Stevenson12e ch8S PDF

Rev.
Confirming Pages
8 SUPPLEMENT TO
CHAPTER
The Transportation
Model LEARNING OBJECTIVES
After completing this supplement you
should be able to:
SUPPLEMENT OUTLINE Unequal Supply and Demand 8S-12 LO8S.1 Describe the nature of a
Degeneracy 8S-12 transportation problem.
8S.1 Introduction 8S-1
Summary of Procedure 8S-14 LO8S.2 Solve transportation problems
8S.2 Obtaining an Initial Solution 8S-3 manually and interpret the
8S.6 Location Decisions 8S-14
The Intuitive Lowest-Cost Approach 8S-3 results.
8S.7 Other applications 8S-14
8S.3 Testing for Optimality 8S-4 LO8S.3 Set up transportation problems in
Evaluating Empty Cells: The Stepping- 8S.8 Computer Solutions 8S-15 the general linear programming
Stone Method 8S-5 Key Terms 8S-16 format.
Evaluating Empty Cells: The MODI Solved Problems 8S-16 LO8S.4 Interpret computer solutions.
Method 8S-8
Problems 8S-21
8S.4 Obtaining an Improved Solution 8S-9 Discussion and Review
8S.5 Special Problems 8S-12 Questions 8S-25
The transportation problem involves finding the lowest-cost plan for distributing stocks of
goods or supplies from multiple origins to multiple destinations that demand the goods. For
instance, a firm might have three factories, all of which are capable of producing identi-
cal units of the same product, and four warehouses that stock or demand those products, as
depicted in Figure 8S.1. The transportation model can be used to determine how to allocate
the supplies available from the various factories to the warehouses that stock or demand those
goods, in such a way that total shipping cost is minimized (i.e., the optimal shipping plan).
Usually, analysis of the problem will produce a shipping plan that pertains to a certain period
of time (day, week), although once the plan is established, it will generally not change unless SCREENCAM TUTORIAL
one or more of the parameters of the problem (supply, demand, unit shipping cost) changes.
8S-1 INTRODUCTION
Although Figure 8S.1 illustrates the nature of the transportation problem, in real life manag- LO8S.1. Describe the nature
ers must often deal with allocation problems that are considerably larger in scope. A beer ofa transportation problem.
maker may have four or five breweries and hundreds or even thousands of distributors, and
an automobile manufacturer may have eight assembly plants scattered throughout the United
States and Canada and thousands of dealers that must be supplied with those cars. In such
cases, the ability to identify the optimal distribution plan makes the transportation model very
important.
8S-1
ste24102_ch08s_08S1-08S25.indd 8S-1 1/24/14 7:20 PM

Rev.Confirming Pages
8S-2 Supplement to Chapter Eight The Transportation Model
FIGURE 8S.1
D
The Transportation problem
(demand)
involves determining a minimum-
cost plan for shipping from S
multiple sources to multiple (supply)
destinations D S
(demand) (supply)
D
(demand)
S D
(supply) (demand)
x
The shipping (supply) points can be factories, warehouses, departments, or any other place
e cel from which goods are sent. Destinations can be factories, warehouses, departments, or other
mhhe.com/stevenson12e points that receive goods. The information needed to use the model consists of the following:
1. A list of the origins and each ones capacity or supply quantity per period.
2. A list of the destinations and each ones demand per period.
3. The unit cost of shipping items from each origin to each destination.
This information is arranged into a transportation table (see Table 8S.1).
The transportation model is one of a class of linear programming models, so named
because of the linear relationships among variables. In the transportation model, transporta-
tion costs are treated as a direct linear function of the number of units shipped.
Use of the transportation model implies that certain assumptions are satisfied. The follow-
ing are the major ones:
1. The items to be shipped are homogeneous (i.e., they are the same regardless of their
source or destination).
2. Shipping cost per unit is the same regardless of the number of units shipped.
3. There is only one route or mode of transportation being used between each origin and
each destination.
TABLE 8S.1 Cost to ship

A transportation table one unit from
factory 1 to
warehouse A Warehouse
A B C D Supply
4 7 7 1
1 100 Factory 1
can supply 100
12 3 8 8 units per period
Factory 2 200
8 10 16 5
3 150
450 Total supply

Demand 80 90 120 160 450 capacity per
period
Warehouse B Total demand

can use 90 units per period per period

Supplement to Chapter Eight The Transportation Model 8S-3
Warehouse FIGURE 8S.2

A B C D Supply Solution for the problem shown
4 7 7 1 in Figure 8S.1
1 10 90 100
12 3 8 8
Factory 2 90 110 200
8 10 16 5
3 80 70 150
450
Demand 80 90 120 160 450
The solution to the problem shown in Figure 8S.1 would look like the one shown in
Figure8S.2.
The transportation model starts with the development of a feasible solution, which is
then sequentially tested and improved until an optimal solution is obtained. The description
of the technique on the following pages focuses on each of the major steps in the process in
this order:
1. Obtaining an initial solution.
2. Testing the solution for optimality.
3. Improving sub optimal solutions.
8S-2 OBTAINING AN INITIAL SOLUTION

To begin the process, it is necessary to develop a feasible distribution plan. A number of LO8S.2 Solve transportation
different methods are available for obtaining such a plan. The discussion here will focus problems manually
on the intuitive approach, a heuristic approach that yields an initial solution that is often
optimal or near optimal. Intuitive approachHeuristic
approach to obtaining an initial
The Intuitive Lowest-Cost Approach solution.
With the intuitive approach, cell allocations are made according to cell cost, beginning with
the lowest cost. The procedure involves these steps:
1. Identify the cell with the lowest cost.
2. Allocate as many units as possible to that cell, and cross out the row or column (or both)
that is exhausted by this.
3. Find the cells with the next lowest cost from among the feasible cells.
4. Repeat steps (2) and (3) until all units have been allocated.
Cell 1-D has the lowest cost ($1) (see Table 8S.2). The factory 1 supply is 100, and the ware-
house D demand is 160. Therefore, the most we can allocate to this cell is 100 units. Since
Warehouse TABLE 8S.2

