Discuss in detail Decision Making under uncertainty using valid real world examples.

[30]
Decision analysis is a systematic approach by decision making that allows managers to solve
problems with uncertainty figures as a prominent factor. A normative model is developed to
represent the decision making problem, facilitate logical analysis, and produce a recommended
course of action. The technique is most useful in managerial situations where risk is significant. The
resulting formal model is capable of generating optimal strategies for multi-stage decision making
problems that involve a variety of contingencies.

According to Pažek (2008), within their research paper five decision rules (criteria) commonly used
in decision process under uncertainty were presented being:
1. Wald’s Maximin criterion
2. Hurwicz’s criterion
3. Maximax criterion
4. Savage’s minimax regret criterion
5. Laplace’s insufficient reason criterion.

Thus, the payoff (or decision) matrix M = { A , S , R , P } formally defines a decision analysis
problem.

Where:
A - the set of decision alternatives Ai (for i = 1, 2, ..., m )
S - the set of events S
j (for j = 1, 2, ..., n )
R - the set of payoffs (rewards) Rij obtained by choosing alternative Ai if state Sj occurs
P - the probability distribution applicable to S (the set of probabilities pj describing the
likelihood that state S
j will occur).

For this discusion Wald’s, Hurwicz’s, Maximax, Savage’s and Laplace’s criteria are calculated and
discussed in the sample of pumpkin oil production and selling as illustrated by Pažek (2008). Three
production business alternatives with different production area of oil pumpkins (A1, A2, A3) and
three different market opportunities for pumpkin oil (S1, S2, S3) were calculated and analysed as
given below:
 Basic data aid for business alternative evaluation

Source: Pažek (2008).

The decision matrix for pumpkin oil (based on financial parameters in €)

Source: Pažek (2008).

Wald’s Maximin Criterion

The decision-theoretic view of statistics advanced by Wald had an obvious interpretation in terms of
decision-making under complete ignorance, in which the maximin strategy was shown to be a best
response against natures’ minimax strategy. Wald’s criterion is extremely conservative even in a
context of complete ignorance, though ultra-conservatism may sometimes make good sense (Wen
and Iwamura, 2008). The Maximin criterion is a pessimistic approach. It suggests that the decision
maker examines only the minimum payoffs of alternatives and chooses the alternative whose
outcome is the least bad. This criterion appeals to the cautious decision maker who seeks ensurance
that in the event of an unfavourable outcome, there is at least a known minimum payoff. This
approach may be justified because the minimum payoffs may have a higher probability of occurrence
or the lowest payoff may lead to an extremely unfavourable outcome.

The maximax criterion is an optimistic approach. It suggests that the decision maker examine the
maximum payoffs of alternatives and choose the alternative whose outcome is the best. This criterion
appeals to the adventurous decision maker who is attracted by high payoffs. This approach may also
appeal to a decision maker who likes to gamble and who is in the position to withstand any losses
without substantial inconvenience. See the table below for an illustration of this criterion.
The Wald’s criterion is an approach which the pessimistic farmer will prefer to apply. In the
framework of the observed criteria the decision maker prefers the highest value of bad conditions.
However, according to Wald's criterion, the farmer should select the maximum of the row minima.
In the presented research the alternative 3 (-1196 €) is selected:
 The computation results for Wald’s criterion

Source: Pažek (2008).

Maximax Criterion
The Maximax criterion is an optimistic approach. It suggests that the decision maker examine the
maximum payoffs of alternatives and choose the alternative whose outcome is the best. This criterion
appeals to the adventurous decision maker who is attracted by high payoffs. This approach may also
appeal to a decision maker who likes to gamble and who is in the position to withstand any losses
without substantial inconvenience. It is possible to model the optimist profile with the Maximax
decision rule (when the payoffs are positive-flow rewards, such as profits or revenue. When payoffs
are given as negative-flow rewards, such as costs, the optimist decision rule is Minimin Note that
negative-flow rewards are expressed with positive numbers.)
Maximax decision rule is followed:
1. For each action alternative (matrix row) determine the maximum payoff possible.
2. From these maxima, select the maximum payoff. The action alternative leading to this payoff is
the chosen decision.

The maximin criterion is a pessimistic approach. It suggests that the decision maker examine only
the minimum payoffs of alternatives and choose the alternative whose outcome is the least bad. This
criterion appeals to the cautious decision maker who seeks to ensure that in the event of an
unfavorable outcome, there is at least a known minimum payoff. This approach may be justified
because the minimum payoffs may have a higher probability of occurrence or the lowest payoff may
lead to an extremely unfavorable outcome. This criterion is illustrated in the table below.
According to Maximax criterion, the farmer (the decision maker) chooses the best among the
conditions determined for each business alternative. The decision maker is optimistic about the
pumpkin oil production and oil selling conditions. The Maximax criterion showed that A1 (5 ha of
oil pumpkin) was the best choice (2475 €) (Tables 2 and 5).

