You are on page 1of 15

# REVISED

M09_REND6289_10_IM_C09.QXD 5/12/08 12:01 PM Page 115

C H A P T E R

Linear Programming: The Simplex Method

9
0 0 9 0

TEACHING SUGGESTIONS
Teaching Suggestion 9.1: Meaning of Slack Variables. Slack variables have an important physical interpretation and represent a valuable commodity, such as unused labor, machine time, money, space, and so forth. Teaching Suggestion 9.2: Initial Solutions to LP Problems. Explain that all initial solutions begin with X1 ϭ 0, X2 ϭ 0 (that is, the real variables set to zero), and the slacks are the variables with nonzero values. Variables with values of zero are called nonbasic and those with nonzero values are said to be basic. Teaching Suggestion 9.3: Substitution Rates in a Simplex Tableau. Perhaps the most confusing pieces of information to interpret in a simplex tableau are “substitution rates.” These numbers should be explained very clearly for the ﬁrst tableau because they will have a clear physical meaning. Warn the students that in subsequent tableaus the interpretation is the same but will not be as clear because we are dealing with marginal rates of substitution. Teaching Suggestion 9.4: Hand Calculations in a Simplex Tableau. It is almost impossible to walk through even a small simplex problem (two variables, two constraints) without making at least one arithmetic error. This can be maddening for students who know what the correct solution should be but can’t reach it. We suggest two tips: 1. Encourage students to also solve the assigned problem by computer and to request the detailed simplex output. They can now check their work at each iteration. 2. Stress the importance of interpreting the numbers in the tableau at each iteration. The 0s and 1s in the columns of the variables in the solutions are arithmetic checks and balances at each step. Teaching Suggestion 9.5: Infeasibility Is a Major Problem in Large LP Problems. As we noted in Teaching Suggestion 7.6, students should be aware that infeasibility commonly arises in large, real-world-sized problems. This chapter deals with how to spot the problem (and is very straightforward), but the real issue is how to correct the improper formulation. This is often a management issue.

1st Iteration
Cj l b Solution Mix S1 S2 Zj Cj Ϫ Zj 3 X1 1 1 0 3 9 X2 4 2 0 9 0 S1 1 0 0 0 0 S2 0 1 0 0 Quantity 24 16 0

2nd Iteration
Cj l b Solution Mix X2 S2 Zj Cj Ϫ Zj 3 X1 ⁄4 1 ⁄2
1 9 3

9 X2 1 0 9 0

0 S1 ⁄4 Ϫ1⁄2
1 9 ⁄4 Ϫ9⁄4

0 S2 0 1 0 0

Quantity 6 4 54

⁄4 ⁄4

This is not an optimum solution since the X1 column contains a positive value. More proﬁt remains (\$C\v per #1). 3rd/Final Iteration
Cj l Solution b Mix 9 3 X2 X1 Zj Cj Ϫ Zj 3 X1 0 1 3 0 9 X2 1 0 9 0 0 S1 ⁄2 Ϫ13⁄2
1

0 S2 Ϫ ⁄2 23⁄2
1

Quantity 4 8 60

⁄2 Ϫ ⁄2
3 3

⁄2 Ϫ3⁄2
3

This is an optimum solution since there are no positive values in the Cj Ϫ Zj row. This says to make 4 of item #2 and 8 of item #1 to get a proﬁt of \$60. Alternative Example 9.2: Set up an initial simplex tableau, given the following two constraints and objective function: Minimize Z ϭ 8X1 ϩ 6X2 Subject to: 2X1 ϩ 4X2 ജ 8 3X1 ϩ 2X2 ജ 6 The constraints and objective function may be rewritten as: Minimize ϭ 8X1 ϩ 6X2 ϩ 0S1 ϩ 0S2 ϩ MA1 ϩ MA2 2X1 ϩ 4X2 Ϫ 1S1 ϩ 0S2 ϩ 1A1 ϩ 0A2 ϭ 8 3X1 ϩ 2X2 ϩ 0S1 Ϫ 1S2 ϩ 0A1 ϩ 1A2 ϭ 6

ALTERNATIVE EXAMPLES
Alternative Example 9.1: Simplex Solution to Alternative Example 7.1 (see Chapter 7 of Solutions Manual for formulation and graphical solution).

