Management Science

1
Network Modeling

Module 10 Network Modeling

At the end of this module you are expected to:

1. Evaluate the Transportation Model
2. Discover the Assignment Model
3. Interpret the Shortest Path Method
4. Infer the Network Flow Model

Arranged paths connected in different points where things move from one point to
another is called a Network. Highway systems, railroad systems, television networks, and
telephone networks are examples of a network that virtually everyone is familiar with. An
example is a railroad network consisting of a number of paths or routes of the rails conjoined by
terminals in different intersections of the rail routes. The use of network models has in recent
years, become an in-demand technique of management science for the following important
reasons:

1) Network models literally provide a depiction of the system to be analyzed because it is

drawn as a diagram.
2) A large number of real-life systems can be modeled as networks.

Network Components
Illustrated as diagrams, there are two main components of networks: nodes
and branches. Junction points are represented by nodes, an example is an intersection
of several streets. Connecting the nodes on the other hand are branches. Branches
reflect the flow from a part of the network to another. Nodes are denoted by circles in
the diagram of a network, while lines denote the branches that connect the nodes.
Typically, nodes represent localities, intersections, railroad or airport
terminals and cities. Similar to roads that connect cities and junctions or air routes or
tracks of railroad connecting terminals are paths called branches that connect the
nodes. Representing distance, time, or cost are values usually assigned to branches.

Course Module
 Management Science
2
Network Modeling

Therefore, determining the shortest length of time, the shortest distance, or

lowest cost between network points is the purpose of a network.

A. Transportation Model

The following are unique characteristics of a class of problems where the

transportation model is formulated for:

1) Products or items are shipped from several possible sources to several possible
destinations at the lowest cost possible
2) A fixed number of units of the products or items can be supplied by each source, and
there is a fixed demand for the item or the product in each destination.

Generally, there is a wide variety of problems where the transportation model can be
applied. This particular application of the transportation of goods, meaning from the two
unique characteristics, that the problem draws its name from. In other words, the model’s aim
is to find the optimum transportation schedule at the minimum cost possible.

Now, what is a transportation problem? It is a problem of the special kind, a kind of

programming that is linear. The objective of the transportation problem is to minimize the
distribution cost of a product from a number of origins or sources to a number of end points. It
is important to note that the usual simplex method cannot be used or is not suitable to solve
transportation problems.

There are two types of Transportation Problems:

1) Balanced Transportation Problem where the total supply is equal to the total demand;
and
2) Unbalanced Transportation Problem where the total supply is not equal to the total
demand.

There are also certain applications of the Transportation Model like the following:

1) To minimize shipping costs

2) To determine a low-cost location
3) To find the minimum cost production schedule
4) To utilize a military distribution system
 Management Science
3
Network Modeling

The solution to a transportation problem follows or happens in 2 phases:

Phase 1 – obtaining the initial basic feasible solution

Phase 2 – obtaining the optimal basic solution

To better understand the formulation of a transportation model, let us take the following
example:

In the Midwest United States, Wheat is harvested and is stored in three

different cities where grain elevators are located namely, Kansas City, Omaha, and
Des Moines. Three mills of flour found in Chicago, St. Louis, and Cincinnati are
supplied by these grain elevators. Railroad cars are used to ship the grains to the
mills, with each transport able to carry a ton of wheat. On a per month basis, every
elevator of grain is able to supply these tons (railroad cars) of wheat to the mills:

The following monthly tons of wheat is required by each mill:

Course Module
Management Science
4
Network Modeling

The transporting cost of 1 ton of wheat from each source (grain elevator) to each
destination (mill) is different from each. It will depend on the rail system and
distance. The following are our cost assumptions or estimates per location:

Our problem is determining how much wheat to ship from the grain elevators to
each mill on a per month basis to lower the transportation cost. Figure 1 shows the
diagram of the various directions with supply and demand.

Figure 1 Network of Transportation Routes for Wheat Shipments

For this problem the following can be the formulated linear

programming model (starting with the objective function as we have learned in
previous lessons):
 Management Science
5
Network Modeling

Representing the tonnage of wheat transported from each grain

elevator to each mill are the decision variables Xij, in this model. This is where
subscript i representing the grain elevator is equal to 1,2,3 and subscript j
representing the mill is equal to A, B, C.

The total cost of transportation for every route is represented by the

objective function. The tonnage cost transferred from one route is reflected by each
term in the objective function. An example is if mill A gets 20 tons transported to it
from elevator 1, the cost of $6 gets multiplied by X1A = 20 which would equal to
$120.

Represent the supply at each elevator. An example is the first supply

constraint X1A + X1B + X1C = 150, which represent the tonnage of wheat shipped from
the City of Kansas to all the other mills, Chicago (X1A), St. Louis (X1B), and Cincinnati
(X1C). 150 tons is the limit to the tonnage transported from Kansas City.