The results of Maximax criterion calculation

Source: Pažek (2008).

Savage’s Minimax Regret

The Savage Minimax Regret criterion examines the regret, opportunity cost or loss resulting when a
particular situation occurs and the payoff of the selected alternative is smaller than the payoff that
could have been attained with that particular situation. The regret corresponding to a particular
payoff Xij is defined as Rij = Xj(max) – Xij where Xj(max) is the maximum payoff attainable under
the situation Sj. This definition of regret allows the decision maker to transform the payoff matrix
into a regret matrix. The minimax criterion suggests that the decision maker looks at the maximum
regret of each strategy and selects the one with the smallest value. This approach appeals to cautious
decision makers who want to ensure that the selected alternative does well when compared to other
alternatives regardless of the situation arising. It is particularly attractive to a decision maker who
knows that several competitors face identical or similar circumstances and who is aware that the
decision maker’s performance will be evaluated in relation to the competitors. This criterion is
applied to the same decision situation and transforms the payoff matrix into a regret matrix.
The Minimax Regret criterion focuses on avoiding the worst possible consequences that could result
when making a decision. Although regret is an emotional state (a psychological sense of loss) which,
being subjective, can be problematic to assess accurately, the assumption is made that regret is
quantifiable in direct (linear) relation to the rewards Rij expressed in the payoff matrix. This means
that an actual loss of, say, an euro (an accounting loss) will be valued exactly the same as a failure to
take
advantage of the opportunity to gain an additional euro (an opportunity loss, which is disregarded in
financial accounting). In other words, the Minimax Regret criterion views actual losses and missed
opportunities as equally comparable. Regret is defined as the opportunity loss to the decision maker
if action alternative Ai is chosen and
state of nature Sj happens to occur. Opportunity loss (OL) is the payoff difference between the best
possible outcome under Sj and the actual outcome resulting from choosing Ai given that Sj occurs.
Thus, if the decision alternative secures the best possible payoff for a given state of nature, the
opportunity loss is defined to be zero. Otherwise, the opportunity loss will be a positive quantity.
Negative opportunity losses are not defined. Savage’s Minimax Regret criterion is formally defined
as:
OL
ij = (column j maximum payoff) - Rij - for positive-flow payoffs (profits, income)
OL
ij = Rij - (column j minimum payoff) - for negative-flow payoffs (costs)
where R
ij is the payoff (reward) for row i and column j of the payoff matrix R.
Opportunity losses are defined as nonnegative numbers. The best possible OL is zero (no regret), and
the higher OL value, the greater the regret.
Minimax Regret decision rule is defined as:
1. Convert the payoff matrix R = { Rij } into an opportunity loss matrix OL = { OLij }.
2. Apply the minimax rule to the OL matrix.

The Savage minimax regret criterion examines the regret, opportunity cost or loss resulting when a
particular situation occurs and the payoff of the selected alternative is smaller than the payoff that
could have been attained with that particular situation. The regret corresponding to a particular
payoff Xij is defined as Rij = Xj(max) – Xij where Xj(max) is the maximum payoff attainable under
the situation Sj. This definition of regret allows the decision maker to transform the payoff matrix
into a regret matrix. The minimax criterion suggests that the decision maker look at the maximum
regret of each strategy and select the one with the smallest value. This approach appeals to cautious
decision makers who want to ensure that the selected alternative does well when compared to other
alternatives regardless of what situation arises. It is particularly attractive to a decision maker who
knows that several competitors face identical or similar circumstances and who is aware that the
decision maker’s performance will be evaluated in relation to the competitors. This criterion is
applied to the same decision situation and transforms the payoff matrix into a regret matrix. This is
shown below.
Regret criterion minimizes the probable regrets for decision maker. The regret values for specific
scenario were determined according to all selling scenarios whereas minimax or Savage’s criterion
was applied to these values

The demonstration of Savage criterion calculation matrix

 Source: Pažek
(2008).
In the observed research the regret of the farmer will be by pumpkin oil production and 50% selling
effectiveness 1463 € (Table 7). Alternative 2 is chosen.

The results of Minmax criterion calculation

Source: Pažek (2008).

Hurwicz’s Optimism – Pessimism Criterion

The most well-known criterion is the Hurwicz criterion, suggested by Leonid Hurwicz in 1951,
which selects the minimum and the maximum payoff to each given action x. The Hurwicz criterion
attempts to find a middle ground between the extremes posed by the optimist and pessimist criteria.
Instead of assuming total optimism or pessimism, Hurwicz incorporates a measure of both by
assigning a certain percentage weight to optimism and the balance to pessimism. However, this
approach attempts to strike a balance between the maximax and maximin criteria. It suggests that the
minimum and maximum of each strategy should be averaged using a and 1 - a as weights. a
represents the index of pessimism and the alternative with the highest average selected. The index a
reflects the decision maker’s attitude towards risk taking. A cautious decision maker will set a = 1
which reduces the Hurwicz criterion to the maximin criterion. An adventurous decision maker will
set a = 0 which reduces the Hurwicz criterion to the maximax criterion.

