You are on page 1of 76
near Programming: 5 plex Method That hiking trip to Yosemite is still going to happen. But when you told the others wha are going about your trail mix of unsalted peanuts and raisins, they all said “borrring” It was decided to use the unsalted peanuts and raisins mixed with M&Ms and salted mini-pratzels. Now its more complicated to termine in what proportions these four ingredients should be mixed, and there remain the issues of carbohydrates, protein, calories, and fat, Fortunately, you alse finished the simplex method before midterm, so you can handle the problem just fine ‘The Chapter Project at the end of the chapter will guide you. CUUTEINTEE 5.1 the Simplex Tableau: Pwvoting 5.2 The Simplex Method: Solving Maximum Problems in Standard Form 5.3 Solving Minimum Problems Using the Duality Princiole 5.4 The Simplex Method for Problems Not in Standard Form * Chapter Review * Chapter Project * Mathematical Questions from Professional Exams A Look Back, A Look Forward In Chapter 4 we described a geometric method (using pis) The simplex method isa way’to sole linear programming for solving linear prograntming problems, Unfortunately this problems involving many equalities and variables. method is useful only when there are no more than two Developed by George Dantzig in 1946, is particularly wel variables and the number of constraints small sited for computerization, In 1984, Narendra Kermarkar of Since most practical linear programming problems involve Bell Laboratories discovered a way of solving large linear systems of several thousand linear inequalities containing programming problems that led to metheds that compete with several thousand variable, more sophisticated techniques need anuloften outperform the simplex method. to be used, One of these methods is the simpler method the A discussion of LINDO, a software package that performs subject ofthis chapter the simplex method, is provided in Appendix B. 217 218 —Chapter'5 Linear Programming: Simplex Method 5.1 The simplex Tableau; Pivoting PREPARING FOR THIS SECTION Before getting started, review the following ‘+ Row Operations (Sections 22. pp. 71-73) ‘No low Work THE’ARE YOU PREPARED?’ PROBLEMS ON PAGE 232 OBJECTIVES, Determine a maximum problem is in standard form (p. 219) ‘Set up the initial simplex tableau (p. 220) Perform pivot operations (p. 223) Analyze a tableau (p. 225) Introduction If we have a large number of variables or constraints itis still true that if an optimal solution exists twill be found at a corner point of the set of feasible points. In fact, we could find these comer points by writing all the equations cerrespending te the ineguah ities of the problem and then proceeding ta salve all possible combinations of these equations. We would, of course, have to discard any solutions that are not feasible (because they donot satisfy one or more of the constraints}. Then we could evaluate the 0 ands» = 0. In other words, Pis not yet a maximum, > Analyzing a Tableau: Finding Current Values The currer are found as fol ues of the objective function and the basic variables in a tableau lows + Fromt the tableu, site the equation corresponding to each ro¥ + Solve the bottom equation for P and the remaining equations for the basic variables + Set each nonbasic variable equal to zero to obtain the current values of Pand. the basic variables ce eure et Perform another pivot operation on the tables g phot element in the tableau below. Then analyze the new tableau Pivot column xt in Display (6), using the circled BV Pivot row —> S) RHS Current values 8740 40 P10 SOLUTION STEP 1 The pivot element, isin row 1. Use the row operation R=sn 5.1 The Simplex Tableau: Pivoting 227 to obtain Fivat column RES 4 4 L Revised pivot row Ry > eos 20 Il 1 f get O'S in the pivot column x2, use the row operations STEP 2 RR RR bes hen RIS Current values 1-20 . 120 P10 In she tableau given in Display (7), the basic variables are 242%,» and P. The variables sy and & are the nonbasic variables. The result of pivoting caused Variable and s; to become « nonbasic variable, Finally, the equations represented by Display (7) can be written as pe Let the nonbasic variables s, and s equal 0. Then the current values of the basic P= 140 “The second pivot has improved the value af Pom 120 t0 140 Because P= > 1 — 2 + 1MDand 1 = 0,9 = 0, the value of Pcannot increase her than 0 reduce the value of P. This beyond 140 because any values of s) and = meas we have maimized Ps 228 Chapter 5. Linear Programming: Simplex Method Look at ve sequence of tableaus from: Examples 4 and 5 using a graph id interpret each tableau Initial Tableau ater T° Pivot After 2 Pivot © RHS BV Py ss. RKS BVP x os 8 RHS BVP on x 4 p the initial Lableau, the three rows represent 820 Wt 4 = 120 = ae tay Set up graphs using ‘each of these tableaus, a the horizontal ee Figure as the vertical axis to illustrate FIGURE 1 2x, 4-80 Dey Ay = 120 The nti Teo TheTableau After PP vot Set the robes 1 ot te ronbase varia 1p as 6 eal oval 120. nen th w0.Then P=ay +4 =310 iy + diy = 3120) 4 420) =740 As Figure | shows the pivoting process takes us from one corner point 0, 0, to mother sarner point (4,0), endingat the corner point (20, 20), whe 5.1 The Simplex Tableau: Pivoting 229 Saou ketes eeenene un cur ekeer SOLUTION ‘The following maximum problem is in standard form Maximize subject to the constraints, {a) Set up the initial simplex tableau Ab) Using the indicated pivot element, perform s pivot operation (6) Analyze the resulting simplex tableau (a) Rewrite the objective function in the form Introduce two slack variables, sj and sand abtain the system of equations 200 where The initial simplex tableau is By P 5 afo 3 1 elo 4 0 Pll 0 b) In the table: are marked below, the pivot element is circled aad the pivot row and pivot column Fivat colunn BVP Pivot row > sj [0 T P 230 Chapter 5 Linear Programming: Simplex Method L STEP 1 To pivot, first use the row operation Ry =, to obtain a | in the pivot row end pivot column: Fivot column STEP 2 use the row operations RSS 2K the result isthe tableau BV Pom RUS Current values 200 3 250, " {e) In the tableau above, the basic variables are x;, 25 and P, The nonbasic variables are xyand 5. The equations represented by the tableau a 200 Let the nonbasie variables variables are ind $; equal 0, Then the current values of the basic 400 400 Because P = 5, ond xy = 0 and s; =O, we see that P cam be further inereased by choosing a pesitive value for the nonbasic variable ay and leaving, = 0. : ‘Noy now work PROBLEM 351) AND 256 COMMENT The pivot operation ean be dane using @ graphing utility such as theT 84 Plas The augmented matrix contained in the inital tableau is entered into the esleulator. pivot element has been identified, the elementary row operations built into the graphing Utility are used to obtain a new taoleau with a pivot element of one and all other elements in the calumnn replaced with zeros. This is best seen using an example . 