115

000 Variable x1 x2 Constraint C1 C2 Value 8.000 0.000 0.000 0.3: Referring back to Hal.000 0.4: Levine Micros assembles both laptop and desktop personal computers. in Alternative Example 7. desktops assembled daily Right-Hand-Side Ranges Lower Constraints C1 C2 Limit 16. Each laptop yields \$160 in proﬁt.500 A minimal.000 Current Values 3.000 Upper Limit 32.000 Objective Coefﬁcient Ranges Lower Variables x1 x2 Limit 2.000 x2 1.000 0.000 24. optimum cost of 17 can be achieved by using 1 of a type #1 and C\x of a type #2.000 Slack/Surplus 0.500 Ϫ1.REVISED M09_REND6289_10_IM_C09.000 Allowable Decrease 0.500 3.000 4. we can perform a variety of sensitivity analyses on this solution.000 Basis x2 x1 Zj Cj Ϫ Zj Final Optimal Solution Z ϭ 60.000 3.250 6.000 8. each desktop \$200. laptops assembled daily X2 ϭ no.000 Shadow Price 1.000 8.000 4.000 Current Values 24. X2 ϭ 4.500 Ϫ1.500 1. Alternative Example 9.000 3.000 Allowable Decrease 8.500 Ϫ1. The ﬁrm’s LP primal is: Maximize proﬁt ϭ \$160X1 ϩ \$200X2 subject to: 1X1 ϩ 2X2 р 20 labor hours Upper Limit 4.000 .000 9X1 ϩ 9X2 р 108 RAM chips 12X1 ϩ 6X2 р \$120 royalty fees where X1 ϭ no.500 12.000 1.000 s2 Ϫ0.000 9.000 Reduced Cost 0. Alternative Example 9.000 60.000 x1 0.000 1.000 12.750 3.000 Allowable Increase 1.000 Allowable Increase 8.000 9.1. Using software (see the printout to the left).500 0.500 Bi 4.000 1.500 2.000 3.QXD 5/12/08 12:01 PM Page 116 116 CHAPTER 9 LINEAR PROGRAMMING: THE SIMPLEX METHOD The ﬁrst tableau would be: Cj l b M M Solution Mix A1 A2 Zj Cj Ϫ Zj 8 X1 2 3 5M 8 Ϫ 5M 6 X2 4 2 6M 6 Ϫ 6M 0 S1 Ϫ1 0 ϪM M 0 S2 0 Ϫ1 ϪM M M A1 1 0 M 0 M A2 0 1 M 0 Quantity 8 6 14M The second tableau: Cj l b 6 M Solution Mix X2 A2 Zj Cj Ϫ Zj 8 X1 1 6 X2 1 0 6 0 0 S1 Ϫ1⁄4 1 ⁄2 Ϫ3⁄2 ϩ 1⁄2M 3 0 S2 0 Ϫ1 ϪM M 3 M A1 ⁄4 Ϫ1⁄2 1 M A2 0 1 M 0 Quantity 2 2 12 ϩ 2M ⁄2 2 3 ϩ 2M 5 Ϫ 2M ⁄2 Ϫ 1⁄2M 3 3 ⁄2 Ϫ ⁄2M 1 Ϫ ⁄2 ϩ ⁄2M The third and ﬁnal tableau: Cj l Solution b Mix 6 8 X2 X1 Zj Cj Ϫ Zj 8 X1 0 1 8 0 6 X2 1 0 6 0 0 S1 Ϫ3⁄8 1 ⁄4 Ϫ1⁄4 1 0 S2 ⁄4 Ϫ1⁄2 1 M A1 ⁄8 Ϫ1⁄4 3 1 M A2 Ϫ1⁄4 1 ⁄2 5 Quantity 3 ⁄2 1 Ϫ5⁄2 5 ⁄4 1 ⁄2 5 17 ⁄4 ⁄2 M Ϫ ⁄4 M Ϫ ⁄2 Printout for Alternate Example 9-3 Simplex Tableau : 2 \Cj Cb\ 9. Proﬁt ϭ \$60.000 16. we had a formulation of: Maximize Proﬁt ϭ \$3X1 ϩ \$9X2 Subject to: 1X1 ϩ 4X2 р 24 clay 1X1 ϩ 2X2 р 16 glaze where X1 ϭ small vases made X2 ϭ large vases made The optimal solution was X1 ϭ 8.000 s1 0.000 9.000 0.

