FIN-3314 Management Science Applications-2023

2. Modeling in Management Science

_______________________________________________________________________________

Introduction

In management science, we approach various managerial problems in the context of models. We

attempt to scale down real complex problems to approximate simple relationships so that we can
understand and analyze them. Because models play an essential role in management science, this
chapter is devoted to modeling and its functions.

What is a Model

There are all types of models around us. The models we see and use most often are photographs,
road maps, model airplanes, and Raggedy Anns. Mathematical equations, chemical molecule
relations, conceptual theories, and our mental images are also some examples of models. Every
model has one basic purpose of representing some aspects of reality by means of a simple object. So
let us state it one more time. The primary purpose of modeling is to understand a complex problem
through a simpler, less expensive, and less cumbersome object.

Model Classification

There are many different ways to classify models. However, the most widely accepted taxonomy for
models is abstract and exact. Abstract models are extremely fuzzy and ill-structured representations
of reality. As with your mental model discussed earlier, they have no physical or symbolic
configuration. Exact models, on the other hand, have some physical characteristics that resemble
the reality under study. Physical models include such things as model airplanes, dolls, and an
architect's scaled-down buildings. Between these two extremes, there are symbolic models such as
theories, written statements, and mathematical relationships.

As we discussed previously, mental models are the most abstract representation of reality, such as
imagination
.
Verbal Models. Verbal models use written models to represent mental models with Worlds, such as
poetry, plays, novels, theories, and a police officer's traffic accident report.

Mathematical Models. Mathematical Models are also symbolic but consist of mathematical
relationships rather than words. Most management science models we will study in this course are
mathematical models.

Analog Models. Analog Models are also physical models that may look like the reality they
represent. Typically, they focus on certain basic functions or relationships. Maps, blueprints, and
organizational charts are good examples of analog models.

Iconic Models. Iconic Models are physical replicas of reality, usually smaller or bigger in scale than
the actual object. Many three-dimensional models, such as model airplanes, buildings, dolls, or two-
dimensional paintings and photographs, are good examples of iconic models.

7
Management Science Modeling

Among all the models discussed above, only mathematical models enable us to communicate
precisely with the model builder without misunderstanding. For example, we cannot always
comprehend accurately what others have in mind when they verbalize their mental models, but if
someone presents his or her annual income as I = 12x, where x = monthly income of Rs.20,000/=
there is no communication problem as to what the person's yearly income is.

Since mathematical models employ the language of mathematics in describing the system under
study, we can also manipulate the models so that we can test the behavior of a system under
varying conditions. For example, in the income model I = 12x, we can easily determine a new annual
income when "x" is increased from Rs. 20,000 to Rs. 25000. We can easily expand the model by
adding other variables such as interest earned, dividends, and other income.

Management science models are good examples of mathematical models. They are formulated and
used to understand or predict certain management systems. One obvious disadvantage a
management science model is that it is something less than the real system it represents. However,
this aspect of a model is also an advantage. Since a model is simple, by analyzing it, we can
understand how the complex system functions. Thus, we can test the behavior of the system under
different conditions. The real challenge of modeling is to build as simple a model as possible while
including all pertinent attributes of the system.

The Model Structure

Management science models are usually in the form of mathematical relationships such as
equations or inequalities. For example, your disposable income can be expressed as a relationship
between your gross income (wages, bonuses, dividends, interest income and so on) and deductions
(taxes, insurance premiums, pension payments and so on) as follows:

DI = GD

where

DI = disposable income
G = total gross income
D = deductions

The above model is perhaps the simplest example of a mathematical model. The complexity as well
as the nature of the model will, of course, depend on the nature of the problem under study and the
purpose of the analysis.

The Model Components

When we attempt to construct a model, the first step is to abstract important components that
explain the behavior of the system we are about to analyze. Such components are usually referred
to as variables. In your disposable income model, G and D are good examples of variables. In a broad
sense, we classify variables into two categories: dependent and independent.

The Dependent Variables The dependent variables are sometimes referred to as "criterion
variables." The value taken on by a dependent variable reflects the level of performance
achievement of the system, for example, in the disposable income model.

8
The Independent Variables The independent variables are those that are not dependent on other
variables in the model. In general, there are two types of independent variables: the decision
(controllable) variables and the exogenous (uncontrollable) variables.

The Decision Variables The decision variables are often the most important components in
management science models. It is the values of these variables for which a solution is sought.

The exogenous variables The second type of independent variables, the exogenous variables, affect
the outcome of the model and the system, but we have very little control over them. For a
production problem, exogenous variables may include the cost of materials and human resources
and pollution-control regulations.

Model Parameters The Parameters are the remaining components of the model that are essential in
developing the relationships among the variables. Parameters are generally classified into two
categories: constant and random.

Constant Parameters The Parameters are often assumed to be constant; that is, a static state exists.
Such an assumption may be valid for a short time before major environmental or organizational
factors change.

