Chapter Four

Decision Theory

4.1 Introduction
Hundreds of decisions are made every day in the operations activity. Each minor decision
determines the company's success or failure. It ranges from simple judgmental to complex
analysis which can also involve judgment (past experience and common sense). They involve a
way of blending objective and subjective data to arrive at a choice. The use of quantitative
methods of analysis adds to the objectivity of such decisions.

Quantitative Approaches to Decision Making

Quantitative approaches to problem solving often embody an attempt to obtain mathematically
optimum solutions to managerial problems. The functions are commonly used quantitative
approaches like, linear programming, Queuing techniques, Inventory models, Forecasting
techniques, Statistical models.

Operation decision become more complex when: it involves many variables, the variable are
highly interdependent or related, and the data describing the variables are incomplete or
uncertain. The necessity of working with incomplete and uncertain data has always been a
problem for decision maker. The following figure depicts the information environment decisions.

Fig. 1.4 Quantitative methods available to operations managers

How much certainty exists?

Completely certain Extremely uncertain

Risk Situation

Objective information Subjective Information

.

(All information) (Some information)

information) (No information)

Page 1 of 7
The following are quantitative tools used under the three situations.
Certainty Risk Uncertainty
- Algebra, Breakeven analysis, - Statistical analysis - Game theory
- Cost benefit analysis, - Quelling theory - Decision theory
- Calculus, mathematical - Simulation
- Programming, linear and - Net Work analysis;
- Nonlinear, integer, dynamic - PERT/CPM
- Programming etc. - Decision tree, Utility theory, etc

Framework for Decisions (Process)

An analytical and scientific framework for decisions implies several systematic steps for decision
makers. They are:
Step 1. Define the problem and its parameters.
Step 2. Establish the decision criteria and set the objective.
Step 3. Formulating a relationship (model) between the parameters and the criteria.
Step4. Generate alternatives by varying the values of the parameters.
Step 5. Choose the course of action, which most closely satisfies the organization.
Step 6. Implement the decision and monitor the result.

In the following section of this unit you will learn how to apply the quantitative tools in solving
operations related problems. Since it is difficult to illustrate the entire models only one model,
which can be used under the three situations of decision-making are discussed.

A. Decision Making under Certainty

As it is shown in the figure 1.4 above different approaches to decision making are available to
decision makers. The most widely used decision rule under the certainty situation is break-even
analysis (cost-profit-volume analysis), cost-benefit analysis and mathematical programming. In
the next section break-even analysis in illustrated.

1. Breakeven Analysis (BEA)

BEA is a widely used decision making tool. It helps to make a decision whether to produce or
not a certain level of output. Breakeven point is a level of output at which, profit is zero or no
loss or no gain. Production or operation level below this point results in a loss, whereas level of

Page 2 of 7
output above this point helps the company to enjoy some level of profit. Therefore, knowing this
point helps the company to take appropriate action.

Example 1: The cost and revenue information of ABC Company are as follows:
Fixed Cost = Br 120,000
Unit Price = Br. 50
Variable Cost/unit (VC/unit) = Br. 30
Find the break-even point in terms of unit and sales in Birr.
Solution:
Solution:
Fixed Cost
Break Even Point-BEP (quantity) =
Price/unit  VC/unit
120,000 120,000
BEP = =
50  30 20
= 6,000 units
BEP (In Birr) = 6,000 X Br. 50 = Br. 300,000.
This can be depicted graphically as follows:
TR

A
TC

Br. 300,000 BEP

Br. 120,000 FC
B

6,000 Quantity
Where: TR = total revenue A= represents the profit region
TC = total cost B= represents the loss region
BEP = breakeven point VC/unit = Variable cost per unit

Interpretation
The company, if it wants to be profitable, should produce and sale more than 6, 000 units of
output. For example, If the external and internal environment force it to produce only 3,000 units

Page 3 of 7
of the product the company will incur a loss. So it has to take some short-term as well as long
term measures to correct the situation.

B. Decision Making Under Risk

Decision Tree
Decision tree is a schematic diagram used to determine expected value. It shows the alternative
outcomes and independence of choice. It is used in risk situation where there is only probabilistic
information stated in probabilistic value.