The purpose of the simplex method is to ﬁnd the optimal solution to LP problems in a systematic and efﬁcient manner. The procedures are described in detail in Section 9. respectively. A shadow price is the value of one additional unit of a scarce resource. with a \$2. It should be noted that while there will be eight basic variables. In the primal. 9-6. have a zero coefﬁcient in the objective function. in minimization problems. The variable with the largest objective function coefﬁcient should enter as the ﬁrst decision variable into the second tableau for a maximization problem. Similarities: (1) Both methods ﬁnd the optimal solution at a corner point.. variables in the solution) is always equal to the number of constraints. Cjl b 200 160 0 Solution Mix X2 X1 S3 Zj Cj Ϫ Zj \$160 X1 0 1 0 160 0 \$200 X2 1 0 0 200 0 0 S1 1 Ϫ1 6 40 Ϫ40 0 S2 Ϫ1⁄9 2 ⁄9 Ϫ2 131⁄3 Ϫ131⁄3 0 S3 Quantity 0 0 1 0 0 8 4 24 \$2. X2 ϭ 8. that is. an infeasible solution may occur. so they are quickly removed from the initial solution. simplex checks a lesser number of corners. 9-3. A nonbasic variable is one that is not currently in the solution. The right-hand-side values in the primal become the dual’s objective function coefﬁcients. The Zj entry in the “quantity” column stands for proﬁt contribution or cost. . objective function and constraints. but it may take more iterations before the optimum is reached. 9-9. 9-7. (2) Both methods require a feasible region and the same problem structure. They. The graphical method is preferable when the problem has two variables and only two or three constraints (and when no computer is available). The person formulating the problem should look for the cause. in maximization and minimization problems. simplex can handle any dimensions. not listed in the solution mix column of the tableau. S3 ϭ \$24 in slack royalty fees paid Proﬁt ϭ \$2. Pivot number: Deﬁned to be at the intersection of the pivot column and pivot row.240/day Here is the dual formulation: Minimize Z ϭ 20y1 ϩ 108y2 ϩ 120y3 subject to: 1y1 ϩ 9y2 ϩ 12y3 ജ 160 2y1 ϩ 9y2 ϩ 6y3 ജ 200 Here is the dual optimal solution and ﬁnal tableau. the negatives of the Cj Ϫ Zj values in the slack variable columns are the shadow prices. In the minimization problem. 9-14.5 objective coefﬁcient. that is. (3) Simplex method can be automated and computerized. the Cj Ϫ Zj row is the main difference. They represent a quantity of unused resource and have a zero coefﬁcient in the objective function. Without the rule. 9-15. Pivot row: Select the row with the smallest quantity-tocolumn ratio that is a nonnegative number. The solutions to the Ui dual variables are the primal’s shadow prices. 9-5. Cj l b 108 20 Solution Mix y2 y1 Zj Cj Ϫ Zj 20 y1 0 1 20 0 108 y2 1 0 108 0 120 y3 2 Ϫ6 96 ϩ24 0 S1 Ϫ2⁄9 12⁄9 Ϫ42⁄9 ϩ42⁄9 0 S2 ⁄9 Ϫ1 Ϫ8 ϩ8 1 Quantity 131⁄3 40 \$2. 9-11.33 y3 ϭ marginal value of one more \$1 in royalty fees ϭ \$0 SOLUTIONS TO DISCUSSION QUESTIONS AND PROBLEMS 9-1. it’s the smallest negative Cj Ϫ Zj. They carry a high coefﬁcient. Pivot column: Select the variable column with the largest positive Cj Ϫ Zj value (in a maximization problem) or smallest negative Cj Ϫ Zj value (in a minimization problem). the greatest positive Cj Ϫ Zj indicates the new pivot column. (4) Simplex method involves use of surplus. 9-12.QXD 5/12/08 12:01 PM Page 117 CHAPTER 9 LINEAR PROGRAMMING: THE SIMPLEX METHOD 117 Here is the primal optimal solution and ﬁnal simplex tableau. and artiﬁcial variables but provides useful economic data as a by-product. or X1 ϭ 4. The Cj Ϫ Zj value is the net change in the value of the objective function that would result from bringing one unit of the corresponding variable into the solution. we ensure feasibility at the next iteration. So in this case there will be eight basic variables. Hence X3 (with a value of \$12) will enter ﬁrst. 9-13. An optimal solution will still be reached if any positive Cj Ϫ Zj value is chosen. usually conﬂicting constraints. Differences between graphical and simplex methods: (1) Graphical method can be used only when two variables are in model. By choosing the minimum ratio. The number of basic variables (i.240 This means y1 ϭ marginal value of one more labor hour ϭ \$40 y2 ϭ marginal value of one more RAM chip ϭ \$13. The minimum ratio criterion used to select the pivot row at each iteration is important because it gives the maximum number of units of the new variable that can enter the solution. slack. the problem is infeasible. This procedure will result in a better (more proﬁtable) solution at each iteration. The Zj values indicate the opportunity cost of bringing one unit of a variable into the solution mix. Maximization and minimization problems are quite similar in the application of the simplex method. If an artiﬁcial variable is in the ﬁnal solution. 9-2. Minimization problems usually include constraints necessitating artiﬁcial and surplus variables. (2) Graphical method must evaluate all corner points (if the corner point method is used).240 Artiﬁcial variables have no physical meaning but are used with the constraints that are ϭ or у. 9-8.e. X1 will enter ﬁrst. In terms of technique. 9-4. the values of some of them may be zero. The dual will have 8 constraints and 12 variables. too. In maximization problems. Surplus variables convert у constraints into equalities and represent a resource usage above the minimum required. Slack variables convert р constraints into equalities for the simplex table.REVISED M09_REND6289_10_IM_C09. 9-10.6. the least-cost coefﬁcient is X1.