Random Parameters in a more dynamic or complex situation. The model parameters may not have a
constant relationship. But rather the parameters themselves vary according to some probability
distribution.

Relationship Within the Model

The essence of management science models is the representation of the relationships among the
various components of the model. The relationship can be illustrated graphically. For example, your
wage model W = 5H can be presented graphically.

Decision Making Environment

Management science is primarily concerned with searching for information about feasible courses of
action. The type of management science models used may be based both on the mature of the
problem at hand and on the decision environment. There are four different states of decision
environment: certainty, risk, uncertainty, and conflict. We will discuss each decision environment
in greater detail.

The Process of Management Science Modeling

Management science is the application of scientific approaches to management decision-making.

There is no one correct way to apply management science. All practitioners of management science
have their ways of doing things. However, there are several major steps that almost everyone agrees
are important in management science application:

1. Formulation of the problem

A. Orientation period
B. Definition of problem components

(1) The decision maker

(2) The decision objectives
(3) The decision environment
(4) Alternative courses of action

9
2. Development of the model
3. Model validation
4. Solution of the model
5. Implementation of the solution

Let us discuss these major phases as outlined above to investigate the role of modeling in
management science.

Formulation of the Problem

Although management science application begins with the formulation of a problem, this step must
be a continuous process until the completion of the project. Once an initial formulation of the
problem is completed and subsequent step proceed, the problem under study is subjected to
continuous modification. Consequently, a continuous updating or modification of the problem is
necessary to assure the validity of the model solution.

Each decision-making problem may require a unique approach to formulate it However, there are
several generalized steps in the problem formulation phase.

Orientation Period The orientation period provides us with an opportunity to assess the overall
picture of the problem. During this period, we can obtain a broad understanding of the
organizational climate, the objectives of the organization, and the purpose of the management
science analysis. The orientation period can be used to specify the conditions required to carry out
the analysis, such as time and resource requirements, administrative arrangements, and the scope
of the application.

Definition of Problem Components Before the problem can be formulated, and its components
must be clearly defined. The first component to be defined is the decision maker who is not satisfied
with the current state of affairs. The second component to be analyzed is the objectives of the
decision-maker. What are the things that the decision-maker is trying to achieve through the
analysis? The third component is the decision environment or system that embraces the problem.
Finally, a problem cannot be evaluated unless the decision maker has an alternative course of action.

Development of the Model

Modeling is the crux of the management science approach. It is as important as laboratory

experiments in physical science. We are getting into the fine details of the problem in this phase.
Developing a model allows a comprehensive analysis using a logical expression of the complexities,
unique characteristics, and possible uncertainties of the problem. The logical expression requires a
mathematical formula to represent the interrelationships among the system's elements.

The model serves as a convenient vehicle that helps analyze a complex reality concisely and
relatively simply. A model clarifies the feasible decision alternatives, the economic and non-
economic consequences of these alternatives and the optimum alternative for the problem. A
systematic analysis is possible only when the relationships among the components and their
objective criteria are expressed in a comprehensive but manageable mathematical model.

10
Model Validation and Data Collection

Once the model is developed, it is important to test model concerning the validity of its
assumptions. The problem at hand may change over time and thus a continuous updating and
validating of the model components, their relationships, and the objective criteria may be necessary.
It is imperative that the objective criteria be continuously evaluated, as management goals often
change in time with the changes in the decision environment.

Frequently it is necessary to modify a model, even when the model is correctly developed because
required data are not available or are too costly to obtain. Such data are often uncontrollable inputs
or parameters to the model. Sometimes, raw data must be analyzed and rearranged by statistical
techniques before they become useful input to the model.

Although the initial model development may be simple, its modification may be extremely difficult.
Thus, it may be necessary to go through several cycles of model validation and search for more data.
Each model leads to a search for data that may only sometimes yield the required information.
However, the search process often reveals how the model should be modified to use the available
data.

Solution of the Model

Once a mathematical model is formulated, the next phase is deriving an optimum solution from the
model. The optimum solution presents the model result that optimizes the given decision criterion
(objective). As pointed out earlier, a mathematical model is a simple abstract of reality. If the model
represents almost every characteristic of the problem, it may be too complex to allow easy
formulation, manipulation, and solution. However, if the model is too simple, it may only represent
some of the relevant aspects of the problem. Therefore, the model solution phase may determine
whether the model is appropriate for the problem. This phase may give rise to a continuous review
of the problem-formulation/model-development/model-solution/model-validation cycle until a
satisfactory solution is found.

Implementation of the Solution

The true value of management science is realized when the model solution is put to actual use. This
phase usually requires a translation of the solution into a set of management policies or operating
procedures that can be easily implemented by the operating personnel. A problem that plagues
most management science practitioners is that the system under investigation keeps changing.
Consequently, a control process must be established over the solution so that the model can be
updated continuously.

11