5.1 The Simplex Tableau: Pivoting 231 USING TECHNOLOGY aomr Rees eee n eR uP In the initial tableau below, the pivot element is circled and the pivot row and pivot column are marked. Use a graphing utility to perform a pivot operation Pivot column BV Pivot row —> sy T matrix has 3 100s Using the pivot element $ as indicated SOLUTION First enter the initial tableau into the graphing utility. See Figure 2 and 6 colunans and has been named matrix [A bove, follow the steps helow using the elementary row operations in the caleulator to stain the new tableau. FIGURE 2 he wit a0. —_ nt STEP 1 Make the pivot element one by dividing the pita row by 3, On a TESA Plas, th operation is matrix math Es‘row( (hit enter *row( 1/3, [A 1) STEP 2 Obtain DSclsewherein the pure pivot row trom Step 1. To obtain 0s nd add it to row 2, then multiply aperations bby performing row operations using the revised in the pivot column, multiply row Zand add itto row 3. On a TI84 Plus, the matrix math F:trow + thitenter “row + (1, ANS, 1,2) matrix math F:*row + € (hit enter *row + (2,ANS, 1,3 Figure 3 shows the result Reour(2 TAG 33) frrour br rac. pFrac. ta 273 173. 173, 19973 is] ff 1% e873 237] [2a 232 Chapter Lineer Programming: Simplex Method EXERCISE 5.1 ‘Are You Propared?’ Problems Answers ave wiven 2 the endl ofthese exercises, IP you get aon as 1. On the matty below, perform the ra operation: Ry = Any. " ¥ pp. 71-73 [: spare | 8 1 38 1 150 12a ae 4 Concepts and Vocabulary. 3, Trueor False For a masinnam prvblenn in standatd orn, each fof the constraints, with the exception of the nonaegativity coastzaints, isiwetten witha = symbol 4."The consteaints of muasinen prbleat in Sandare form are changed Hoa inequalities to equations. by intrxuei ead the pages Usted in rel 2.On the mntrix. below, perform the row operations: R= rand Ry = Wp + rylpp. 7173 fo 8 21 ww 1 1 o 2 pod so 3 3 1 ©. True oF False bor a wavinuon problens in standard fosm, the slack variables are sometimes negative 6. True or False Once the pivot cement is hlenttod in tablet, pivot operations cause the pivot cement to become 2 Tan ‘couse the remaining entries in the pivot column lo became 8s Skill Building, Its Problens 7-4 de nara prob Ne 7 Maximize 8, Masimize 9. Masinive 10. Moxinrize P= yy bd: subject to the a subject tothe constraints wi D6 Qn b= 2 aeay =o . M1, Masimiize 12, Maximize Pray tas oa subject to tke constraints atat mes txt dy 20 =O 15. Masiinize Dy =k +a subject to the constraints n=O PEW Peay subject to the constraints ait wee [> hs += 10 13, Moxinie 14, Mosirize P= Dy ba Poy subject to the constraints subject to the constraints ee say dye sd nea m=O 16, Mosinive PH xy +2 ay subject to the constraints seuss usa w= NEO HzO 5.1 The Simplex Tableau: Pivoting 233 ii J Pro Hin, write the ion pe tori staal foren, Determine ifthe problem can Be adi so as to i snl form id verso No T? Maximize 18, Moxinive 19, Maximize Pexyty Pays Peyoate subject to the constraints subject to the constraints subject to the constraints woes 6 my $28 = mtoos4 noose 4 =o y= a =o 20. Masiinize 21, Masimize 22, Maximize P= dy ayaa Pomyruray Paxysm tay subject to the constraints stubjoct to the constraints stubjct to the constraints be 6 6 EO NEO eo =o eo J Problems 28-0, ect mexivanen probe iin staudnd fore. For ach one inure slack nriables end setup the sta Spe ae Se 23. Maximize 24, Maximize 25, Maximize PH 2x, bay + ey Paiytdyte Pay ay subject to te constraints subject to the constraints subject to the constraints bys 2ye S10 yo mS 12 dys Ne m=6 28, Masinize 27. Masini 28, Maximize Parytoy Point teres Paras t as subject to te constr subject to the constants subject to the constraints 1 Ape ye ay s50 way t= 10 Hay) boty ayy ssn mee n=O neo me " 28. Masiinize 30, Moximive Peay tay Padyouche subject to tke constraints subject to the constraints Meo os 234 Chapter Linear Programming: Simplex Method Jn Probleris 31 ble can be vodified so as tbe a staadand form, Wete the maria version & fo ect ons introduces 31. Masiinize 32, Masimive subject to the constants subject to the constraints =m + wt 33, Masiinize 34, Masimive PH dy ty tat oy Ps pstaints subject to the constraints 1 BN EA = Hy subject to the — ay tes way = Ho a we Att wa xP bs =O K=O MEO YEO MEO =O Iie Prob Lal perfor a pivot operation on ec eae. The pivot elem Lib) Using the iy $50 +t obained 4 wes BV nfo 2 1 af 300 6 h BP wou bh bow W RIS BY P hob by RIS 8 nfo ee ee “0 nfo 1 oof 6 Loa Cr a o 10 af aos wloow as bo 1 of28 o v1 fs ula o 3 9 1 9 4 [4 ufo 6 o ifs a Phe se 7 oof a Application: In Seton 5.2, yo wl ead 0 sole Probans 41-46 41, AToy Story Part | 4 0 sone hassel space hold at oust maine its prt by ondering te appropriate mabe of Garba dolls ot dace diferent yes Sleeping luby dalseoxt ea typeof dol RememnbersProfit = Rewente ~ Cos.) S6 ex talking baby doll cos 82.50 cal and walking baby a) Define the variables, dlls cost $9 each The store cam afford te spend up to S05 for Write the maximum problem in stand form, the dolls. The sleeping aking, axl walling baby dls resell fc} Lutroduce slack variables S850 and S17 each, respectively, The sone wishes to (ch) Setup the intial tableau, 42, RToy Story, Part Il Concerned avout having variety of vay dolls for customers, the manager ofthe toy store From Protem 41 decides oe ordering reyuitements as aby dolls orlered should cha loss The number of sleepin beat most 6 ban the aumnber of talking dolls ordered, und the nunber of talking dolls ordered should be at most 9 rove than the nu lea with these addtional constraints Clothing Steve oivns 9 company that produces mens shies jackets, and pants, each of hich require time for cutting and sewing, One shirt requites 1 oinutes of cutting and 20 nutes cutting and 40 und oie pair of pants requites 20 minutes of col walking dlls, Rework Px 43 cf sewing, one jacket requires 20- mutes minutes of sein cutting and 30 minutes of sewing, Steven has 500 enunutes of cutting time avilable eal day and 0 minute Fis company mikes profit ot S19 foreach shirt, 841 foreach jacket aud $15 oa exh pai his sewing tine pats, Steven wants to masimize it, but can produce at mest 30 ens none day {a} Define the variables, () Write the maximo peoblen in standard form, Lc} Inteadce sack sariables ul) Set up the initial tableau, FruitTrees Susanne ins vldivionaD fut tees 6 use at most 4 clu and wants te plant sone Join maximum revenue. She will suite feet of land and can plant pear al ofpeach cans peach, and plun tees, She wants the combined Tuees aad plum tes tbe