QXD 5/12/08 12:01 PM Page 118 118 CHAPTER 9 LINEAR PROGRAMMING: THE SIMPLEX METHOD The primal objective function coefﬁcients become the righthand-side values of dual constraints.4 X2 2 1 7 Ϫ8M 0. Once the right-hand side went beyond 240. g. e.600 X1. Maximize earnings ϭ 0. S2 ϭ 9.REVISED M09_REND6289_10_IM_C09. S2 will leave next.000 Table for Problem 9-18 Cjl b 0 0 Solution Mix S1 S2 Zj Cj Ϫ Zj \$900 X1 14 10 0 900 \$1.60) and the proﬁt is 7.200.2 Ϫ 2M 9-19. S1 ϭ 150. See the table below.500 X2 4 12 0 1.1 X4 Ϫ5 Ϫ8 Ϫ1 Ϫ7M Ϫ0. and there is now slack for the ﬁrst constraint.600 0 20 X1 120 b. X2 у 0 c.1X4 ϩ 0S1 ϩ 0S2 Ϫ MA1 Ϫ MA2 subject to X1 ϩ 2X2 ϩ X3 ϩ 5X4 ϩ S1 ϭ 150 X2 Ϫ 4X3 ϩ 8X4 ϩ A1 ϭ 70 6X1 ϩ 7X2 ϩ 2X3 Ϫ X4 Ϫ S2 ϩ A2 ϭ 120 c.400)/(240 Ϫ 80) ϭ 4. h. First tableau: \$3 X1 0 3 \$0 \$3 \$5 X2 1 2 \$0 \$5 \$0 S1 1 0 \$0 \$0 \$0 S2 0 1 \$0 \$0 Quantity 6 18 \$0 Cj l Solution b Mix \$0 \$0 S1 S2 Zj Cj Ϫ Zj . with constraint inequality signs reversed.360.500)(800) ϭ 1.800/160 ϭ 30 Table for Problem 9-19b Cj l b 0 ϪM ϪM Solution Mix S1 A1 A2 Zj Cj Ϫ Zj 0. another constraint prevented any additional proﬁt.8X1 ϩ 0. Proﬁt will increase by (Cj Ϫ Zj)(units of variable entering the solution) ϭ (1. With the additional change.1 ϩ 7M 0 S1 1 0 0 0 0 0 S2 Ϫ0 Ϫ0 Ϫ1 ϪM ϪM ϪM A1 Ϫ0 Ϫ1 Ϫ0 ϪM Ϫ0 ϪM A2 Ϫ0 Ϫ0 Ϫ1 ϪM Ϫ0 Quantity 150 70 120 Ϫ190M 9-20. 800 units of X2 will be in the solution at the second tableau. a.4 ϩ 8M 1. The shadow price ϭ (increase in proﬁt)/(increase in right-hand side value) ϭ (7.2 X3 Ϫ1 Ϫ4 Ϫ2 2M 1.600. 9-18. Maximize proﬁt ϭ 900X1 ϩ 1. X2 60 d. A2 ϭ 120. The transpose of the primal constraint coefﬁcients become the dual constraint coefﬁcients.200. a.500 \$0 S1 1 0 0 0 \$0 S2 0 1 0 0 Quantity 3. 9-16. the optimal corner point in part B is still the optimal corner point. all other variables ϭ 0 Ϫ0.500X2 d.8 ϩ 6M 0. a. The student is to write his or her own LP primal problem of the form: maximize proﬁt ϭ C1X1 ϩ C2X2 subject to A11X1 ϩ A12X2 р B1 A21X1 ϩ A22X2 р B2 and for a dual of the nature: minimize cost ϭ B1U1 ϩ B2U2 subject to A11U1 ϩ A21U2 у C1 A12U1 ϩ A22U2 у C2 9-17. b.200 Ϫ 2. A1 ϭ 70.360 9. Basis is S1 ϭ 3. Proﬁt doesn’t change.2X3 Ϫ 0.360 10X1 ϩ 12X2 р 9.8 X1 1 0 6 Ϫ6M 0.4X2 ϩ 1. The new optimal corner point is (0. c. f. 14X1 ϩ 4X2 р 3. X2 should enter basis next.

0) on the graph. The entering variable is X1. X2 ϭ 6. Ratios: Row 1: 80/4 ϭ 20 Row 2: 50/1 ϭ 50 These represent the points (20. ⁄3 \$1 Ϫ\$1 X1 ϭ 2. e.5 0 S1 0.3333 Ϫ0. If the largest ratio had been selected. so 20 units of the entering variable (X1) will be brought into the solution. Isoprofit line 20 X1 50 (X2 ϭ 0. Second iteration Cj l Solution b Mix 10 X1 S2 Zj Cj Ϫ Zj 10 X1 1 0 10 0 8 X2 0. The leaving variable is the solution mix variable in row with the smallest ratio.0). S2 ϭ 0) Second tableau: A1 ϭ 55 X1 ϭ 25 10. X2 = 6.25 0 S2 0 1 0 0 Quantity 20 30 200 X2 0 1. The smallest ratio is 20.REVISED M09_REND6289_10_IM_C09.0). the next tableau would represent an infeasible solution since the point (50. S1 ϭ 0.1667 2 Ϫ2 0 S2 Ϫ0. Profit = \$36) 6 b c g. 20. S2 ϭ 0.0) is outside the feasible region. X2 40 a.25 5 3 2. Constraints 25 h. The pivot column is the X1 column.6667 2 Ϫ2 Quantity 10 20 260 9-21. The value of this will be 0 in the next tableau. The third (and ﬁnal) iteration represents the point (10. Thus. X2 ϭ 0. S2 ϭ 0. f. c. Quantity 6 2 \$36 d. S1 ϭ 0.20). Cj l Solution b Mix 0 0 S1 S2 Zj Cj Ϫ Zj 10 X1 4 1 0 10 8 X2 2 2 0 8 0 S1 1 0 0 0 0 S2 0 1 0 0 Quantity 80 50 0 Quantity 6 6 \$30 Third and optimal tableau: Cj l Solution b Mix \$5 \$3 X2 X1 Zj Cj Ϫ Zj \$3 X1 0 1 \$3 \$0 \$5 X2 1 0 \$5 \$0 \$0 S1 1 Ϫ2⁄3 \$3 Ϫ\$3 \$0 S2 0 1 This represents the corner point (0. Basis for ﬁrst tableau: A1 ϭ 80 A2 ϭ 75 (X1 ϭ 0.5 3 First Corner Point of Simplex 0 a 0 3 X1 6 9 Third iteration Cj l Solution b Mix 10 8 X1 X2 Zj Cj Ϫ Zj 10 X1 1 0 10 0 8 X2 0 1 8 0 0 S1 0.5 Ϫ2.3333 0. 9-22. S1 is the leaving variable.QXD 5/12/08 12:01 PM Page 119 CHAPTER 9 LINEAR PROGRAMMING: THE SIMPLEX METHOD 119 Second tableau: Cj l Solution b Mix \$5 \$0 X2 S2 Zj Cj Ϫ Zj \$3 X1 0 3 \$0 \$3 \$5 X2 1 0 \$5 \$0 \$0 S1 1 Ϫ2 \$5 Ϫ\$5 \$0 S2 0 1 \$0 \$0 b. A2 ϭ 0) . S1 ϭ 0. The second iteration represents the corner point (20.5 Ϫ0.0) and (50. and proﬁt ϭ \$36 Graphical solution to Problem 9-20: 9 Second Corner Point of Simplex (Optimal Corner Point of Simplex) (X1 = 2.

