UNIT 7 production center should supply and which

DISTRIBUTION AND NETWORK MODELS distribution center to minimize cost.

(TRANSPORTATION AND ASSIGNMENT PROBLEMS) − Typically, the quantity of goods available at each
supply location (origin) is limited, and the quantity of
TRANSPORTATION, TRANSSHIPMENT, AND ASSIGNMENT goods needed at each of several demand locations
PROBLEMS (destinations) is known.
− The models discussed in this unit belong to a special Therefore, when we formulate the mathematical
class of linear programming problems called network model, the origin constraints would usually utilize less
flow problems. than or equal to inequality sign < since the supply is
− Example: Supply chains, transportation and limited. The destination constraints would have equal
transshipment problems, assignment problems, to = signs since the demand is known.
shortest-route problems, and maximal flow − The usual objective in a transportation problem is to
problems. minimize the cost of shipping goods from the origins
− In each case, we present a graphical representation to the destinations.
of the problem in the form of a network. We will
answer transportation and assignment problem using TRANSPORTATION SAMPLE PROBLEM
a special purpose solution procedure. Foster Generators wants to transport its product from
Since network problems deals with problems with three plants to four distribution centers. Foster
multiple variables, we cannot use the graphical Generators operates plants in Cleveland, Bedford, and
method to solve optimization problems. Furthermore, York. Production capacities over the next three-month
simplex method can take quite some time to compute planning period for one particular type of generator are
without the aid of computer software, we are to utilize as follows:
a special-purpose solution procedure specifically Three-Month Production
Origin Plant
intended to solve transportation and assignment Capacity (units)
optimization problems. 1 Cleveland 5,000
2 Bedford 6,000
SUPPLY CHAIN MODELS 3 York 2,500
− A supply chain describes the set of all interconnected Total 13,500
resources involved in producing and distributing a
product. The firm distributes its generators through four regional
Example: If you are in a business of printing and distribution centers located in Boston, Chicago, St. Louis,
selling textbooks, you need to correspond with the and Lexington; the three-month forecast of demand for
suppliers of the materials needed to print the books the distribution centers is as follows:
like paper and ink. After procuring these materials and Distribution Three-Month Demand
printing the books, you need to find a way to distribute Destination
Center Forecast (units)
the book to the bookstores so that your products can 1 Boston 6,000
be solved. This entire process is called a supply chain 2 Chicago 4,000
and it costs money to maintain this process. 3 St. Louis 2,000
− Main Goal in Supply Chain: In general, supply chains 4 Lexington 1,500
are designed to satisfy customer demand for a Total 13,500
product at minimum cost. *Note that the 3 plants of the company are able to
− Those that control the supply chain must make produce 13,500 generators to supply the 13,500
decisions such as where to produce a product, how generators demanded by the 4 distribution centers.
much should be produced, and where it should be
sent. The transportation cost per unit (in $) for Foster
Generators from each point of origin to each point of
TRANSPORTATION PROBLEM destination are as follows:
− The transportation problem arises frequently in Destination
planning for the distribution of goods and services Origin Boston Chicago St. Louis Lexington
from several supply locations to several demand Cleveland 3 2 7 6
locations. Bedford 7 5 2 3
Example: If you have 3 production centers and 3
York 2 5 4 5
distribution centers, you need to determine which
*If Cleveland will send generators to Boston, it will cost TRANSPORTATION PROBLEM: MATHEMATICAL MODEL
$3/generator. A linear programming model can be used to solve this
*If York will send generators to Chicago, it will cost transportation problem. We use double-subscripted
$5/generator (and so on). decision variables.
xij = number of units shipped from
Management would like to determine how much of its origin i to destination j
production should be shipped from each plant to each Destination (j)
distribution center in order to minimize costs. Boston Chicago St. Louis Lexingto
Origin (i)
1 2 3 n4
Here is a diagram showing the 12 distribution routes 1-Cleveland 3 2 7 6
Foster can use. 2-Bedford 7 5 2 3
• Such a graph is called a network. 3-York 2 5 4 5
• The circles are referred to as nodes and the lines
connecting the nodes as arcs. Objective Function:
• Note that the direction of flow (from origin to • Transportation costs for units shipped from Cleveland
destination) is indicated by the arrows. = 3x11 + 2x12 + 7x13 + 6x14
• Each origin and destination is represented by a node, • Transportation costs for units shipped from Bedford
and each possible shipping route is represented by an = 7x21 + 5x22 + 2x23 + 3x24
arc. • Transportation costs for units shipped from York
• The amount of the supply is written next to each = 2x31 + 5x32 + 4x33 + 5x34
origin node, and the amount of the demand is written
next to each destination node. Supply Constraints:
• The goods shipped from the origins to the • x11 + x12 + x13 + x14 < 5,000 Cleveland Supply
destinations represent the flows in the network. • x21 + x22 + x23 + x24 < 6,000 Bedford Supply
• x31 + x32 + x33 + x34 < 2,500 York Supply
The objective is to determine the routes to be used and
the quantity to be shipped via each route that will provide Demand Constraints:
the minimum total transportation cost. • x11 + x21 + x31 = 6,000 Boston Demand
• x12 + x22 + x32 = 4,000 Chicago Demand
The cost for each unit shipped on each route is shown on • x13 + x23 + x33 = 2,000 St. Louis Demand
each arc. • x14 + x24 + x34 = 1,500 Lexington Demand

Complete Mathematical Model of Foster Generators

Min 3x11 + 2x12 + 7x13 + 6x14 + 7x21 + 5x22 + 2x23 + 3x24 + 2x31
+ 5x32 + 4x33 + 5x34

Subject to Constraints portrayed above.

