoff criterign For each possible decision alternate, find the minimum payoff over all possible states of nate, Next, find the maximum of these juinimum payoffs. Choose the al- __lemative whose miniinum payoff gives this maximum, — Table 15.3 shows the apifitation of this criterion to the prototype example, Thus, since the minimum payoff for selling (90) is larger than that for dilling (~100), the former alternative (sell the land). will be chosen ‘The rationale for this critézion is that it provides the best guarantee of the payoff that will be obtained Regardless of what the tre state of nature tums oot to be for the example, the ayo! 8m selling the land cannot be less than 90, which provides the best available guarantee. Thus ths criterion takes the pessimistic iewpoint that, regardless of sich alternative is selected, the worst slate of nature for, that alternative is tty 10 occur, so we should choose the alternative which provides the best payoff with is.worst state of. nature. jt —) ty * BEV TUPIanef Stey Desdinati'§ TABLE 15,3 Application of ‘the maximin payoff criterion to the first 2 Goferbroke Co, problem State of Nature : oi Dry Minimum J, Dail for cil 7° -100 =100 2, Sell the land 90 90, 90 & Prior probability | 025 075 v i vii . ‘ : i 8 TABLE 15.4 Application of the maximum likelihood criterion to the first. coeal Goferbroke Co. problem ‘ as State of Nature OOo 1 i Alter: oa Dry } : 1. Drill for off “sm 900 “100 —100° 2, Sell the land 90 90° ‘90 Maxi int his eon * . | Prior probability <-> | 0.25 ars | Ea eae | eee er 4 } I iy i i Maximum ‘This rationale js.quite.valid- when one is comipeting against a rational ahd-malevolent opponent. Hov- ever, this crteri¥SAt Gfien used in bames against nature because it is én extremely.conservative rite- | rion in-this.context. In effect, i assumes that nature is a conscious opponent that wants to inflict 36 mich I damage as possible on the decision maker, Nature is not a malevolent opponent, and the decision maker i E does not need to focus solely on the Worst possible payoff from each alternative. This is especially we i i when the worst possible payoff from an alternative comes from_a relatively {unlikely state of aati. E | ‘Thus, this criterion normally is of interest only to 2 very cautious decision maker. i The Maximum Likelihood Criterion |. The next criterion focuses on the mast-fkely-siate oF nature, of suminarized belo | _ Maximum likelihood criterién: Identify the most likely state of nature (the one with the Lapest | prior probability). For this state of nature, find the dfcsion alternative withthe maximar:payof. | Choose this decision alternative ‘Applying this criterion to the eample, Table 15.4 indicates that the Dryfstate has the lagest prior probability. In the Dry column, the sell atemative has the thaximum payoff, so the choice is to sell the land. The appeal of this:criterion is that the most important state of nature is ihe most likely one, s0.the alk temative chosen is the best one for this particularly important state of nature. Basing the decision on the ___ | assumption that this state of nature will occur tends 0 give a better chance jf @ favorable outcome than as- suming any other state of nature. Furthermore, the criterion does not rely dn questionable subjective e#l- | mutes ofthe probabilities of the respective states of nature other than identifying the most likely site ! The maior drawback of the criterion is that i comely ipnoses much relevant information. No state 2 sport likely one. Ip a nrohtem sith ¢ oa vite Brely one 2 F VTUPlanetthe probability of the most likely one may be quite small, so focusing on just this one state of nature. is quite unwarranted. Even in the example, where the prior probability of the Dry state is 0.75, this criterion ignores the extremely attractive payoff of 700 if tie company drills and finds oil."In effect, the criterion does not permit gambling on a low-probability big payoff, no matter howvattractive the gamble may be. Bayes’ Decision Rule’ | | (Our third criterion, and the one commonly aot is Bayes’ decision rule, described below. Bayes’ decision rule: Using the best available estivates of the probabilities of the-respective states, —— of nature (currently the f i probabilities), calculate the expected value of the payoff for each of the possible decision altf natives. Choose the decision altemative with the maximum expected payoff | For thie prototype example, these expected payoffs are calculited directly from Table 15.2 as follows: rf SEER > & son 1S pe sass [Payoff (drill)] = 0.25(700) + 0.75(~-100) 00. 