Variable X2 should enter the solution next. 9-24. X2 ϭ 0. S2 ϭ 0. Since there is no nonnegative ratio. But the ratios are as follows: At this point. A2 ϭ 0) Cost ϭ 221 at optimal solution 9-23. X2 = 75) 60 b.REVISED M09_REND6289_10_IM_C09. A1 ϭ 0. The optimal solution using simplex is X1 ϭ 3. a. the following simplex tableau is found: Cj l Solution b Mix 6 0 X1 S2 Zj Cj Ϫ Zj 6 X1 1 0 6 0 3 X2 Ϫ1 0 Ϫ6 9 0 S1 1 (X 1 = 3/7. indicating an alternative optimal solution exists by inserting X2 into the basis. 9-25. X2 = 12/7 ) (X1 = 3. X2 should enter the basis next. This problem is infeasible. 9-26. Z ϭ \$6. At the second iteration. X2 = 0) 0 S2 0 1 0 0 (X1 = 1. ROI ϭ \$6. But the two ratios are 1/Ϫ1 ϭ negative and 2/0 ϭ undeﬁned. X2 = 0) 0 0 c 1 X1 2 b 3 ⁄2 ⁄2 1 3 Ϫ3 Alternative optimum at a and b. but an artiﬁcial variable remains in the basis. 0 3 0 20 40 X1 60 80 9X1 + 3X2 ≥ 9 Third tableau: X1 ϭ 14 X2 ϭ 33 2 a X2 6X1 + 9X2 ≤ 18 1 Feasible Region Quantity 1 2 6 (S1 ϭ 0.42. the problem is unbounded.7. X2 = 0) 2 X1 0 1 2 0 3 X2 1 0 3 0 0 S1 1 0 S2 2 ϪM A1 Ϫ2⁄7 3 Quantity 12 X2 40 (X1 = 14. X2 = 33) (Optimal Solution) 3 2 ⁄7 ⁄21 ⁄7 Ϫ ⁄21 1 1 Ϫ ⁄7 1 ⁄7 3 ⁄7 ⁄3 0 0 0 ϪM \$6 20 Ϫ1⁄3 d. ROI ϭ \$6. This is illustrated in the problem’s ﬁnal simplex tableau: Tableau for Problem 9-25a 5 X 3 row ϭ 5 1 X 1 row 12 ϭ unacceptable Ϫ3 S2 row Cj l Solution b Mix 0 2 S1 X1 Zj Cj Ϫ Zj 2 X1 0 1 2 0 3 X2 7 0 S1 3 0 S2 1 0 0 0 ϪM A1 Ϫ1 0 0 ϪM 10 ϭ5 2 Quantity 6 3 \$6 ⁄2 ⁄2 ⁄2 ⁄2 3 1 3 0 13⁄2 Ϫ13⁄2 . All Cj Ϫ Zj are zero or negative. Tableau for Problem 9-25c Cjl b Solution Mix X2 X1 Zj Cj Ϫ Zj (X1 = 80.QXD 5/12/08 12:01 PM Page 120 120 CHAPTER 9 LINEAR PROGRAMMING: THE SIMPLEX METHOD Graphical solution to Problem 9-22: 80 (X1 = 0. The graphical solution is shown below. This problem is degenerate. c. The variable X2 has a Cj Ϫ Zj value of \$0. X2 ϭ ZX\m ϭ 1. The alternative optimal solution is found in the tableau in the next column to be X1 ϭ C\m ϭ 0.