A B C D Supply
4 7 7 1
1 100 100
0
12 3 8 8
Factory 2 200
8 10 16 5
3 150
160 450
Demand 80 90 120 60 450

the supply of 100 is exhausted, we cross out the costs in the first row along with the supply
of 100. In addition, we must adjust the column total to reflect the allocation, which leaves
60units of unallocated demand in column D.
The next lowest cost is $3 in cell 2-B. Allocating 90 units to this cell exhausts the column
total and leaves 110 units for the supply of factory 2 (see Table 8S.3). Also, we cross out the
costs for column B.
The next lowest cost (that is not crossed out) is the $5 in cell 3-D. Allocating 60 units to
this cell exhausts the column total and leaves 90 units for row 3. We now cross out the costs
in column D (see Table 8S.4).
At this point, there is a tie for the next lowest cell cost: cell 3-A and cell 2-C each has a
cost of $8. Break such a tie arbitrarily. Suppose we choose cell 3-A. The demand is 80 units,
and the remaining supply is 90 units in the row. The smaller of these is 80, so that amount is
placed in cell 3-A (see Table 8S.5). This exhausts the column, so we cross out the cell costs in
column A. Also, the remaining supply in factory 3 is now 10.
The other cell in the tie, cell 2-C, still remains. The remaining supply in the row is 110 and
demand in the column is 120. The smaller of these two amounts, 110, is assigned to the cell,
and the row costs are crossed out. The row total is changed to zero, and the column total is
changed to 10 units. The only remaining cell with a cost that has not been crossed out is cell
3-C. Both supply and demand are 10 units, so 10 units are assigned to the cell. This completes
the initial allocation. (See Table 8S.6.)
Now lets determine if this initial solution obtained using the intuitive method is optimum.
TABLE 8S.3 TABLE 8S.4

Warehouse Warehouse
A B C D Supply A B C D Supply
4 7 7 1 4 7 7 1
1 100 100 1 100 100
12 3 8 8 12 3 8 8
Factory 2 90 200 Factory 2 90 200
110 110
8 10 16 5 8 10 16 5
3 150 3 60 150
90
160 450 160 450

Demand 80 90 120 Demand 80 90 120
60 450 60 450
TABLE 8S.5 TABLE 8S.6 Initial allocation using intuitive approach

Warehouse Warehouse
4 7 7 1 4 7 7 1
1 100 100 1 100 100
12 3 8 8 12 3 8 8
Factory 2 90 200 Factory 2 90 110 200
110
8 10 16 5 10 8 10 16 5
3 80 60 150 3 80 10 60 150
90
160 450 450
Demand 80 90 120 Demand 80 90 120 160
60 450 450
8S-3 TESTING FOR OPTIMALITY

Testing for optimality and revising sub optimal solutions involves analysis of each unused
(empty) cell to determine the potential for reducing the total cost of the solution, This is
accomplished by transferring one unit into an empty cell and noting its impact on costs. If

costs are increased, that implies that using the cell would increase total costs. If costs remain
the same, that implies the existence of an alternative option with the same total cost as the
current plan. However, if analysis reveals a cost decrease, the implication is that an improved
solution is possible. The test for optimality requires that every unused cell be evaluated for
potential improvement. Either of two methods can be used: stepping-stone or MODI.
Evaluating Empty Cells: The Stepping-Stone Method

In the stepping-stone method, cell evaluation proceeds by borrowing one unit from a full cell Stepping-stone method:Cell
and using it to assess the impact of shifting units into the empty cell. For instance, if a shift evaluation by borrowing one
of one unit causes an increase of $5, total costs would be increased by $5 times the number unit from a full cell to assess the
of units shifted into the cell. Obviously, such a move would be unwise since the objective is impact of shifting units into the
to decrease costs. empty cell.
The name stepping-stone derives from early descriptions of the method that likened the
procedures to crossing a shallow pond by stepping from stone to stone. Here, the occupied
cells are the stones; shifting units into empty cells requires borrowing units from occupied
cells. To maintain the balance of supply and demand for every row and column, a shift of one
unit into an empty cell requires a series of shifts from other occupied cells.
The best way to understand the evaluation process is to consider a simple example.
Suppose we want to evaluate cell 1-A (see Table 8S.7). We must shift one unit into that cell,
which implies that one unit will be shipped from factory 1 to warehouse A. However, factory
1 has only 100 units it can ship, and all are allocated to warehouse D. Therefore, to ship one
unit to warehouse A, we borrow a unit from cell 1-D. This creates two problems: Column A
now has one extra unit, and Column D is short one unit (i.e., A has 1 80 81 units but
needs only 80 units, and D has 9960159 but needs 160 units). We remedy these prob-
lems by subtracting one unit from cell 3-A and adding it to cell 3-D; this gives column A a
total of 17980 and column D a total of 9961160. (Note: Instead of making each
addition or subtraction of the single unit, a or sign is simply inserted in the appropriate
cell to signify the operation.)
Lets see what effect such a shift would have on costs. For each cell to which a unit was
added, the cost will increase by the transportation cost for that cell; for cells that gave up a
unit, costs will decrease by the cells transportation cost. We can summarize this with a T
account as follows.
Cell 1-A

(1A) 4 (1D) 1
(3D) 5 (3A) 8
9 9
0
Thus, shifting units around this path would have no impact whatsoever on transportation
costs: use of cell 1-A would be an equivalent-cost alternative. Management may prefer to use
it for other reasons, or management may prefer to stay with the original solution. In any case,
knowledge of such alternatives adds a certain degree of flexibility to decision making.
A B C D Supply Evaluation path for cell 1-A
4 7 7 1
1 (+) () 100
100
12 3 8 8
Factory 2 90 110 200
8 10 16 5
3 () (+) 150
80 10 60
450
Demand 80 90 120 160 450

TABLE 8S.8 Warehouse

Evaluation path for cell 2-A A B C D Supply
4 7 7 1
1 100 100
12 3 8 8
Factory 2 (+) () 200
90 110
8 10 16 5
3 () (+) 60 150
80 10
450
Demand 80 90 120 160
450
So far, we have not discovered any improvement potential, but five unused cells still must
be analyzed. Before we evaluate the remaining empty cells, it is important to mention that
constructing paths using the stepping-stone method requires a minimum number of occupied
cells, and unless that number exists, it will be impossible to evaluate all the empty cells in this
manner. The number of occupied cells must equal the sum of the number of rows and columns
minus 1, or R C 1, where R Number of rows and C Number of columns. In this
example, R3 and C4, so we must have 3416 used or completed cells (which
we do have). If there are too few occupied cells, the matrix is said to be degenerate. A method
for overcoming this problem is explained later in this supplement.
For now, lets continue evaluating the unused cells. Suppose we now consider cell 2-A.
Begin by adding a unit to the empty cell. Moving to the right in row 2, we have what seems
to be a choice: borrow a unit from 90 or from 110. However, borrowing from 90 will leave
column B one unit short, and since adding and borrowing must involve occupied cells and
there are no others in the B column, the 90 cannot be used. Instead, we must borrow from
the 110 and add one unit to the 10 in cell 3-C. We can complete our path back to the original
cell by subtracting a unit from the 80 units in cell 3-A. The / path is shown in Table 8S.8.
The impact on total cost for the path associated with cell 2-A would be:
Cell 2-A

(2A) 12 (2C) 8
(3C) 16 (3A) 8
28 16
12
This means that for every unit shifted into cell 2-A, the total cost would increase by $12.
Thus we should avoid shipping units from factory 2 to warehouse A.
At this point, lets consider some helpful rules for obtaining evaluation paths, since you
may still be unsure of how to do it.
1. Start by placing a sign in the cell you wish to evaluate.
2. Move horizontally (or vertically) to a completed cell (a cell that has units assigned to it).
It is OK to pass through an empty cell or a completed cell without stopping. Choose a
cell that will permit your next move to another completed cell. Assign a minus sign ()
to the cell.
3. Change direction and move to another completed cell. Again, choose one that will per-
mit your next move. Assign a plus sign () to the cell.
4. Continue this process of moving to completed cells and alternating and signs until
you can complete a closed path back to the original cell. Make only horizontal and ver-
tical moves.
5. You may find it helpful to keep track of cells that have been evaluated by placing the
cell evaluation value in the appropriate cell with a circle around it.