25(905 + 0.75(90) _ | = 90. a Elpayotf en) Since 100 ig larger than, 90, the alternative selected is to drill for oil: Note that this choice coptrasts with the selection of the self altemative under each of the two preceding criteria, ' The big-advantage of Bayes’ decision rule is that it incomporates all the available information, includ- - ing all the payoffs and the iss available estimates of the probabilities of the respective'states of nature, It is sometimes argued 0 are too shaky to be trusted. There is no accurate way of predicting the future, including a future state of natuie, even in probability terms. This argument has some validity. The reasonableness of the estimates off the probabilities should be assessed in each individual situation, . Nevertheless, under many circumstances, past experience and current evidence: enble one. to develop reasonable estimates of the pfobabilities. Using this information should provide better grounds for a sound decision than ignoring jt. Furthermore, experimentation frequently can be condhicted to improve these estimates, a8 described in the next section. Therefore, we will be tsing only Bayes’ decision rule throughout the remainder of the chapter. i To BSSESS tie EHEC oF poSIOTe THACCUTAGIES i Une prioT probabjlfies, 1 offen is helpful to conduct sen sitivity analysis, as described below. i 3 Sensitivity Analysis with Bayes’ Decision Rule « Sensitivity analysis commonly is used with various applications of operations research to study the effect if some of the numbers included in the mathematical model are not correct. In this case, the mathematical ‘model is represented by the payoff table shown in Table 15.2. The mumbeis in this*table that are most ques~ that these estimates, of the’ probabilities necessarily are largely Subjective and, ca | i VTUPlanetues | a9 tionabfe are the prior probabilities. We will focus thé sensitivity analysis on these murmbers, although a similar approach could be’ applied to the payoffs given in the table, ‘The sum of the two-prior probabilities must equ 2, so ineréasing one of these probabilities, autonvgt- ically decreagés the ‘other one by the same amount, ard viee versa’ Goferbroke’s management feels that tht true chances of having oil on-the tract-of land are likely to lie somewhere between 15 and 35 percent. In other-words, the true prior probability of having oil is likely 10 be in the range from 0.15 to 0.35, 0 the comesponding prior probability of the land being dry would range from 0.85 to 0.65. Letting AS 99728 stiri 99727 31411 98436 a2iea p= prior probability of oi, the expected payoff from diilling for any p is E{Payoft (drill) = 700p - to - = 800p 5 ‘The stntng line in Fig, 15.1'shows the plot of this expesed payoff versus p. Since the payof from sling the land would be 96 for any p, the fat line in Fig. 15.1 gives E[Payoft ‘The four dots in Fig: 15..show the expected payoff for the two de 10.35. Wien p = 0:15, the decision swings’over to seling the land by a wide margin (an expected payoff of 90 versus only 20 Sor tilling). However, when p = 0.35, te decision i8 to dill by a wide margin (expected payoff = 180 versus only 90 for selling). Thus, the decision is Very sensitive to p. This sensitivity analysis has re- vealed that itis important to do mofe, if possible, to develop 2 inore precise estimate of the trae valve of p. @ FIGURE 15.1 - Graphical display of how the expected 10 payolt for-eaci- IF ! Dail for it decision alternative. E =} changes whén the & oor fo a prior probabiity of 5 I * oil changes fov 5 pee ena frst Goferbrte.Co. Bn ection toate ok problem shulate’ | ——- stl! etn, | _ ' 4 300 1 le i 1 i } i aha i | 1 } 1 ‘00 ¢ Sell tel 0 | it ‘The point in Fig. 15.{ where the two lines intersect is the crossover point where the decision shifts ‘fom one-altératve (sll he Land) tothe oti (dil for ol) as the prior probability increases. To find this point, we set EXPayoff (drill)} = ElPayoff (sell) : 800p — 100-= 90. 190 Ea P= R09 = 0.2375 ' rentals hey Syn Conclusion: Should sell the Jand if p < 0.2375." Should dill for oil if p > 0.2375. ' ‘Thus, when trying (o refine the estimate of the tre vatue of p, the key question is whether itis smaller or lager than 0.2375. For other problems that have more than two decision alternatives, the same kind of analysis can be Zip plied, The main difference is that there now would be more than two lines (one per altériative) in the graph- | display corresponding jo Fig: 15:1. However, the top line for any particular value of the prior probability still indicates which alternative should be chosen. With: more, than wo lines, there might be more than one crossover point where she decision shifts