000 0X1 ϩ 2X2 ϩ 2X3 ϩ 0S1 ϩ 0A1 ϩ 1A2 ϭ 2. It is X1 ϭ 27. The optimal solution is shown below.000M ϩ 75.500 1.QXD 5/12/08 12:01 PM Page 121 CHAPTER 9 LINEAR PROGRAMMING: THE SIMPLEX METHOD 121 Since X3 and S2 are tied. X2 ϭ 5.000 1X1 ϩ 0X2 ϩ 0X3 ϩ 1S1 ϩ 0A1 ϩ 0A2 ϭ 1. proﬁt ϭ \$177.000 1.500 3. X3 ϭ 0.REVISED M09_REND6289_10_IM_C09.000 2.000 500 \$125. MA2 Minimum cost ϭ 50X1 ϩ 10X2 ϩ 75X3 ϩ 0S1 ϩ MA1 ϩ subject to 1X1 Ϫ 1X2 ϩ 0X3 ϩ 0S1 ϩ 1A1 ϩ 0A2 ϭ 1. Cj l Solution b Mix \$5 \$6 \$3 X3 X1 X2 Zj Cj Ϫ Zj 6 X1 0 1 0 6 0 3 X2 0 0 1 3 0 5 X3 1 0 0 5 0 0 S1 1 0 S2 Ϫ ⁄2 1 3 0 S3 7 Quantity 0 27 5 3 ⁄2 ⁄2 ⁄2 3 ⁄2 3 ⁄2 ⁄2 3 Ϫ ⁄2 1 1 1 Ϫ1⁄2 13 ⁄2 3 13 ⁄2 3 8 ⁄2 3 \$177 Ϫ13 ⁄2 Ϫ8 ⁄2 Ϫ13 ⁄2 9-27. we can select one at random.000 ⁄2 1. in this case S2.000M Second iteration: Cj l b M 75 0 Solution Mix A1 X3 S1 Zj Cj Ϫ Zj 50 X1 1 0 1 M ϪM ϩ 50 10 X2 Ϫ1 1 0 ϪM ϩ 75 M Ϫ 65 75 X3 0 1 0 75 0 0 S1 0 0 1 0 0 M A1 1 0 0 M 0 M A2 0 1 Quantity 1.500 First iteration: Cj l b M M 0 Solution Mix A1 A2 S1 Zj Cj Ϫ Zj 50 X1 1 0 1 M ϪM ϩ 50 10 X2 Ϫ1 2 0 M ϪM ϩ 10 75 X3 0 2 0 2M Ϫ2M ϩ 75 0 S1 0 0 1 0 0 M A1 1 0 0 M 0 M A2 0 1 0 M 0 Quantity 1.000 ⁄2 1.000 0 37 ⁄2 1 M Ϫ 371⁄2 Third iteration: Cj l b 50 75 0 Solution Mix X1 X3 S1 Zj Cj Ϫ Zj 50 X1 1 0 0 50 0 10 X2 Ϫ1 Ϫ1 Ϫ1 Ϫ25 Ϫ15 75 X3 0 1 0 75 0 0 S1 0 0 1 0 0 M A1 1 0 Ϫ1 50 M Ϫ 50 M A2 0 1 Quantity 1.000 1.000 0 37 ⁄2 1 M Ϫ 371⁄2 .

500.500 9-28.QXD 5/12/08 12:01 PM Page 122 122 CHAPTER 9 LINEAR PROGRAMMING: THE SIMPLEX METHOD Fourth and ﬁnal iteration: Cj l b 50 75 10 Solution Mix X1 X3 X2 Zj Cj Ϫ Zj 50 X1 1 0 0 50 0 10 X2 0 0 1 10 0 75 X3 0 1 0 75 0 0 S1 1 Ϫ1 1 Ϫ15 15 M A1 Ϫ0 Ϫ1 Ϫ1 65 M Ϫ 65 M A2 0 1 Quantity 1. X2 Ն 0 . Z ϭ \$117. cost ϭ \$300 9-29. X3 ϭ 500. X1 ϭ number of mattresses X2 ϭ number of box springs Minimize cost ϭ 20X1 ϩ 24X2 subject to X1 ϩ X2 у 30 X1 ϩ 2X3 у 40 X1.500 ⁄2 500 500 \$117. X2 ϭ 20 kg. X1 ϭ number of kilograms of brand A added to each batch X2 ϭ number of kilograms of brand B added to each batch Minimize costs ϭ 9X1 ϩ 15X2 ϩ 0S1 ϩ 0S2 ϩ MA1 ϩ MA2 subject to X1 ϩ 2X2 Ϫ S1 ϩ A1 ϭ 30 X1 ϩ 4X2 Ϫ S2 ϩ A2 ϭ 80 Cj l b M M Solution Mix A1 A2 Zj Cj Ϫ Zj \$9 X1 1 1 2M Ϫ2M ϩ 9 \$15 X2 2 4 6M Ϫ6M ϩ 15 \$0 S1 Ϫ1 0 ϪM M \$0 S2 0 Ϫ1 ϪM M M A1 1 0 M 0 M A2 0 1 M 0 Quantity 30 80 110M First iteration: Cj l b 15 M Solution Mix X2 A2 Zj Cj Ϫ Zj 15 \$9 X1 1 \$15 X2 1 0 15 0 15 \$0 S1 Ϫ ⁄2 1 \$0 S2 0 Ϫ1 ϪM M 15 M A1 1 M A2 0 1 M 0 Quantity 15 20 225 ϩ 20M ⁄2 ⁄2 Ϫ1 ⁄2 Ϫ M ⁄2 ϩ M 2 Ϫ ⁄2 ϩ 2M 15 Ϫ2 ⁄2 Ϫ 2M 15 3 ⁄2 Ϫ 2M 3M Ϫ ⁄2 Second iteration: Cj l b 15 0 Solution Mix X2 S1 Zj Cj Ϫ Zj \$9 X1 Ϫ ⁄4 1 \$15 X2 1 0 15 0 \$0 S1 0 1 4 0 \$0 S2 Ϫ ⁄4 1 M A1 Ϫ0 Ϫ1 Ϫ0 M M A2 1 Quantity 20 10 \$300 15 ⁄4 ⁄2 ⁄4 Ϫ1⁄2 Ϫ ⁄4 15 Ϫ1⁄2 Ϫ ⁄4 15 1 15 Ϫ ⁄4 21 Ϫ ⁄4 15 M Ϫ ⁄4 Third and ﬁnal iteration: X1 ϭ 0 kg. X2 ϭ 500.REVISED M09_REND6289_10_IM_C09.500 1 0 37 ⁄2 1 M Ϫ 37 ⁄2 X1 ϭ 1.

