6634 CH05 UG

8/23/02

2:10 PM

Page 26

C H A P T E R

Transportation, Assignment, and Network Models

SOLUTIONS TO DISCUSSION QUESTIONS AND PROBLEMS

5-1. The transportation model is an example of decision making under certainty since the costs of each shipping route, the demand at each destination, and the supply at each source are all known with certainty. 5-2. A balanced transportation problem is one in which total demand (from all destinations) is exactly equal to total supply (from all sources). If a problem is unbalanced, either the demand or the supply constraints must be inequalities. 5-3. The enumeration method is not a practical means of solving 5 5 or 7 7 problems because of the number of possible assignments to be considered. In the 5 5 case, 5! ( 5 4 3 2 1) 120 alternatives need to be evaluated. In the 7 7 case, there are 7! 5,040 alternatives. 5-4. The minimal-spanning model is one that will nd the best way to connect all the nodes in a network together while minimizing the total distance between nodes or the total cost of connecting the nodes together. A number of decision modeling problems can be solved using this model: an example was given connecting water and power to a real estate development project. This model can also be used to determine the best way to deliver cable TV to households, connect computers on a computer network, install an oil pipeline, develop a natural gas network, and more. 5-5. The maximal-ow model can be used to determine the maximum number of cars that can ow through a road system, the number of gallons of chemicals that can ow through a chemical processing plant, the barrels of oil that can go through a pipeline network, the number of people that can use public transportation to get to work, the number of pieces of mail that can go through a mail service, and more. Any time that material or items ow through a network, the maximal-ow model can be used. 5-6. The shortest-route model can be used to nd the best way to install a phone cable between two major cities. Any time items must be moved from one place to another or something, like a cable, must be used to connect two points, the shortest-route model can be used. 5-7. A ow balance constraint calculates the net ow at a node (that is, the difference between the total ow on all arcs entering the node and the total ow on all arcs leaving the node). At each source node, the net ow is expressed as a negative quantity, and represents the amount of ow created at the node. At each destination node,

the net ow is expressed as a positive quantity, and represents the amount of ow consumed at the node. At each pure transshipment node, the net ow is zero. 5-8. To set up a maximal-ow problem as an LP problem, we create a unidirectional dummy arc going from the destination node to the source node, and set the capacity of this arc at innity. 5-9. For many network models, the number of arcs (each of which corresponds to a decision variables) could be quite large. It may therefore be convenient to model these problems in Excel in such a way that decision variables are in a tabular form. For example, rows in the table could denote starting nodes for the arcs, and columns could denote ending nodes. 5-10. To specify the entire table as the Changing Cells in Solver for a maximal-ow network model, we assign a capacity of zero for all arcs that do not actually exist. This will prevent any ow on those arcs. 5-11. To specify the entire table as the Changing Cells in Solver for a shortest-path network model, we assign a unit ow cost of innity (or some arbitrarily high number) for all arcs that do not actually exist. This will prevent any ow on those arcs. 5-12. Let Xij number of students bused from sector i to school j 5XAB 0XBB 4XCB 7XDB 12XEB subject to XAB XBB XCB XDB XEB XAB XAC XAE All Xij XAC XBC XCC XDC XEC XBB XBC XBE 0 XAE XBE XCE XDE XEE XCB XCC XCE 700 (number students in sector A) 500 (number students in sector B) 100 (number students in sector C) 800 (number students in sector D) 400 (number students in sector E) XDB XDC XDE XEB XEC XEE 900 (school B capacity) 900 (school C capacity) 900 (school E capacity) 8XAC 4XBC 0XCC 2XDC 7XEC 6XAE 12XBE 7XCE 5XDE 0XEE Objective: minimize total travel miles

Note that this is an unbalanced problem since total students total school capacity

26

6634 CH05 UG

8/23/02

2:10 PM

Page 27

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

27

Solution: XAB XAE XBB XCC XDC XEE Distance

400 (See le P5-12.XLS) 300 500 100 800 400

Des Moines to Job 2 2,750 tons Total cost $31,750 (b) The formulation should be expanded to include two additional decision variables: Davenport to Des Moines, and Dubuque to Des Moines. The revised solution (see worksheet b in le P5-13.XLS) is: Dubuque to Des Moines 1,500 tons Davenport to Job 3 1,000 tons