As you can see there are 12 variables and 7 constraints. It

is close to impossible to solve this problem without the aid
of computer software. But as discussed previously
specifically for transportation and assignment problems,
we can apply a special-purpose procedure to come up with
the optimal solution.
TRANSPORTATION SIMPLEX METHOD: A SPECIAL- Problem Variations
PURPOSE SOLUTION PROCEDURE 1. Total supply not equal to total demand.
In this section we describe a special-purpose solution − Oftentimes, supply is not equal to demand.
procedure, called the transportation simplex method, − If total supply exceeds total demand, no
that takes advantage of the network structure of the modification in the linear programming
transportation problem and makes possible the solution formulation is necessary.
of large transportation problems efficiently on a − Excess supply will appear a slack in the linear
computer and small transportation problems by hand. programming solution.
− Slack for any particular origin can be interpreted
The transportation simplex method, like the simplex as the unused supply or amount not shipped from
method for linear programs, is a two-phase procedure; it the origin.
involves first finding an initial feasible solution and then − If total supply is less than total demand, the linear
proceeding iteratively to make improvements in the programming model of a transportation problem
solution until an optimal solution is reached. will not have a feasible solution. In this case, we
modify the network representation by adding a
To summarize the data conveniently and to keep track of dummy origin with a supply equal to the
the calculation, we utilize a transportation tableau. difference between the total demand and the
total supply.
− With the addition of the dummy origin and an arc
from the dummy origin to each destination, the
linear programming model will have a feasible
solution.
− A 0 cost per unit is assigned to each arc leaving
the dummy origin so that the value of the optimal
solution for the revised problem will represent the
shipping cost for the units actually shipped.
− When the optimal solution is implemented, the
destinations showing shipments being received
from the dummy origin will be the destination
experiencing a short fall or unsatisfied demand.

2. Maximization objective function.

− Usually for transportation problems, we want to
• The twelve cells correspond to the twelve routes from find out a solution that minimizes transportation
one origin to one destination. Thus, each cell in the cost.
transportation tableau corresponds to a variable in − However, in some transportation problems, the
the linear programming formulation. objective is to find a solution that maximizes profit
• The entries in the right-hand margin of the tableau or revenue.
indicate the supply at each origin. − Using the values for profit or revenue per unit as
• The entries at the bottom margin indicate the demand coefficients in the objective function, we simply
at each destination. solve a maximization rather than a minimization
• In this problem, total supply = total demand of 13,500 linear programming.
units. − This change does not affect the constraints.
• Each row corresponds to a supply node and each
column correspond to a demand node in the network 3. Route capacities or route minimums.
model of the problem. − The linear programming formulation of the
• The number of rows (3) plus the number of columns transportation problem also accommodate
(4) equals the number of constraints which is 7. capacities or minimum quantities for one or more
• The entries in the upper right-hand corner of each cell of their routes.
show the transportation cost per unit shipped over the − Example: For the Foster Generator problem, the
corresponding route. York to Boston route which is origin 3 to
destination 1 had a capacity of 1000 units
because of limited space availability on its normal
 mode of transportation. With x31 denoting the
amount shipped from York to Boston, the route
capacity constraint would be x31 < 1000.
− Similarly, route minimums can be specified. For
example: x22 > 2000. This means that it would
guarantee that a previously committed order for
Bedford-Chicago delivery of at least 2000 units
would be maintained in the optimal solution.

4. Unacceptable routes.
− Establishing a route from every origin to every The first thing to do in finding the initial feasible solution
destination may not be possible. To handle this is to allocate as much flow as possible to the minimum
situation, we simply drop the corresponding arc cost arc.
from the network and remove the corresponding • Cleveland-Chicago
variable from the linear programming • Bedford-St.Louis
formulation. • York-Boston
− Example: If the Cleveland-St. Louis route were
unacceptable or unusable, the arc from They each qualify as the minimum cost arc because each
Cleveland-St. Louis could be dropped, and x13 of them has a transportation cost of $2/unit. There is a
could be removed from the linear programming triple tie that happened here. When ties occur, we will
formulation. select the arc to which the most flow could be allocated
− Solving the resulting 11 variable, 7 constraint and work with that route first.
model would provide an optimal solution while • In Cleveland-Chicago route, there will be a flow of
guaranteeing that the Cleveland-St. Louis route is 4000 units since the supply from Cleveland is 5000
not used. units, while the demand of Chicago is 4000 units.
Therefore, 4000 units can be shipped from Cleveland
to Chicago.
TRANSPORTATION SIMPLEX METHOD: A SPECIAL- • For the Bedford-St. Louis route, there will be a flow of
PURPOSE SOLUTION PROCEDURE 2000 units since the supply from Bedford is 6000 units,
Phase 1: Finding an Initial Feasible Solution while the demand of St. Louis is 2000 units. Therefore,
− Such a solution provides arc flow that satisfy each 2000 units can be shipped from Bedford to St. Louis.
demand constraint without shipping more from any • For the York-Boston route, there will be a flow of 2500
origin node than the supply available. units since the supply from York is 2500 units, while
− The procedure most often used to find an initial the demand of Boston is 6000 units. Therefore, 2500
feasible solution to a transportation problem are units can be shipped from York to Boston.
called heuristics.
− A heuristic is a commonsense procedure for quickly Since we need to allocate as much possible to the
finding a solution to a problem. minimum cost arc, then the order of priority will be:
− The heuristic we describe for finding an initial feasible (1) Cleveland-Chicago – 4000 units
solution to a transportation problem is called the (2) York-Boston – 2500 units
minimum cost method. This heuristic strikes a (3) Bedford-St. Louis – 2000 units
compromise between finding a feasible solution
quickly and finding a feasible solution that is close to The next thing we need to do is to reduce the row supply
the optimal solution. and the column demand by the amount of flow allocated
Although some heuristics can find an initial to the cell identified previously.
feasible solution quickly, often the solution they find is
not especially good in terms of minimizing total cost.
Other heuristics may not find an initial feasible
solution as quickly, the solution they find is often good
in terms of minimizing total cost.
If demand/supply is exhausted cross it out which is shown
in red lines above.