TABLE 8S.9 Evaluation path for cell 3-B TABLE 8S.10 Evaluation path for cell 2-D
Warehouse Warehouse
4 7 7 1 4 7 7 1
1 0 100 100 1 0 100 100
12 3 8 8 12 3 8 8
Factory 2 +12 (2) (1) 200 Factory 2 +12 90 (2) (1) 200
90 110 110
8 10 16 5 8 10 16 5
3 80 (1) (2) 60 150 3 80 21 (1) (2) 150
10 10 60
450 450
Demand 80 90 120 160 Demand 80 90 120 160
450 450
Lets try another one, say, 3-B. Start by placing a sign in cell 3-B. Move to the 90 in cell
2-B, and place a sign there. Move to the 110 in cell 2-C, and place a sign there. Move to
cell 3-C, and give it a sign. The path is shown in Table 8S.9.
This one was fairly easy. Lets see what the impact on cost would be:
Cell 3-B

(3B) 10 (2B) 3
(2C) 8 (3C) 16
18 19
1
The 1 indicates that for each unit we can shift into cell 3-B, our cost will decrease by $1. If
we could shift 100 units, we could save $100 over the present solution. However, at this point
we are not ready to make any changes since some other empty cell might yield even greater
savings, and we can only make one shift per table. Evaluation of cell 2-D is similar to the
evaluation of cell 3-B. The evaluation path is shown in Table 8S.10.
The cost impact would be:
Cell 2-D

(2D) 8 (3D) 5
(3C) 16 (2C) 8
24 13
11
Thus, no improvement is possible.
There are two cells remaining to be evaluated: 1-B and 1-C. The evaluation of cell 1-B is a lit-
tle more involved than the previous paths. Begin by placing ain cell 1-B. Move to cell 1-D
(100 units) and give it a. Move vertically to 3-D (60 units) and give it a. Move to the left
and put a in cell 3-C (10 units). Move up to 110 and put a in that cell. Move to the left to
90 units (cell 2-B) and give it a sign, completing the path. The path is shown in Table 8S.11.
The cost impact of the path would be zero.
Cell 1-B

(1B) 7 (1D) 1
(3D) 5 (3C) 16
(2C) 8 (2B) 3
20 20
0

TABLE 8S.11 Evaluation path for cell 1-B TABLE 8S.12 Evaluation path for cell 1-C
Warehouse Warehouse
4 7 7 1 4 7 7 1
1 0 (1) (2) 100 1 0 0 (1) (2) 100
100 100
12 3 8 8 12 3 8 8
Factory 2 +12 (2) (1) 200 Factory 2 +12 90 +11 200
90 110 110
8 10 16 5 8 10 16 5
3 80 21 (2) (1) 150 3 80 21 (2) (1) 150
10 60 10 60
450 450
Demand 80 90 120 160 Demand 80 90 120 160
450 450
Thus, no improvement is possible using this cell path.

The last empty cell is 1-C. Its evaluation path is shown in Table 8S.12 (above). Begin with
1-C as , 1-D as , 3-D as , and 3-C as . The cost impact would be:
Cell 1-C

(1C) 7 (3D) 1
(3D) 5 (3C) 16
12 17
5
Hence, for each unit we can transfer into this empty cell, we improve the cost by $5.
At this point, each of the unused cells has been evaluated. The resulting costs are
summarized below.
Cell........ 1-A 2-A 3-B 2-D 1-B 1-C
Cost........ 0 12 1 11 0 5
Cells that have a positive or zero evaluation present no opportunities for improvement. The
fact that 1-A and 1-B are both zero indicates that at least one other equivalent alternative
exists. However, they are of no interest because two cells have negative cell evaluations, indi-
cating that the present solution is not an optimum. In a later section, you will learn how to
develop an improved solution.
Evaluating Empty Cells: The MODI Method

Modified distribution Another method for evaluating empty cells is the modified distribution method (MODI).
method (MODI)Evaluating It involves computing row and column index numbers that can be used for cell evaluation.
empty cells by computing row In many respects, it is simpler than the stepping-stone method because it avoids having to
and column index numbers. trace cell evaluation paths. Nevertheless, the cell evaluations it produces are identical to those
obtained using the stepping-stone method. Note, however, that if a solution is not optimal, one
stepping-stone path must be traced to obtain an improved solution. This is explained in detail
in the next section.
The MODI procedure consists of these steps:
1. Make an initial allocation using the intuitive method.
2. Obtain an index number for each row and column. Do this using only completed cells.
Note that there will always be at least one completed cell in each row and in each column.
a. Begin by assigning a zero to the first row.
b. Determine the column index for any completed cells in row 1 using the relationship:
Column indexCell costRow index. For example, if a cell cost is $8 per unit, the
column index will be 808.

TABLE 8S.13 MODI index numbers TABLE 8S.14 MODI cell evaluation
Warehouse
A 4 B 7 C 12 D 1 Supply Cell Evaluation
0 4 7 7 1
1 100 100
1-A 4(04)0
24 12 3 8 8 1-B 7(07)0
Factory 2 90 110 200 1-C 7(012)5
4 8 10 16 5 2-A 12(44)12
3 80 10 60 150 2-D 8(41)11
3-B 10(47)1
450
Demand 80 90 120 160
450
c. Each new column value will permit the calculation of at least one new row value,
and vice versa. Continue until all rows and columns have index numbers, using only
completed cells.
3. Obtain cell evaluations for empty cells using the relationship:
Cell evaluationCell cost(Row indexColumn index)
Comments:
1. Row or column values may be positive, negative, or zero.
2. A reallocation requires that new row and column cost values be calculated.
Lets see how the MODI method can be used for cell evaluation.
We begin by assigning a zero to row 1 (see Table 8S.13). Since we can only work with
occupied cells and since cell 1-D is the only such cell in row 1, we focus on it. Since the sum
of the row and column index numbers must add to the cell cost (which is 1), we can see that
the index number for column D must be 1. There are no other occupied cells in row 1, so we
shift to Column D, cell 3-D. Again, the sum of the row and column index numbers must add
to the cell cost (which is 5). Since the column index for D is 1, the index number for row 3
must be 4, as 514.
We can use this row 3 index number and the other two occupied cells in row 3 to obtain
index numbers for columns A and C. For column A, the A index number plus 4 must equal the
3-A cell cost of 8. Hence, the index for A is 4: 844. For C, the index plus 4 must equal
16; hence, the index number is 12.
Next, for row 2, using cell 2-C and a C index of 12, the row index is 4. Then, using cell
2-B, the column index is 7:3 (4)7. This completes the row and column index numbers.
The cell evaluation for each empty cell is determined by subtracting from the cost of the
empty cell the sum of the row and column index numbers:
Cell evaluationCell cost(Row indexColumn index)
Table 8S.14 shows all of the evaluations. Note that these agree with the evaluations obtained
using the stepping-stone method. Because some evaluations are negative, the solution is
not optimal.
8S-4 OBTAINING AN IMPROVED SOLUTION