5,400 student miles 7X13 7X21 11X22

8X12 5-13(a). Minimize 9X11 6X23 4X31 3X32 12X33 subject to X11 X21 X31 X11 X12 X13 All Xij X12 X22 X32 X21 X22 X23 0 X13 X23 X33 X31 X32 X33 1,500 1,750 2,750 2,000 3,000 1,000

Davenport to Des Moines 750 tons Des Moines to Job 1 2,000 tons Des Moines to Job 2 3,000 tons Total cost $27,500. Marc Smith saves $4,250 by consolidating shipping at Des Moines. 5-14. Krampfs problem is a balanced transportation problem since total supply of cars equals the total demand. The Excel layout and solution is shown in le P5-14.xls. The optimal solution is: Morgantown to Coaltown 35 cars Youngstown to Coal Valley 30 cars Youngstown to Coaltown 5 cars Youngstown to Coal Junction 25 cars Pittsburgh to Coaltown 5 cars Pittsburgh to Coalsburg 20 cars Total distance 3,100 miles 5-15. The optimal solution to the Hall Real Estate decision is shown in the table below. (See le P5-15.XLS)

The solution (see worksheet a in le P5-13.XLS) is: Dubuque to Job 1 250 tons Dubuque to Job 2 250 tons Dubuque to Job 3 1,000 tons Davenport to Job 1 1,750 tons
TO FROM First Homestead 9% Commonwealth Washington Federal Loan Needed $60,000 $60,000 9% Hill St. 8%

Banks St. 8% $40,000 10% 11% $40,000

Park Ave. 10% $40,000 12% 10% $90,000 $130,000

Drury Lane 11%

Max. Avail. $ 80,000

10% $40,000 9% $30,000 $70,000

$100,000 $120,000 $300,000

The total interest cost would be $28,300, or an average rate of 9.43%. An alternative optimal solution exists. It is First HomesteadHill Street First HomesteadBanks Street First HomesteadPark Avenue CommonwealthHill Street CommonwealthDrury Lane Washington FederalPark Avenue 30,000 40,000 10,000 30,000 70,000 120,000

5-16. Mehtas production smoothing problem is a good exercise in the formulation of transportation problems and applying them to real-world issues. The problem may be set up as a transportation model as shown in the table. All squares with Xs represent nonfeasible (backorder) solutions. In applying an LP model to solve such a problem, a very large cost (say about $5,000) would be assigned to each of these squares. This would assure that they would not appear in the nal solution. The optimal solution has a cost of $65,700. (See le P516.XLS)

6634 CH05 UG

8/23/02

2:10 PM

Page 28

28

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

Table for Problem 5-16

Destination (Month) Sources Beginning inventory Regular prod. (month 1) Overtime (month 1) Regular prod. (month 2) Overtime (month 2) Regular prod. (month 3) Overtime (month 3) Regular prod. (month 4) Overtime (month 4) 150 Outside purchases Demand 120 160 150 30 240 1 10 40 100 80 130 50 100 90 130 50 10 20 140 110 2 20 3

unit cost

4 30 120 150 110 140 100 40

Capacity 40 130 100 160 50 120 100 150 50 110 100 140 50 100

100 130 50 100 130

100 50

150 100

150 450 1,090

5-17. To determine which new plant will yield the lowest cost for Ashley in combination with the existing plants, we need to solve two transportation problems. We begin by setting up a transportation table that represents the opening of the third plant in New Orleans (see the table). You should note that the cost of each individual plant to distribution center route is found by adding the distribution costs to the respective unit production costs. Thus the total production plus shipping cost of one auto top carrier from Atlanta to Los Angeles is $14 ($8 for shipping plus $6 for production). Table for Problem 5-17
unit cost TO FROM Atlanta $9 Tulsa $9 New Orleans Demand 800 1,200 $10 500 2,000 $12 900 Los Angeles $14 New York $11 600 Production Capacity

If Ashley selects to open the New Orleans plant, the rms total distribution system cost will be $20,000. (See le P5-17.XLS) If the Houston plant site is chosen, the table is as follows:
TO FROM Atlanta $9 Tulsa $7 Houston Demand 800 1,200 $9 500 2,000 $12 900 Los Angeles $14 New York $11 600 Production Capacity

If Ashley selects to open the Houston plant, the total cost will be $19,500 (See le P5-17.XLS). Upon comparing total costs for the Houston option ($19,500) to those for the New Orleans option ($20,000), we would recommend to Ashley that all factors being equal, the Houston site should be selected.

