Sei sulla pagina 1di 17
Linear Programming: The Simplex Minimization Method ‘Adthoush > and = symboldre oeeasienally wed in constraints of muxinization probletss, ‘hme ure srace common ammong rminlnizatinn problems. We are nov gong to disewss how 1 “hag these consdraints with = and = symbols 1c equations. ub¢rection of slack variables = ‘potmltted in minimirsion, but Hot i maximisation, becasive if we intend te minimise, tia but (bgleal to subtract, but if we intend to maximbs, iw otherwise Common sensetaty usthat to change nconatraint with > sign erequetion, ally subtract « siadkyarlable trom he lef ember: Byt i this particular cane, ifother variables are veo, the ‘ack variable wil have a negative value. For instance, X + ¥ 2 10, if converted to equation it ‘bwcomen X + ¥ —§1 = 10. Now fx = 0 and y =0, since the objective is to minimize, then 31 = 10, Butitis ageneral rule is line programming that all varisbiesinust egreater then oF ‘equate uero,dietee theresa need to Use Enother Kn«tof variable het wilsave the alc variatle ‘Poin becomming negative. Let thia be kncwn ay an ‘artficial” variable (An) An artificial variable ‘haa large value which is used ae. computations! doviee that prevents an equality enstraint ‘om equating meonstaattozeo, and preventstheslack variables from becoming negative.Thas, in u > aymbel, we convert he inequality fe equation by subtracting asleck varabis, but atthe ‘same te add an arte! variable. In a constraint containing = symbol_add an artificial variable o he tft member, co save Uke right menber from beng equated to vers, i other solution yarisbles ure alt 2270. ‘Example: 10x + 2y = 10, ifx = andy 0, ten we wil PaveO = 10, which iéale. once we os an tical variable and have 10x 42+ hy = 10. ‘itinueatzatton problems commianly desl with cost. Slack vastsbles do not contribute any Amowin to cost, but artifical variablos sented te biggest amount to ses: ina minknisation problem. [¢ contributes to the objeetive an amoua’ crealer then eny of the coetfinienta of solution variables. ‘For convenience, in representing the contribution ofthe artificial variable to the dbjective, ‘womay use a quantity which s apower often, grater than any ofthe coefficients found in the Constraints and obeetive Powers often sie nsmnerais such as 10,100, 1000, ete. ‘Sepe to oolving # minimlzation yrcbiers are similar to matatzation except for processes 1) TheCjoolumn ofthe list tablebegins mith the coniiiente or rtiieat variables and ff lack variables is the oajeetive with positive onefficienta in the constraints 2) Instead of loking for the teat pantie quantity Ix ©) 2, row for the optimum column, fk for the moat negative entry. 1s) Tueoptimumeableve final tate haseatries in the ~ row whieh are ether em or positive Summary of Converting Constraints to Rquationsina Minimization Problem 1) Add an atic variable ithe sito! =. (2) Addn slueic variable the myo fs 2) Subtract aaluck aridblebut add un erfcal variable if Une symbols 2. eawple 1: Conalder the lowing linear progr: Minimise 2 = 6+ Sty Galject tax Se + 5S 15 de haye 720, y20 Sobaion: ‘The new program wh ick and rif vartsbiet: Mirkines 2 00x + Bly = O81 +082 + 10001 Subjectuc: Sx + 8y+ 5, =15 arhay-S2 A= 16 x20, YEO, S20, S220 ‘Note that tbe ata asiabina nave ede couclnts in the objective fanetion while the aceit variable contribates 100, whieh ls & power often yreater thus any coeffileat isthe program, ‘To Initial table haw, and A eniziesin x product cokine, wth ther resperthy eit ‘butias tothe objective inthe Geeta, ‘Chones the most negative fe the ~ 2 ry. ~380 beg the most nigatve, belongs to the ‘optunin eoluain Tolocate the pivotal row divide iS bybandiOty.Since 15 + 8 has axmaller ‘uate, fuen 3} row is the pvetal rom. ysett “eminent Ue tn nl da the pn hy Aig the pal row bo 18 8810.)+ Sequtst 21 L 0 0 wicnbowonanthareplacingrow Thoasarng ‘ane, wah i soreepondlg cmc tin af bere wwe of the eco ae 3 - oy Comping or Ay ent sy o=e $eyei=t Ao ta=o tonto=e ote oat tet ‘Tablet (with Ay row entice) Computing 25 wet t 90) = 10 om wo 0 0 woe oh vy ae mea ‘Agiin choose the mace negative entry in the ast row. Negative one hundred thirty (180) teh the mew epee shoal ancld. Then eats th peta rv by dng 8 by 2 sn oy sine 4 has male que wich 5 ten Ay rete palo. ‘Take © ein iow ncaa parvo yap wich a@hoganef-atepanFd “PSB e te pes st os ££ OF tt coming rye i epi =P and a th own be eiae ‘aa ota come nota $ era ra oy x 7 ay Pp oe Et Fe or £ Oot tt [= yj ms ow 3 Oe RY ‘Since theres no more negative entey iy the tx; row, the table is optimum, Deckion X= Sor 85 are v=doris Minimum 2 = 225 cs 2 Sample “The Phippne Peas Me. profucsapetiy ented eed soppment Has an eee eooene ics Protein and C ‘The fos ingracant pound Th second ingrodiantO, corte PIS & pound Memixtare’ 15Opounds Pandit met hase at heast200 pouidsC Thecompeayspmbless todeterminehawmochof ach ingredient touse fe mininae com, but aay te requrersent, ‘Weting the Linear programm Minupise = 60+ 160: Autjectta: PF C= 400 totalerder Psi Reeew 2200 ¢ content r: ozo ‘Tae new program with sack and arial variables: Wate: T= OP +160 +081 +05 + 100h4 + 10042 Subject ion PEC HAL = 40 PES = 160 o=Set Ag 8, P29, C24, 512%, 2z0 ‘Setup he ina cable and compte fr 3 and G~ 2 rows 5 nla Past (Tablet) “The eniries in the product column ofthe ini table re arial varabin and sek ‘varabies that éo not have negate cootickents a te contrat ‘Tscattrie uti eth atic arth the objective fcdonts 100 since the highest ceetteent i the roasts Tyemons negative ey G 164 hence lt non tthe epetnm estuma.ia ‘ne motion ofche prota row, db ot cade be econ eons 5] row becnae ¥e ‘anata by era. Th chants two mad Agrowe Ay iv hn nr quoter the put Gide ey mh pan cn Hence kate peal ow Te phot ‘Table ‘The replacing row nasi jaa le pruotalrow (since the photis 1), except that she outgoing variable must be replaced by the entering variable atthe left ofthe tab. _cTusgeons consrunin abel Som) evmatarna seme nne tet sbowethe pot ‘Table 1 (with the replacling and 9) rows) aa Prod 6 P 1a ai os ioc 3 4 1 20, Computing for AL row: 200(-1) +400 = 200 (1) +4=1 Wy +1=0 OI) +0=9 =I) tost a+ t= a In +o= Cormpsituig tne Gj— 2) rei 6160 100 16 0 “4 O08 OTe =1) = 20090 100 © 0 100: 100 =160 = 000000 0 0-16 016 = 8200 _0. 18. 