Stam one alternative to another. ‘You-can.see-an-example-of'perfontiing this-kind of analysis-with three decision altematives in the ‘Worked Examples section of the CD-ROM. (This same example also illustrates the application of all'threes decision criteria considered in this section.) For a problem with more than two possible states of nature, the most straightforward approach is to focus the sensitivity analysis on only two states ata time as described above, This again would involve in- Yestigating what happens when ‘the prior probability of, one. state increases as the prior probability of the other state decreases by the same amount, holding fixed the prior probabilities of the remaining states. This procedure then can be repeated for as matly otfer pairs of states as desired Practitioners sometimes use software to astistthein in performing this kind of sensitivity analysis, including generating the graphs. For example, aj Excel add-in in your OR Courseware called Sensit is de- signed specifically for conducting sensitivity-analysis with probabilistic models such as when applying Bayes’ decision rule. Compjete documentation for SensIt is included on your CD-ROM. Section 15.5 will describe and illustrate she application of Senslt. Because the’ decision te Goferbroke Co. should nicke depends so critically on the true probability of oil, serious consideration should be given torconducting a seismic survey to estimate this probability more closely. We will explore this option in the next two sections. 15.3, DECISION MAKING WITH EXPERIMENTATION Frequently, additional testing (experimentation) canbe done to improve the preliminary estimates of the Probabilities of the respective states of nature provided by the prior probabilities. These improved estimates are called posterior probabilities. We first update the Go.erbroke Co. example to incorporate experimentation, then describe how to de- tive the posterior probabilities, and finally discuss how to decide whether it is worthwhile to conduct ex- petimertation. Continuing the Prototype Example As mentioned at the end of Sec. 15.1, an available option before making a decision i to conduct a detailed y 7 ismir cnrvey of the land dn obtain a hetter extimate-of the probability of el, The cost is $30.000. | © 9, 7 rorall viru needs; A seismic survey obtains seismic soundings that indicate whether the geological structure is favorable to the presence of oil. We will divide the possible findings.of the survey into the following two categories: USS: Unfavorable seismic soundings; iti fairly unfikely. i FSS: Favorable seismic soundings; ols fairy Hey, y Based on past experience, if therg is oil, then the probability of unfavorable seismic soundings is PCUSS|State = Oil) = 04, so PORSS| State = Oil) = 1-04 = Similarly, if there-is no oil (.e., the true state of nature is Dry), then the probability of unfavorable soundings is estimated to be {Tt PUSS |State=Dry) = 0.8, iso PCFSS|State = Dry) = 1 - 0.8 = 0.2. ‘We soon will use these data to find the posterior probabilities of the respective states of nature given the seismic soundings. Posterior Probabilities Proceeding now in general terms, we let i n= number of possible states of nature; ior probability that tue slate of nature is state i fori 2M | Finding = finding frdm éxpetimentation (a ridon ‘variable Finding j = one possible value of findihg: P¢State = state i] Finding = finding f) = posterior probability tha uve state of nature is state i, given that Finding = finding j, for i= 1, 2,..., ‘The question currénily being atest is the following: Given P(State = state i) and P(Finding = finding j| State = state’i), for i= 1,2,... .m what if P(State = state‘i| Finding = finding j)? This question is answered by pine the following standard formulas of probebility theory: P(State = slate i, Finding = finding j) na d= PeFinging = mane | P(State = state.i| Finding’: P(Finding' = finding j) = ¥ P(Staie = state k, Finding = finding j} ites, el P(State = state i, Finding = finding j) = PF finding j| State = state #) P(State = state i), Therefor, for each i= 1,2,... 7 the desired formiba for the conespating posterior probability is 5 (TOR Dy a. sone sist “WB gonraiitt" (This formula often js referred to as'Bayés? theovem vecause it was developed by Thomas Bayes, the same 18th-ceitury mathematician who is credited with developing Bayes" decision mule.) ” Now let us retumm to the prototype example and apply this formula, If the finding of the seismic survey is unfavorable seismic soundings (USS), then the posterior probabilities are . 