Study Reminders
Text Version

Introduction to Decision Support Systems

Set your study reminders

We will email you at these times to remind you to study.
  • Monday




























What is the decision support system? Right. A decision support system is basically a
man computer interactive system which aids in making decision. This helps the
managers to make decisions. So, there is an interaction between the manager and the
computer. In decision support system what happens is that the computer’s ability to
process databases and models is combined with the manager’s ability to draw on his
experience and judgment.
And, thereby also it capitalizes manager’s intuition in taking a decision ok. So, once
again decision support systems are basically man computer interactive system. It aids in
management in making decision. It combines computers ability to process databases and
models with the manager’s ability to draw on his experience and intuition and thereby
exercise his judgment.

(Refer Slide Time: 02:18)

So, the scope of this particular module is basically we will discuss about the role played
by decision support systems in solving managerial decision problems. Then we will be
discussing about the basic components of a decision support system. And, thereafter we
will deal with the basic steps that are involved in the design of a decision support system.
(Refer Slide Time: 02:50)

Now, when we discuss about decision support systems, the first thing that we need to
know is what are the different types of managerial decision problems. First of all what is
basically a decision making that we need to know? In any decision making environment

we have a problem at hand and then we need to find out or identify different alternatives
to solve that particular problem.
Then we have to analyze those alternatives and select one of the out of the several
alternatives to solve the problem. And, that alternative is supposed to be the best
alternative as per the manager who is taking the decision. There are various criteria based
on which those alternatives are adopted or selected, also it also depends on the mental
makeup of the managers or the decision maker.
Now, we will discuss in detail at a later point in time: what are the basis for selection of
those alternatives for solving a problem? Now, coming back to the different types of
managerial decision problems, we need to know about three different types. One is
structured decision problem, second one is unstructured decision problem and the last
one is semi structured problem.
The category in which a decision problem falls depends on how well the objectives of
the problem can be defined and whether there are well established solution procedures
for solving that problem ok.
(Refer Slide Time: 05:13)

Now, problems where we have a well defined objective and well established solution
procedures they belong to the class of problem which are basically structured decision
problems. Typically, you see in any inventory replenishment situation the type of
problems that we encounter are basically the structured ones.
Problems where the objective is vague and there are no well established solution
procedures for solving those problems they are basically unstructured problems. For
example, say when you are trying to find out what are the new products to be
manufactured then that belongs to an unstructured problem and semi-structured problems
combine the features of both this structured as well as unstructured. So, they lie in
(Refer Slide Time: 06:24)

The structured problems can be completely delegated to the computer, because they are
the objective is well known, well defined, there are established solution procedures for
solving those problems. And, hence it can be completely delegated to the computer
whereas, the unstructured problems have to be solved by the manager alone; where it
requires his intuition, his judgment, his experience to solve those kind of problems.
And, semi-structured problems basically lie between the structured and unstructured
problems. It can neither be wholly solved by a computer nor can be solved by the
manager alone. Hence, in solving a semi structured problem it requires a man computer
interaction, it requires a man computer team for solving those problems. It is therefore, in
handling the semi-structured problems that the decision support systems can be most

(Refer Slide Time: 07:49)

Now, we need to know: what are the different phases of a decision making process? The
first thing is the intelligence phase; in the intelligence phase, that decision-maker scans
the environment, processes the raw data and identifies the problem areas. Basically, he
will be scanning this entire environment; will try to see: what is the nature of data; what
kind of data is available in the environment?
And, then he would try to understand: what is the problem area, and identify their
problem. In the design phase what happens is that the problem that has been identified in
the intelligence phase for that problem they try to find out various alternatives. So,
various alternatives are formulated for solving the identified problem and each of these
alternatives are analysed in this particular phase.
And, then we come to the choice phase, in the choice phase a selection is made from the
proposed alternatives and the chosen alternative is implemented. So, that is what I said
early that what decision making process is. In a decision making process, we are trying
to select the best alternative out of the set of alternatives which have which have been
formulated to solve the given problem in hand.
And, this selection of this best alternative is also dependent on the mental makeup of the
manager, what he wants to achieve and there are other criteria which I said that we will
discuss at a later point in time.