Course Module
Management Science
6
Network Modeling

Represent the demand at each mill. It is important to note that all the
constraints are equalities (=) rather than inequalities (≤) because the total demand
of 600 tons need to be met by all the tons of wheat available.

Simply put, the demand coming from the three mills totals 600 tons
which is the amount that can be supplied exactly by the three elevators of grain.
Therefore, all that can be supplied will match whatever amount is demanded. This
type of problem or model is what we call the Balanced Transportation Problem or
Balanced Transportation Model.

In reality though, an Unbalanced Transportation Problem or

Unbalanced Transportation Model, in which demand exceeds supply or supply
exceeds demand is more likely to happen.

In the example, if demand at Cincinnati will be increased to 350 tons

from 300 tons, the situation will be that total demand 650 tons is higher than total
supply 600 tons. Our linear programming model will then change to look like the
following:
 Management Science
7
Network Modeling

Since there is a great lack of total supply to fulfill total demand, one of
the constraints of demand won’t be fulfilled. If supply has great excess over demand,
then the constraints of supply will be ≤.

There would be times that one or more routes may be “prohibited” in

the transportation model. This is when items can’t be transferred from one source
or a specifc source to another or specific end point. If this circumstance happens, it
should be made sure that the variable which represents the route doesn’t have a
value in the solution which is optimal. In order to accomplish this, a huge relative
cost must be assigned as prohibited variable coefficient in the objective function. In
the example of transporting wheat, if the way from the City of Kansas to Chicago is
unallowed (perhaps due to a rail strike), 100 is the given coefficient to variable X1A
in the objective function instead of 6, so X1A will be zero in the optimal solution
because it has a high relative cost. As an alternative, the prohibited variable can be
removed in the formulation of the model.

B. The Assignment Model

Similar to the Model of transportation is a special form of programming that is

linear in nature called the Assignment Model. However, it has differences with the
Transportation Model. The demand at each destination and the supply at each source is
limited to one unit in the assignment model.

Let us take this example to demonstrate the assignment model. The ACC which
stands for Atlantic Coast Conference holds 4 games of basketball on a certain night. The
Office of the Commissioner wants to assign 4 teams of officials to the 4 games in such a way
that the total travelled distance by the officials is at its minimum. One team of officials will
always be the supply, and the demand is will only be for one team of officials at every game.

Now let us show how far in miles will be for each team of officials to every location
of game.

Course Module
 Management Science
8
Network Modeling

Similarly the transportation model formulation is the formulation of the linear

programming for the assignment model but the difference is that all values of the supply
for every source is equal to 1, and at each destination all the demand values is also equal to
one. Thus, the following is our linear programming formulation for the assignment model:

A balanced assignment model is shown by the previous linear programming model.

Again, a model which is unbalanced exists when supply is greater than demand, or vice
versa. It is also important to note that the assignment model’s application is not only in
Network Modeling, but it is mainly used in solving one to one assignment problems of
assigning employees to jobs, employees to machines, machines to jobs, etc. The objectives
in these other types of assignment problems can either be to minimize or to maximize.

C. Shortest Path Model

 Management Science
9
Network Modeling

The shortest path or route problem wants a determination of the shortest distance
between the point of origin and different end or final points. For example, six trucks of the
company Stagecoach Shipping transport oranges from city of Los Angeles to 6 cities in the
Midwest and West part of the USA. The cities of destination, and the length of time in hours a
truck needs to travel from Los Angeles in the different routes are shown in Figure 2.

Figure 2 Shipping routes from Los Angeles

The determination of the best routes to take for the trucks in terms of the shortest
travel time to get to their destination is what the shipping manager does. By using the
shortest route or shortest path technique, we can solve this problem. Representing the
system of truck routes as a network makes applying this technique convenient.

Figure 3 shows the said network.

Course Module
 Management Science
10
Network Modeling

Figure 3 Network of Shipping Routes

Therefore, in this network, we would like a determination of the shortest route from
the origin or the first node to the destinations, node 2 to 7.

Shortest Path Solution Approach

Our technique obviously will start at the origin or node 1 and determining
the shortest time required to get to an adjacent or directly connected node. As
shown in Figure 3, there are 3 nodes connected directly to node 1 and these are
nodes 2, 3, and 4. Out of the 3 nodes, the route with shortest time is to node 3 which
is 9 hours.