The presence of negative cell evaluations is evidence that an improved solution is possible. By
the same token, if no negatives appear, an optimum has been achieved.
In this case, there are two cells with negative evaluations: 3-B, with a value of 1, and
1-C, with a value of 5. Select the value that implies more improvement (i.e., 5), and
ignore the other.
The implication of the 5 is that a savings of $5 per unit will be realized for every unit
that can be shifted into cell 1-C. Recall the path used to evaluate the cell; it is reproduced in
Table 8S.15A.

TABLE 8S.15 Evaluation path for cell 1-C (see A) TABLE 8S.16 Revised solution
and reallocation of 10 units around the path (see B)
Warehouse
C D C D A B C D Supply
A. 7 1 B. (1) 0 7 (2)100 1 4 7 7 1
210 10 90 100
1 110
(2) 1
1 (1)
100 10 90
8 8 8 8 12 3 8 8
2 2 110 Factory 2 90 110 200
110
16 5 (2) 10 16 (1) 60 5 8 10 16 5
3 (2) (1) 3 210 110 3 80 70 150
10 60 0 70
450
Demand 80 90 120 160
450
In the course of the evaluation, a single unit was added or subtracted to certain cells, as
denoted by the and signs. We want to repeat that procedure, but instead of shifting one
unit, we want to shift as many as possible. The quantities in the cells that have signs (i.e.,
10 in cell 3-C and 100 in cell 1-D) are potential candidates for shifting, since the sign
implies units are subtracted from those cells. It is the smaller of these quantities that is the
limiting factor. That is, we can subtract 10 from both numbers (10 and 100), but if we try to
subtract 100 from both, we end up with 0 and 90. The negative value would imply shipping
from the warehouse back to the factory! Hence, 10 units can be shifted. Always shift the mini-
mum quantity that occupies a negative position in the path.
To accomplish the shift, add 10 units to each cell that has a sign and subtract 10 units
from each cell with a sign. (Because the signs alternate, the row and column totals will still
be satisfied,) This is shown in Table 15 (above).
The improved solution is shown in Table 8S.16 (above). The total costs are:
Total Cost
1-C . . . . . . . . . . . . . . . . . . . . 10($7) $ 70
1-D . . . . . . . . . . . . . . . . . . . . 90($1) 90
2-B . . . . . . . . . . . . . . . . . . . . 90($3) 270
2-C . . . . . . . . . . . . . . . . . . . . 110($8) 880
3-A . . . . . . . . . . . . . . . . . . . . 80($8) 640
3-D . . . . . . . . . . . . . . . . . . . . 70($5) 350
$2,300
Note that the revised solution has a total cost that is $50 less than the previous solution
($2,300 versus $2,350), which is exactly equal to the cell evaluation value times the number
of units shifted: 510units$50.
Is this the optimum? Once again, we must evaluate each empty cell to determine if further
improvements are possible. Recall that we must check to see if there are enough completed
cells to allow all empty cells to have evaluation paths. Thus,
RC13416
Since there are six completed cells, the second round of cell evaluations can proceed.
The cells paths and their values are shown in Tables 8S-17 to 8S-22. Since all are either
zero or positive, the solution is the optimum. (Note: MODI could have been used instead of
stepping-stone.)


A B C D Supply Evaluation of cell 1-A
4 7 7 1 Cell 1-A
1 (1) (2) 100
10 90 1 2
12 3 8 8 4 1
Factory 2 90 110 200 5 8
1 9 29
8 10 16 5
3 (2) (1) 150 0
80 70
450
Demand 80 90 120 160
450

A B C D Supply Evaluation of cell 1-B
4 7 7 1 Cell 1-B
1 0 (1) (2) 90 100
10 1 2
12 3 8 8 7 7
Factory 2 (2) (1) 200 8 3
90 110
115 210
8 10 16 5
3 80 70 150 15
450
Demand 80 90 120 160
450

A B C D Supply Evaluation of cell 2-A
4 7 7 1 Cell 2-A
1 0 15 (1) (2) 100
10 90 1 2
12 3 8 8 12 8
Factory 2 (1) (2) 200 7 1
90 110 5 8
8 10 16 5
124 217
3 (2) (1) 150
80 70 17
450
Demand 80 90 120 160
450

A B C D Supply Evaluation of cell 2-D
4 7 7 1 Cell 2-D
1 0 15 (1) (2) 100
10 90 1 2
12 3 8 8 8 8
Factory 2 17 90 (2) (1) 200 7 1
110
115 29
8 10 16 5
3 80 70 150 16
450
Demand 80 90 120 160
450


Evaluation of cell 3-B A B C D Supply
4 7 7 1 Cell 3-B
1 0 15 (2) (1) 100
10 90 1 2
12 3 8 8 10 3
Factory 2 17 (2) (1) 16 200 8 7
90 110 1 5
8 10 16 5
119 215
3 80 (1) (2) 150
70 14
450
Demand 80 90 120 160
450

Evaluation of cell 3-C A B C D Supply
4 7 7 1 Cell 3-C
1 0 15 (2) (1) 100
10 90 1 2
12 3 8 8 16 7
Factory 2 17 90 16 200 1 5
110
117 212
8 10 16 5
3 80 14 (1) (2) 150 15
70
450
Demand 80 90 120 160
450
8S-5 SPECIAL PROBLEMS

Not all transportation problems are as straightforward as the one just presented. Various irreg-
ularities may occur, which require certain adjustments before a solution can be obtained. Two
of the most common irregularities are unequal supply and demand, and degeneracy, a term
previously noted as referring to an insufficient number of completed cells to allow evaluation
of every empty cell.
Unequal Supply and Demand

A case in which supply and demand are not equal is illustrated in Table 8S.23. Total supply
(capacity) of the two sources is 200 units, but total demand of the destinations is only 170.
dummyImaginary number Thus, the deficiency in demand is 30 units. This condition can be remedied by adding a
added. Equal to the difference dummy destination with a demand equal to the difference between supply and demand. Since
between supply and demand there is no such destination, no units will actually be shipped, and unit-shipping costs for each
when these are unequal. dummy cell are $0. Table 8S.24 illustrates this.
TABLE 8S.24 A dummy column with a