04029) 0.4(0.25) + 0.8(6.75), P(State = Oil | Finding = USS) = P(State = Dry |Finding'= USS) Similarly, if the seismic survey gives favorable seismic soundings (FSS), then f 0.6(0.25) i P(State = Oil [Finding = FSS) = FENN é PGSiate = Dry Binding = P89) = 1- = 4 ‘The probability tree diagram in Fig. 15.2 shows a-nice way of organizing these calculations in an in- twitive mennér. Thé prior probabilities in the: fifst\colums and the conditional probabilities in the second col. ‘am are pat of the input dat! for the problem. MalGplylid ah probability ip the fist column by a probability in thé, second column gives the corresponding joint probability in the thisd column. Each joint probability then becomes the. numerator in.the, epson of the corresponding posterior probability in the ey aed FIGURE 15.2 Prior inion Toiat Posterior Probability tree Probabilities Probabilities: Prebabilities Probabilities diagram for the full P(state) —” P(findinglstate) P(state andfinding) _F(statelfinding) _ Gofesbrake Co, : problem showing all | the probabilities leading to the ', calculation of each * posterior probability = of the state of oe nature given the : st finding of the j Usp 04 _ i seismic survey "Bigg 04504) =01 —o. __0.75(0,8) = 0.6 0: Dry and USS Dry, given USS Oil, given USS I VTUPianet * PP nsfourth column, Cumulating the joint probabilities with the same finding (as shawn at the bottom-of the figure) provides the denominator for each posterior probability with this finding, (Ifyou would like to see another example’ of using a probability tree, diagram to determine the posterior probabilities, one is included in the Worked Examples section of the CD-ROM) ; . Your OR Courseware also includes an Excel template for computing these posterior probabilities, as shown in Fig, 15.3 Aer these computations have been complete, Bayes’ decision rule can be applied just as before, with the posterior probabilities now replacing the prigr probabilities. Again, by using the payofs.(in units of thousands of dollars) from Table’ 15.2 and sublcacting the cost of the experimentation, we obtain the results shown below. ‘ : ; Expected payoffs if finding is unfavorable ee (USS): 1 E_Payoft (érilt| Finding ~ USS)} (700) Ate 100) ~ 30 e 15,7. | %, sei E(Payoff (sell | Finding = USS)] = +60) + S00) -30 i " = 60. B IGURE 15.3 als = [op This posterior 1 | Template for Posteri probabilities ZT tt “fF template in your or |-3-} [ate Courseware enables [7 efficent calculation 5. of posterior fs. probabilities, as 7 illustrated here for 8 the full Goferbroke : Co. problem. i Sas oe Se Posteriorl P(State Finding) = Probabilities: ————— ‘Stare ot nature pf — Find e z o i 12 [Posterior — Fatale]. [713 [Probab Soxeotnawwe [| | finding | : i foie, .Expected payoffs if finding is favorable seisniie soundings (FSS): EXPayoff (drill! Finding = FSS)] = 300) + te 100) ~ 30° § ' = 270. (Payot (cil Finding = FS) = 00 + 100) — 30 =.60. Since the objective is to maximize the expected payoff, these results yield the optimal policy shown in Table 155. Hoviévét, what this analysis does not answer is whether it is worth spending’ $30,000 to conduct the experimentation (the seismic survey). Perhaps {t would be batter to forgo this major expense and just,use the optinial solution without experimentation (Ail! for oil, with an expected payoff of $100,000). We ad- dress this issue next : : The Value of one Before performing any exp}riment, we should determine its potential value, We present two complemen. tary methods of evaluating its potential value, ‘The first method assunjes (unrealistically) that the’ experiment will remove all uncertainty about what the true state of nature is, dad then this method makes a very quick calculation of what the resulting in- provement in the expected phyoff would be (ignoring the cost of the’ experiment). This quantity, calied the expected value of perfect information, provides an upper bound on.the. potential value of the experiment Therefore, if this upper bound is less than the cst of the experiment, the experiment definitely should be forgone 1 However, if this-upper bound exceeds. the cost of the experiment, then the second (slower) method should be used next. This method calculates the actual imptovement in the expected payoff (ignoring the cost of the experiment) tha would result from performing the experiment. Comparing this improvemeet with the cost indicates whether the experiment should be performed. Expected Value of Perfect Information. Suppose now that the experiment could definitely identify what the true state of naturel, tiereby providing “perfect” information. Whichever state of nature is iden- life, you naturally choise the action svth the maximum. payo-fr-that-state-We do not Know in advance which state of nature will be identified, soa calculation of the expected payoff with perfect inforanation (ignoring the cost of the experiment) requires weighting the maximjm payotf for eachie of nature by the prior probability of that state of nature. This calculation is show) atthe bottom of Table 15.6 for the full Goferbroke Co. problem, where the expected value, of perfect infiirmation is 242.5. 