According to Sharma (2016), “the success or failure that an individual or
organization experiences, depends to a large extent, on the ability to make
acceptable decisions on time.” You are what you are today because of the
decisions you made in the past. This is not about blaming anyone but more
about learning from the experience.
So what differentiates a good from a bad decision? According to
Render, et. al. (2015), “A good decision is one that is based on logic,
considers all available data and possible alternatives, and applies the
quantitative analysis.” Otherwise, it is a bad decision.
Decision theory is the systematic manner of arriving with decisions. It
follows certain procedures that consider the degree of knowledge and
compare it to expected consequences from varied strategies. It considers
both quantitative and qualitative perspectives as shown in Figure 6.
Figure 6. Decision Analysis
To
have a better understanding of decision theory, let us define the
usual terms we will encounter in this module:
An alternative pertains to a course of action or choice or strategy.
The states of nature refer to the circumstances where the decision maker
has little or no power to control.
The conditional values pertain to the payoffs (profits or costs) following
from each possible combination of alternatives and states of nature. These
are numerical in nature.
The decision table also known as payoff table as illustrated in table 2
depicts the alternatives and states of nature. While the conditional values
are represented by the outcomes.
Table 2. Payoff Table
Steps in Decision Making
It doesn’t matter how big or small your decision is. To arrive at a good decision,
you will undergo the same steps as follows:
For this illustrative example of the decision theory steps, we will use the LMS
Food Café Case.
Step 1. Ms. Santos is considering the re-designing of the operations of the café.
Step 2. Ms. Santos identified three alternatives namely, (1) open 24-hours (2) open
usual 8-hours, or (3) do not open.
Ms. Santos has to make sure she identifies all important courses of action. Failure to
identify some may be crucial. While some strategies may seem of little value, to your
surprise it could be your best choice.
Step 3. Ms. Santos determines that there are only two possible outcomes: the market of
the café could be favorable, meaning that there is a high demand for the product, or it
could be unfavorable, meaning that there is a low demand of the market.
Just like the alternatives, it is important that Ms. Santos consider all possible
circumstances. One has to learn to carry all the hats. Be optimistic and pessimistic as
well.
Step 4. Because Ms. Santos wants to maximize her profits, she can use the profit to
evaluate each outcome. She has already evaluated the potential profits associated with
the various outcomes.
With a favorable market, she thinks opening for 24-hours would result in a net profit
of P60,000 a month. This P60,000 is a conditional value. The conditional value if the
market is unfavorable would be a P20,000 net loss. Opening the usual 8-hours
would result in a net profit of P30,000 in a favorable market, but a net loss of P10,000
would occur if the market was unfavorable. Finally, not opening would result in P0 profit
in either market.
Having all the available information, the best way to show an understanding of the
problem is to create the payoff table. Table 3 shows the payoff table for LMS Food
Café.
Table 3. Payoff Table of LMS Food Café
Steps 5 and 6. The last two steps are to select and apply a decision theory model to
the data to come up with a decision. Choosing the model depends on the knowledge of
the environment and the amount of risk and uncertainty.
Types of Decision Making Environments
As mentioned in the previous lesson, model selection depends on the environment
and the amount of risk and uncertainty involved. According to Sharma (2016), there are
three types of decision-making environments as follows:
Type 1: Decision making under certainty
Decision maker has the complete knowledge of the consequences of every alternative
or decision choice. An example is saving your money in a bank with a known interest.
Type 2: Decision making under uncertainty
The decision maker is unable to specify the probability for occurrence of particular state
of nature. An example is probability of a person to become the President of a country.
Type 3: Decision making under risk
The decision maker does not have perfect knowledge about possible outcome of every
decision alternative. This could be due to having more than one states of nature. In a
such a case the decision maker makes an assumption of the probability for occurrence
of particular state of nature.
Decision Making Under Uncertainty
There are several models available for making decisions under conditions of uncertainty
as follows:
1. Optimistic (Maximax or Minimin);
2. Pessimistic (Minimax or Maximin);
3. Criterion of realism (Hurwicz);
4. Equally likely (Laplace); and
5. Minimax regret (Salvage).
Optimistic (Maximax or Minimin)
The Optimistic criterion ensures that the decision-maker selects the alternative
that represents the largest possible profit (maximax) for maximization problems or the
lowest possible cost (minimin) for minimization problems. The analytical method is
summarized as follows:
(a) Determine the maximum (or minimum) conditional values corresponding to each
decision alternative.
