Republic of the Philippines

UNIVERSITY OF EASTERN PHILIPPINES

University Town, Northern Samar
Web: http://uep.edu.ph Email: [email protected]

COLLEGE OF AGRICULTURE, FISHERIES, AND NATURAL RESOURCES

(Agribusiness and Agricultural Economics Department)

Subject: Introduction to Management Science

Instructor:PRECIOUS JADE R. VILLEGAS, MBA
Source: Jan Fábry (UNIVERSITY OF ECONOMICS PRAGUE
Faculty of Informatics and Statistics)
__________________________________________________________________________

INTRODUCTION TO MANAGEMENT SCIENCE

Chapter 1

1. Introduction
Management is a process used to achieve certain goals through the utilization
of resources (people, money, energy, materials, space, time). Management Science (MS),
an approach to managerial decision making based on the scientific method, makes extensive
use of quantitative analysis. The alternative name for quantitative approaches to decision
making is Operations Research (OR). We shall treat both the terms as synonyms throughout
the text. Other names for more or less the same area are: operational research, operations
analysis, quantitative analysis, quantitative methods, decision analysis and decision science.

The significant development of the Operations Research disciplines and techniques

started during World War II in form of military applications (strategic and tactical tasks).
After this period many more methodological developments followed and at the end
of 20th century the information technology explosion created new possibilities for
management science. It is nowadays hardly realizable to carry out an analysis and make
decisions without computers. Many software products are used in management science. Some
simple decisions can be made using standard spreadsheets (MS Excel), whereas the complex
real problems require professional software (Lindo, Lingo, Xpress, AIMMS, CPLEX, etc.).

1.1 Management Science – Definition and Characteristics

Two classical definitions of management science (operations research), according to
Turban and Meredith [8] are:

1. MS/OR is the application of scientific methods, techniques and tools to

problems involving the operations of systems so as to provide those in control
of the operations with optimum solutions to the problems.
MS
2. MS/OR is the application of the scientific method to the study of the operations
of large, complex organizations or activities.
 Management science can be briefly defined as: The application of the scientific
method to the analysis and solution of managerial decision problems. The major
characteristics of management science are:

1. A primary focus on managerial decision making.

2. The application of the scientific approach to the decision making process.
3. The examination of the decision situation from a broad perspective; i.e. the
application of a system approach.
4. The use of methods and knowledge from several disciplines.
5. A reliance on formal mathematical models.
6. The extensive use of computers.

Decision making is a process of choosing between two or more available

alternative courses of action for the purpose of attaining a specific goal. The conclusion
of the process is a decision. Decision making is a systematic process and can be simply
described by the following steps:
a) Defining the problem.
b) Searching for alternative courses of action.
c) Evaluating the alternatives.
d) Selecting one alternative.

This process is, of course, just one of many possibilities how to approach the
problem and is typical for such discipline as Multicriteria Decision Making, etc.

The structure of systems consists of three related parts: inputs, processes and
outputs. Inputs enter the system (people, raw materials, money). Processes convert the
inputs into outputs (processes may use energy, operating procedures, employees,
machines). Outputs leave the system (products, served customers). In many applications
the outputs are connected with the goal of the management (profit maximization, total
costs minimization). We consider closed and open systems. A closed system is totally
independent of and fully isolated from influences of the environment (elements outside
the system). An open system is significantly interconnected with the environment; this
type of system exchanges information, material, energy, people with the environment.

Systems approach involves finding and defining all of the mentioned elements
and relations between them in the real world. This approach is the necessary assumption
for using management science techniques, as they require the exact description of a
problem. Only system approach enables the formulation of the problem in the
mathematical way.
1.2 Models and Modeling
The model is a fundamental term of management science. Modeling is the process
of handling real-world problems and describing them in mathematical terms.

Real-World Problem Definition of

 the Problem

Implementation Formulation and

 Construction of the
Mathematical Model

Model Validation and Solution

Sensitivity Analysis of the Model

Fig. 1.1 Modeling Process

Nevertheless, the model is just simplified representation of the real world and therefore only
important and relevant items or properties can be used in the model.

The first step is the recognition of the problem in reality and its definition. If we, for
example, produce different products, the question is: how to set the production, i.e. how many
of each product should be produced? In this step it is necessary to define goals that the
management wants to achieve. In the example, it could be maximization of revenue
or minimization of total cost. It is obvious that we have to know the details of production
process and all the necessary information from marketing, finance, accounting, etc.:
1. what kinds of material are used and what is the available stock of them,
2. how many people can be involved in the production,
3. what is our budget,
4. how much material, energy, time and money is necessary for producing one unit
of each product, how many people are needed,
5. unit prices of the final products (revenue maximization), unit cost of each product (total
cost minimization),
6. the level of demand for each product.