(Refer Slide Time: 10:09)

Now, already we said that decision support systems are mostly suited for semi-structured
type of problem where there is an interaction between the manager and the computer. In
the process the problem solving is iterative and interactive in nature; that means, once the
problem output is interpreted by the manager, he might revise the input.
And, in the process again he gets an output, he interprets that output; if he is not happy or
not satisfied then again he can modify the data and then this entire process is repeated
several time. And hence we said that the problem solving process is not only interactive,
but iterative in nature. The computer provides some output to the manager, the manager
then interprets the output, gives new input, modifies the existing data.
And, then he might asks new questions and this process is repeated. The computers
major tools in solving the problems are models, both simple models as well as
sophisticated models. Whereas, the manager mostly uses his experience and judgment in
interpreting the model output and then trying to modify the basic input to solve the

(Refer Slide Time: 12:04)

The models that are used in decision support systems by the computer basically consists
of finding the relationships between the major elements governing the system which is
being studied. And, in developing that relationship it might require certain degree of
abstraction. The output of the models will basically be an approximation of the
behaviour of the real system.
And the models will be valid only if certain assumptions about the real system are
satisfied; that means, each and every models that we are using in decision support
systems will basically depend on various assumptions that we make. And, if those
assumptions are valid, if those assumptions are satisfactory then the models will also be
giving valid output.

(Refer Slide Time: 13:20)

Now, in this respect another important thing that needs to be discussed is that, what are
the differences between a management information system and a decision support
system. Because, management information systems also provides the managers with the
right information at the right time to take certain decisions. But then what is this salient
or the basic difference between MIS and DSS?
DSS designer views effectiveness as more important than the efficiency. In MIS, the
managers are more bothered about efficiency whereas, in decision support systems
managers are more keen to find out an effective solution.
Now, what is this difference between effectiveness and efficiency? When we are talking
about effectiveness, we are basically trying to find out that whether we are doing the
right thing. And, when we are talking about efficiency, we are mostly emphasizing
whether we are doing it in the right manner.
So, effectiveness basically questions that whether the given problem that we are trying to
solve is the right one or not and given a problem whether we are solving it in the right
manner or not is what basically efficiency means. So, effectiveness involves identifying
what should be done, am I doing the right thing and ensuring that the chosen criteria is a
relevant one. On the other hand, efficiency is basically I am doing it in the right manner
or not.

So, efficiency involves some kind of minimization of cost says time or effort in
performing a specified activity. In indecision support systems we are more bothered about
effectiveness rather than efficiency and one thing we have to remember then that in DSS
the computer is used not to replace the decision maker, but to enhance his or her decision
making abilities ok. It is not replacing the decision maker.
(Refer Slide Time: 16:39)

In DSS, the manager plays a very active role because; we already said that is a man
computer interactive system. And, the manager uses his or her experience and judgment
wherever necessary and controls as well as directs the computer in its tasks. The manager
might input a given set of data at a point in time, interpret the output then a may again
revise the input and he may give different set of commands ok.

(Refer Slide Time: 17:23)

So, here the manager’s experience and judgment plays a very important role, while
talking about the range of capabilities of a decision support system we also need to
remember that not all decision support systems include models in order to be effective. A
decision support system can be a simple data retrieval facility and yet it can be useful in
aiding decision making.
(Refer Slide Time: 17:52)

The range of facilities offered by decision support systems can be therefore, listed as
retrieving information from a database. Decision support systems provide adhoc data
analysis, it has an inbuilt mechanism for doing those kind of ad hoc data analysis. It has
the capability of aggregating data in the form of reports. Decision support systems can
also help in estimating the consequences of proposed decisions and thereby it proposes
the right kind of decisions. So, these are the range of capabilities of a decision support
(Refer Slide Time: 18:52)

Need for decision support system approach is very relevant when we are basically trying
to solve a particular problem. We have to remember that a problem must be semi-
structured ok. Once again I repeat semi-structured problems are those kinds of problems
where the objective is not well defined; neither there are well-established solution
So, in here all the parameters of the problem are not clearly defined and the parameters
are firmed up as one proceeds through the solution process. Here the inputs which are
provided to the model for solving the problem are dynamic in nature; that means, the
input data for the problem are not fixed. These data change from time to time for
example, incorporation of new customers or change in demand from a customer ok. So,
these are highly dynamic data.

(Refer Slide Time: 20:16)

Now, we will give an example of a decision support system for determination of price of
a product ok. So, when you are trying to determine the price of a product is basically a
pricing decision involving several factors. Some of these factors can be quantified and
some of these factors cannot be quantified. What are those kind of factors? Number 1:
competitor’s prices for products similar to the one that we are considering.
Influence of price, quality and advertising expenditure on demand for that product. Risk
of legal suits from competitors in case the price that I set is too low. Another factor
maybe the cost of production of or for that particular product because the price that, I set
for that product must be more than the cost of production.