TABLE 8S.23 Supply exceeds demand of 30 units is added to the matrix; dummy
demand by 30 units cell costs are zero
To: A B Supply To: A B Dummy Supply
From: 5 9 From: 5 9 0
1 100 1 100
4 2 4 2 0
2 100 2 100
200 200
Demand 80 90 Demand 80 90 30
170 200

Once the dummy has been added, the transportation method can then be applied in the
usual manner. The final solution will have a total number of units assigned to the dummy that
equals the original difference between supply and demand (see Table 8S.25). These numbers
indicate which source(s) (e.g., factories) will hold the extra units or will have excess capacity.
In the example shown, factory 1 will ship only 70 units even though it has the capacity to ship
100 units.
The total transportation costs for this distribution plan is:
Total Cost
1-A . . . . . . . . . . . . . . . . . . 70($5) $350

1-Dummy . . . . . . . . . . . . . 30($0) 0
2-A . . . . . . . . . . . . . . . . . . 10($4) 40
2-B . . . . . . . . . . . . . . . . . . 90($2) 180
$570
A similar situation exists when demand exceeds supply. In those cases a dummy row (e.g.,
factory) with a supply equal to the difference between supply and demand must be added.
Again, units will not actually be shipped, so transportation costs are $0 for each cell in the
dummy row. Units in the dummy row in the final solution indicate the destinations that will
not receive all the units they desire and the amount of the shortage for each.
Note: When using the intuitive approach, if a dummy row or column exists, make assign-
ments to dummy cell(s) last.
Degeneracy
The transportation method involves evaluation of empty cells using completed cells as step-
ping stones. Degeneracy exists when there are too few completed cells to allow all neces- degeneracyThe condition
sary paths to be constructed. The condition occurs if an allocation (other than the final one) of too few completed cells to
exhausts both the row (supply) and the column (demand) quantities. It can occur in an initial allow all necessary paths to be
solution or in subsequent solutions, so it is necessary to test for degeneracy after each iteration constructed.
using RC1.
An example of a matrix with too few completed cells is shown in Table 8S.26. Note that
there are only four completed cells, although the necessary number is five:
3315. Because of this, it is impossible to develop an evaluation path for cells 1-B,
1-C, 2-A, or 3-A. The situation can be remedied by placing a very small quantity, represented
by the symbol , into one of the empty cells and then treating it as a completed cell. The quan-
tity is so small that it is negligible; it will be ignored in the final solution.
The idea is to select a position for that will permit evaluation of all remaining empty
cells. For example, putting it in cell 3-A will permit evaluation of all empty cells (see
Table 8S.27). Some experimentation may be needed to find the best spot for , because not
every cell will enable construction of evaluation paths for the remaining cells. Moreover,
TABLE 8S.26 There are too few completed

cells; some empty cells cannot be evaluated
TABLE 8S.25 Units in the dummy column A B C Supply
are not shipped; factory 1 will ship only 70 units 3 2 5
1 40 40
To: A B Dummy Supply
From: 5 9 0 8 1 4
1 70 30 100 2 50 10 60
4 2 0 7 7 6
2 10 90 100 3 20 20
200 120
Demand 80 90 30 Demand 40 50 30
200 120

TABLE 8S.27 A B C Supply

Cell 3-A becomes a completed 3 2 5
1 40 40
cell with the addition of a very
small quantity, represented by the 8 1 4
symbol 2 50 10 60
7 7 6
3 e 20 20
120
Demand 40 50 30
120
avoid placing in a minus position of a cell path that turns out to be negative because real-
location requires shifting the smallest quantity in a minus position. Since the smallest quantity
is , which is essentially zero, no reallocation will be possible.
The cell evaluations (not shown) are all positive, so this solution is optimal. (Note that cell
3-A is considered to be a completed cell and is therefore not evaluated.)
Summary of Procedure
1. Make certain that supply and demand are equal. If they are not, determine the difference between
the two amounts. Create a dummy source or destination with a supply (or demand) equal to the
difference so that demand and supply are equal.
2. Develop an initial solution using the intuitive, lowest-cost approach.
3. Check that the number of completed cells is equal to RC1. If it isnt, the solution is degen-
erate and will require insertion of a minute quantity, , in one of the empty cells to make it a com-
pleted cell.
4. Evaluate each of the empty cells. If any evaluations turn out to be negative, an improved solution
is possible. Identify the cell that has the largest negative evaluation. Reallocate units around its
evaluation path.
5. Repeat steps (3) and (4) until all cells have zero or positive values. When that occurs, you have
achieved the optimal solution.
8S-6 LOCATION DECISIONS

The transportation model can be used to compare location alternatives in terms of their
impact on the total distribution costs for a system. The procedure involves working through
a separate problem for each location being considered and then comparing the resulting total
costs.
If other costs, such as production costs, differ among locations, these can easily be included
in the analysis, provided they can be determined on a per-unit basis. In this regard, note that
merely adding or subtracting a constant to all cost values in any row or column will not affect
the optimum solution; any additional costs should only be included if they have a varying
effect within a row or column.
8S-7 OTHER APPLICATIONS

We have seen how the transportation model can be used to minimize the costs associated with
distributing goods, and we have seen how the model can be used for comparing location alter-
natives. The model is also used in a number of other ways. For example, in a slight variation
of the model, profits can be used in place of costs. In such cases, each of the cell profits can be
subtracted from the largest profit, and the remaining values (opportunity costs) can be treated
in the same manner as shipping costs.

Some of the other uses of the model include production planning (see Chapter 11), prob-
lems involving assignment of personnel or jobs to certain departments or machines (see Chap-
ter 16), capacity planning, and transshipment problems.1
The use of the transportation model for capacity planning parallels its use for location
decisions. An organization can subject proposed capacity alternatives to transportation analy-
sis to determine which one would generate the lowest total shipping cost. For example, it is
perhaps intuitively obvious that a factory or warehouse that is close to its marketor has
low transportation costs for some other reasonshould probably have a larger capacity than
other locations. Of course, many problems are not so simple, and they require actual use of
the model.
8S-8 COMPUTER SOLUTIONS

Although manual solution of transportation problems is fairly straightforward, computer solu-
tions are generally preferred, particularly for moderate or large problems. Many software
packages call for data input in the same tabular form used in this supplement. A more general
approach is to format the problem as a standard linear programming model (i.e., specify the
SCREENCAM TUTORIAL
objective function and a set of constraints). That approach enables one to use the more gen-
eral version of an LP package to solve a transportation problem. Lets consider this general
approach. LO 8S.3 Set up transportation
The decision variables for a transportation model are the quantities to be shipped. Because problems in the general linear
each cell represents a potential transportation route, each must have a decision variable. We programming format.
can use the symbol x1A to represent the decision variable for cell 1A, x1B for cell 1B, and so
on. The objective function consists of the cell costs and these cell symbols:
Minimize 4x1A7x1B7x1C1x1D12x2A 3x2B 8x2C