The following step is the key issue of the modeling process – formulation and
construction of a mathematical model. The problem defined above is being transformed
into the mathematical form, i.e. into the world of numbers, symbols, parameters, variables,
functions, equations, inequalities, etc. In this case, variables correspond to the amounts
of products, parameters are the material stock, budget, unit prices, etc. There are several
constraints (equations, inequalities) in the model, which do not allow us to exceed
production possibilities and which assure that the produced amount is on the demanded level.
 The model includes the objective function for revenue (or total cost). All these terms
are treated in detail in Chapter 2.

Construction of the model is doubtless the most difficult and the most important part
of the whole process. Remember, the model is just the simplification of reality. The simpler
the model, the easier its manipulation and solution will be. However, the simplest models
could be far from reality and it has no sense to implement the results because of their low
significance. The most complex model, e.g. with exact (nonlinear) relationships between
variables, can be perfect representation of reality; it is hardly solvable. Whenever we start to
construct a new model, it is necessary to remember an important rule:

Finding a proper balance between the level of simplification of the

model and the good representation of reality.

There are two kinds of models: deterministic and probabilistic. In deterministic

models all aspects are known with certainty. In the above example, we can e.g. assume that
the stock level, budget or used energy are exact numbers. Similarly, we exactly know how
much of raw material will be used per each unit of a specific product. Deterministic models
are not certainly perfect, but they may offer a reasonably good approximation of reality.
Probabilistic models contain a specific level of uncertainty. Low level of uncertainty can be
ignored and deterministic models can be used instead of the probabilistic models. In case of
high level of uncertainty we should consider random variables instead of constants. In the
illustrative example, we should assume that the demand for our products cannot be
determined exactly. Probabilistic models, of course, require different (mostly more complex)

approach to their solutions, the methods therefore differ from the methods used for
deterministic models.

Solution to the proposed model is more or less a technical job. When we make
a decision on the form of applied model, we should consider methods, techniques
or principles used for the solution. Each area of management science gives typical examples
of problems and suggests specific ways of solution. Knowledge of these typical examples
facilitates the selection of the model form and the used methods. The typical goal of the most
problems is to find an optimal solution, i.e. the best of all feasible solutions (solutions that
satisfy all the constraints). In the example, the optimal solution is the one, which gives the
highest revenue (the lowest total cost), while respecting all the constraints. This solution
determines how many items of each product we should produce and what revenue (total cost)
we can expect.

Expansion of computers brought new opportunities for solving wide range of real
problems and nowadays it is hardly possible to get results without the use of computers and
efficient software. Professional software offers the user-friendly environment with easy
control, enables flexible entering of inputs and provides detailed outputs with possibility
of experiments with the model. From this point of view very important step is the model
validation and sensitivity analysis. Validation of the model is its comparison with the real
system. The better a model represents the reality, the more valid it is. General method for
testing the validity of a model is to compare its performance with some past data available
for the actual system.
 If we want to examine the impact of changes in inputs on changes in outputs, we have
a convenient tool for this purpose – sensitivity analysis. In the example, there is specific stock
of raw material. The level of stock is mostly the limitative factor for the production and hence
for the attainable revenue. The interesting question is: what will happen to the proposed
production (and revenue) if we avail of one additional unit of any material? In case all the
material is used for our production, we can reasonably expect that, using one additional unit,
we are able to produce more products and gain more money.

The final and very important step of the modeling process is the implementation of the
proposed solution to the real system. This is actually the main goal of management and the
original purpose of the whole process - not the model itself, but adjustment of reality
according to the recommendations ensuing from the results of the modeling process. In case
we do not use the results in the real production process, all our effort was absolutely vain.
On the contrary, if we constructed the model in a wrong way and we did not validate it, the
applied results could seriously harm the real system. In order to achieve the best management
results, each step must be carefully considered and cannot be skipped.