(b) Choose a decision alternative with best conditional value (maximum for profit and
minimum for cost).
Following the case of LMS Food Cafe that wants to maximize profit, table 4 shows
the payoff table for a positive manager:
Table 4. Payoff Table of LMS Food Café - Optimistic
Therefore, the decision will be to operate for 24 hours.
Pessimistic (Minimax or Maximin)
The Pessimistic criterion ensures that the decision-maker selects the alternative
that represents that he would earn no less (or pay no more) than some specified
amount. The analytical method is summarized as follows:
(a) Determine the minimum (or maximum in case of profit) conditional value in case of
loss (or cost) data corresponding to each decision alternative.
(b) Choose a decision alternative with the best conditional value (maximum for profit
and minimum for loss or cost).
Following the case of LMS Food Cafe, payoff table 5 will be for a negative thinker
manager:
Table 5. Payoff Table of LMS Food Café - Pessimistic
Therefore, the decision will be to not operate.
Criterion of realism (Hurwicz)
The Criterion of realism (Hurwicz) ensures that the decision-maker selects the
alternative with a combination of optimism and pessimism. A coefficient of realism,
designated by the symbol α , is chosen at the beginning depending on the decision
maker’s state of positivity and negativity. The lowest α that can be chosen is 0, which is
an indication that the decision maker is 100% pessimistic about the future. While the
highest α is 1, indicating that the decision maker is 100% optimistic about the future.
According to Render, et.al., (2015), for a maximization problem, the best payoff
for an alternative is the highest value in the row, and the worst payoff is the lowest value
in the row. This value is computed for each alternative, and the alternative with the
highest weighted average is then chosen. In contrast, for minimization problems, the
best payoff for an alternative would be the lowest payoff and the worst would be the
highest payoff. The alternative with the lowest weighted average is then selected.
The analytical method is summarized as follows:
(a) Determine the coefficient of optimism (α) and then coefficient of pessimism (1 – α).
(b) Calculate the weighted average with the formula:
Hurwicz = α (Maximum in column) + (1 – α ) (Minimum
in column)
(Eq.2)
(c) Choose an alternative with highest weighted average conditional value.
For the case of LMS Food Cafe, Ms. Santos is a more positive manager. So we
will be using α closer to 1. Using a = 0.7, please find below the computation :
For the 24-hours alternative: (0.7)(60,000) + (1 – 0.7)(–20,000) = 36,000
For the 8- hours alternative: (0.7)(30,000) + (1 – 0.7)(–10,000) = 18,000
Table 6 presents the payoff table for a manager who compromise between
optimism and pessimism.
Table 6. Payoff Table of LMS Food Café – Hurwicz
Therefore, the decision will be to operate for 24 hours.
Equally likely (Laplace)
The criterion of equally likely considers equal probabilities of occurrence for the
states of nature. This is following the idea that probabilities of states of nature are
unknown. Equal probability value to each state of nature using the formula: 1 ÷
(number of states of nature).
The analytical method is summarized as follows:
(a) Determine the average payoff for each alternative
For the 24-hours alternative: [(60,000) + (–20,000)]/2 = 20,000
For the 8- hours alternative: [(30,000) + (–10,000)]/2 = 10,000
(b) Choose the alternative with the highest average
Result of this analysis is depicted in Table 7.
Table 7. Payoff Table of LMS Food Café – Laplace
Therefore, the decision will be to operate for 24 hours.
Minimax regret (Salvage)
The salvage criterion is based on opportunity loss (regret). Such loss is
computed by subtracting the actual payoff received from the best payoff. Basically, this
is finding the alternative that minimizes the maximum opportunity loss (regret).
The analytical method is summarized as follows:
(a) From the given payoff matrix, construct an opportunity-loss table as follows:
(i) Find the best payoff corresponding to each state of nature.
(ii) Subtract all other payoff values in that row from this best payoff.
(b) For each decision alternative identify the worst (or maximum regret) payoff value.
Record this value in the new row.
(c) Select a decision alternative resulting in a smallest anticipated opportunity-loss
value.