no more tha 150, but ae sure that the nual number of peach trees by mote than LD, The table provides addtional information we af pear trees dows mot esceed the {Define the variables {b) Wiite the masizmars problem in standard form. 5.1 The Simplex Tableau: Pivoting 235 {ec} Inteaduce ack variables, (ul) Set up the initial table, Pex i TAR Phim 20 2 12k 45 Source: UUS.D.A, National Agtculture Statistics Service, Mare 2010 Investments. Kani bs pt it among theee investments: (1) a money market account that pas 1.05% ausally, 2) mnt fund that pays 5% anual land (3) acertlicate of Geposit (CD that pass 1.86% annul She wants to masinnize her retur p00 Lo inves and can divide Fou docs not yan to invest tne than Sib between the mutual fund and the CD. bn tion, the antount invested in the CD cannot exceed the “amount invested in the money market by more than $1500 (a) Detine the variables {b) Write the maxinsum problem in standard form, {c1 Intraduce slack variables (4) Setup the initial tableau, Source: moweyratescon publishing ¢ and 15 sales Tiavel Arrangements (esl 10 sales representatives stationed in. Chica representatives stationed in Boston. ‘The company wants 10 seid representatives ti ti dtferent conventions ing held iv in Atlanta and Sezitle. The plo L2 reps tothe Allanta convention and up (08 a Hip ticket fon sinnltanes Tike toh reps to the Seaille convention. 31 Chicago to Atlanta oximately $6, from Chicago to Seattle costs approximately $454, from Boston to Atlant cosisapproninsely S48), and ron Boston to Seale costs approsimately $475, Source: priceline.com) The total ight Costs forthe trips cannot exceed $77 ts ap oo, The company needs idly trom Seattle, from Boston to tes determine bes rman sles representatives sh Chicage to Atlanta from Chicig slant, ana fom Boston to Seat inorder to mavimnize the ‘aumnber of representatives atte conferences, (a) Detine the variables {b) Write the maximum problem in standard form, {c1 Intraduce slack variables aly Setup the initial tableau, ‘Are You Propared? Answers 1, 200 er etost o Fan afo 1 sar tal 2 bao a 4 eogie 236 — Chapter'5._ Linear Programming: Simplex Method 5.2 The Simplex Method: Solving Maximum Problems in Standard Form OBJECTIVES ill Solve maximum problems in standard form using the simplex method (p. 236) (© Determine if a tableau is final, requires additional pivoting, or indicates ro solution (p. 240) Solve applied problems involving maximum problems in standard form (p. 243) Solve Maximum Problems in Standard Form Using the Simplex Method he steps given on pa to solvea maxim problem in standard form. We return to Examples tard 5 of the previous section to help see the rationale foreach st Look back at Example 4 on page 224. The initial simplex tableau given theres repeated in the farlettcoluran, The tableau obtained ater one pivot s given nest, and then the tableau ‘obtained after another pivot, this one being te final tableau, he pivot element is circled. Why did we select this particular entry? Remember, we want to maximize and the objective row tells us P= 3x) -F 4x. We can improy the value of Phy increasing ether X or x:. We'll choose x}. This vorresponds to nd identities the pivot 1 we will eventually leave the ve ist negative entry in the objective row choos column, But we cannot increase x; without bound the constraints 2ej + dey = 120 and 2x; + 2p = 80 Ix, = 0, then x =60 and 4 = 40 so X; = 40.This number { = isthe smallest quotient obtained when the KES (80) is divided by the corresponding entry in the pivot column {2} and identifies the pivot row Why do we pivot again? Now Since xy = and sy = 0,we can improve Py increasing x9 and lettin This corresponds to choosing the first negative entry in ths establishing the pivot column x3. Do you see why the pivot row is si? jective rov Why is this final tableau? Here 140. ive, Pismaximized correspond to no Since 5) = Oands; = O,and their coefficients are ne ands: = 0. The negative coefficients of ative entries in the ebjective ro 5.2. The Simplex Method: Salving Maximum Problems in Standard Form — 237 1 The Simplex Method for Solving a Maximum Problem in Standard Form STEP 1 cae iipestia ste 2 aT SFT he Sew Chose es ne See Peete Tae aT thn Viscount feraer eer ieee a STEP3 Look at the entfies i Nhe itive roe et oenepiine, STOP. Th [ernie Rs ty agement Re eat eee ere ee [hero q STEP 4 (Tianhe cane STEPS Ppt STHS% 5 wud tl peered PRs eou) Maximize subject to the constraints 238 Chapter. Lineer Programming: Simplex Method SOLUTION STEP 1 This isa maximum problem in standard form, To obtain the initial simplex tableau proceed as follows: The objective function is rewritten in the form Po 3x — aes Afier introducing slack variables sy and s, the constraints take the form 30 where he initial simplex tableau is WP oy oss His I STEP 2 ‘The list negative entry in the objective row is —3. The pivot column is column xy. STEP 3 For cach positive entry above the abjective row in the y cof the corresponding RHS entry divided by the positive pivor colunnn, form the quot entry Sere (re 120 0 = 2 80+ 2=40 “The smallest nonnegative value is 40, 50 the pivot raw is row STEP 4 he tableau below shows the pivot element (ci Fivat alumna id the current values HS. Current values 120 Pivot row—» 80 20 0 P 0 ° he pivot element is 2. Next, operations: 1 QR =H RAR HY After pivoting, we have the tableau BV Pox 5 Current values 0 0 1 si 40 " 2) a) 0 Notice that is now a basic variable, replacing 5.2. The Simplex Method: Salving Maximum Problems in Standard Farm — 239 STEP 5 STEP 2 The first negative entry in the objective row is “1. The pivot column is column peat Step. STEP 3 For cach positive entry above the objective row in the pivot column, form the quotient of the corresponding RHS entry by the positive entr (ire 40 10 x 1 Ei 10+ 1=40 EES The smallest nonnegative value is shows the pivot element ci 19 the pivot row is row ys The tableau b Pivor column RHS Currant values Pivot row— HO] = 40 4 n= 40 120 Rtn RK BVP om os) RHS. Curentvalues (3) STEP 5 Repcat STEP? STEP 2 Thereare non ntries in the abjective row so we STOP The maximum value of Pis, pao anil it occurs at x= 5 os . ‘The tableau in Display (3) is called the final tableau because with this tableau the maximum value is foun: 240 Chapter 5 Linear Programming: Simplex Method @ Determine If a Tableau Is Final, Requires Additional Pivoting, or Indicates No Solution Poon Determine whether each tableau below (1) isa final tablean, (Fis, give the solution {2) requires additional pivoting. [If so, dently the pivot element (3) indicates no solution BVP om ms mio 1-0 pli oo ifn wv Px SOLUTION ative entry in the abject The pivot column is colunn X3 ntries in the pivot column are zero or negative, the problem is