We need to do the same procedures again until all the

nodes will be exhausted. Look at the costs, we note that
the least cost among those cells are the following:
This solution is feasible because all the demand is satisfied
• Cleveland-Boston with $3 cost.
and all the supply is used. The total transportation cost
• Bedford-Lexington with $3 cost. associated with this solution is $42,000. Although it is a
Bedford-Lexington is given priority first since feasible solution, it is not necessarily the optimal solution.
Lexington demand can be satisfied with the units
available for Bedford, unlike Cleveland-Boston. Phase 2: Iterating the initial feasible solution in Phase 1 to
the optimal solution.
Recall that each cell in the transportation tableau
corresponds to an arc or route in the network model of the
transportation tableau.
• Identify an incoming arc. An incoming arc is the
currently unused route (unoccupied cell) where
making a flow allocation will cause the largest per-
unit reduction in total cost.
• Assign a flow to the incoming arc.
Flow is then assigned to the incoming arc and the
Since Bedford still has 2500 units to supply, and Boston amounts being shipped over all other arcs to which
has an unfulfilled demand of 2500 units, we can allocate flow had been previously assigned or occupied cell are
it. adjusted as necessary to maintain a feasible solution.
• Identify and drop an outgoing arc from the solution.
Thus, at each iteration in phase 2, we bring a
currently unused arc or unoccupied cell into the
solution and remove an arc to which flow had been
previously assigned or what we call the unoccupied
cell from the solution.
• The Modified Distribution (MODI) Method is used to
identify the incoming arc.

TRANSPORTATION SIMPLEX METHOD PHASE 2:

ITERATIONS
Getting rid of the lines, we can see how many units can
The table shows the initial feasible solutions, and we will
be transferred from a point of origin to a point of
be applying the MODI Method from hereon.
destination.
The MODI Method requires that we define an index.
• Ui for each row of the tableau.
• Vj for each column of the tableau.

Multiply the units shipped with the cost per unit to arrive
at the total cost. Computing these rows and column indexes requires that
the cost coefficient for each occupied cell is: ui + vj
Thus, when Cij is the cost per unit from origin i to What do these numbers mean?
destination j, we can establish that (Cij) = ui + vj for each • This means that shipping one unit over the route
occupied cell. from origin 1 to destination 3 (Cleveland-St. Louis
Route) will increase total cost by $9.
Let us assign numbers to each row and column indexes. • This means that shipping one unit over the route
Use algebra to assign numbers in the other indexes. from origin 1 to destination 4 (Cleveland-
Lexington Route) will increase total cost by $7.
Occupied Cell Cij = ui + vj • This means that shipping one unit over the route
Cleveland-Boston u1 + v 1 = 3 from origin 2 to destination 2 (Bedford-Chicago)
Cleveland-Chicago u1 + v 2 = 2 will decrease total cost by $1.
Bedford-Boston u2 + v 1 = 7 • And so on…
Bedford-St. Louis u2 + v 3 = 2 On the basis of net evaluation indexes, the best arc to
Bedford-Lexington u2 + v 4 = 3 choose in terms of cost reduction is associated with
York-Boston u3 + v 1 = 2 Bedford-Chicago route. In fact, it is the only cell that can
reduce cost since it is the only that is negative. Therefore,
the cell in row 2, column 2 is chosen as the incoming cell.
Total cost decreases by $1 for every unit of flow assigned
to this arc.

How much flow should be assigned to this arc? Because

the total cost decreases by $1 per unit assigned, we want
to allocate the maximum possible flow. To find that
maximum, we must recognize that to maintain feasibility,
each unit of flow assigned to this arc will require
adjustments in the flow over the other currently used arcs.
The Stepping Stone Method can be used to determine the
adjustments necessary and to identify an outgoing arc.
Net Evaluation Index for each Unoccupied Cell:
eij = cij – ui – vj The incoming arc is the Bedford-Chicago route since it is
the best in terms of cost-reduction.

We need to compute for Net Evaluation Index for each

Unoccupied Cell.

Note that there are 6 unoccupied cells.