1.3 Management Science Techniques

In the previous section it was claimed that the kind of problem indicates the form
of model and used techniques. The following list [8] presents typical managerial problems
where management science techniques could be applied:

🞄 Inventory control.
🞄 Facility design.
🞄 Product-mix determination.
🞄 Portfolio analysis.
🞄 Scheduling and sequencing.
🞄 Merger-growth analysis.
🞄 Transportation planning.
🞄 Design of information systems.
🞄 Allocation of scarce resources.
🞄 Investment decisions (new plants, etc.).
🞄 Project management – planning and control.
🞄 New product decisions.
🞄 Sales force decisions.
🞄 Market research decisions.
🞄 Research and development decisions.
🞄 Oil and gas exploration decisions.
🞄 Pricing decisions.
🞄 Competitive bidding decisions.
🞄 Quality control decisions.
🞄 Machine setup problems in production.
🞄 Distribution decisions.
🞄 Manpower planning and control decisions.
🞄 Credit policy analysis.
🞄 Research and development effectiveness.
For solving these and many other problems the following techniques can be used:

Linear Programming
It is one of the best-known tools of management science. This approach mostly defines the
problem as the maximizing (minimizing) a linear function, respecting the set of linear
constraints. The example mentioned in Section 1.2 is a typical task that can be successfully
solved by the techniques of linear programming.

Integer Linear Programming

Further requirements for the values of variables are added to the original linear model (i.e. the
model with the linear function and linear constraints). All or some of the variable values have
to be an integer. The special type of those variables is binary variable (0-1 variable), value of
which can be 0 or 1. In such case we speak about Binary Integer Linear Programming.
If only some variables in the model are defined as integer (binary), we speak about Mixed
Integer Linear Programming.

Goal Programming
When several competing objectives have to be considered simultaneously, more powerful tool
is needed. Goal programming is a special technique for dealing with such cases, usually
within the framework of linear programming.

Distribution Models
A distribution problem is a special type of linear programming problem. There are two main
types of distribution problems: the Transportation Problem and the Assignment Problem.
The transportation problem deals with shipments from a number of sources to a number
of destinations, whereas the assignment problem deals with finding the best one-to-one match
for each of a given number of possible “candidates” to a number of proposed “positions”.

Nonlinear Programming
Models used in this area of management science are similar to the models of linear
programming; however there is an important difference between them: nonlinear
models contain nonlinear objective function and/or some nonlinear constraints. Methods
used for solving tasks from this area of management science are, of course, rather
different from the linear programming methods.

Network Models
Some problems can be described graphically as a network (the set of nodes and arcs).
Typical situation is a transportation network: cities (nodes) are connected to each other
by roads (arcs). If we evaluate the network (in this case we are interested e.g. in
distances between all the cities), the task is often to find the minimal distance from one
city to all other cities. Some types of networks may be evaluated by capacities instead
of distances and then the problem of maximal flow can be solved. In many problems,
solved with use of the network models, the most important value is unit cost and the
goal is to find the minimal total cost.

Project Management
In many situations the managers are responsible for planning, scheduling and
controlling projects that consist of many separate jobs or tasks performed by a variety
of departments or individuals. An execution of each job takes specific time. There are
two basic methods for solving those problems: CPM (Critical Path Method) and PERT
(Program Evaluation Review Technique). Both methods require the network
representation of the problem.

Inventory Models
Inventory control is one of the most popular techniques, which helps managers to
determine when and how much to order. The main goal is usually to find a proper
balance between the inventory holding cost and the cost of executing an order. There
are many various inventory models because of many various real inventory systems.
We consider two separated classes of models: deterministic and probabilistic. In
deterministic models the rate of demand is constant through the time, whereas in
probabilistic inventory models the demand fluctuates through the time and can be
expressed only in probabilistic terms.

Waiting Line Models

This area of management science techniques deals with the situations where
units (e.g. customers) need to be served by a number of channels (e.g. vendors). As
the number of vendors is limited, some of the units have to wait for the service in a
queue. Hence the alternative name for the waiting line models: Queuing Models. In
real situations both the process of units’ arrivals and the service times are random and
the probabilistic approach is necessary. Simple waiting line models can be solved
analytically (exact solution using derivative formulas), whereas for complex queuing
systems the technique of simulation is required. The main managerial goal in waiting
line models is the decision about the number of service lines (finding a proper balance
between size of the queue and the total service cost).

Simulation
When managerial problems become more complex, they are often
impossible (or non-effective because of spent time and cost) to be solved using standard
techniques. For this purpose, simulation approach is advantageous and in many cases
it is the only way how to manage the problem. Simulation is a computer
experimentation with a simulation model aimed at describing and evaluating the real
system’s behavior - the computer program simulates the real system. The typical
situations for successful use of simulation are complex waiting line models and
inventory models.