X2 Ն 0 1 6 3 .QXD 5/12/08 12:01 PM Page 123 CHAPTER 9 LINEAR PROGRAMMING: THE SIMPLEX METHOD 123 Initial tableau: Cj l b M M Solution Mix A1 A2 Zj Cj Ϫ Zj \$20 X1 1 1 2M Ϫ2M ϩ 20 \$24 X2 1 2 3M Ϫ3M ϩ 24 \$0 S1 Ϫ1 Ϫ0 ϪM M \$0 S2 0 Ϫ1 ϪM M M A1 1 0 M 0 M A2 0 1 M 0 Quantity 30 40 70M Second tableau: Cj l b M \$24 Solution Mix A1 X2 Zj Cj Ϫ Zj 1 \$20 X1 1 \$24 X2 0 1 24 0 \$0 S1 Ϫ1 Ϫ0 ϪM ϪM 1 \$0 S2 1 M A1 1 0 0 0 1 M A2 Ϫ ⁄2 1 1 Quantity 10 20 10M ϩ 480 ⁄2 ⁄2 ⁄2 1 Ϫ1⁄2 ⁄2M Ϫ 12 Ϫ1⁄2M ϩ 12 ⁄2 ⁄2M ϩ 12 Ϫ ⁄2M ϩ 12 3 Ϫ1⁄2M ϩ 12 ⁄2M Ϫ 12 Final tableau: Cj l b \$20 \$24 Solution Mix X1 X2 Zj Cj Ϫ Zj \$20 X1 1 0 20 0 \$24 X2 0 1 24 0 \$0 S1 Ϫ2 Ϫ 1 Ϫ16 Ϫ16 \$0 S2 Ϫ1 Ϫ1 Ϫ4 Ϫ4 M A1 Ϫ2 Ϫ1 Ϫ16 M Ϫ 16 M A2 Ϫ1 Ϫ1 4 MϪ4 Quantity 20 10 \$640 X1 ϭ 20. proﬁt ϭ \$96 9-31. Maximize proﬁt ϭ 8X1 ϩ 6X2 ϩ 14X3 subject to 2X1 ϩ X2 ϩ 3X3 р 120 2X1 ϩ 6X2 ϩ 4X3 ϭ 240 X1. X2 ϭ 10. cost ϭ \$640 9-30.REVISED M09_REND6289_10_IM_C09. X2 ϭ 2. X2 Ն 0 Initial tableau: Cj l Solution b Mix \$0 \$0 S1 S2 Zj Cj Ϫ Zj \$9 X1 1 1 0 9 \$12 X2 1 2 0 12 \$0 S1 1 0 0 0 \$0 S2 0 1 0 0 Quantity 10 12 \$0 Final tableau: Cj l Solution b Mix \$4 \$12 X1 X2 Zj Cj Ϫ Zj \$9 X1 1 0 9 0 \$12 X2 0 1 12 0 \$0 S1 2 Ϫ1 6 Ϫ6 \$0 S2 Ϫ1 1 3 Ϫ3 Quantity 8 2 \$96 Second tableau: Cj l Solution b Mix \$0 \$12 S1 X2 Zj Cj Ϫ Zj \$9 X1 1 \$12 X2 0 1 12 0 \$0 S1 1 0 0 0 \$0 S2 Ϫ1⁄2 Ϫ1⁄2 Ϫ6 Ϫ6 Quantity 4 6 \$72 ⁄2 ⁄2 X1 ϭ 8. Maximize proﬁt ϭ 9X1 ϩ 12X2 subject to X1 ϩ X2 р 10 X1 ϩ 2X2 р 12 X1.