5200 100 180 84 100, 94 ‘Tadie IF (with complete extsiea) 180(—1) + 200 = 60 AH ti=0 %-1) +0=0 Ma) to=— HD +1=1 ON +1=1 : 11 1) = G00 9 0-109 160 100 —100 eas 2 1 OUdmeeNe 6 56. c 1820001 0-10 1) = 800 9, is. R100 18 5a To Developing Tabte 1 “The sow negation entry nC) ~ row (tn Table Ii) lb ~84 The plore! ow le nane other than de Meat constralie row (Ax row) nines the quantity eannotbe divided lyy—T and Oi the ‘optimum colar, “ue pivotal row heoomos the Feplacing row in Table {V after replacing the outgoing var~ tel tha tne whe tne eet nae Ve ie re a te "There nochange the P-row, nce the ent in the optraam etna iy already ero, ‘Table IV Gish explacing row and P eo) a ¢ ie 8 iw i Paty PC AA ae oe oe tt cP te 168 6 ark : Computing fee Crowe #23) 4 200— 250 qa) +0—0 nays i=t 1a) 0-1 14-10 14 0=4 aiti=o iw a 190 100 1) cs & seats Sacesary of Tbe ef Pattippine Feeds Ine. 8 Bg ering ot sey Ie ‘A. Ineaeh of the fovowing pes mnlation abe, compute fr 2 0m a mind hepmu np ie sae na, ee onions » a, ee) Prod Gy XY 4 & mw ox we 1 ¢ fe 0% 9 of p14 4 +4 2) y ‘jo 200 10 v0 000 | Prd Qy PF CC Ar 31 See a a re ar a a 0 & © 1 0,0 1 0 @ mic wm 6 i oi y ory a ary g 280° °86 100 0 0 tao Prod Qty X12 AL SiS A cr a a 2% 9 £0 0 1 0 6 8m M ot 9 Oat oa -—-——————— —_———_ — Tinie % days fret 6 0) SHIR TAG 1 Nia a Ke ‘Toe pure Yast oi ‘Add hw uppeoprinte slack and artificial variables t cnr each exp constraint ‘tw equation Do net solve i), Minutes, Z= 40K, 9 2Kp 9, New Program with Sack Variables. Subjetter XL + X9E10 Minimise x Sabot 2) Minune, 2 = 40%) 1 ke New Prograsa with Stack Varinbes Sublette 2X, +X.218 Minimise % OE Subject ts M20, 20 8) Minkminns = 21K) + NOKR +9ING New Programa with Sack Variables: Sujet, M+XZ+XGSS Minimize BR) 4Kg 4 3Q=20 Subjects: ax +X50 HL X0.X028 4) Mines = 40% + 60K + OXY New Prope th Sack Vasile Subjesttx 4+ 2KFaty 1S Minnie, ATOZ TARO Subewt en ey + ong F4ty S18 Ma eS 5) tiles: 2 = 80%) + 10K2 +S New Prograsn with Sick Variable Suiiectte, 28 FARE Mime oh Be go cl: 1. %F oO Na a ease ‘Tine Moectaas ‘Camplece te olowirgsimpex aio 1) Minimise 25 Say + fag © New Progra: Sidjertx xt42xp= 0 Member, Z= Sey + Orn + 08;-+ #CA1 + 10Ar Met Suber +E +h1=8 i ssn thee nna tame Bacrcie 4 Sai te apn etd iy neaper area Ys wis w iin oi120% + 10%, fal annie ce Seren Decent ipod gai? eh connote cided soe, 2), "The Neen Dog Feds Corporation ant vis $00 pounds oF pecs! J food Tite 3 a eae We ma, acs of we: X md 3 The ft ‘aiaree ‘Cort PIO a pound and the wicond, ¥ costs P00 por pomwd Choiical one tata he mi ect ore Wes 400 pends af an eet ‘contain ot Lewst 2.000 pounds of ¥, How many yourds ef ench ingsedicw must be utiived i etter bina con 16 Ye al8 2 ae be prodeed W cents 2 edosien, go iam sd co's PQ pe wit 3 Zentans no enor, 2 nds af vinmne and coe F400. Y contains 1 unt a en posi ne inter gid feet preempebe mle berated ‘yeas ad | The cininwm calcric requirement iy 1S units und tee ‘tana ian 8 twice How tex af ent foe sould be produced to Tans thee Sl te mini the ot ceth ‘A.owpany wihee sw» minimize coo poductinneprseied by vu 222 + 33¥ + 112, 2 Supeabe : eve a z5s eer teers) XYE20) om way uso, Yank dd esd? ea eo i oT

Potrebbero piacerti anche