Decision Analysis
These techniques can be used to select optimal strategies out of several decision
alternatives. Managerial problems and appropriate tools are divided, according to the
kind of manager’s information, into three classes: decisions under certainty
(deterministic), decision under risk (probabilistic) and decisions under uncertainty. We
consider special tools for this purpose: decision tables and trees.

Theory of Games
This area is an extension of decision analysis to the situations with two or more decision
makers. Simultaneous decisions (selected strategies) of all managers initiate an action
that affects all decision makers (players), i.e. their profit, cost, etc. In some
conflicts, there is a possibility for two or more decision makers to cooperate, while
competing with the others. In economic theory we can find a typical case of strategic
game - oligopoly model.
Forecasting
Forecasting methods support the manager’s prediction of future aspects of a business
operation. Statistics and econometrics offer many techniques based on time-series and
regression analysis. The main managerial goal is to project future trends following the
previous behavior of the system. The well known are the methods of Moving Averages, Least
Squares, Exponential Smoothing, etc. Since the statistical significance is very important for
these models, the manager’s experience with hypotheses validation and statistical tests is
necessary.

Multicriteria Decision Making

In many managerial problems the decision maker needs to consider multiple criteria. If we
find a solution improving one criterion, it mostly worsens some of other criteria. It is usually
impossible to optimize simultaneously all the criteria. The reasonable opportunity for the
management is to find a suitable compromise. If the count of alternatives is limited, we use
methods of alternatives evaluation. Some problems are described with the set of constraints
and the set of objective functions. In this case the solution is provided by multiobjective
programming techniques. The special category of this management science technique is goal
programming.

Markov Analysis
This technique can be used to describe the behavior of a system in a dynamic situation
(evolution of the system throughout the time). If - at a given time point - the system is in one
of possible states, at following time point the system can remain in current state or can move
into any other state. Remaining in current state or movement to another state are set by
transition probabilities. The manager can be interested in the probability with which the
system will be in the specific state at the specific time. Markov analysis is a very powerful
tool of management science with many real applications.

Dynamic Programming
Management must frequently consider a sequence of decisions where each decision
significantly affects future decisions. Dynamic programming helps managers to solve certain
types of such sequential decision problems. There is no single model for solving dynamic
programming problems and the problems are therefore classified into many groups. One
possible classification considers deterministic and probabilistic models. Models often use
network representation of the sequential problems. Markov analysis can be considered as
a probabilistic model of dynamic programming.

The survey presented by Anderson, Sweeney, and Williams [1] indicates that the most
frequently used techniques are statistical methods, computer simulation, PERT/CPM, linear
programming and queuing theory. Frequency of using all mentioned techniques, of course,
depends on the specialization of the firm and its size. With respect to the limited space in this
textbook and its purpose, we cannot go through all the management science techniques in detail
and we will therefore point out the key methods and applications.
 1.4 Glossary
Binary Integer Linear Programming
Special case of linear programming in which all the decision variables are binary.

Constraints
Restrictions on the problem solution arising from limited resources, policy requirements, etc.

Decision Making
A process of choosing between two or more available alternative courses of action for
the purpose of attaining a specific goal.

Deterministic Model
A model in which the functional relationships and parameters are known with certainty.

Feasible Solution
A solution that respects all the constraints.

Goal Programming
A problem approach when several objective functions are considered simultaneously.
The objective is to minimize the undesirable deviations from the goals.

Integer Programming
A programming approach that assumes the indivisibility of all the decision
variables (e.g. products, people, etc.). The values of the variables must be integers.

Linear Programming
A mathematical procedure for optimizing the linear objective function, respecting
the set of linear constraints.

Management Science
see Operations Research.

Mathematical Model
A system of symbols and expressions aimed at representing a real situation.

Maximization
Optimization of objective function, which looks for the highest objective value (e.g.
profit, revenue).

Minimization
Optimization of objective function, which looks for the lowest objective value (e.g. cost,
loss).

Mixed Integer Linear Programming

Special case of integer linear programming in which some but not all of the decision variables
are integer.

Model
An abstraction of reality.

Nonlinear Programming
A problem approach used when the objective function and/or one or more constraints are
nonlinear.

Objective Function
A mathematical function expressed in terms of decision variables, which is to be optimized
(maximized or minimized).

Operations Research
The application of the scientific method to the study and analysis of problems involving large
and complex systems, organizations or activities.

Optimal (Optimum) Solution

A feasible solution that maximizes or minimizes the objective function. The best of all
feasible solutions.

Probabilistic Model
A model that incorporates uncertainty in its functional relations and uncontrollable variables.

Sensitivity Analysis
Measuring the effect of a change in one input parameter on a proposed solution.