unbounded and has na solution, (b} ‘The objective row contains no negative entries, so this isa final tableau, The solution is when S49 y= 29 = 0 =O {6} ‘The first nogatve entry inthe objective row is “4. "The pivot column is column x. Since both entries in the pivot column are positive, compute the quotients of the RHS clivided by the corresponding entries inthe pivot column, 60 +1=60 Since 41% < 60,3 is the pivor row and 3 isthe pivot element. . ‘No nowwork pRosiem 5 REE er Sn ORS) Maximize oxy 8 boy SOLUTION Notice that the maximum problem is in standard 5.2 The Simplex Method: Salving Maximum Problems in Standard Form — 241 subject to the constraints ax aed mids bys 9 6a +20 + Sey = 30 sel Sy the constraints take the form where Rewriting the objective function as Pa 6x, — 8 — the initial simplex tableau takes the form Pivot column d BP 4 om ms 8 RIS. Current values RHS Piva row—+ he pivot column is the column containing the first negative entry in the objective row in column xy). The pivot row is obtained by dividing each entry in the RHS column by the corresponding entry in the pivot colunin and selecting the smallest nonnegative quotient. The pivot raw is row s;, The pivot element is 6, which is circled Pivot: LR DRS RH REAR +H ROR HH Current values RSH tp Pivot row» 155 242 Chapter 5 Linear Programming: Simplex Method ative entry, “6, in the ebjective row indieates, ihe vive of Pas improved 1030, burt the further improvementis posible, We determine the next pot element to be which fs 1 Pivot LR = tri 2 R= SR ny R= 5 ROR oe BVP» : Current values 1 his is a final tableau since there are no negative entries in the objective row he solution fs found by looking at the current values of the final tableau, The Frands = 0 . maximum value of Pis F aimed when x ‘Ne nowwork prosten 1s Reenter Sn Ree) Maximize yy + ans + 3 subject to the constraints bat HSE SOLUTION The maximum problem is in stangaré form. Introduce slack variables 51.5355 and 52. The Simplex Method: Salving Maximum Problems in Standard Farm — 243 Since 4 isthe frst row is row si ative entry in the objective row, the pivot column isxy + The pivot Do you see why?) The pivot element is 2, which is circle Pivot 1 Ry= AR bt R= AR or s RNS Currentvalues RHS- 1p naa 3 e=2 2dee paw ‘The pivot element is 2, which is circled Pivot 1. Rs 2R= ' bb aR HY BY Current values 12 me te a 6 5 Solve Applied Problems Involving Maximum Problems in Standard Form* Wermeur bans Mike's Famous Toy" Trucks specializes in making four kinds of toy trucks a delivery truck, dump truck, a garbage truck, and a gasoline truck. Three machines—a metal casting machine, 2 paint spray machine, and a the caging machine—are used in the production of trucks, The maximum time available per week fr each machine i given as metal cast 009 hours paint spray 1806 hours and packaging, 1000 hours. The mein hours each machine works to make exch type of truck and the profi for esch truck are given in Table 1 {a} How many of each type truck should be produced weekly to masimize profit Assume that every truck made is sok (b) Analyze the solution oroantger programming problems S 244 Chapter 5 Linear Programming: Simplex Method TABLE 1 Metal Casting 2 hours hours Paint Spray Thou LS hours Do hours Packaging 05 hour 05 hoor ours Profit S050 S100 the number of delivery trucks, dump trucks, garbage pectively, to be made. It P denotes the protit to be SOLUTION (a) Let ay.¥9,¥%, and a4 dene twucks, and gasoline truck: maximized, we have this prob Maximize PHOS 4 a 4 1306 2 subject t0 the constraints Qu 42 ny 1S + at 2ny = 1800 Wy Fay xy = 1000 oF Qu b 2xy = 4000 This maximum problem is in standard form so we introduce slack variables 55 nd ss,and write the initial simplex tableau. To help in doing the required arithmetic, BV RES Currentyaluas RUS +14 s 100 & 1800 1800 1000 2000 P=0 The pivot element 1 fs circled Pivots =r 2 R= OR bry R= aR brs R= SR > BoP % ss RIES. Currantyalues RAS +e 100) =400 x 1800} m= 18001200 100 900) p=900 5.2. The Simplex Method: Salving Maximum Problems in Standard Form Pivots 1. Ro 265 1 R=tk ty R= Gk + BY Py SRS. Gurren values. RAS 1 rfo 4 o| 1000 se nfo 2 ° este 1800 o i 1| 400 600 é T [1 § 0 P= 1200 Div te Rs nada soe wom evalnes RHS +14 : ‘ ot 59 oH 6 p to fo So wlo 44 x + 600 1200 aha pir go 1700} p=t700 The pivot element 1 Pivots Ree Frets R= BP oy & x 1 4 This isa final tableau. Xo SS RS Curentvalves 51-2000, x1=800 200 1800 P= 1809 he maximum profits P = $1900, obtained when 200 300 246 Chapter Lineer Programming: Simplex Method {b) Observe that for maximum profit, ne delivery trucks andl no dump trucks are produced, The practical considerations of this are that delivery trucks and dump trucksare ether too costly to produce or too litle profitis being gained from their sae. Since the slack variable 5; hasa value of 2000 for maximum Pend, since represents the number of hours the metal casting machine isidle, it may be possible to release this machine for other duties. Also note that both the paint spray and packaging machines are operating at Tull capacity. This means that to inerease productivity, more paint spray and packaging capacity is required . ‘No wowwork prostem 2, Analyzing the Simplex Method ‘To justify some ofthe stepsin the simplex method, we analyze more carefully Example 5 In the inital tableau, the objective function is P= 05x +x 4 15x IF we set the nonbs = sy = xy = 0, ne get the profit P = 0. OF course this prot can easily be improved by manufacturing some trucks. We decide to increase xy. the number of delivery trucks. But by how much can xy be increased? ‘That is determined by the constraints eferringto Table L we se that each delivery truck takes 2 hours to cast, | hovr to paint, sic variables equal to zero, %) = and 5 hour fo package, The limited time available for each task restricts the number of ag truck that ean be made, So we look atthe constraints for prosiveing delivery trucks: metal casting: 4000 hours available at 2 hours per track mean «an be manufactures paint spraying: 1800 hours available at | hour per truck means 1800 delivery trucks 5 2000 delivery trucks can be manufactured packaging: 1000 hours available at hour per truck means 2000 delivery trucks can bbe manufactured hee paint spraying constraint indicates that we ean make st most 1800 delivery trucks, So we choose row sto be the pivet row. This will increase X} to as large an amount as possible, namely 1800. After pivoting, the tableaw indicates that if we make 1 = 1800 delivery trucks, but no other trucks (= 5 = y= 0), we will make « profit P of $900. However, the negative entries in the objective rovr indicate that the profit can be improved Each tetation ofthe simplex method consistsef choosing an “entering” variable