• For u1, v3 = 7 – 0 – (-2) = 9
• For u1, v4 = 6 – 0 – (-1) = 7
• For u2, v2 = 5 – 4 – 2 = -1
• For u3, v2 = 5 – (-1) – 2 = 4
• For u3, v3 = 4 – (-1) – (-2) = 7
• For u3, v4 = 5 – (-1) – (-1) = 7
 • The incoming arc is the Bedford-Chicago route. We
place 2500 on that cell. Then, we are going to deduct
2500 from the Cleveland-Chicago route making it
1500 units. Then we are going to add 2500 in
Cleveland-Boston to make it 3500. Deduct 2500 to
Bedford-Boston route making it 0.

Here are the new allocations for the units to be shipped

from each point of origin to each point of destination.

Since Bedford-Boston route is the last one to be deducted,

this is referred as the outgoing arc.

In summary:
• Incoming Arc = Bedford-Chicago
• Outgoing Arc = Bedford-Boston Perform MODI Evaluation again!
Occupied Cell ui + yj = Cij
Is it possible that we are going to deduct the York-Boston Cleveland-Boston u1 + v 1 = 3
by 1 instead of the Bedford-Boston? No, because if we Cleveland-Chicago u1 + v 2 = 2
deduct the York-Boston route by 1, the supply for the Bedford-Chicago u2 + v 2 = 5
Bedford-Boston will not equal to the units allocated from
Bedford-St. Louis u2 + v 3 = 2
Bedford to the three distribution centers.
Bedford-Lexington u2 + v 4 = 3
The important thing to determine in the stepping stone York-Boston u3 + v 1 = 2
method is to determine which cells to add or deduct.

STEPPING STONE METHOD

Net Evaluation Index for each Unoccupied Cell:

eij = cij – ui – vj

• From the cells with a minus sign in the stepping stone

path, choose the smallest number from the stepping
stone path.
• Thus, to determine the maximum amount that may be
allocated to the incoming cell, we simply look at which
cells in the stepping stone path identify with a minus
(-) sign. The minus sign cells with the smallest amount • For u1, v3 = 7 – 0 – (-1) = 8
allocated to it will determine the maximum amount • For u1, v4 = 6 – 0 – 0 = 6
that can be allocated to the incoming cell. After • For u2, v1 = 7 – 3 – 3 = 1
allocating to the incoming cell, we make all the • For u3, v2 = 5 – (-1) – 2 = 4
adjustments necessary on the stepping stone path to
• For u3, v3 = 4 – (-1) – (-1) = 6
maintain feasibility.
• For u3, v4 = 5 – (-1) – 0 = 6
Since none of the net evaluation index is negative, this − The objective in the transshipment problem is to
means that we cannot expect for any more cost reduction. determine how many units should be shipped over
In fact, since all of the numbers are positive it means that each arc in the network so that all destination
for every unit we add for each of these cells, it would result demands are satisfied with the minimum possible
to additional costs. transportation cost.
− Supply is limited and demand is specific.
If any number is 0, it means that there will neither be an
additional cost nor a reduction in cost. TRANSSHIPMENT SAMPLE PROBLEM
Ryan Electronics is an electronics company with
Once all of the net evaluation indexes are 0 or positive, an production facilities in Denver and Atlanta. Components
optimal solution has been reached. We don’t have to do produced at either facility may be shipped to either of the
the stepping stone path anymore. firm’s regional warehouses, which are located in Kansas
City and Louisville. From the regional warehouses, the
OPTIMAL SOLUTION AND MINIMUM COST firm supplies retail outlets in Detroit, Miami, Dallas, and
New Orleans.

Shows below are the production capacities of Ryan’s two

plants.
Origin Plant Units
1 Denver 600
2 Plant 400
TOTAL 1000

And shown below are the demand for Ryan’s retail

outlets.
Destination Distribution Units
Center
5 Detroit 200
6 Miami 150
Cleveland-Boston 3500 x 3 $10,500 7 Dallas 350
Cleveland-Chicago 1500 x 2 3,000 8 New Orleans 300
Bedford-Chicago 2500 x 5 12,500 TOTAL 1000
Bedford-St. Louis 2000 x 2 4,000
Bedford-Lexington 1500 x 3 4,500 Difference: Transportation and Transshipment Problems
York-Boston 2500 x 2 5,000 In transshipment problems, we do not ship the
$39,500 components directly from the plants to the distribution
centers. We have to pass through an intermediary. In this
Shipping the goods in this manner would allows Foster problem, the intermediaries are either of the firm’s
Generator to achieve a minimum cost of $39,500. regional warehouses which are located in Kansas City and
Louisville.
TRANSSHIPMENT PROBLEM
− The transshipment problem is an extension of the The transportation cost per unit (in $) for Ryan Electronics
transportation problem in which intermediate nodes, from each point of origin to each warehouse, and the
referred to as transshipment nodes, are added to transportation cost from each warehouse to each point of
account for locations such as warehouses. destination are as follows:
− In this general type of distribution problem,
shipments may be made between any pair of the Warehouse
three general types of nodes: origin nodes, Plant Kansas City Louisville
transshipment nodes, and destination nodes. Denver 2 3
− As was true for the transportation problem, the Atlanta 3 1
supply available at each origin is limited, and the
demand at each destination is specified.
 Retail Outlet And shown below are the demand for Ryan’s retail
New outlets.
Warehouse Detroit Miami Dallas
Orleans Destination Distribution Units
Kansa City 2 6 3 6 Center
Louisville 4 4 6 5 5 Detroit 200
6 Miami 150
7 Dallas 350
8 New Orleans 300
TOTAL 1000

Origin Node Constraints:

1) x13 + x14 < 600
2) x23 + x24 < 400
*since there are 2 origin nodes, we should also have 2
origin constraints.