2 TABLE 15.5 The optimal policy with experimentation, under Bayes’ dectsion f rule, for the full Goterbroke Co. problem Expected Payoff | | Expected Payoff Including Finding from | Optimal ' Excluding Cost of Survey Seismic Survey} Alternative Sell the fond FSS | Dail for ci tm |i + “ 4 ‘RL TABLE 15,6 Expected payoff with perfect information for the full Goferbroke Co. problem, L State of Nature Alternative oil Sey 1. Drill for ofl 700 ~190 2. Sell the land 90 90 ene te Maximum payoff Wns =~ Prior probability er 0.75 Expected payoff with perfect inforfation = 0.25(700) + 0,75(90) = 242. ‘Thus, if the Goferbroke Co. could lear before choosing its action lee the land contains oil, the ex- petted payoff es of naw (before acquiring this information) Would be'8242,500 (eetning the cost of the experiment generating the information.) ! “Torevaluate wliethér the experiinent should be conducted, we now’ fe this quantity to _ the ex: pected value of perfect information, ‘The expected value of perfect information, abbreviated RVPI, is-catculated as BVPI = expected payott with perfect information -j expect pajott withiut ekperieataton.” ‘Thus, since experimentation usually cannot providelperfect information, EVPI provides ai upper bound on ‘he expected vale of experinientation, } For this same example, we found in, Sec,,.15.2 that the expected payoff without experimentation (under Bayes’ decision ule) is 100. Therefore, EVP = 242.5 100 = 142.5. | i Since 142.5 far exceeds 30, the cost of experimentdtion (a seismic survey), it may be worthwhile to proceed with the seismic survey, To find out for sure, we now go-to the second method of evaluating the potential benefit of experimentation. 1 Expected Value of Experimentation, » Rather tian just obtain an upper bound on the expected increase in payoff (excluding the cost of the experiment) dine te etformiig experimentation, we how will do some- what mote work te callie this expected increase ditectly. This quantity is called the expected value of experimentation. : Calculating this quantity requires first computing the expected payoff with experimentation (excluding the cost of the experiment), Obtaining this latter quantity requires doing all the work described earlier to ficd = all the posterior probabilities, the resblting optimal policy with experimentation, and the cortesponding expected payoff (excluding the cost of the experiment) for each possible finding from the-experiment! Then each | Of these expected payoffs needs to’be weighted by the probability of the cotresponding finding, that is, i Expected payoff with experimentation’ =) P(Finding = finding j) 7 i . a {payoff | Finding = finding j], ‘where the summation is taken over all possible values of j. i fae4 ; oe | For the prototype example, we have already done all the work to obtain the terms on the right side of this etuation. The values of P(Finding = finding f) for the two possible findings from the seisinic survey—~ unfavorable (USS) and favorable (FSS)—were calculated-at the bottom of the probability tree diagram in Fig. 5.28, PUSS) = 0.7, PUFSS) = 03, For the,optimal policy with experimentation, the corresponding expected payoff (excluding the cost of the | seismic survey) for each “ was obtained inthe third colum of Table 15.5 as § B(Payatt| Finding = USS) = 99, ae i + (Payoff Finding = FSS) = 270. | & “ 5 t Raa? G © With these numbers, Sa wasn = amt Expected payoff with experimentation = 0790) + 0.3000) | oO tee i = 153, ! Now we are ready to cae the expected:value of experimentation: The expected value of eperimentation, abbreviated EVE, is calculated as : ; ' _ EVE = expected payoff with experinicatation ~ expected payoff without experimentation. — Thus, EVE identifies the potential value of experimientation.. ; iad j | For the Goferbroke Co,, t EVE = 153 - 100 = 53. ‘ 4 ' “Since this value exceeds 20, the cost of conducting a detailed seisnic'survey (in units of tiousiidsof dol- i lars), this experimentation should be done. ‘1 i . tt 5.4 DECISION TREES ; -etison ets provide a sel way of sual paying te totem and then organiing the computa ; onal work already described in the preceding two sections, ‘Theseltsees are