6634 CH05 UG

8/23/02

2:10 PM

Page 29

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

29

5-18.

Considering Fontainebleau, we have (See le P5-18.XLS)

shipping quantity Canada 60 South America 70 4,000 55 55 2,000 60 50 5,000 75 80 4,000 4,000 5,000 10,000 90 5,000 5,000 60 9,000 24,000 65 70 5,000 40 70 2,000 Pacic Rim 75 Europe 75 8,000 unit cost

Capacity

Waterloo Pusan Bogota Fontainebleau Market Demand

4,000

Optimal cost

$1,530,000.
shipping quantity Canada 60 South America 70 4,000 55 55 2,000 60 50 5,000 70 75 4,000 4,000 5,000 10,000 85 5,000 5,000 65 9,000 24,000 65 70 5,000 40 70 2,000 Pacic Rim 75 Europe 75 8,000 unit cost Capacity

Considering Dublin, we have the following nal solution: (See le P5-18.XLS)

Waterloo Pusan Bogota Dublin Market Demand

4,000

Optimal cost

$1,535,000. (See le P5-18.XLS)

Carbondale to Ciro 100 units Carbondale to Des Moines 50 units St. Louis to Blue Earth 50 units St. Louis to Ciro 100 units Total cost = $17,250 Therefore, St. Louis is $150 per week cheaper than East St. Louis. 5-20. The Excel set up and solution for the revised problem is shown in le P5-20.XLS. The revised solution for the East St. Louis plant is: Decatur to Blue Earth 50 units Decatur to Des Moines 250 units Minneapolis to Blue Earth 200 units Carbondale to Ciro 200 units Carbondale to Des Moines 100 units East St. Louis to Carbondale 150 units Total cost = $16,850 Therefore, with the new shipping option, East St. Louis is $400 per week cheaper than St. Louis.

There is no difference in the routing of shipments, but the Fontainebleau location is $5,000 less expensive than the Dublin location. As a practical matter, changes in exchange rates, subjective factors, or evaluation of future intangibles may overwhelm such a small difference in cost. 5-19. See le P5-19.XLS. Decatur to Blue Earth 50 units Decatur to Des Moines 250 units Minneapolis to Blue Earth 200 units Carbondale to Ciro 150 units East St. Louis to Ciro 50 units East St. Louis to Des Moines 100 units Total cost = $17,400 If we open the new plant in St. Louis, the optimal solution is: Decatur to Des Moines 300 units Minneapolis to Blue Earth 200 units

If we open the new plant in East St. Louis, the optimal solution is:

6634 CH05 UG

8/23/02

2:10 PM

Page 30

30

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

5-21.

Assignment can be made as follows (See le P5-21.XLS)

5-23

Optimal Solution: (See le P5-23.XLS)

Assignment Hawkins to cardiology Condriac to urology Bardot to orthopedics Hoolihan to obstetrics Total cost scale Rating 18 32 24 12 86

Job A12 to machine W Job A15 to machine Z Job B2 to machine Y Job B9 to machine X Time 10 12 12 5-22. Let Xij where i j

16

50 hours

1 if pitcher i is scheduled to go against opponent j, 0 otherwise 1, 2, 3, 4 stands for Jones, Baker, Parker, and Wilson, respectively, and 1, 2, 3, 4 stands for Des Moines, Davenport, Omaha, and Peoria, respectively. sum of 5-24.

Optimal assignment: (See le P5-24.XLS)

Objective: maximize overall probability of winning probability of winning each game 0.6X11 0.7X21 0.9X31 0.5X41 subject to X11 X21 X31 X41 X11 X12 X13 X14 All Xij 0 1, X12 X22 X32 X42 X21 X22 X23 X24 X13 X23 X33 X43 X31 X32 X33 X34 X14 X24 X34 X44 X41 X42 X43 X44 1 (Dead-Arm Jones) 1 (Spitball Baker) 1 (Ace Parker) 1 (Gutter Wilson) 1 (Des Moines) 1 (Davenport) 1 (Omaha) 1 (Peoria) 0.8X12 0.4X22 0.8X32 0.3X42 0.5X13 0.8X23 0.7X33 0.4X43 0.4X14 0.3X24 0.8X34 0.2X44

taxi at post 1 to customer C taxi at post 2 to customer B taxi at post 3 to customer A taxi at post 4 to customer D Total distance traveled 5-25. squad 1 to case C squad 2 to case D squad 3 to case B squad 4 to case A squad 5 to case E Total person-days projected using this assignment 8 8 28 days. 5-26 Optimal Solution: (See le P5-26.XLS)
Assignment Andersonnance Sweeneyeconomics Williamsstatistics McKinneymanagement Rating 95 75 85 380 Total rating 335

18 miles.