from the nenbasie variables and a “leaving” variable from the basic vari criterion so that the vahte of the objective function is not decreased. The iterative procedure stops when no variable can enter from the nonbasie variables that improves profit. This fsillustrated in the tina tableau, which shows the snaxinum profits obtained by producing x, = 200 garbage trucks and xy = 800 gasoline trucks, he reasoning behind the simplex method for standard maximum problemsis not too complicated, and the provess of “moving to ab by following the rules Briefly, the pivoting strategy works like ths: sing a sslective ter solution” is made easy simply Othe choice of the pivot column forces us to pivot a variable that improves the value of the objective function. ‘The choice of the pivot row prevents us from making this variable foo large to be feasible 5.2. The Simplex Method: Salving Maximum Problems in Standard Farm — 247 Geometry of the Simplex Method ‘The maximum value of the objective function, if it exists, will occur at one of the corner points of the set of feasible paints. The simplex metiod is designed to. move from comer point to corner paint of the set of feasible points, at each stage improving the value of the objective function until a solution is found or no solution is indicated. More precisely, th etry behind the simplex method is outlined below 1. Agiven tableau corresponds to a corner point of the set of feasible points, 2. The operation of pivoting moves us to anadjacent corner point, where the objective function has a value at least as largeas it did atthe previous corner point 3. The process continues untl the final tableau is reached—which produc point that maximizes the objective function—or no solution is indicated ‘The Unbounded Case Sa arin our discussion, it has always been possible to contimate ta choose pivat elements until the problem hasbeen solved. Bus it may urn our thatall he entriesin 2 pivat column ofa tablean are 0 or negative at some sage. In this case, the problem isi maximum solution does not exist For example, consider the initial simplex tableau BP oxy x randed nd a he pivot element is 1 in colu Afler pivoting, the tableau becomes BVP omy om 4S RES FIGURE 4 en mio 0 Pr os Now the only negative entry in the objective row is ~3 in column xy. Notice that it is impossible to choose « pivot element in. that calump, To see why; we go back tothe initial tableau. The objective function is P = 2x; and the constraints are: iy 4X S Dx xy = 20m; = Ovand ay = 0. The set of feasible points, shown in Figure 4, is unbounded and the objective function is heading in the direction of the unbounsedness ‘Summary of the Simplex Method “The general procedure for solving a maximum problem in standard form using the simplex method can be outlined as follows: 4. The maximum pro Maximize lems stated in standard form as Prax ten subject t0 the constraints hy + bees: ah a 248 Chapter8. Linear Programming: Simplex Method 2. Introxtuce slack variabless;y 35+» +» $90 tha the constraints take the form of equalities: Wixi tb bate xe + mt aX + ty =m =O, S208 = 0.55 8. Write the objective function in the form BP 5. Pivot until {a} All the entries in the objective row are nonnegative. ‘This isa final tableau trom ‘which a solution can be read, Or until (byhe pivot column is a column whose entries are n problem is unbounded and there is no solution ative or zero, ln this ease the The flowshart in Figure 5 illustrates the steps to be used in solving standard maximum linear programming problems, FIGURE 5 prthan'hs rcpt, NOy pola scasiohy Tne "eat yes yrs Isso enter stipsmplarian “ sae thse cite SO, cr ai ves S Get vo ves 5.2. The Simplex Method: Salving Maximum Problems in Standard Farm — 249 USING TECHNOLOGY Cariac ae SOLUTION STEP 1 Use Excel te solve Exam Maximize P= 055, + +13 + 2ey subject to the constraints Dey + 25a Bes + Ay = 4000 Mit 1Sey + xy + Ixy = 1800 05x, +0.5r) + xy + xy = 1000 =0 “=o x Set up an Excel spreadsheet containing the variables, the objective function, and the constraints 3s follows iN Ej Ge 1. Variables 2. Delivery Truck xy [0 3 Dump Truck [0 4, Garbage Tucks x5 [0 5. Gasoline Truck ay [0 7. Objective 9, Maximize Profit = 0.5°B2 4 PBS 15° BA + 2°BS TI. Constrai 12, Amount Used ‘Maxim 13, Metal Casting’ 2°21 25B3 4 2B 4 "BS 14, Paint Spray B13 15, Packaging = x1 should obtain ‘Dump Tuck, 32 ° (Gabage Tuck, x 0 (Geena Tuck ° z 3 4 5 a C nl jective 1 Waxinize Prost p 10 12 Atpunt Used Maru 13 sal Casing oa 1 Pie Sixay ‘0 45 Peckagng ojo 16 250 Chapter8 Linear Programming: Simplex Method STEP 2 STEP 3 STEP 4 Click on Tools and then Solver to get the screen below. + Sot Target Cell + Equal To: + By Changing Cells:The variable eels + Subject to the Constraints: The system of inequalities Input Set Target Cell B9s Equal To: Max: By Changing Cells: B2 to BS. Just cliek on the appropriate eels the cel ofthe optimization function, ows you to maximize or minimize the objective function. seragecet [ioe J] Fade: Fe Com Cie Pg, | Add the constraints, + Click cursor into Subject to the Constraints entry bex. + Press Add button. You should see: + Cell Reference is the cell containing the formula Constraint is the cell containing the masimuum, You must also enter the constrain that the variables must be greater than or qual tozer. To enter the Metal Casting constraint, put the cursor in the Cell Reference box and in the middle box. + Type in zero in the constraint box, ee ax Giesesess f= ele + When you are finished, lick on OK. STEP 5 Your Solver box showk’ look lise this STEP 7 Find the solution: Click on Solve and highlight Answer. 252 Chapter 5 Linear Programming: Simplex Method he solution is inthe Excel spreadsheet below Variables Delivery Truck, ° Dump Truck, xs 0 Garbage Truck, % zuu Gasoline Truck, » 800 Objective Maxime From 1800 Constraints Amount Used — Maximum Metal Casting 2000 4000 Paint Spray 4200 1800 Packaging 4000 1000 ‘To get a maximum profit of $1909, you should produce 0 delivery trucks, 0 dum trucks, cline trucks. Only 2000 hours of metal cat are used. The other constraints are used at full capacity ‘Noy nowwork PROBLEM 29 USING EXCEL ¢ trucks, and 800 garb, Sarna Concepts and Vocabulary. 