Transshipment Node Constraints:

1) -x13 – x23 + x35 + x36 + x37 + x38 =0
2) -x14 – x24 + x45 + x46 + x47 + x48 = 0
*since there are 2 transshipment nodes, we should also
have 2 transshipment constraints.
*for the intermediaries/warehouses, we should make sure
that the number of units that the warehouses sent to retail
outlets must equal to the number of units that the
A linear programming model can be used to solve this warehouses received from the plants.
transshipment problem. We use double-subscripted
decision variables. Destination Node Constraints:
Xij = number of units shipped from node I to node j 1) x35 + x45 = 200
2) x36 + x46 = 150
Warehouse 3) x37 + x47 = 350
Plant (i) Kansas City 3 Louisville 4 4) x38 + x48 = 300
Denver 1 2 3 *since we have 4 destinations, then we must have 4
Atlanta 2 3 1 destination constraints (lol)

COMPLETE MATHEMATICAL MODEL OF RYAN

Retail Outlet
New ELECTRONICS
Warehouse Detroit Miami Dallas
Orleans
(i) 5 6 7
8
Kansa City 3 2 6 3 6
Louisville 4 4 4 6 5

Objective Function:
Min 2x13 + 3x14 + 3x23 + 1x24 + 2x35 + 6x36 + 3x37 + 6x38 + 4x45 Min 2x13 + 3x14 + 3x23 + 1x24 + 2x35 + 6x36 + 3x37 + 6x38 + 4x45
+ 4x46 + 6x47 + 5x48 + 4x46 + 6x47 + 5x48
s.t.
Origin Plant Units 1) x13 + x14 < 600
1 Denver 600 2) x23 + x24 < 400
2 Plant 400 3) -x13 – x23 + x35 + x36 + x37 + x38 =0
TOTAL 1000 4) -x14 – x24 + x45 + x46 + x47 + x48 = 0
5) x35 + x45 = 200
6) x36 + x46 = 150
7) x37 + x47 = 350
8) x38 + x48 = 300
For transshipment problems, we will only discuss up to the corresponding project completion times. With three
formulation of the mathematical model. Special solutions project leaders and three clients, nine assignment
procedures do not cover transshipment problems, it only alternatives are possible. Here are the estimated project
covers transportation and assignment problems. completion times (days) for each possible project leader-
client assignment:
ASSIGNMENT PROBLEMS
The assignment problem arises in a variety of decision- Client
making situations; typical assignment problems involve Project Leader 1 2 3
assigning jobs to machines, agents to tasks, sales Terry 10 15 9
personnel to sales territories, contracts to bidders, and so Carle 9 18 5
on. McClymonds 6 14 3
Assume that you have a group project that If Terry is assigned to Client 1, it will require him 10 days
consists of 5 requirements and there are 5 people in your to complete. If Client 2, it will take him 15 days. If Client 3,
group. Allocating which tasks should be done by which it will take him 9 days to complete. And so on…
person is considered to be an assignment problem. McClymonds – the most productive since he can perform
• A distinguishing feature of the assignment problem is the task in the fewest number of days out of the three
that one agent is assigned to one and only one task. project leaders. But again, we cannot assign more than
Although in reality, there are people who can one client to him since 1 client = 1 project leader.
multitask. Let us say in a group project, student A is
very productive and can do 3 tasks out of the 5 tasks
by himself more efficiently than if one task is assigned
to each person. However, in assignment problems,
student A should be assigned only one task, and the
other members of the group should likewise be
assigned to do one task each.
• Specifically, we look for the set of assignments that
will optimize a stated objective, such as minimize
cost, minimize time, or maximize profits.