500X4 ϩ 0S1 ϩ 0S2 ϩ 0S3 ϩ 0S4 ϩ 0S5 ϩ 0S6 ϩ 0S7 ϩ 0S8 Ϫ MA1 Ϫ MA2 subject to 1.000X2 ϩ 5.3X2 Ϫ 0. a.29. Maximize proﬁt ϭ 8.1 Ϫ ⁄7 30 ϪM Ϫ ⁄ 7 X 1 ϭ 0.000X2 ϩ 600X30. X3ϭ . ϩ 200X4 ϩ S4 X1 ϩ X2 X1 ϩ X2 ϩ X3 ϩ X3 ϩ X4 ϩ X4 ϩ S5 ϭ 35. a.7X4 Ϲ 0 a .000 1. Variable X5 will enter the basis next. ϩ 400X30.QXD 5/12/08 12:01 PM Page 124 124 CHAPTER 9 LINEAR PROGRAMMING: THE SIMPLEX METHOD Initial tableau: Cj l b 0 ϪM Solution Mix S1 A2 Zj Cj Ϫ Zj \$8 X1 2 2 Ϫ2M 8 ϩ 2M \$6 X2 1 6 Ϫ6M 6 ϩ 6M \$14 X3 3 4 Ϫ4M 14 ϩ 4M 0 S1 1 0 0 0 ϪM A1 Ϫ0 Ϫ1 ϪM Ϫ0 Quantity 120 240 Ϫ240M Second tableau: Cj l b \$0 \$6 Solution Mix S1 X2 Zj Cj Ϫ Zj \$8 X1 5 \$6 X2 0 1 6 0 \$14 X3 7 0 S1 1 0 0 0 ϪM A1 Ϫ1⁄6 Ϫ ⁄6 1 Quantity 80 40 \$240 ⁄3 ⁄3 ⁄3 ⁄3 1 2 2 6 4 10 Ϫ1 ϪM Ϫ 1 Final tableau: Cj l b \$14 \$6 Solution Mix X3 X2 Zj Cj Ϫ Zj \$8 X1 Ϫ ⁄7 5 \$6 X2 0 1 6 0 \$14 X3 1 0 14 0 0 S1 Ϫ ⁄7 3 ϪM A1 Ϫ ⁄14 1 Quantity 240 ⁄7 ⁄7 Ϫ1⁄ 7 Ϫ ⁄7 64 Ϫ2⁄ 7 Ϫ ⁄7 30 Ϫ3⁄14 ϩ ⁄7 2 120 \$582 ⁄ 7 6 2 Ϫ1.000X2 ϩ 600X30.600X2 ϩ 1.7X3 Ϫ 0.200X3 ϩ 900X4 ϩ S3 1.6X1 ϩ 0. (Its Cj Ϫ Zj value is the smallest negative number.4X4 м 0 0.7X4 ϩ S8 ϩ A2 9-33.600X2 ϩ 1.000 2.4X4 Ϫ S7 ϩ A2 0.000 700X10. X1 ϭ number of deluxe one-bedroom units converted X2 ϭ number of regular one-bedroom units converted X3 ϭ number of deluxe studios converted X4 ϭ number of efﬁciencies converted Objective: maximum proﬁt ϭ 8.4X3 Ϫ 0.200X3 ϩ 900X4 р \$45. The initial formulation is minimize cost ϭ \$12X1 ϩ 18X2 ϩ 10X3 ϩ 20X4 ϩ 7X5 ϩ 8X6 subject to X1 2X1 ϩ X2 Ϫ 3X3 25X2 ϩ X3 ϩ 2X4 ϩ 8X5 ϩ 4X4 ϩ 18X1 Ϫ 15X2 Ϫ 2X3 Ϫ X4 ϩ 15X5 2X4 ϩ 6X5 ϭ 100 р 900 X6 у 250 у 150 25X6 р 300 у 70 b.000 X1 ϩ X2 X1 ϩ X2 ϩ X3 ϩ X3 ϩ X4 ϩ X4 р 50 у 25 Ϫ S6 ϩ A1 ϭ 25 0.000 ϭ 28.000X1 ϩ 400X20.) Variable A3 will leave the basis because its ratio (150/15) is the smallest of the three positive ratios.000X1 ϩ 1. X1 ϩ X2 у 0.000 ϭ 19. that is.6X2 Ϫ 0. ϩ 400X30. ϩ 900X30.3X1 ϩ 0.REVISED M09_REND6289_10_IM_C09. ϩ 600X20. ϩ 600X20.3X2 Ϫ 0. X3 ϭ 34. ϩ 300X4 р \$28.000X3 ϩ 3. ϩ 300X4 ϩ S2 2.000X3 ϩ 3.000 ϭ 45.000X1 ϩ 6. Ϫ21M ϩ 7.000X1 ϩ 6. ϩ 200X4 р \$19.000 ϭ 50 ϭ0 ϭ0 (which is X1 ϭ 0. ϩ 500X4 р \$35.6X1 ϩ 0.86) 9-32. proﬁt ϭ \$582.40(X1 ϩ X2 ϩ X3 ϩ X4) X1 ϩ X2 р 0.70(X1 ϩ X2 ϩ X3 ϩ X4) The above constraints can be rewritten as: 0. X 2 ϭ 120 240 . profit = \$582 N\m 7 7 b.000X1 ϩ 400X20.7X3 Ϫ 0.000X2 ϩ 5.3X1 ϩ 0. ϩ 500X4 ϩ S1 700X10. ϩ 900X30.500X4 subject to 1. X2 ϭ 17.100X1 ϩ 1.6X2 Ϫ 0.100X1 ϩ 1.14.4X3 Ϫ 0.000X1 ϩ 1.