Professional Documents
Culture Documents
Introduction:
Decision theory is primarily concerned with helping people and organizations in
making decisions. It provides a meaningful conceptual frame work for important
decision making. The decision making refers to the selection of an act from amongst
various alternatives, the one which is judged to be the best under given
circumstances.
The management has to consider phases like planning, organization, direction,
command and control. While performing so many activities, the management has to
face many situations from which the best choice is to be taken. This choice making
is technically termed as “decision making” or decision taking. A decision is simply
a selection from two or more courses of action. Decision making may be defined as
- “ a process of best selection from a set of alternative courses of action, that course
of action which is supposed to meet objectives up to satisfaction of the decision
maker.”
Meaning
Decision theory is a general approach to decision making when the outcomes
associated with alternatives are often in doubt. It helps operations managers with
decisions on process, capacity, location, and inventory because such decisions are
about an uncertain future.
Acts (or courses of action): Decision making problems deals with the selection of
a single act from a set of alternative acts. If two or more alternative courses of action
occur in a problem, then decision making is necessary to select only one course of
action. Let the acts or action be a1, a2, a3,… then the totality of all these actions is
known as action space denoted by A. For three actions a1, a2 a3; A = action space =
(a1, a2, a3) or A = (A1, A2, A3). Acts may be also represented in the following matrix
form.
Events (or States of nature): The events identify the occurrences, which are
outside of the decision maker’ s control and which determine the level of success for
a given act. These events are often called ‘ States of nature’ or outcomes. An
example of an event or states of nature is the level of market demand for a particular
item during a stipulated time period.
A set of states of nature may be represented in any one of the following ways:
In a tree diagram the places are next to acts. We may also get another act on the
happening of events as follows:
Pay -off may be also in terms of cost saving or time saving. In general, if there are
k alternatives and n states of nature, there will be k × n outcomes or pay-offs. These
k × n payoffs can be very conveniently represented in the form of a k × n pay -off
table.
A technique in which quantitative data are used along with the principles of
mathematics is known as mathematical quantitative techniques. Mathematical
quantitative techniques involve
1. Permutations and Combinations
2. Set Theory:
3. Differentiation:
4. Differential Equation.
Statistical techniques involve: It is a mathematical equation which involves the
differential coefficients of the dependent variables. Statistical Quantitative
Techniques Statistical techniques are those techniques which are used in conducting
the statistical enquiry concerning to certain Phenomenon. They include all the
statistical methods beginning from the collection of data till interpretation of those
collected data
1. Collection of data
2. Measures of Central Tendency.
3. Correlation and Regression Analysis:
4. Index Numbers
5. Time series Analysis:
Types of decision making: Decisions are made based upon the information
data available about the occurrence of events as well as the decision situation. There
are two types of decision making situations: certainty and uncertainty
Minimax criteria
This criterion is the decision to take the course of action which minimizes the
maximum possible pay-off. Since this decision criterion locates the alternative
strategy that has the greatest possible gain. The working method is:
Example
Consider the following pay-off (profit) matrix Action States
Solution:
Example
A business man has three alternatives open to him each of which can be followed by
any of the four possible events. The conditional pay offs for each action - event
combination are given below:
Solution:
Example
Consider the following pay-off matrix
Using minmax principle, determine the best alternative.
Solution:
min( 27, 25, 23, 32) = 23. Since the minimum cost is 23, the best alternative is E3
according to minimax principle.
Problem: Risk
Suppose, a grocer is faced a problem of how many cases of Milk to stock to meet
tomorrow’s demand. All the cases of milk left at the end of the day are worthless.
Each case of Milk is sold tk. 8 and purchased for tk. 5. Hence, each case sold bring
a profit of tk. 3 but if it is not sold at the end of the day, then, it must be loss of tk.
5.