ASSIGNMENT SAMPLE PROBLEM

Let us consider the case of Fowle Marketing Research,
which has received requests for market research studies
from 3 new clients.
• The company faces the task of assigning a project
leader (agent) to each client (task).
• Three individuals have no other commitments and
are available for the project leader assignments.
• The time required to complete each study will Client (j)
depend on the experience and ability of the project
Project Leader (i) 1 2 3
leader assigned.
Terry 1 10 15 9
• The three projects have approximately the same
priority, and management wants to assign project Carle 2 9 18 5
leaders to minimize the total number of days required McClymonds 3 6 14 3
to complete all three projects. Use the double subscripted variable to help us construct
• A project leader is to be assigned to one client only. the complete mathematical model.
• Fowle’s management must first consider all possible Xij
project leader-client assignments and then estimate
the corresponding project completion times. With Objective Function:
three project leaders and three clients, nine Min 10x11 + 15x12 + 9x13 + 9x21 + 18x22 + 5x23 + 6x31 + 14x32
assignment alternatives are possible. + 3x33
*minimize the combined number of days required for the
Fowle’s management must first consider all possible project leaders to complete their assignments
project leader-client assignments and then estimate the
Days required for Terry’s assignment ASSIGNMENT PROBLEM: A SPECIAL-PURPOSE SOLUTION
= 10x11 + 15x12 + 9x13 PROCEDURE
− The assignment problem is a special case of the
Days required for Carle’s assignment transportation problem. Thus, the transportation
= 9x21 + 18x22 + 5x23 simplex method can be used to solve the assignment
problem.
Days required for MyClymonds’s assignment − The assignment problem has an even more special
= 6x31 + 14x32 + 3x33 structure: All supplies and demands equal to 1.
Example: One person is assigned to one job.
Constraints (s.t.): − The Hungarian Method is a method specifically
1) x11 + x12 + x13 < 1 Terry’s assignment designed to solve assignment problems.
2) x21 + x22 + x23 < 1 Carle’s assignment − The Hungarian method involves what is called matrix
3) x31 + x32 + x33 < 1 McClymonds’s reduction. Subtracting and adding appropriate values
assignment in the matrix yields an optimal solution to the
4) x11 + x12 + x13 = 1 Client 1 assignment problem.
5) x12 + x22 + x32 = 1 Client 2
6) x13 + x23 + x33 = 1 Client 3 STEPS IN HUNGARIAN METHOD
*since there are 3 project leaders and 3 clients, there 1. Reduce the initial matrix by subtracting the smallest
should be 6 constraints. element in each row from every element in that row.
Then, using the row-reduced matrix, subtract the
The constraints reflect the conditions that each project smallest element in each column from every element
leader can be assigned to at most one client and each in that column.
client must have one assigned project leader. 2. Find the minimum number of straight lines that must
be drawn through the rows and the columns of the
In the team leader’s assignment, we use > 1 since each current matrix so that all the zeros in the matrix will
project leader can be assigned to at most one client. be covered. If the minimum number of straight lines
For the client, we use the = 1 since each client should be is the same as the number of rows (or equivalently,
assigned to one project leader. columns), an optimal assignment with a value of zero
can be made. If the minimum number of straight lines
COMPLETE MATHEMATICAL MODEL is less than the number of rows, go to step 3.
3. Subtract the value of the smallest unlined element
from every unlined element and add this same value
to every element at the intersection of two lines. All
other elements remain unchanged. Return to step 2
and continue until the minimum number of lines
necessary to cover all the zeros in the matrix is equal
to the number of rows.

Objective Function: ASSIGNMENT PROBLEM: HUNGARIAN METHOD

Min 10x11 + 15x12 + 9x13 + 9x21 + 18x22 + 5x23 + 6x31 + 14x32 STEP 1: REDUCE THE INITIAL MATRIX BY SUBTRACTING
+ 3x33 THE SMALLEST ELEMENT IN EACH ROW FROM EVERY
ELEMENT IN THAT ROW. THEN, USING THE ROW-
Constraints (s.t.): REDUCED MATRIX, SUBTRACT THE SMALLEST ELEMENT
1) x11 + x12 + x13 < 1 IN EACH COLUMN FROM EVERY ELEMENT IN THAT
2) x21 + x22 + x23 < 1 COLUMN.
3) x31 + x32 + x33 < 1 Initial Matrix
4) x11 + x12 + x13 = 1 Client
5) x12 + x22 + x32 = 1 Project Leader 1 2 3
6) x13 + x23 + x33 = 1 1 – Terry 10 15 9
2 – Carle 9 18 5
xij > 0 for i = 1, 2, 3 and j = 1, 2, 3
3 – McClymonds 6 14 3
Choose the smallest number in each row.
Subtract it from every element in the row.

Client
Project Leader 1 2 3
1 – Terry 1 6 0 Choose smallest of the unlined elements.
2 – Carle 4 13 0 Subtract it from every unlined element. Add it to the
3 – McClymonds 3 11 0 intersection.
Choose the smallest number in each column.
Subtract it from every element in the column.

Client
Project Leader 1 2 3
1 – Terry 0 0 0 Lined numbers (not intersection) will remain the same.
2 – Carle 3 7 0
3 – McClymonds 2 5 0 STEP 2

STEP 2: FIND THE MINIMUM NUMBER OF STRAIGHT

LINES THAT MUST BE DRAWN THROUGH THE ROWS AND
THE COLUMNS OF THE CURRENT MATRIX SO THAT ALL
THE ZEROS IN THE MATRIX WILL BE COVERED. IF THE There is already an optimal answer since the # of lines = #
MINIMUM NUMBER OF STRAIGHT LINES IS THE SAME AS of rows or columns.
THE NUMBER OF ROWS (OR EQUIVALENTLY, COLUMNS),
AN OPTIMAL ASSIGNMENT WITH A VALUE OF ZERO CAN Optimal solution. Locate any row or column that contains
BE MADE. IF THE MINIMUM NUMBER OF STRAIGHT LINES only one zero.
IS LESS THAN THE NUMBER OF ROWS, GO TO STEP 3. If all the row/column have more than 1 zero. Choose the
row/column with the fewest zeroes.

No. of straight lines = 2 Row 2 has only 1 zero. Therefore, it is the chosen row.
No. of Rows/Column = 3 Assign Carle to serve Client 3.
Therefore, no optimal solution yet. Go to Step 3.