(Refer Slide Time: 21:40)

So, if these are all these factors suppose there may be a several other factors, let us limit
our discussion with respect to these factors only. So, some of these factors determine to a
certain extent the lower and the upper bound for the price. But, having decided that this
is the lower bound of the price and this is the upper bound of the price, there is a range
over which the price can be varied.
One of the objectives of the decision support system is to predict the outcomes when the
price, quality or advertising expenditures are varied over some ranges. So, there is a
flexibility in that and we have to find out that to what extent the price can be changed
when other factors are varied over this ranges.

(Refer Slide Time: 22:47)

So, that system that we are talking about basically involves division of tasks between
manager and the computer and this is very clearly delineated. The computer’s main job is
to determine the values of parameters in equations that define the relation between a
demand price index and price when quality and advertising expenditure are held
Again the parameters need to be determined in the equation that, define the relation
between demand for the product or demand quality index and quality when price and
advertising expenditure are held constant.

(Refer Slide Time: 23:48)

The relation 3 can be relation between demand advertising index and advertising
expenditure when quality and price are held constant. And the 4th relation can be the
relation between demand and all the three variables price, advertising expenditure and
Now, the computer cannot determine the parameters of these relations unless it has at
least a few inputs; that means, for each value of the dependent variable, the independent
variables value have to be also supplied. And, these given set of data need to be input
first by the manager in order that the computer can determine the values of these

(Refer Slide Time: 24:43)

So, it is the manager’s task to supply these inputs; that means, for a given set of input
variables this is the value of the dependent variable; like this a complete set has to be
given. And after receiving these inputs from the manager, the parameters of the
previously mentioned relations; all those four relations that I had mentioned will be
determined by the computer. And, then this relationship will be displayed in graphical
form to the manager. The manager can at this stage incorporate any desired
(Refer Slide Time: 25:22)

In each of these first three relations, the effect of one of the variables price, quality and
advertising on an index is considered, while keeping the other two variables fixed. And,
in relation 4 the effects of varying all the three variables simultaneously on the demand is
being studied. So, the parameters of the relation are initially determined by the computer
using the earlier relations 1, 2 and 3.
(Refer Slide Time: 26:00)

Then, the computer displays the effect of varying two of the three independent variables
and then see: what is the effect on the dependent variable? The manager has at this stage
an opportunity to tell the computer whether the above estimate is too high, accurate or
too low according to his judgment.
And depending on the feedback from the manager, the computer either freezes the values
of parameters in the 4th relation or modifies them and invites feedback from the manager
again. In as in our next module we will also discuss how the different parameters are

(Refer Slide Time: 26:44)

The process of adjusting the parameters of relation 4 is basically called balancing. Once
the balancing phase is over, the manager can fix values for two of the independent
variables and specify a range for the third variable. The computer computes and displays
the demand and profit corresponding to different values of the third variable over its
specified range. The manager can stop here or reinitiate the process of modifying any of
these relationships.
(Refer Slide Time: 27:17)

In this system, the following points are worth noting: number 1, the various tasks in the
decision making process are distributed between the manager and the computer and
neither of the two is replacing the other. And, the second one the manager uses his
experience in predicting the demand volumes for different values of one of the three
variables price, quality and advertising.
(Refer Slide Time: 27:46)

The computer can then fit the above into three equations and predict by means of that
fourth equation the combined effect of all the three variables on demand and then plot
the profits over a given range of any of the three variables.
The objective of using the computer is to improve the effectiveness of decision making.
The focus of this entire system or the whole system is a decision problem, namely price
determination which is semi-structured in nature.

(Refer Slide Time: 28:23)

There are several other problems for which decision support systems have been reported
to have been successfully employed, some of these problems are decisions on merger of
companies. Basically semi-structured in nature. Budget planning, portfolio management
in banks, corporate planning, also capacity planning in production; these are examples of
highly popular decision support systems.
(Refer Slide Time: 28:53)

So, relationship between the various parameters when they are not clearly defined then
we need a DSS approach for solving the parameter problem. Mostly DSS is suited for
situations where the objective is not clearly defined. Problem is one with multiple
objectives and the trade offs are not clearly known; through a decision support system
the decision maker can know the effect of changing the various parameters and their
effect on the different objectives and decide the course of action to be followed by him.