Number of cases of Number of times of Probability of each
Milk Demanded Milk Demanded event
0-12 0 0.00
13 5 0.05
14 10 0.10
15 20 0.20
16 30 0.30
17 25 0.25
18 10 0.10
Over 18 0 0.00
Grocer has the perfect knowledge
Requirement:
a. What should be the optimal decision concerning the number of cases of milk
to stock?
b. How much be expected opportunity loss will for taking optimum decision?
What item should be stock?
c. Make a decision on what item should be stored using marginal profit analysis.
d. What will be the marginal probability of milk cases? Make decision using
marginal analysis.
Solve:
Req. a.
Sale value per case = 8
Marginal loss =5
Marginal Profit =3
Possible action concerning stock =Earning -loss
Calculation of EMV
Stock 13 = 39×0.05+39×0.10+39×0.20+39×0.30+39×0.25+39×0.10 = 39
Stock 14 = 34×0.05+42×0.10+42×0.20+42×0.30+42×0.25+42×0.10
= 41.6
Stock 15 =43.4
Stock 16 = 43.6
Stock 17 = 41.4
Stock 18 = 37.2
Calculation of EOL
Stock 13 = 0×0.05+3×0.10+6×0.20+9×0.30+12×0.25+15×0.10
= 8.70
Stock 14 = 5×0.05+0×0.10+3×0.20+6×0.30+9×0.25+12×0.10
= 6.10
Stock 15 =4.30
Stock 16 = 4.10
Stock 17 = 6.30
Stock 18 = 10.50
Decision: The lowest EOL is 4.10 corresponding to action of stock 16 cases of milk.
The optimal course of action under the given condition of risk is to stock 16 cases
of Milk.
Req. c
Stock Profit Conditional value Probability Expected
Value
13 3 39 0.05 1.95
14 3 42 0.1 4.2
15 3 45 0.2 9
16 3 48 0.3 14.4
17 3 51 0.25 12.75
18 3 54 0.1 5.4
EPPI 47.7
From req. a, we get max EMV = 43.60
EVPI =EPPI - max (EMV)
= 47.70-43.60
= 4.10
Hence, the expected value of perfect information ECPI = tk. 4.10 being the value of
transformation risk into certainty.
Req. d
𝑀𝑎𝑟𝑔𝑖𝑛𝑎𝑙 𝐿𝑜𝑠𝑠
P≥
Marginal Profit +Marginal Loss
5
P≥ = 0.625
3 +5
The flowchart structure includes internal nodes that represent tests or attributes at
each stage. Every branch stands for an outcome for the attributes, while the path
from the leaf to the root represents rules for classification.
Decision trees are one of the best forms of learning algorithms based on various
learning methods. They boost predictive models with accuracy, ease in
interpretation, and stability. The tools are also effective in fitting non-linear
relationships since they can solve data-fitting challenges, such as regression and
classifications.
Types of Decisions
There are two main types of decision trees that are based on the target variable, i.e.,
categorical variable decision trees and continuous variable decision trees.
A categorical variable decision tree includes categorical target variables that are
divided into categories. For example, the categories can be yes or no. The categories
mean that every stage of the decision process falls into one category, and there are
no in-betweens.
ple, the income of an individual whose income is unknown can be predicted based
on available information such as their occupation, age, and other continuous
variables.
Planning a day
Let’s start with a basic decision tree about planning the events of the day. As you
can see, if a friend would visit, then we can just visit a restaurant. If not, then it would
depend on the weather. If it is rainy, then it is better to stay indoors while on a sunny
day, we can go out and play. In case if, it is windy, and then we can, go either
shopping or movies, depending on the available money.
Then, we can write the expected gross revenues next to each outcome. For example,
if the market impact of creating a new advertising campaign is high, then the
expected revenues for the company will reach $500,000. If the outcome is
moderated, the expected revenues will be $300,000.
After adding the probability value of each outcome and its expected revenues, the
total outcome value is calculated by multiplying the probability value times the profit
value. For example, the value for the new marketing idea option would be as follows:
5% X 500,000 = $25,000
3% X 300,000 = $9,000
2% X 150,000 = $3,000.
THE END