8x2D8x3A10x3B16x3C 5x3D
Because the amounts allocated in any row or column must add to the row or column total,
each row and column must have a constraint. Thus, we have
Supply (rows) x1Ax1Bx1Cx1D100

x2Ax2Bx2Cx2D200
x3Ax3Bx3Cx3D150
Demand (columns) x1Ax2Ax3A 80
x1Bx2Bx3B 90
x1Cx2Cx3C120
x1Dx2Dx3D160
We do not need a constraint for the total; the row and column constraints take care of this.
If supply and demand are not equal, add an extra (dummy) row or column with the neces-
sary supply (demand) for equality to the table before writing the constraints.
Another approach to transportation problems is to use spreadsheet software. The Excel LO 8S.4 Interpret computer
templates can also be used to solve transportation problems. Figure 8S.3 illustrates the Excel solutions.
worksheet for the preceding problem.
1
Transshipment relates to problems with major distribution centers that in turn redistribute to smaller market des-
tinations. See, for example, W. J. Stevenson and C. Ozgur, Introduction to Management Science with Spreadsheets
(New York: McGraw-Hill, 2006).

FIGURE 8S.3 Excel template for the computer solution.
KEY TERMS degeneracy, 8S-13 modified distribution method

dummy, 8S-12 (MODI), 8S-8
intuitive lowest-cost approach, stepping-stone method, 8S-5
8S-3
SOLVED PROBLEMS
Problem 1 Use the intuitive lowest-cost approach to develop an initial solution to the following problem, and
then determine if the solution is optimal. If it is not, develop the optimal solution.
To: A B Supply
From: 9 6
1 75
5 3
2 75
Demand 80 90
Solution a. First check to see if supply and demand are equal. Total demand170 and total supply150;
therefore, it will be necessary to create a dummy origin (supply) with zero transportation costs in
each of its cells (see following table).

To: A B Supply
From: 9 6
1 75
5 3
2 75
0 0
Dummy 20
170
Demand 80 90
170
b. Find the cell that has the lowest unit transportation cost. When a dummy is present and you are
using the intuitive method, leave the dummy allocation until last. Thus, aside from the $0 costs in
the dummy row, the lowest cost is $3 for cell 2-B. Assign as many units as possible to this cell.
The column total is 90, and the row total is 75. Because 75 is the smaller of the two, assign 75
to cell 2-B. This exhausts the row total, so cross out the 75, and cross out the cell costs for the
second row. Revise the column total to 15.
To: A B Supply
From: 9 6
1 75
5 3
2 75 75
0 0
Dummy 20
170
Demand 80 90
15 170
The next smallest cell cost is $6 in cell 1-B. We want to assign as many units as possible to this
route (cell). The remaining demand for column B is 15, and the row total is 75. We assign the
smaller of these, 15, to cell 1-B, thereby exhausting the column and reducing the row supply to
60. Next we draw a line through the cell costs of column B.
To: A B Supply
From: 9 6
1 15 75
60
5 3
2 75 75
0 0
Dummy 20
170
Demand 80 90
15 170
Still ignoring the dummy row, the next lowest cell cost is $9 in 1-A, where a supply of 60 and a
demand of 80 exist. We assign the smaller of these, 60 units, to cell 1-A, so row 1 costs must be
crossed out. Also, the revised column A total becomes 20.

To: A B Supply
From: 9 6
1 60 15 75
60
5 3
2 75 75
0 0
Dummy 20
170
Demand 80 90
20 15 170
At this point, only the dummy row remains. Cell dummy-A has a supply and a demand of 20;
assigning this amount to the cell exhausts both the supply and the demand, finishing the initial
allocation.
To: A B Supply
From: 9 6
1 60 15 75
60
5 3
2 75 75
0 0
Dummy 20 20
170
Demand 80 90
20 15 170
c. The next step is to evaluate this solution to determine if it is the optimal solution. First we must
check for degeneracy: R C 1 3(rows) 2(columns) 1 4, the required number of
filled or completed cells. Because there are four completed cells, the solution is not degenerate.
The evaluation can use either the stepping-stone method or MODI. Suppose we use MODI. We
always begin by placing a zero at the left edge of row 1. This is the row 1 index number. We can
use it to find the column index number for any cell in row 1 that is a completed cell. Because
both cell 1-A and cell 1-B are completed, we can obtain both of their column index numbers.
For completed cells, the sum of the row and the column index numbers must equal the cell cost.
For cell 1-A, 0Column index9, so the column A index must be 9. Similarly, for cell 1-B,
0Column index6, so the column index must be 6. Next, these index numbers are used to
find values of row index numbers for completed cells in their respective columns. Thus, for cell
2-B, 6Row index3, so the row index must be 3. For dummy-A, 9Row index0, so
the row index must be 9.
To: A 9 B 6 Supply
From: 9 6
0 60 15 75
1
23 5 3
2 75 75
29 0 0
Dummy 20 20
170
Demand 80 90
170
Next, we use these index numbers to evaluate the empty cells. We determine a cell evaluation by
subtracting from the cell cost the sum of its row and column index numbers.
Thus, for cell 2-A, we have 5(39)1, and for cell dummy-B, we have
0(96)3.

To: A 9 B 6 Supply
From: 9 6
0 60 15 75
1
23 5 3
2 21 75 75
29 0 0
Dummy 20 13 20
170
Demand 80 90
170
The appearance of a negative cell evaluation tells us this solution can be improved on. When
negatives appear, we select the cell with the largest negative as the local point of our effort to
improve the solution. Because there is only one such cell, it is the one we select.
d. The 1 in cell 2-A tells us that for every unit we can move into this cell, we can reduce the total
cost of the solution by $1. Obviously, we would like to put as many units as possible into this cell,
but we are limited by the row and column totals (row 2 and column A). We are also constrained
by the need for all row and column totals to be maintained. To put units into cell 2-A, we must
take them from some other cell, which will be in another row or column. Fortunately, this matter
is simple to resolve: Determine the stepping-stone path for cell 2-A, and use it to guide the real-
location. That path is illustrated in the following table.
To: A B Supply
From: 9 6
1 (2) (1) 75
60 15
5 3
2 (1) (2) 75
75
0 0
Dummy 20 20
170
Demand 80 90
170
The amount that can be shifted into cell 2-A is the smallest quantity that appears in a negative
position of the path. There are two such quantities, 60 and 75. Because 60 is the smaller of these,
it is shifted around the entire path (i.e., added where there is a sign and subtracted where there
is a sign). The result is our revised solution:
To: A B Supply
From: 9 6
1 75 75
5 3
2 60 15 75
0 0
Dummy 20 20
170
Demand 80 90
170
e. This improved solution may be optimal, or it may need further revision. To find out, we must
evaluate it.