especialy helpful when 2 se i uence of decisions must be made. I -onstructing the Decision Tree ~ The prototype example involves a sequence of two decisions: i 1, Should a seismic stirvey be conducted before an action ie chosen? | 2. Which action (drill for oil or sell the land) should be chosen’? The comesponding decision tree (before adding numbers-and performing ‘compatations) is displayed in Fig. 7 54 ‘The junction points in the decision tree are referred to @s nodes (or forks), and the lines are called “ranches, ! A decision node, represented by a square, indicates that a decision needs t6 be made at that point in the process. An event node (or chance node), represented by, 2 ci tes that 9 retom FUP las|___node.d:represent-the-two-possiéle outcomes of ee Lge po acon siti y/ ey @ FIGURE 15.4 ‘The decision tree (before including any numbers) for the full Gofesbroke Co. problem. Sely i \ ‘Ths, in Fig. 15.4, the first decision is represented by decision dode a. Node b is dit event node repre- senting the random event of the gutcome of oe survey) The two branches emanating from event @) with its two possible choices-Tf the decision isto drill for oil, then'we come to another event figde (i J. 8, and h), where its two branches correspond th the'two possible states of nature. survey. Next gomds the second decision (nodes ¢, d, and, | Note thatthe path followed from node a to reach an terminal branch (except the bottom One) is de- = termined both by the decisions made and by catidom events that are outside the control of the decision maker. This is characteristic of problems addressed by decision analysis. ‘The next step in constructing the decision the is to insert numbers into the tree as shown in'Fig, 15.5. ‘The numbers under or over the branches that are piot in!parentheses are the cash flows (in thousands'of dol- Jars) that occur at those branches. For each path] through the tree from node @ to 2 cerminal branch, these same numbers then are.added to obtain the branch. The last set of numbers is the probabilities of random ever: In particular, since each’ branch emanating from an event node represents a possible fandom event, the probability of this event occurring from this node has been inserted in parentheses along this branch. From event node fh, the probabilities are the, iting ioral payoff shown in boldface to the right of that}: 7130 Be Bs NE == : on k —Bo : e f 700 3B FIGURE 15.5. sdding bioth the wrobabiltes of andom events and: 2 payotts. ven-to dell). Therefore, the probabilities from these event nodes are the posterior probabilities of the states nature, given-the finding from the seismic survey, where these niinbers are. given in Figs. 15.2 and 15.3. Fi ally, we have. the two branches. emanating from-event-node-b-Phe-nembers-here-are the Protabititer of se findings from the seismic survey, Favorable (FSS) or Unfavorable (USS), as given undemeath the sobability tree diagram-in Fig. 15.2 or in cells C15:C16 of Fig. 15.3. Performing the Analysis faving. constructed the decision tree, including ifs numbers, we now are ready to analyze. the problem by sing the following procedure. \ i “. Start atthe righ side ofthe dodision tee and mote left one column at «time. For each cola, perform either step 2 or step 3 depending upon whether}he nddes in that column axe event nodeb or decision nodes, = ——™ For each event node, calculate its expected piryaf by multiplying the expected payoff of each bean (Showa in boldface tothe right of the branch) by he probability ofthat branch and thea summing these Products. REcon lis expectenl pay tne arn dsti vs seaacensetametalimeminnter te 1 i } { this quantity as also being the expected payoff for the bcanch leading to this node. 3.-For each decision node, ccimpare the expected payoffs of its branches and choose the alternative whose ‘branch has the largest expected payoff. In each case, recosd the choice on the decision tree by insert: ing a double dash as a barrier through each rejécted branch. . To begin the procedure, consier tne rightmost Golunn of nodes} fameln event nodes fg, and A. Ap- plying step 2, their expected payoffs (BP) are calculated as PP FOF FD Si tar nodef ( EP= Lowy + te 130) = = po for node g, , r= Le) + 3¢-100) = 100, for nc i A : ‘These expected payoffs then are placed above these jiodes, as shown in Fig. 15.6. FIGURE 15.6 The final decision tree that records the’ analysis for the full Goferbroke Co. problem when using 4 monetary payots oynas sunt 99227 21111, pase s2i4e

OR - Notes PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

OR - Notes PDF

Uploaded by

Copyright:

Available Formats

You might also like