If minimum straight lines = columns/rows, optimal

assignment with a value of 0 can be made.
If minimum straight lines < number of rows, go to Step 3.
Draw a line to eliminate them from further consideration.
STEP 3: SUBTRACT THE VALUE OF THE SMALLEST
UNLINED ELEMENT FROM EVERY UNLINED ELEMENT AND
ADD THIS SAME VALUE TO EVERY ELEMENT AT THE
INTERSECTION OF TWO LINES. ALL OTHER ELEMENTS
REMAIN UNCHANGED. RETURN TO STEP 2 AND
CONTINUE UNTIL THE MINIMUM NUMBER OF LINES Row 1 still has 2 zeroes, Row 3 has only 1 zero. Therefore,
NECESSARY TO COVER ALL THE ZEROS IN THE MATRIX IS it is assigned to McClymonds.
EQUAL TO THE NUMBER OF ROWS. Eliminate the row and column after.
Assign McClymonds to serve Client 1.
 number of days required by the assignments actually
made.

If the assignment alternatives are evaluated in terms of

Last remaining unlined zero. revenue or profit rather than time or cost, the linear
Assign Terry to serve Client 2. programming formulation can be solved as a
maximization rather than a minimization problem.
In summary, the optimal solutions are as follows to assign − We need to convert the matrix into opportunity cost.
project leaders to the clients in order to minimize the
number of days to complete their assignments. If one or more assignments are unacceptable, the
• Assign Carle to serve Client 3. corresponding decision variable can be removed from the
linear programming formulation. This situation could
• Assign McClymonds to serve Client 1.
happen if an agent did not have the experience necessary
• Assign Terry to serve Client 2.
for one or more of the tasks.
− Suppose that you have 3 employees to assign 3 clients.
Original Matrix
But let us say that one client refuses to be served by
client 3. This is an example of unacceptable
assignment.

ASSIGNMENT PROBLEM 2
Salisbury Discounts has just leased a new store and is
Assigning Terry to serve Client 2 will take 15 days. attempting to determine where various departments
Assigning Carle to serve Client 3 will take 5 days. should be located within the store. The store manager has
Assigning McClymonds to serve Client 1 will take 6 days. four locations that have not yet been assigned a
The optimal (minimum) time it will take to serve the 3 department and is considering five departments that
clients is 26 days. might occupy the four locations. The departments under
consideration are shoes, toys, auto parts, housewares,
ASSIGNMENT PROBLEM and videos. After a careful study of the layout of the
Because the assignment problem can be viewed as a remainder of the store, the store manager has made
special case of the transportation problem, the problem estimates of the expected annual profit (in thousands of
variations that may arise in an assignment problem dollars) for each department in each location.
parallel those for the transportation problem.

Specifically, we can handle:

1. Total number of agents (supply) not equal to the total
number of tasks (demand).
2. A maximization objective function.
3. Unacceptable assignments.
Note! There are 4 columns and 5 rows here.
ASSIGNMENT PROBLEM VARIATIONS There are two problem variations present here:
If the total number of agents (supply) are not equal to the • Number of rows is not equal to number of columns.
total number of tasks (demand): • Maximization problem instead of minimization.
• This is analogous to total supply not equaling total
demand in a transportation problem. Inserting a Dummy Location:
• The extra agents simply remain unassigned in the Location
linear programming solution. Department 1 2 3 4 5
• If the number of tasks exceeds the number of agents, Shoe 10 6 12 8 0
the linear programming model will not have a feasible Toy 15 18 5 11 0
solution. By adding dummy project leaders, we can Auto parts 17 10 13 16 0
create a new assignment problem with the numbers Housewares 14 12 13 10 0
of project leaders equal to the number of clients. The Videos 14 16 6 12 0
objective function coefficients for the assignment of We do this in order to apply the Hungarian Method since
dummy project leaders would represent the total it requires that # of column = # of rows.
Converting the maximization assignment problem to a Therefore, there is no optimal solution yet. Go to Step 3.
minimization assignment problem: Since it has to be the straight lines = # of rows/columns
We can obtain an equivalent minimization assignment
problem by converting all of the elements in the matrix to STEP 3
opportunity losses. We do so by subtracting every element
in each column from the largest element in the column.
Finding the assignment that minimizes opportunity loss
leads to the same solution that maximizes the value of the
assignment in the original problem. Thus, any
maximization assignment problem can be converted into
a minimization problem by converting the assignment
matrix to one in which the elements in which represents
opportunity losses. Hence, we begin the solution to this
maximization assignment problem by developing an 2 is the smallest number. Subtract it to unlined numbers.
assignment matrix in which each element represents the Add to intersections.
opportunity loss for not making the best assignment.

Find the largest element for each column.

Subtract it to the rest of the elements in the column.
Lined numbers (not intersection) remain the same.

STEP 2

It becomes an opportunity loss because if you assign the

Auto Parts department to Location 1, you will be able to
earn $17,000. However, if you assign the Shoes
department to Location 1, you will have a profit of only
$10,000. You will have a loss opportunity of $7,000.

STEP 2 # of Straight Lines = 4

For maximization problems, no need to perform Step 1. In # of Rows/Columns = 5
lieu of reducing initial matrix, we converted the Therefore, there is no optimal solution yet. Go to Step 3.
maximization to opportunity loss minimization problem. Since it has to be the straight lines = # of rows/columns

STEP 3

# of Straight Lines = 4
# of Rows/Columns = 5
1 is the smallest number. Subtract it to unlined numbers.
Add to intersections.