Optimal assignment: (See le P5-25.XLS)

Solution: X12 1, X23 1, X34 1, X41 Total P 2.9 (See le P5-22.XLS)

5-27 Minimal-spanning tree model. The optimal solution is shown by the bold arcs. Total length 4500 feet.
3 1 1 4 2 5 5 2 2 4 4 5 7 6 7 10 3 6 3 6 9 4 6 5 7 8 3 7 5 12 4 14

13 3 11 3

5-28 Optimal solution is 500 cars per hour. See le P5-28.XLS. (Assuming all gures are in hundreds of cars)
Flow (Cars/Hour) 200 200 100 500

Path 12578 1368 148 Total

6634 CH05 UG

8/23/02

2:10 PM

Page 31

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

31

5-29 The shortest route is 13571013. The distance is 430 miles. See le P5-29.XLS. 5-30. Minimal-spanning tree model. This is the only optimum solution to this problem (177 units of length).
4 26 1 37 50 2 6 23 3 41 5

5-34. The impact of the construction project to increase the road capacity around the outside roads from International Drive to Disney World would increase the number of cars per hour to 1,700 per hour (17). The increase is 400 cars per hour as would be expected. The solution shown in le P5-34.XLS, is as follows:

Flow 12 13 14 15 26 37 48 58 69 710 810 911 1011 5 8 2 2 5 8 2 2 5 8 4 5 12

5-31. There are several possible solutions for this maximal ow problem: One solution is presented in le P5-31.XLS. The solution may be interpreted as: 146 1256 1356 1456 40 55 45 27 167 widgets per day

Total maximum ow: 17.

Alternative solutions: Substitute 1246 for 32 in lieu of 146 or 1456 (or for some portion of the 32). 5-32. See le P5-32.XLS. No, the changes do not have an impact on the nal solution. With the changes, the optimal solution still has a shortest distance of 430 miles. The nal network is given below. Note that we have increased the value for the paths 69 and 89 to a very high relative number (5,000 in our Excel model) to ensure that these paths are forced out of the nal solution. 5-33. The maximum number of cars that can ow from the hotel complex to Disney World is 13 (1,300 cars per hour). See le P5-33.XLS for the Excel solution. The solution is: 5-35. Solving this maximal ow problem results in a situation where 3,000 gallons per hour (3) will be owing from the origin to the nal network node. The solution is shown in le P5-35.XLS and is as follows:
Flow 12 13 14 25 36 48 59 610 811 910 1012 1113 1314 1 1 1 1 1 1 1 1 1 1 2 1 3

Flow 12 13 14 26 37 48 69 710 810 911 1011 3 8 2 3 8 2 3 8 2 3 10

Total maximum ow: 3. Alternate solutions are possible.

Total maximum ow: 13.

6634 CH05 UG

8/23/02

2:10 PM

Page 32

32

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

5-36. The impact of the emergency repair is that nodes 6 and 7 cannot be used. All ow in and out of these nodes is 0. As a result, the ow from the origin to the nal network node has been reduced to 2,000 gallons per hour (2). The solution is shown in the following table. Note that ows leading to and from nodes 6 and 7 have been changed to 0. See le P5-36.XLS.
Flow 12 14 25 48 59 811 910 1013 1113 1314 1 1 1 1 1 1 1 1 1 2

5-39. Grey can use the minimal-spanning tree model to determine the least-cost approach to connect all houses the cable TV. The bold arcs in the gure indicate the selected arcs. The total cost is $3,400.
2 5 1 6 6 5 4 4 5 8 3 5 7 2 7 6 1 7 6 9

Total maximum ow: 2.