1, True or False‘the pivot chmeat is sometimes located in the 2, True or False Whe a tableau is obtained that has waly objective to nonnegative entries inthe objective oi final table 3. Trueor False a tableau isobtained and thepivelcolurm bus 4, For a maximum problem an standard form, the pivot entries that are positive or ger, the atasimaun pevblemn will is lovated by selecting the ist neyative entry in Thave no solution the objective to, Skill Bui 4 Probie satements i W Lo) te sequites, {} Ie iicutes no Ne BBV P 6. cn 13, Maximize Paix, +7 subject to he constants dye te ee me een 16. Maximize Pain tay: subject to the constraints 18. Masiniize Pan tates subject to the constants yrs 22. Masiinize P= oy HS HIN subjett to the constraints dy $y = Byte MEL a. 44, Maximize subject to the constraints die es ede 7. Maximize subject to the constraints 2s 20. Maximize P= dy + 2y + Say subject tothe constraints nds + y=M wyotin nee eo 23. Maximize Pex, tata tay subject tothe constraints The Simplex Method: Salving Maximum Problems in Standard Farm — 253 HS 15, Maximize subjett to he constraints ny 426 ait x 18. Maximize Pain tx: subject to the constraints ib osd hese meso wee 21. Maximize bay + aks subject to he constraints ay eo med 24, Masiinize Peony tay tay subject to the constraints ays de bMS sis ba ES nase tine 254 Chapter8 Linear Programming: Simplex Method 25. Masiinize 26, Masinive dy bans Payot oy subjett to he constraints subject to the constraints wt dest= 20 ays hax sd 60 wy dy} meson 2. Masinize 28, Slasimive Pent Pexsien subjett to he constraints subject to the constraints sans + oxy = 2 By sha Hay = Oxy 24 2x By 40 fey + da 4 sd Application: 29, Process Utilization A jeu wnanutisturer muses three pes of snpany’ as available 60 unsts of rubber, 800 units of jeans ca of which goes throu ree manu toring pases plastic, ane! 400 units of alumnioumn, The company anakes a ‘cutting, sewing. and finishing, “Le aumber of minutes exch type profit of $4, $3, and S2 om toys A, Be amd C, respecte ‘of prodiael roguirs in exc of the three phases is f Assuring all toys mamutictured ean be sold, deterring production order so that profit is. maximum, {) Pormulatea linewrproramuming problem that models the problem given above, Be suse to wei al variables use Uo) Solve the Linear programming problem, 1 18 24 12 fo) Analyze the solution. S200 minutes of cutting tine, 60010 aninates of 31. Maximizing Profit Three kinds of wrapping paper ane sold sewing time, and 2200 auinutes of tnishing time each day. The nppany can sell ll ce jeans i mtakes and makes a pric of ‘om each Jean [, $4.80 on cach fean TI, and 86 on each Tear TIL. What number of jeans in each cate should be made ‘enn that anal the fo) Posmubites tineirprogeann problem Ito) Solve the nica programming problems, Lc} Analyze the solution, 30, Process Utilization ‘company niamutietures three types of loys A,B, and C. Each requires rubber, plastic, andl alumninun islisted below Toy i ‘ B 1 2 E 1 2 verge) 10 minutes to sell, and costs $0.30 t customer The hola weapping paper coss $8.4 rol, takes 15 rnimates to sl ands ef with the customer atthe tine of sale 1a roll takes 12 minutes the customer, Busi The wedding wrap to sal Jelver 1 finan, the: Reis al the three types oy at mst a tal of 8D ours sel vee sell up fo S80 worth rapping paper and ceweed 875, The Team members times cin be ses nol 1 curing the weexbong fund-raiser. ‘Their pris ate $1 for etc oll ofbirlhdey paper and holiday paper they selland 82 foreach rll of weading paper What combination of sales oF exh type paper til ead fo a {) Pormulatea linewprogramming problem that models the probe given above. Besure to sentiy al variables use Uo) Solve the Linear programming problem, hss 5.2. The Simplex Method: Salving Maximum Problems in Standard Form 32 Resource Allocations lisse hospital chasis sts sursial ‘operations inte thiee categories according to their length and changes fee of S120, SIS and $2400, spectively, reach of the categories, The average time of the operations i the Hhree catagories fs nunutes, | os, an 2 hous, respectively The hospial hus four operating rooms, each of which can be used for 10 oss per da Hrthe total number of cannot exceed 6), haw many of each type should the hospital erations schedule to mxintize is revenues? ‘enn that anal the ‘dental variables used [) Formulate a line progeanming problem given above, Be sure Io) Solve the Lina programming problem fe) Analyze the solution, 33, Mixture Problem Nutt’s Nut Conpsny has pemuts, 100 pounds of pecans, and 50 pounds of cashews om 0 pounds of Inaad. They package thiee types of S-pound caus of nuts Can [contains 3 pouncis of peanuts, | pound of pecans, and 1 pound of cashews; Can IL contains & pounals of peanuts, + pound of pecans, and 4 pound of cashews; and Can I contains Foy Can L S24 foe any cans of ech Kind should pounds of peanuts The sll (Can Ulan $21 for Can I He bbe made to maximize evenne? ‘enn that anal the ideal variables use J) Posmubitea linerprogeann problem given above, Be sure tb So the linear prograrmmying peublent Le} Analyze theslution, 34,Maximizing Profit One of the methods used bythe Alexandes Company to separate copper lead, and zine fon ‘ores isthe Hatin separation cof three steps oiling, ising, and separatio mast be applies tor 22, and L hows espectively 288, This process consists These steps produce uid hous respective to produce Lunt Lunt of copper: 2 8 and 1 1 oni of 7inc of ead nd. 3 hours, respectively, to produce Tae oiling amd separation phases of the pperation for a riexirnum ef 10 hours a mining phase can be in operation far cr amaximun of LY ours aay, The Alexander Company ratkes per, $40 per unit of lead, and pprscess cam be i Hit of S45 per unit of c hese metas is should be use af the latation process te achieve the sold monde ay Inyghest profit Ws many units of cach meta ‘enn that anal the ‘dental variables used () Formulate line programing b shove. Be sure Io) Solve the Linear programming problem ‘Nem give Lc} Analyze the solution, A Toy Story, Part | (lefer te Pwubleny ln Section 3.1.1. toy store as shelf space to hold at most ot aby dels of thzee cltfeeut types. Slesping baby dolls cost $6 each talking baby Sa 1 25 dlolls cost $7.30 gach, and walking baby dolls cost $9 each, The stone can aff to spend up te $405 tor the dolls, Te sleeping. talking, and walking baby dls resel for cand SI7 each, zespectively. The stone wishes to. masianize sts profit by onder i doll. Remembs shu, opriate number of each type Prolit = Revenue = Cost Ga) How many of etch type of doll should be ordered 16 hat is the maxismum prot 6 37 AToy Story, Part Il Solve Probes 12 in Section 5.1 Clothing (Relor Problem 13 in Sostiow 2¥ company that produces men’s shirts, jackets, and pants and seving. One sir reuizes 10 punues of cutting amd 20 nunutes of sewing, one jacket requites 2 culling and 40) migutes of tnd one pair of pants requites 20 minutes of cutting ) Steven owns cael of which rogue Fine for cuting cand 30 minutes inutes of cut serving, Steven I ine waiale ea dt and 900 inte seg ie His company makes a profit of S1¥ for each shit, $3 cael jackal and S15 on exch pair of ants, Seven wants to ronximnize his proit, but cam produce at most 30 items in cone day {4} Delerorine bows many of each