Assign Video Department to Location 1

Lined numbers (not intersection) remain the same

STEP 2

Assign Toy Department to Location 2

Assign Auto Parts Department to Location 4

In summary, we assign the departments to these locations

# of Straight Lines = 5 to achieve maximum profit…
# of Rows/Columns = 5 • Assign Housewares Department to Location 3
There is already an optimal answer since the # of lines = # • Assign Shoe Department to Location 5 (Dummy
of rows/columns. Location)
• Assign Video Department to Location 1
Optimal Solution. Locate any row or column that contains • Assign Toy Department to Location 2
only one zero. • Assign Auto Parts Department to Location 4

Original Profit Matrix

Location
Department 1 2 3 4
Shoe 10 6 12 8
Toy 15 18 5 11
Auto parts 17 10 13 16
Housewares 14 12 13 10
Videos 14 16 6 12
Assign Housewares Department to Location 3 • Shoe department will not be assigned.
• Toy department will be assigned to Location 2,
earning a profit of $18,000.
• Auto parts department will be assigned to Location 4,
earning a profit of $16,000.
• Housewares department will be assigned to Location
3, earning a profit of $13,000.
• Video department will be assigned to Location 4,
earning a profit of $14,000.
• This assignment will enable Salisbury to earn a
maximum profit of $61,000. (Optimal Answer)
Assign Shoe Department to Location 5 (Dummy Location)
UNACCEPTABLE ASSIGNMENT PROBLEM :( STEP 2
Assume the same problem for Salisbury Discounts except
that the store manager believed that the toy department
should not be considered for location 2 and that the auto
parts department should not be considered for location
4.

Assign values to these unacceptable assignments:

-M = Unacceptable Maximization Problems
M = Unacceptable Minimization Problems
# of straight lines = 4
Where M is an arbitrarily large value in fact, we assume # of rows/columns = 5
M to be so large that M plus or minus any value is still Therefore, there is no optimal solution yet. Go to Step 3.
extremely large. Thus, an M-valued cell in an assignment It has to be # straight lines = # rows/columns.
matrix retain its M value throughout the matrix reduction
calculation. An M-valued cell can never be 0. So, it can STEP 3
never be an assignment in the final solution.

Location
Department 1 2 3 4 5
Shoe 10 6 12 8 0
Toy 15 -M 5 11 0
Auto parts 17 10 13 -M 0
Housewares 14 12 13 10 0
Videos 14 16 6 12 0
-M since this is a maximization problem. Assign it to those 1 is the smallest number. Subtract it to unlined numbers.
that are deemed unacceptable. Add to intersections.

Converting the maximization problem to a minimization

problem:

Convert to opportunity loss matrix. As you convert it, it

becomes a minimization problem.
Subtract highest value of each column to the values in that Lined numbers (not intersection) remain the same.
column.
STEP 2

-M will be changed to M.

There is already an optimal answer since the # of lines = #

of rows/columns.
Optimal solution. Locate any row or column that contains Original Profit Matrix
only one zero. Location
Department 1 2 3 4 5
Shoe 10 6 12 8 0
Toy 15 -M 5 11 0
Auto parts 17 10 13 -M 0
Housewares 14 12 13 10 0
Videos 14 16 6 12 0
• Shoes department will not be assigned.
• Toy department will be assigned to Location 4,
Assign Housewares Department to Location 3 earning a profit of $11,000.
• Auto parts department will be assigned to Location 1,
earning a profit of $17,000.
• Housewares department will be assigned to Location
3, earning a profit of $13,000.
• Video department will be assigned to Location 2,
earning a profit of $16,000.
• This assignment will enable Salisbury to earn a
maximum profit of $57,000. (Optimal Answer)

SHORTEST-ROUTE PROBLEM
Assign Shoes Department to Location 5 (Dummy)
The shortest-route problem is concerned with finding the
Assign Auto Parts Department to Location 1
shortest path in a network from one node (or set of
nodes) to another node (or set of nodes).
• If all arcs in the network have nonnegative values,
then a labeling algorithm can be used to find the
shortest paths from a particular node to all other
nodes in the network.
• The criterion to be minimized in the shortest-route
problem is not limited to distance even through the
term “shortest” is used in describing the procedure.
Other criteria include time and cost. (Neither time nor
Assign Toy Department to Location 4 cost are necessarily linearly related to distance).

MAXIMAL FLOW PROBLEM

The objective in a maximal flow problem is to determine
the maximum amount of flow (vehicles, messages, fluid,
etc.) that can enter and exist a network system in a given
period of time.
• In this problem, we attempt to transmit flow through
all arcs of the network as efficiently as possible.
• The amount of flow is limited due to capacity
restrictions on the various arcs of the network. For
Assign Video Department to Location 2 example, highway types limit vehicle flow in a
transportation system, while pipe sizes limit oil flow
In summary… in an oil distribution system. Best used when solving
• Assign Housewares Department to Location 3 traffic problems. Example, how to program the
• Assign Shoes Department to Location 5 (Dummy) vehicular flow that maximizes the number of cars
• Assign Auto Parts Department to Location 1 being able to pass through thereby improving traffic
• Assign Toy Department to Location 4 conditions. In this type of problems, we deal with
• Assign Video Department to Location 2 constraints. Even though how efficient our solution is,
but if our roads have two lanes for example, then we
 cannot expect that there will be a considerable easing
up of traffic problems.
• The maximum or upper limit on the flow in an arc is
referred to as the flow capacity of the arc.
• Even though we do not specify capacities for the
nodes, we do assume that the flow out of a node is
equal to the flow into the node.