Let us first recall the pay-off presented in table 3 in the previous pages:
Source: Table 3. Payoff Table of LMS Food Café
The best payoff for is P60,000 and P0 for favorable market and unfavorable market
respectively. Hence, to construct an opportunity-loss table, we will do the following
computations as presented in Table 8.
Table 8. Opportunity Loss Computation Table of LMS Food Café
Resulting from Table 8, we can now create our payoff table for the minimax
regret criterion. This is depicted in Table 9.
Table 9. Payoff Table of LMS Food Café – Minimax Regret
After getting the maximum in each row, we will choose the minimum in the
column. Therefore, the decision will be to operate for 24 hours.
Decision Making Under Risk
Under this decision-making environment, the decision maker does not have
perfect knowledge about possible outcome of every decision alternative. This could be
due to having more than one states of nature. In a such a case the decision maker
makes an assumption of the probability for occurrence of particular state of nature.
According to Sharma (2016), "knowing the probability distribution of outcomes (states of
nature), the decision-maker needs to select a course of action resulting a largest
expected (average) payoff value. The expected payoff is the sum of all possible
weighted payoffs resulting from choosing a decision alternative."
The commonly used criterion for evaluating alternatives under risk is the Expected
Monetary Value (EMV).
Expected Monetary Value (EMV)
The expected monetary value (EMV) for given alternatives is computed by
getting the summation of the payoff values multiplied by the probabilities associated
with each state of nature. For the analytical method to solve for EMV, we can use the
formula below :
EMV (alternative i) = (payoff of first state of nature) x (probability of first state of
nature) + (payoff of second state of nature) x (probability of second state of nature) +
… + (payoff of last state of nature) x (probability of last state of nature)
(Eq.3)
This can also be written as:
(Eq.4)
Following the case of LMS Food Cafe, Ms. Santos can redesign their operations
by opening for 24 hours which will earn P60,000 or may lose P20,000. If she decides
to operate for 8 hours, she will earn P30,000 or may lose P10,000. There is a 60%
chance that the market is favorable. Below shows the computation for the EMV:
For the 24-hours alternative: [(60,000*.60) + (–20,000*.40)] = 28,000
For the 8- hours alternative: [(30,000*.60) + (–10,000*.40)] = 14,000
The result is presented in table 10.
Table 10. Payoff Table of LMS Food Café – EMV
Therefore, the decision will be to operate for 24 hours.
Decision Trees
Decision trees can be used to present decision tables graphically. This is
preferred to be used when the decision maker intends to make a sequence of
interrelated decisions. To have a better understanding of decision trees, let us provide
the following guides on this topic:
(a) A tree starts from left to right that contain decision (act) nodes and state-of-nature
(chance) nodes. Refer to figure 7.
Figure 7. Decision Tree
Please take note that a certain decision may lead to a state of nature or another
decision. A state of nature may lead to making another decision or another state of
nature.
(b) Squares represent decision nodes from which one of several alternatives may be
chosen
(c) Circles represent states of nature nodes out of which one state of nature will occur
(d) Lines or branches connect the decisions nodes and the states of nature. It is
called decision branch if it branches away from a decision node and represents an
alternative that can be chosen at a decision point. It is known as chance branch if it
branches away from a states of nature node and represents a set of chance events.
(e) Conditional Values or Payoff can be positive or negative. It can be computed
following either decision or state of nature branches.
Render, et. al (2015) identified five steps to analyzing problems with decision trees:
Let’s try creating a decision tree for LMS Food Café.
Analyzing LMS Food Café using a Decision Tree
1. Define the problem.
The problem identified is the same ---- that is the presence of COVID -19.
2. Structure or draw the decision tree.
Figure 8. Decision Tree Structure of LMS Food Cafe
3. Assign probabilities to the states of nature
Let us recall our EMV computation where there is a 60% chance that the market is
favorable.
Figure 9. Decision Tree Probabilities of LMS Food Cafe
4. Estimate payoffs for each possible combination of alternatives and states of nature.
Figure 10. Decision Tree Payoffs of LMS Food Cafe
5. Solve the problem by computing expected monetary values (EMVs) for each state of
nature node.
Figure 11. Decision Tree EMVs of LMS Food Cafe
Comparing the calculated EMVs of P28,000, P14,000 and P0, the decision will be
to open for 24 hours.