5-40. The solution to the minimal-spanning tree problem results in a minimum distance of 21 (2,100 yards). The nal network follows. Arcs in bold should be selected.
4 4 2 3 4 5 2 3 3 7 6 5 4 2 8 5 3 4 8 1 9 3 7

5-37. The shortest route from node 1 to node 16 is 74 kilometers. The solution along with the nal network is shown in the following table and in le P5-37.XLS.
Distance 13 37 711 1114 1416 Shortest path: 137111416 Total shortest distance: 74. 15 11 18 16 14 1

SOLUTION TO OLD OREGON WOOD STORE CASE

1. The assignment algorithm can be utilized to yield the fastest time to complete a table with each person assigned one task. See sheet #1 in le P5-Oregon.XLS.
Person Tom Cathy George Leon Job Preparation Assembly Finishing Packaging Total time Time (Minutes) 100 70 60 210 240

5-38. The impact of closing two nodes (nodes 7 and 8) is to increase the shortest route from 74 to 76 kilometers. Note that all paths into and from nodes 7 and 8 have their values changed to a very high relative number (100) to force these paths out of the nal solution. The solution along with the nal network is given in the following table and in le P5-38.XLS.

Distance 12 26 69 913 1316 Shortest path: 12691316 Total shortest distance: 76. 20 10 12 16 18

2. If Randy is used, the optimum assignment would be (see sheet #2 in le P5-Oregon.XLS)

Time (Minutes) 80 60 80 210 230

Person George Tom Leon Randy

Job Preparation Assembly Finishing Packaging Total time

This is a savings of 10 minutes with Cathy becoming the backup.

6634 CH05 UG

8/23/02

2:10 PM

Page 33

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

33

3. If Cathy is given the preparation task, the solution of the assignment with the remaining three workers assigned to the remaining three tasks is (see sheet #3A in le P5-Oregon.XLS)
Person Cathy Tom George Leon or Randy Job Preparation Assembly Finishing Packaging Total time Time (Minutes) 120 60 60 210 250

Nine tables per day could be achieved by having Tom prepare and assemble 3 tables, George prepare and nish 3 tables, Cathy assemble 6 tables, Leon nish 6 tables, and Randy prepare 3 tables and package all 9. George, Cathy, and Randy would each have 60 minutes per day unutilized and could build 0.6 table having George do preparation (80 minutes), Cathy assembly and packaging (95 minutes), and Randy the nishing (100 minutes).

SOLUTION TO CUSTOM VANS, INC. CASE

To determine whether the shipping pattern can be improved and where the two new plants should be located, the total costs for the entire transportation system for each combination of plants, as well as the existing shipping pattern costs, will have to be determined. In the headings identifying the combination being discussed, Gary and Fort Wayne will be omitted since they appear in every possible combination. Total costs and optimal solutions for each combination are given on succeeding pages. A summary of the total costs and the respective systems is listed below: DetroitMadison MadisonRockford DetroitRockford $10,200 $10,550 $11,400

If Cathy is assigned to the nishing task, the optimum assignment is (See sheet #3B in le P5-Oregon.XLS)
Time (Minutes) 80 60 100 210 250

Person George Tom Cathy Leon or Randy

Job Preparation Assembly Finishing Packaging Total time

4. One possibility would be to combine the packaging operation with nishing. Then, George could build an entire table by himself (in 230 minutes) and Tom could do preparation (100 minutes), Randy the assembly (80 minutes), and Leon the nishing and packaging (90 minutes). This crew could build 4.8 tables in a 480minute workday, while George himself could build 2.09 tablesa total of almost 7 tables per day. To utilize all ve workers, George and Tom could each build entire tables, 2.09 and 1.75 per day, respectively. Letting Randy do preparation (110 minutes), Cathy the assembly (70 minutes), and Leon the nishing and packaging (90 minutes) allows an additional 4.36 tables per day for a total of 8.2 per day.

Since the total cost is lowest in the GaryFort Wayne DetroitMadison combination ($10,200), the new plants should be located in Detroit and Madison. This system is also an improvement over the existing pattern, which costs $9,000, on a cost-perunit basis. Status quo: $9,000/450 units Proposed: $10,200/750 units $20/unit $13.60/unit

Thus the two new plants would denitely be advantageous, both in satisfying demand and in minimizing transportation costs. Existing Shipping Pattern (See sheet #1 in le P5-Custom.XLS)

SHOP PLANT Gary Fort Wayne Demand 300 Chicago 10 200 20 50 100 150 30 Milwaukee 20 100 50 100 200 15 150 750 Minneapolis 40 Detroit 25 300 Capacity

Total costs

200(10) $9,000

50(30)

100(40)

100(15)

The costs for the additional plants are shown on the next page.