First, we check for degeneracy: RC13214. Because there are four completed
cells, the solution is not degenerate.
Next, we develop MODI index numbers using the completed cells. We always begin by
assigning an index of 0 to the first row. Then, for any completed cell in the first row, the
column index number is the same as the cell cost. Because cell 1-B has a cell cost of 6,
the column B index number is 6. Using this value and the cell cost for completed cell 2-B
enables us to find the index number for row 2: Cell costColumn indexRow index. Thus,
363. This row 2 index number can be used with the cell cost of completed cell 2-A to
find the column A index number, 5(3)8, and the index of 8 can be used with the cell
cost of dummy-A to find the dummy index number: 088. These index numbers are
shown in the following table.
To: A 8 B 6 Supply
From: 9 6
0 75 75
1
23 5 3
2 60 15 75
28 0 0
Dummy 20 20
170
Demand 80 90
170
Evaluations for the empty cells can now be determined.

Cell 1-A: 9(08)1
Cell dummy-B: 0(86)2
Because there are no negative evaluations, the current solution is optimal: Ship 75 units from 1
to B, 60 units from 2 to A, and 15 units from 2 to B. The 20 units in dummy-A means that desti-
nation A will be 20 units short of satisfying its demand of 80 units.
The total cost of the optimal solution is $795:
75(6)450
60(5)300
15(3) 45
20(0) 0
795
Problem 2 A firm that specializes in nonferrous casting must decide between locating a new foundry in Chicago
or Detroit. Transportation costs from each proposed location to existing warehouses and transporta-
tion costs for existing locations for a certain type of casting are shown in the following table. Each
will be able to supply 2,500 units per month. Explain how to determine which location should be
chosen on a transportation cost basis.
From Cost From Cost

Detroit to per unit Chicago To per unit
A ..... $10 A ..... $12
B ..... 8 B ..... 13
C ..... 15 C ..... 5

Supply
Warehouse
(hundreds of
To: A B C units/month)
From: 17 10 6
1 30
Foundry
7 12 14
2 20
Demand 25 10 40
(hundreds of
units/month)
This decision requires solving two transportation problems, one using Detroit costs and the other Solution
using Chicago costs.
a. Set up the two transportation matrices.
Warehouse
A B C Supply
17 10 6
1 30
7 12 14
2 20
10 8 15
Detroit 25
Demand 25 10 40
Warehouse
A B C Supply
17 10 6
1 30
7 12 14
2 20
12 13 5
Chicago 25
Demand 25 10 40
b. Solve each problem

c. The problem that yields the lowest total cost will indicate the preferred location.
1. What information do you need to use the transportation model? DISCUSSION

2. What check must you make before proceeding to develop an initial solution? What corrective action AND REVIEW
might need to be taken?
QUESTIONS
3. Would it ever make sense to have a situation that required both a dummy row and a dummy
column? Explain briefly.

4. What is linear about the objective function in a transportation problem?

5. Why do the and signs alternate in cell evaluation paths?
6. How do you know if a solution is optimal?
7. What does a zero value for an evaluation path indicate?
8. If a solution is not optimal:
a. How do you decide which empty cell to shift units into?
b. How do you know which cells to include in the shifting operation?
c. How do you decide how many units to shift?
9. What is meant by the term degeneracy? How is it overcome?
10. How is the transportation model useful in location decisions? How does use of the transportation
model in location decisions differ from its use as a tool to develop an optimal distribution plan for
a given set of sources and destinations?
11. How are total costs determined for a given distribution plan?
12. What interpretation is given to quantities allocated to dummy destinations? Dummy sources?
13. What is the MODI method? How does it differ from the stepping-stone method?
PROBLEMS 1. Solve this linear programming (LP) problem using the transportation method. Find the optimal
transportation plan and the minimum cost. Also, decide if there is an alternate solution. If there is
one, identify it.
Minimize 8x112x125x132x21x22
3x237x312x326x33
Subject to x11x12x13 90
x21x22x23105
x31x32x33105
x11x21x31150
x12x22x32 75
x13x23x33 75
All variables 0
2. A toy manufacturer wants to open a third warehouse that will supply three retail outlets. The new
warehouse will supply 500 units of backyard playsets per week. Two locations are being studied,
N1 and N2. Transportation costs for location N1 to stores A, B, and C are $6, $8, and $7 per unit,
respectively; for location N2, the costs are $10, $6, and $4, respectively. The existing system is
shown in the following table. Which location would result in the lower transportation costs for the
system?
Store Capacity
To: A B C (Units/week)
From: 8 3 7
1 500
Warehouse
5 10 9
2 400
Demand
(Units/week) 400 600 350
3. A large firm is contemplating construction of a new manufacturing facility. The two leading
locations are Toledo and Cincinnati. The new factory would have a supply capacity of 160 units per
week. Transportation costs from each potential location and existing locations are shown in thefol-
lowing table. Determine which location would provide the lower transportation costs.
From Cost From Cost

Toledo to per Unit Cincinnati to per Unit
A $18 A $ 7
B 8 B 17
C 13 C 13

Supply
A B C (Units/week)
10 14 10
1 210
12 17 20
2 140
11 11 12
3 150
Demand
(Units/week) 220 220 220
4. A large retailer is planning to open a new store. Three locations in California are currently under
consideration: South Coast Plaza (SCP), Fashion Island (FI), and Laguna Hills (LH). Transporta-
tion costs for the locations and costs, demands, and supplies for existing locations and warehouses
(origins), are shown below. Each of the locations has a demand potential of 300 units per week.
Which location would yield the lowest transportation costs for the system?
Supply
TO
A B (Units/week)
From 15 9
Warehouse SCP FI LH 1 660
1 $ 4 $7 $5
10 7
2 11 6 5 340
2
3 5 5 6
14 18
3 200
Demand
400 500
(Units/week)
5. Obtain the optimal distribution plan for the following transportation problem. Develop the initial
solution using the intuitive lowest-cost approach.
Use the stepping-stone method for cell evaluations.
To: A B C Supply
From: 3 4 2
1 40
5 1 7
2 60
8 7 4
3 50
150
Demand 30 45 75
150
6. For the following transportation table:

a. Find the optimal solution. Use the intuitive method for the initial solution and MODI for evaluation.
b. What is the total cost?
c. Is there an alternate optimal solution? If so, what is it?
To: A B C Supply
From: 3 6 2
1 40
3 1 3
2 50
7 6 4
3 65
Demand 55 55 45

