From: Ricky Gonzales and Fralin Widener

To: Dr. Hala Nelson

Subject: Math 340 - Operations Research I: Chapter Ten Summary

Due Date: February 20th, 2025

______________________________________________________________________________

10.0 Summary

In this section is an introduction to the chapter. This section also explains networks and how

useful they are in the areas of transportation, electrical, and communication. It explains that

network representations are widely used in diverse problem areas and that they provide a

powerful aid for portraying the relationships between components of systems used in every field.

There has been an unusual rapid advancement in the methodology and the application of network

optimization models and that many of these models are actually just linear programming

problems. There are five important kinds of network problems and how to solve them. These

kinds are the shortest path problem, the minimum spanning tree problem, the minimum cost flow

problem, and determining the most economical way to conduct a project so that it can be

completed by its deadline.

10.1 Summary

This section introduces a prototype example that will be used to illustrate the first three kinds of

network problems and how to solve them. The example is a problem in Seervada Park (Figure 1)

that has three problems. The problems are that they need to determine the smallest total distance

from the entrance (station O) to station T for the operation of the trams (shortest path problem),

the placements of the installation of the telephone lines (minimum spanning tree problem), and
that there are too many people wanting to ride the tram to station T from the entrance (station O)

than can be accommodated in the park’s peak season (maximum flow problem).

Figure 1: Seervada Park layout

10.2 Summary

This section explains the terminology of networks. Nodes are sets of points within a network.

Lines connecting nodes are called arcs. If the flow through an arc is only allowed in one

direction, this is a directed arc. If the flow is allowed in either direction, then this is an

undirected arc, also referred to as a link. A directed network is a network that has only

directed arcs. And if all arcs are undirected then the network is an undirected network. A path

is a connection of distinct arcs connecting two nodes. A directed path is a path from two nodes

that the direction is specifically towards the ending node, and an undirected path is a path from

two nodes where the direction can be either towards or away from the ending node. Two nodes

are said to be connected if the network contains at least one undirected path between them. A

network is a connected network if in the network every pair of nodes is connected. A tree is a

connected network that contains no undirected cycles. A spanning tree is a connected network

that for all n nodes contains no undirected cycles. The maximum flow that can be carried on a

directed arc is called the arc capacity. A supply node is a node that flow out of the node and

exceeds the flow into the node. A demand node is a node that the flow into the node exceeds the
flow out of the node. A transshipment node is a node where the flow into the node is equal to

the flow out of the node.

10.3 Summary

This section describes the shortest path problem, including the algorithm and applications. The

pseudo-algorithm for the shortest path problem is finding the objective of the nth iteration,

inputting for the nth iteration, getting candidates for the nth nearest node, and calculating the nth

nearest node. Using an algorithm to solve a shortest path problem can be a little complicated and

the software required might not be available to the average person. So, using Excel, which is

generally accessible, is a great tool to help solve shortest path problems. This is because Excel

relies on the general simplex method. For the shortest path problems, there are other applications

other than travel. Optimizing with arcs and links can also help one to minimize cost or time of a

sequence of activities, or even a mix of the three categories.

10.4 Summary

This section describes the minimum spanning tree problem using the example of Seervada Park.

The minimum spanning tree problem is similar to the shortest path problem, except that for the

minimum spanning tree problem there must be a link between each node, not just the origin and

the destination. Given the potential links and the positive length for each, one is to design the

network that satisfies the requirement that there be a path between every pair of nodes,

minimizing the total length of the links inserted into the network without cycles. An important

distinction is that for a network with n nodes there only requires n - 1 links for the minimum

spanning tree problem. There are multiple categories one could solve with the application of the

minimum spanning tree problem, such as designing telecommunication networks, a

transportation network, electrical power line network, wiring equipment, or pipeline networks.
The general algorithm of the minimum spanning tree problem is best executed graphically, but

involves selecting an arbitrary node (it does not matter which), connecting it to the nearest

distinct node, identifying the unconnected node that is closest to a connected node, connecting

those two nodes, and repeating the process until all nodes are connected.

10.5 Summary

This section talks more about the max flow problem for the Seervada Park. They go into detail

about how to select the best route in this network through the various nodes with the specific

constraints. This section also gives some applications about what network optimization could

consist of, like flow for factories, vendors, oil, water, and vehicles. The applications listed can

also depend on if there are various origins or various destinations.

This section also introduces an algorithm called the linear programming problem to solve

network flow models. The augmenting path algorithm allows for the calculation for what the

capacity of the network is. The logic here is that you want to send an amount from the origin and

figure out the minimum amount that can go through which is the maximum amount of flow that

is allowed in the network.

10.6 Summary

This section talks about the minimum cost flow problem, which is similar to the max flow

problem. The process to solving the min cost flow problem consists of having a directed and

connected network with multiple nodes with sufficient capacity. This section also goes into detail

about the applications of this type of optimization technique. The model that results from this

problem arises from classifying the connections as variables and constraining them based on

their capacity. It is a condition that the sum of the total flow generated at the supply nodes is
equivalent to the total flow being absorbed, like kirchoff's laws in Physics. You can use excel to

solve these types of problems, and the section ends with special cases of network problems.

10.7 Summary

This section talks about the Network Simplex Method, which is more efficient than the standard

simplex method for solving min cost flow problems. This method executes the same steps,

except in ways that exploit the structure of the network to solve it more efficiently. The section

then goes into detail of how to use the Network simplex method.

10.8 Summary

This section talks about the network model for optimizing a project's time-cost trade-off. This

pretty much entails how certain deadlines for projects require special attention and may not give

enough time to determine an ideal solution. So this introduces time constraints, which the section

gives an example of. A project network consists of a number of nodes and arcs which define the

structure of the network. The section also talks about crash decisions, and gives an example

using linear programming to resolve it.

10.9 Summary

This section concludes the types of networks that are present in portraying the relationships

between the components of systems. Decisions need to be made and all conditions need to be

met to build a flow network that is optimized in the design of what type of product or material is

flowing through.