Therefore, the first quickest or shortest path is from the first node to the
third node, Los Angeles to Phoenix. Nodes 1 and 3 will now be referred to as the
“permanent set” which indicates that the shortest route to these nodes has been
found. It is automatic that node 1 is part of the permanent set because it has no
route to it or it is the starting point.
 Management Science
11
Network Modeling

Figure 4 Network with nodes 1 and 3 in the permanent set

The determination of the shortest route to nodes 2, 4, and 6, the three

directly connected to the “permanent set” nodes 1 and 3 is the next step in the
technique. As we can see in Figure 4, two branches arise from node 1, 1-2 and 1-4.
There are also 2 branches from node 3, 3-4 and 3-6. Node 2 has the quickest time at
16 hours and now becomes part of the permanent set.

So now, our permanent set consists of nodes 1, 2, and 3. Being in the

“permanent set” means that the shortest routes have been found to these nodes.
Now, we must determine which nodes are connected directly to the “permanent set”
nodes.

There are five branches that lead out of nodes 1, 2, and 3 to their directly
connected nodes. The branch representing the shortest time is 3-4 (the total time
from 1-3-4) which is 24 hours shown in Figure 5. Having found the quickest route to
node 4 it now becomes part of the “permanent set”. The other routes to node 4 are
already longer which means we will no longer consider these routes.

Course Module
Management Science
12
Network Modeling

Now, continuing the process, we need to determine the other nodes (5, 6, and
7) with a direct connection to the “permanent set”. We can see that in Figure 6 the
branches from number 1 and 2 nodes to node 4 have been removed because we
have already found that these are not the shortest routes to node number 4.
 Management Science
13
Network Modeling

In the branches going to nodes 5, 6, and 7, the route with the quickest or
fastest amount of time is branch 3-6 with the shortest accumulated time of 31 hours.
Node 6 is now made part of our “permanent set” and we have determined the
shortest route to 1, 2, 3, 4, and 6 numbered nodes. In the next Figure, notice that
branch 4-6 has be removed because the shortest route to node 6 is through node 3.

Numbered 4-5 branches has the shortest accumulated time from the
branches going to nodes 5 and 7 which means that we now have the shortest times
to numbers 1, 2, 3, 4, 5, and 6 nodes. In Figure 8, node 7 finally joins the permanent
set as branch 4-7 has the shortest cumulative time to node 7 with 43 hours.

Course Module
 Management Science
14
Network Modeling

Figure 8 Network with nodes 1, 2, 3, 4, and 6 in the permanent set

Figure 9 Network with optimal routes (shortest travel time from origin to destination

Figure shows the optimal routes for the six trucks, route 1-2, route 1-3, route
1-3-6, route 1-3-4, route 1-3-4-5, and route 1-3-4-7.
 Management Science
15
Network Modeling

D. Network Flow Model

In general, our first three topics – the transportation problem, the assignment
problem, and the shortest path method can all be classified as network problems. As we
have already seen, a network is a collection of nodes and arcs, arcs being the directed
line segments depicting a flow entering or leaving a node and passing through nodes.

Flow is therefore associated with the network. External flows are flows entering or
leaving a node. A flow that enters a network is the supply, a positive external flow. A
flow that leaves the network is the demand, a negative external flow.

A network model that uses the word flow in its name is the Maximal Flow Problem
or Maximal Flow Model , a network flow model. A maximal flow problem’s goal is to
maximize the total flow amount from a starting point to an end point. Maximal Flow
problems lends itself to a lot of real-world scenarios like the product flow through a
system of production line, the flow of water, gas, or oil through a network of pipelines,
the flow of forms through a paper processing system, or the flow of traffic through a
road network. In the foregoing examples, the branches of the network have limited and
often different capacities. With the said conditions, the decision maker would like to
determine the maximum flow that can be obtained through the system.

Course Module
 Management Science
16
Network Modeling

The system of railway network between Omaha and St. Louis depicted in Figure 10
is an example of a maximal flow problem. Tractor parts shipped from Omaha to St.
Louis by the Scott Tractor Company through a railroad. Unfortunately, due to a
contract, there are limitations to the number of railroad cars that a company can
secure on each branch during a week.

Figure 10 Network of Railway System

With these limiting conditions or constraints, the company wants to

maximize the number of railroad cars containing tractor parts that can be shipped from
Omaha to St. Louis during a week. To the immediate right of each node representing a
rail junction is a number indicating the available railroad cars to the tractor company on
each rail branch. As seen on Figure 10, there are six cars ready from node 1 to node 2, 8
cars are ready from node 2 to node 5, and so on. At the immediate left of the node is a
number representing the available cars for deployment in the other way. For example,
there are no cars available from node 2 to node 1. This would mean that this is a
“directed branch” because there is only one possible direction of the flow (from node 1
to 2 but not from node 2 to 1). When flow is possible in two ways like in nodes 2 and 4,
and nodes 3 and 4, they are referred to as “undirected branches”.
 Management Science
17
Network Modeling

Course Module