clothing em shoul te mace fo) What i the muaxinuan dy pros 38, Maximizing Profit wroweht iron misnufacturer proces chandelirs,stairaitngs and gate laces eac of hits mast hhaveits pars blacksithed, assemble, and crated for shipping Fach chardkier requires ® hours to black, § fours asstnble, and 4.1 hour to erate and returns 2 profit of S10, Thours to blacksmith te and rturas a profit pat to blackinith, 1 bour an retusa prot of 83. Each set of stair railings requites ‘Shours to assemble, al 4 hour ‘of $25, Each gate latch requires tes assetible an 1 hou tera and 120 hours available id crating spectively duct should be manulactuted 1 he mawafichurer bas 8,00, 18 sseuly for Blacksmithing, asmbhi How many units of each pe rnnimize profit [) Formulates linearp au {c) Analyze the solution, samning problem that models the problem given above, Be sure to idem all variables used dem, sve the linear pear arn 39, Mixture ‘Ihe Maatlnu ol efincty lends high ad less octane sztsoline imo three intermediate grades: regula premium snd super premium, The regular grade consis of £09 hgh foctane- and 10% ous octane, the premium consists af 708 high ictane, and the sup oasis of 80% high ovtane and 20% lows octane, The sompany has avaiable 140,00 wae and 120,000 gallons of low octane bul can mix. only: 225.000 Resuliy per gallon, and super premiamt sels for $48 per gallon. Howe tame and 30% low gallons of bi los ws sells for $2.80 per gallon, premiuan sells fo $297 256 Chapter 5 Linear Programming: Simplex Method nay gallons ofeach grade should the vosmpany mix in order ‘enn hat enadels th fo) Postale linear shove, Be sure the near progranning problem, Lo. Mixture Resonk Problems 39 under the that the combined tral nuatber of lution, sional assamption aleas proaluced by the refinery canmot excel 200,000 glans, 41. Investment financial consultant has al ypost SSO to average yields Tor sto bonds aes shoul anvest in eal secu’ to mayamize the ks, corporate bonds, aad muuicipal and 3%, respectively, Determine hos mc, her investments, Shes decided shat her inesanen in socks should not eycest Ina her finals and chat the dtferenes between her investment in corporate bonds aad anunisipal bonds shoul not be mone than 208 her invetnen ta) Foca near rogram prble that nul th problemen ave, Be sure Lit) seve the nea Laz. Investment Rework Psubien 41 under the assumption that Pans rivested in stocks and that the difference etseen her investment in corporate and naaicipal bonds shoul wot be mote than 1D" 43, Crop Planning farmer hus a stitable for cultivating crops of cost or cultivating soybeans, command whet ake SHO, § $8) peracre, respective The fae asa maxinnary of SIKGN) rvalable for land cultivation, Soybeans, cor, and wheat 20,20, anal 15 hours per ate of Labor respectively, and there inaxianamof Le hours of labor available the Fier expects leumtke tpn of S70, $9, and $5 und seat, respectively, how many’ ares of each ctep should he plant in ots to masiznize his pet fo) Posmatlate linear ‘enn that anal the problem given above. Be sure to identi a Lib Solve the lineae ps 44, Crop Planning Repeat Problem 43 wil the Cost $3) 4 sa $12,000 Hours: w 3a Profit 3 $4 si (9) Formulates linear programming problem hat models the snigiven ae, Be sure to ideiy all variables w L4s.Fruit Tees. (Refer « Problem 44 in Seetion 5.1 Lib Solve the lineup {c)) Compare and contrast the problemand solution with that TV manufacturer has warehouse facilities for storing ils 52-inch color TVs in Chicago, New York, and Denver: Each mont the city of Auta is shippe at ‘nost Zour hunved 52-inch'TWS, The cost of transport TV to Atuata feora Chic $50, $19, and S80, respective foe pack S200 is allocated cach month for transportation costs 0, News York, and Dene flee ost labor requized averazes $6, S% and Sf, eespectivel Suppose und S400 fs allocated for labor costs, [the profit ow exc TV stored in Chicago S70, in New York is $80 and i Denver is $10, how should rnonthly shipping atrangemeuts be (1) Formulates linewprosramuming problem that models the problem given above, Be suse to wlemiy al variables use Lib) Sole the Finear pre analyze the solution foisns an orlaad an 13,504 square feet of land and cam plan plum und plum Hees to be te avore than 150 but also wants 10 ants to phint sexe additional fruit see, She will use pear, peach, and ss, She wants the combined total of peach trees raunber of peach tees by ore than Ib (a) Determine how many of eac typeof tee Suzanne should plant and the maxinnen revenue eat 120 2 1 Peach 450 4 Tas Phim 20 Tok Source: Virginia Conpecatve Extension or mse tre Pc ant 1 Kanibsup 1 $2540 19 invest lev taavinize be e nd can divide it amg t sm: (1) a nwney market account that pa 1.059% annually, (2/4 mutual fund that pays 3% annually and G HT leposit (CD) that pays 1.66% army. She sn 10 invest mote than S10 tise the rma Fay and the CD. In ation, the anno inseste inthe CD cannot esceed the amwunt vested in the money matket by snoce than S150 (a) Determine bow much Kami shoul put inte each {(b) What isthe nuaxiemumn returot Source: moncyratesson 53. Solving Minimum Problems Using the Duality Principle 257 1 48. Travel Arrangements (es to Problem 46 in Section 5.1.8 tesibook publishing company has 10 sales representative stationed in Chicago and 15 sales representatines stationed in Beston, The company wants to send representatives to ts slffenent conventions being hed simultaneously in AUanta and Seale. The company would Tike te Mya to" 12 reps te the Allunta conention and upto reps to the Seale convention. & roumibtrip ticket from Chieage to Atlanta costs approsimaely S108 rom Chicago to Seatle cents approsinnately S150, fron Beaton to Atlanta costs approsiastely $100, an from Boston total Milt costs fr the eps cannot esc {a} Determine hows neany sles ropeesentative should 2 fora Chicago to Atlanta, feom Chicago to Seat, from Boston te nuuber of repeesemtaives at the conference, to) What isthe maxionam nusaber of cops that will attend the 5.3. solving Minimum Problems Using the Duality Principle PREPARING FOR THIS SECTION Before gotting started, roview the following: ‘© Matrix Algebra (Section 3.1. pp. 106-116) ‘Ny wow WoRKTHE ‘ARE YOU PREPARED?’ PROBLEMS ON PAGE 266 OBJECTIVES Gi Determine a minimum problem is in standard form (p. 258) Obtain the dual problem of a minimum problem in standard form (p. 259) © Solve a minimum problem in standard form using the duality principle (p. 261) A linear programming problem in which the objective function is to be minimize is relerred lo 98 aminimum problem, There are two ways to solve minimum problems: [J 1. Change the minimum problem to a maximum problem by using the fact that ininimizingis the same as maximizing P= —2, 2. Apply the duality principle which states that the solution of the minimus problem can be obtained by finding the solution of the dual program, a maximum problem, Notice that both of these methodologies require solving a certain maximum problem. The choice of which to use usualy is determined by the size of the initia tableau of the maxianun: problem —the smaller size being preferable We discuss the duality principle inthis section and the second methodology in the next section Theorem [J Duality Principle Ia mininnum problem hasa solution, the minimum value of the objective funtion of the minimum problem equals the maximum value of the objective function of the dual problem, a maximum problem, 258 Chapter 5. Linear Programming: Simplex Method our purpose in this chapter is to introduce the simplex method, we shall only pply the duality principle toa particular class of minimum problems, those that are in standard foray Definition [J Standard Form ofa Minimum Problem A minimum problem i said to be in standard form provided the following three conditions are met CONDITION 1 All the variables are nonnegative CONDITION 2 All other constraints are written as linear ex greater than or equal fo a constant CONDITION 3 The objective function is expressed ass linear expression with nonnegative coefficienss sions that are GW _Determine a Minimum Problem Is in Standard Form Sinn Lar mS ia Determine which of the following minimum problems are in standard form. PROBLEM (a) Minimize SOLUTION (a) Since Conditions 1, 2, and 3 are : va met, this mininsum problen is in oar se standaed form sulbject to the constraints y+ ME PROBLEM (b} Minimize SOLUTION (b} Conditions 1 and 2 are met, but Condition 3 is not met, since the Ca ayy tas coctlicient of x2 in the objective function is negative. This minimum subject to the constraints problent isnot in standard form wimtme? xtatyee 53. Solving Minimum Problems Using the Duality Principle 259 ditions | and 3 are mel, but PROBLEM (c} Minimize SOLUTION (c) « Condition 2 is not met, since the Cay tx bx first constraint subject to the constraints xy ae bas is not written with a = sign, The minimum problem as stated is not * w= in standard form, Notice, however, x =0 0 0 that by tnulliphns by mi ne can write this constraint as mx) o Be - 2 \Weitten in this way, the mininuny problem isin standard fort, PROBLEM (d) Minimize SOLUTION (d) Conditions 1,2, and 3 are each me 5 so this minimum problem is i Cam+x standard form, subject to the constraints mx te tay ‘Noy nowwork PRoBLEM » (@I Obtain the Dual Problem of a Minimum Problem in Standard Form To apply the Duality Principle, we need to find the: een Rue tears eee rec See Obtain she dual p Minimize >the minimum problem. Jem of the following minimum problem: 300%; + 480%) subject to the constraints SOLUTION First notice that the minimum pro that re Constraint: 14 + Sty = 025 Constraint 214 + 2ip = 0.85 Objective function: C= 30014 + 4801» 260 Chapter 5 Linear Programming: Simplex Method Now form the matrix that has as columns the rows of the abave matrix by thi column 1 above and writing it as raw 1 below taking colunmn 2 above and writing it as row 2 below, and taking caliman 3 above and wisi 12 [300 025 045] 0. “This matrix is called the transpose of the fist motrin From this transpose matrix, create the follwing maximum probleny Maximia subject to the conditions his maximum problem is thedual of the given minimum problem, 7 he steps to use for obtaining the dual problem are listed below OF Steps for Obtaining the Dual Problem STEP 1 Write the minisnum problem in standard form, STEP 2. Construct a matrix that represents the constraints and the objective funstion, placing the objective fustction in the bottom row. STEP. Interchange the rows an columns to form the matrix ofthe dual problem, STEP 4 ‘Translate this matrix into a maximum problem in standard form, Poe eats ee ce See eo Find the dual of the following minimum problem: Minimize C= 2x by subject to the constraints Qt = 6 Kydd ed wt oes m=O m=0 SOLUTION STEP 1 The minimum problem isin standard form, STEP 2. The matrix that represents th straints and the objective function is 2 1 [6] constrain: am tte = 6 1 2]4 | Constraine: in + 21m = 4 Constraint ty 1 1p = 53. Solving Minimum Problems Using the Duality Principle 261 STEP 3 Interchanging rows and columns, we obtain the matrix 127] Constraint: 24; + yp + yes 2 3 unction P= By | dys + 5s Il 12 113] Constrsine y+ ays + ¥ 1 s|o} o STEP 4 This matrix represents the following masimum prob Maximize P= oy 4 yy subject t0 the constraints This maximum problem is in standard form and is the dual problem of the minimum problem . Some observations about Exariple 3 4. The variables (x13) of the minimum pro cual problem (44) lem are different from the variables of its, 2, ‘The minisnum problem has three constraints and two variables, while the dual problem has two constraints and three variables. (In general if a minimum problem has arable, its dual problems wil ba constraints andl sevariables, 3. The inequalities defining the constraints are = for the minimum problern and. = for the maximum problem 4, Since the coeiicients in the objective function toe minimiz problem has nonnegative nunibers tothe right of the = d are positive, the dual 5. We follow the custom of denoting an obiective function by C (for Cast, iF i is tb ininaizeds and P (For Profi if itis to be maxims \o Nowwork PROBLEM 15 Solve a Minimum Problem in Standard Form Using the Duality Principle Jn Examples 2 and 3, notice thatthe dul of a minimum problem in standard form i « maximum problem in standard form so is can be salved by using techniques discussed in the previous section enema sen Standard Form by Using the Duality Principle Continue with Example 8 anal solve the maximum problem using the simplex method Obtain the solution of the mininsam problem, 262 Chapter 5 Linear Programming: Simplex Method SOLUTION. ‘The maxisnum probler obtained in Esarnple 3 (STEP 4) is repeated below Maximize P= oy 4 dys + 5p subject to the constraints Introduce slack variables 5 and sto obtain We pb MDS a3. he objective function Pis written as P — 64 = 4y; — 545 = 0. he initial simplex tableau is he pivot element, 2s ezcled ee WP oe ow 8 Currentvalues RAS.» yo ot 3 wet 2 9 i =2 4 == P86 1 Pivot: 1. LR=-SRtn RRs e BOP oy 6S RES Currontyalues AHS + ys 53. Solving Minimum Problems Using the Duality Principle 263 L be pivot element $ is circled Pivot: 1. Rp = 3p 2 Re the pivot element 1 is circled Pivot R= 2 RRA Current values yee ‘This isa final tableau, so an optimal solution has been found. We read from it that the solution to the maximum problen is PH y= he duality principle states thatthe mtinimem value of the cbjective funtion ithe

You might also like