6634 CH05 UG

8/23/02

2:10 PM

Page 34

34

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

Cost Table for Custom Vans, Inc.

SHOP PLANT Existing Gary Fort Wayne Chicago 10 20 Milwaukee 20 30 Minneapolis 40 50 Detroit 25 15 Capacity 300 150

Detroit* Proposed Madison** Rockford Forecast Demand

26 7 5 300

36 2 10 100

56 22 30 150

1 37 35 200

150 150 150

*Since a plant at Detroit could purchase a gallon of berglass for $2 less than any other plant, and one Shower-Ric takes 2 gallons of berglass, a systems approach to transportation warrants that $2(2), $4, be deducted from each price quoted in the case for shipments from Detroit. **Since a plant at Madison could hire labor for $1 less per hour than the other plants, and one Shower-Ric takes 3 labor hours to build, $1(3) or $3 should be deducted from each price quoted for shipments from Madison.

DetroitMadison (see sheet #2 in le P5-Custom.XLS)

SHOP PLANT Gary Fort Wayne Detroit 7 Madison Demand 300 100 2 150 150 200 22 Chicago 10 200 20 100 26 36 56 150 37 150 750 100 30 50 50 1 150 15 150 Milwaukee 20 Minneapolis 40 Detroit 25 300 Capacity

Total cost

$10,200

MadisonRockford (See sheet #3 in le P5-Custom.XLS)

SHOP PLANT Gary Fort Wayne 7 Madison Rockford Demand Total Cost $10,550 300 100 50 5 100 10 150 150 200 30 35 150 750 2 22 Chicago 10 250 20 30 50 150 37 150 Milwaukee 20 Minneapolis 40 50 15 150 Detroit 25 300 Capacity

6634 CH05 UG

8/23/02

2:10 PM

Page 35

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

35

DetroitRockford (see sheet #4 in le P5-Custom.XLS)

10 SHOP PLANT Gary Fort Wayne Detroit 5 Rockford Demand Total cost $11,400 300 100 10 150 150 200 30 Chicago 10 200 20 100 26 36 56 150 35 150 750 100 30 50 50 1 150 15 150 Milwaukee 20 Minneapolis 40 Detroit 25 300 Capacity 30 40 5

SOLUTION TO BINDERS BEVERAGE CASE

This is a shortest-route problem. With the data given in the problem, the shortest-route model can be used to determine the minimum time in minutes required to go from the plant to the warehouse in east Denver. The results are shown in the table, and in le P5Binders.XLS. As you can see, the best route is to take North Street to I-70. At Exit 137, South Street is taken to the warehouse. This route takes one hour (60 minutes).

INTERNET CASE STUDIES

Northwest General Hospital
Optimal Solution (See le P5-Northwest.XLS)
Source From: Station 5A 5A 5A 3G 3G 3G 1S 1S Destination To: Wing 2 3 5 1 3 5 2 4 Number of Trays 55 65 80 80 85 60 65 210

Data Start Node End Node Distance North Street I 70A I 70B High StreetA High StreetB Columbine Street West StreetA West StreetB West StreetC 6 AveA 6 AveB 6 AveC Rose StreetA Rose StreetB South AveA South AveB Shortest Path Total distance 1 2 4 1 3 1 3 5 7 4 5 6 6 7 8 9 2 4 8 3 4 5 5 7 9 5 6 10 7 8 9 10 20 5 10 20 20 30 15 20 15 15 25 40 20 20 10 15

Optimal Cost: 4,825 minutes. Alternal solutions are possible.