7. Given the following transportation problem, do each of the following:

a. Develop an initial feasible solution using the intuitive approach.
b. Use the stepping-stone method to evaluate the solution.
c. Repeat part b using MODI.
d. Find the optimal solution.
Warehouse
To: A B C D Supply
From: 18 12 14 16
1 40
23 24 27 33
2 80
42 34 31 26
3 130
Demand 90 80 30 50
8. Determine the optimal distribution plan for the following problem, and compute the total cost for
that plan. Use the MODI method for cell evaluation.
Warehouse
To: A B C D Supply
From: 14 24 18 28
1 48
17 18 25 16
2 56
30 16 22 30
3 32
Demand 41 34 35 20
9. Refer to problem 7. Due to a temporary condition, origin 3 has experienced a capacity reduction of
20 units per period. Using the intuitive method to develop an initial solution, determine a tempo-
rary distribution plan that will minimize transportation costs.
10. Refer to problem 8. The market supplied by destination D is experiencing a period of rapid
growth. Projected demand is for 60 units per period. A new factory with a capacity of 50 units is
planned for one of two locations, Baltimore or Philadelphia. Transportation costs are as follows.
Which location would result in the lower total cost? Use the MODI method for cell evaluation.
From Cost From Cost

Baltimore to per unit Philadelphia To per unit
A ..... $18 A ..... $31
B ..... 16 B ..... 25
C ..... 22 C ..... 19
D ..... 27 D ..... 20
11. Use the following transportation table to do the following:

a. Find the optimal solution. Use MODI for evaluation.
b. What is the total cost?
c. Is there an alternate optimal solution? If so, what is it?

To:
Rochester Detroit Toronto Supply
From: 6 4 8
Cleveland 100
7 2 7
Chicago 100
4 4 5
Buffalo 80
Demand 70 90 120
12. A soft drink manufacturer has recently begun negotiations with brokers in areas where it intends
to distribute new products. Before making final agreements, however, the firm wants to determine
shipping routes and costs. The firm has three plants with capacities as follows:
Capacity
Plant (cases per week)
Metro 40,000
Ridge 30,000
Colby 25,000
The estimated shipping costs per case for the various routes are:
Demand
Warehouse (cases per week)
RS1 24,000
RS2 22,000
RS3 23,000
RS4 16,000
RS5 10,000
TO
From RS1 RS2 RS3 RS4 RS5
Metro .80 .75 .60 .70 .90
Ridge .75 .80 .85 .70 .85
Colby .70 .75 .70 .80 .80
Determine the optimal shipping plan that will minimize total shipping costs under these
conditions:
a. Route Ridge-RS4 is unacceptable.
b. All routes are acceptable.
c. What is the additional cost of the Ridge-RS4 route not being acceptable?
Stevenson, William J., and Ceyhun Ozgur. Introduction

to Management Science with Spreadsheets.
SELECTED
NewYork: McGraw-Hill, 2006. BIBLIOGRAPHY
AND FURTHER
READING

Stevenson12e ch8S PDF

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Stevenson12e ch8S PDF

Загружено:

Авторское право:

Доступные форматы

Rev.

ste24102_ch08s_08S1-08S25.indd 8S-1 1/24/14 7:20 PM

8S-2 Supplement to Chapter Eight The Transportation Model

TABLE 8S.1 Cost to ship

450 Total supply

Warehouse B Total demand

ste24102_ch08s_08S1-08S25.indd 8S-2 1/13/14 1:45 PM

Supplement to Chapter Eight The Transportation Model 8S-3

Warehouse FIGURE 8S.2

8S-2 OBTAINING AN INITIAL SOLUTION

Warehouse TABLE 8S.2

ste24102_ch08s_08S1-08S25.indd 8S-3 1/13/14 1:45 PM

8S-4 Supplement to Chapter Eight The Transportation Model

TABLE 8S.3 TABLE 8S.4

160 450 160 450

TABLE 8S.5 TABLE 8S.6 Initial allocation using intuitive approach

8S-3 TESTING FOR OPTIMALITY

ste24102_ch08s_08S1-08S25.indd 8S-4 1/13/14 1:45 PM

Supplement to Chapter Eight The Transportation Model 8S-5

Evaluating Empty Cells: The Stepping-Stone Method

ste24102_ch08s_08S1-08S25.indd 8S-5 1/13/14 1:45 PM

8S-6 Supplement to Chapter Eight The Transportation Model

TABLE 8S.8 Warehouse

ste24102_ch08s_08S1-08S25.indd 8S-6 1/13/14 1:45 PM

Supplement to Chapter Eight The Transportation Model 8S-7

ste24102_ch08s_08S1-08S25.indd 8S-7 1/13/14 1:45 PM

8S-8 Supplement to Chapter Eight The Transportation Model

Thus, no improvement is possible using this cell path.

Evaluating Empty Cells: The MODI Method

ste24102_ch08s_08S1-08S25.indd 8S-8 1/13/14 1:45 PM

Supplement to Chapter Eight The Transportation Model 8S-9

8S-4 OBTAINING AN IMPROVED SOLUTION

ste24102_ch08s_08S1-08S25.indd 8S-9 1/13/14 1:45 PM

8S-10 Supplement to Chapter Eight The Transportation Model

ste24102_ch08s_08S1-08S25.indd 8S-10 1/13/14 1:45 PM

Supplement to Chapter Eight The Transportation Model 8S-11

Warehouse TABLE 8S.17

Warehouse TABLE 8S.18

Warehouse TABLE 8S.19

Warehouse TABLE 8S.20

ste24102_ch08s_08S1-08S25.indd 8S-11 1/13/14 1:45 PM

8S-12 Supplement to Chapter Eight The Transportation Model

TABLE 8S.21 Warehouse

TABLE 8S.22 Warehouse

8S-5 SPECIAL PROBLEMS

Unequal Supply and Demand

TABLE 8S.24 A dummy column with a

ste24102_ch08s_08S1-08S25.indd 8S-12 1/13/14 1:45 PM

Supplement to Chapter Eight The Transportation Model 8S-13

1-A . . . . . . . . . . . . . . . . . . 70($5) $350

TABLE 8S.26 There are too few completed

ste24102_ch08s_08S1-08S25.indd 8S-13 1/13/14 1:45 PM

8S-14 Supplement to Chapter Eight The Transportation Model

TABLE 8S.27 A B C Supply

8S-6 LOCATION DECISIONS

8S-7 OTHER APPLICATIONS

ste24102_ch08s_08S1-08S25.indd 8S-14 1/13/14 1:45 PM

Supplement to Chapter Eight The Transportation Model 8S-15

8S-8 COMPUTER SOLUTIONS

Minimize 4x1A7x1B7x1C1x1D12x2A 3x2B 8x2C

Supply (rows) x1Ax1Bx1Cx1D100

ste24102_ch08s_08S1-08S25.indd 8S-15 1/13/14 1:45 PM

8S-16 Supplement to Chapter Eight The Transportation Model