Example:
Example: ABC manufacturing firms wants to meet the excess demand to its products. The
firm’s management is concerning three alternative courses of action.
A. Arrange for subcontracting
B. Begin overtime production
C. Construct new facilities

The correct choice depends largely on future demand, which may be low, medium or high.
Management ranks the respective probabilities as 10%, 50% and 40% to low, medium and high
product demand in the future respectively. A cost analysis reveals the effect on profit of each
alternative under a given state. This is given in the payoff table below.

Pay off table

Alternatives Profit if Demand is (Birr)
Low (0.1) Medium (0.5) High (0.4)
Arrange for sub contract (A1) 10,000 50,000 50,000
Over time (A2) -20,000 60,000 100,000
New facilities (A3) -150,000 20,000 200,000
Required: Which alternative is the viable choice?
We can use expected monetary value approach and decision tree to answer this question.
1. Expected monetary value (EMV) – Determine the expected payoff of each alternative
and choose the alternative that has the best expected payoff.
EMV (A1) =10,000X0.1+0.5X 50,000+0.4X50, 000
=1000+25,000+20,000
= Br. 46,000

Page 4 of 7
 EMV (A2) = -20,000X0.1+60,000X0.5+100,000X0.4
= -2,000+30,000+40,000
= Br. 68,000
EMV (A3) = 150,000X0.1+20,000X0.5+200,000X0.4
= -15,000+10,000+80,000
= Br.75, 000 (highest expected value)
Decision: The best alternative is to construct new facilities because it has the highest expected
value.

You can also use decision tree two depict this information to make a decision as shown below.

Br. 46,000

Br. 75,000 Br. 68,000

Br. 75,000

Decision: Construct new facility because it have a better return than other two alternatives.

C. Decision making under uncertainty

At the opposite extreme is complete uncertainty, no information is available on how likely the
various states of nature are under these condition, four possible decision criteria are :
A. Maximin-
Maximin- determine the worst possible pay off for each alternative, and then choose the
alternative that has the “ best worst”.
B. Maximax-
Maximax- determine the best possible payoff and choose the alternative with that payoff.

Page 5 of 7
 C. Laplace-
Laplace- determine the average payoff, and choose the alternative with the best average.
D. Minimax regret-
regret- determines the worst regret for each alternative, and choose the
alternative with “best worst”.

Example:
Example: Based on the above pay off table and assuming that there is not probability value of
occurrence of each outcome, determine which alternative would be chosen under each of these
strategies.
A. Maximin
B. Maximax
C. Laplace
D. Minimax regret

Solution:
Solution:
A) Maximin criteria
The worst pay off for the alternatives are Br. 10,000 for subcontracting, Br. -20,000 for overtime
and Br. -150,000 for new facilities and 10,000 is the best out of the worst; hence, the decision is
to choose subcontracting as an alternative using the maximin criteria.
B) Maximax criteria
The best pay off for each alternative, that is, for Sub contracting is 50,000, over time Br. 100,000,
and New facilities is Br. 2000,000. The decision is to construct new facility which is an
alternative with the best payoff value i.e., Br. 200,000.
C) Laplace criteria
The average payoff of each alternative is;
A1 = 10,000+50,000+50,000/3 = 36,667
A2 = -20,000 + 60,000 + 100000 /3 = 46,667
A3 == -150,000 + 20,000 + 200000 / 3 = 23,333
Decision: use overtime to absorb the excess demand.
D) Minimax regret
Regret values are computed by subtracting each entry in a column from the largest value in the
same column. For example, assume that probability values are not given in the above example.
For the first column, the maximum value is 10,000, so the regret values are 10,000 - 10,000 = 0,
10,000 - (-20,000) = 30,000, and 10,000 - (-150,000) = 160,000.
160,000.

Page 6 of 7
Regret table
Profit if demand is (Birr)
Alternatives Low Medium High Maximum
regret
Arrange for sub contract (A1) 0 10,000 150,000 150,000
Overtime (A2) 30,000 0 100,000
100,000

New facilities (A3) 160,000 40,000 0 160,000

Decision: the minimax regret value is Br. 100,000. Therefore, use overtime to absorb the excess
demand.

Page 7 of 7