ANDREWCARTER, INC.
This case presents some of the basic concepts of aggregate planning by the transportation method. The case involves solving a rather complex set of transportation problems. Four different congurations of operating plants have to be tested. The costs are: (See le PS-Andrew.XLS)
Total Variable Cost $179,730 188,930 183,430 188,360 Total Fixed Cost $41,000 33,500 34,000 33,000 Total Cost $220,730 222,430 217,430 221,360

Conguration All plants operating 1 and 2 operating, 3 closed 1 and 3 operating, 2 closed 2 and 3 operating, 1 closed

60 Start End Cumulative Node Node Distance Distance

North Street I 70A I 70B South AveA South AveB

1 2 4 8 9

2 4 8 9 10

20 5 10 10 15

20 25 35 45 60

The lowest weekly total cost, operating plants 1 and 3 with 2 closed, is $217,430. This is $3,300 per week ($171,600 per year) or 1.5% less than the next most economical solution, operating all three plants. Closing a plant without expanding the capacity of the remaining plants means unemployment. The optimum solution, using plants 1 and 3, indicates overtime production of 4,000 units at plant 1 and 0 overtime at plant 3. The all-plant optima have no use of overtime and include substantial idle regular time capacity: 11,000 units (55%) in plant 2 and either 5,000 units in plant 1

6634 CH05 UG

8/23/02

2:10 PM

Page 36

36

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

(19% of capacity) or 5,000 in plant 3 (20% of capacity). The idled capacity versus unemployment question is an interesting, nonquantitative aspect of the case and could lead to a discussion of the forecasts for the housing market and thus the plants product. The optimum producing and shipping pattern is
From Plant 1 (R.T.) Plant 3 (R.T.) Plant 3 (O.T.) To (Amount) W2 (13,000); W4 (14,000) W1 (9,000); W3 (8,000); W5 (8,000) W3 (3,000); W4(1,000)

2. Moving footprint number 16 to accommodate the expansion of the pond area has increased the minimum total distance to 10,100 feet (101). A decision now has to be made about whether the increased distance and cost for the water and sewer system is worth the additional expected property prices. The solution follows.
Start Node 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 1 1 2 2 3 3 4 5 5 6 6 7 8 9 10 10 11 11 12 13 13 14 15 16 17 18 19 19 20 21 21 22 22 23 24 25 26 End Node 2 5 3 10 4 8 8 6 10 7 11 12 9 13 11 15 12 16 17 14 18 15 20 22 23 19 20 24 21 22 25 23 25 26 27 27 27

Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Total

Cost 3 2 1 6 1 5 5 2 5 2 4 4 2 7 8 11 2 9 9 4 6 4 7 12 8 2 2 5 4 1 4 6 5 7 11 3 10

Include Y Y Y Y Y Y Y Y Y Y Y

Cost 3 2 1 1 5 2 5 2 4 2 7

There are three alternative optimal producing and shipping patterns, where R.T. regular time, O.T. overtime, and W warehouse.

SOLUTION TO RANCH DEVELOPMENT PROJECT CASE

1. The minimum distance that will connect all houses to the water and sewer lines is 10,000 feet (100). The solution along with the nal network follows:
Start Node 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 1 1 2 2 3 3 4 5 5 6 6 7 8 9 10 10 11 11 12 13 13 14 15 16 17 18 19 19 20 21 21 22 22 23 24 25 26 End Node 2 5 3 10 4 8 8 6 10 7 11 12 9 13 11 15 12 16 17 14 18 15 20 22 23 19 20 24 21 22 25 23 25 26 27 27 27

Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Branch Total

Cost 3 2 1 6 1 5 5 2 5 2 4 4 2 7 8 11 2 8 9 4 6 4 7 8 8 2 2 5 4 1 4 6 5 7 11 3 10

Include Y Y Y Y Y Y Y Y Y Y Y

Cost 3 2 1 1 5 2 5 2 4 2 7

Y Y Y Y Y

2 9 4 6 4

Y Y Y Y Y Y Y Y Y Y

8 2 2 5 4 1 4 6 7 3 101

Y Y Y Y Y

2 8 4 6 4

Y Y Y Y Y Y Y Y Y Y

8 2 2 5 4 1 4 6 7 3 100

The gure for (1) is shown on the next page.

6634 CH05 UG

8/23/02

2:10 PM

Page 37

CHAPTER 5

TRANSPORTATION ASSIGNMENT

AND

NETWORK MODELS

37

Figure for RDPA Case (Part 1)

7 9 4 1 3 1 2 3 1 2 5 2 6 2 7 Bold lines indicate the selected branches 11 4 2 12 9 8 4 5 11 6 10 5 8 5 8 2

13 4

6 18

14 4 15 7

2 5 19 2 20 4 4 25 3 24 11

m Strea

27

Pond

21 1
St

5 10

22
re am 6

8 16 8 17

26

23