The Hurwicz criterion attempts to find a middle ground between the extremes posed by the optimist
and pessimist criteria. Instead of assuming total optimism or pessimism, Hurwicz incorporates a
measure of both by assigning a certain percentage weight to optimism and the balance to pessimism.
A weighted average can be computed for every action alternative with an alpha-weight α, called the
coefficient of realism. "Realism" here means that the unbridled optimism of Maximax is replaced by
an attenuated optimism as denoted by the α. Note that 0 ≤ α ≤ 1. Thus, a better name for the
coefficient of realism is coefficient of optimism. An α = 1 denotes absolute optimism (Maximax)
while an α = 0 indicates absolute pessimism (Maximin). The α is selected subjectively by the
decision maker. Selecting a value for α simultaneously produces a coefficient of pessimism 1 - α ,
which reflects the
decision maker's aversion to risk. A Hurwicz weighted average H can now be computed for every
action alternative Ai in A as follows:
H (Ai ) = α (row maximum) + ( 1 - α ) (row minimum) - for positive-flow payoffs (profits, revenues)
H (Ai ) = α (row minimum) + ( 1 - α ) (row maximum) - for negative-flow payoffs (costs, losses)
Hurwicz decision rule is followed:
1. Select a coefficient of optimism value α .
2. For every action alternative compute its Hurwicz weighted average H.
3. Choose the action alternative with the best H as the chosen decision ("Best" means Max {H} for
positive-flow payoffs, and Min {H} for negative-flow payoffs).

This approach attempts to strike a balance between the maximax and maximin criteria. It suggests
that the minimum and maximum of each strategy should be averaged using a and 1 - a as weights. a
represents the index of pessimism and the alternative with the highest average is selected. The index
a reflects the decision maker’s attitude towards risk taking. A cautious decision maker will set a = 1
which reduces the Hurwicz criterion to the maximin criterion. An adventurous decision maker will
set a = 0 which reduces the Hurwicz criterion to the maximax criterion.
According to the Hurwicz’s criterion, the farmer is between pessimistic and optimistic attitude. Each
result has been weighted according to optimistic coefficient (k = 0.7). The highest and the lowest
values of each business alternative has been multiplied by optimistic coefficient (k = 0.7) and
pessimistic coefficient (1-k = 0.3). The highest calculated average value is selected, as seen in Table
4, by alternative 1 (495.5 €).

The computation results of Hurwicz’s criterion (calculated by k = 0.7)

Source: Pažek
(2008).

Laplace’s Criterion
The Laplace’s insufficient reason criterion postulates that if no information is available about the
probabilities of the various outcomes, it is reasonable to assume that they are likely equally.
Therefore, if there are n outcomes, the probability of each is 1/n. This approach also suggests that the
decision maker calculate the expected payoff for each alternative and select the alternative with the
largest value. The use of expected values distinguishes this approach from the criteria of using only
extreme payoffs. This characteristic makes the approach similar to decision making under risk. The
Laplace’s criterion is the first to make explicit use of probability assessments regarding the
likelihood of occurrence of the states of nature. As a result, it is the first elementary model to use all
of the information available in the payoff matrix. The Laplace’s argument makes use of Jakob
Bernoulli's Principle of Insufficient Reason. The
principle, first announced in Bernoulli's posthumous masterpiece, Ars Conjectandi (The Art of
Conjecturing, 1713), states that “in the absence of any prior knowledge, we should assume that the
events have equal probability". It meas that the events are mutually exclusive and collectively
exhaustive. Laplace posits that, to deal with uncertainty rationally, probability theory should be
invoked. This means that for each state of nature (Sj in S), the decision maker should assess the
probability of pj that Sj will occur. This can always be done - either theoretically, empirically or
subjectively. Laplace decision rule is followed:
1. Assign pj = P (Sj ) = 1/n to each Sj in S, for j = 1, 2, ..., n.
2. For each Ai (payoff matrix row), compute its expected value: E (Ai ) = Σj pj (Rij ).
for i = 1, 2, ..., m. Since pj is a constant in Laplace, E (Ai ) = Σj pj (Rij ) = pj Σj Rij .
3. Select the action alternative with the best E (Ai ) as the optimal decision. "Best" means max for
positive-flow payoffs (profits, revenues) and min for negative-flow payoffs (costs)
The Laplace insufficient reason criterion postulates that if no information is available about the
probabilities of the various outcomes, it is reasonable to assume that they are equally likely.
Therefore, if there are n outcomes, the probability of each is 1/n. This approach also suggests that the
decision maker calculate the expected payoff for each alternative and select the alternative with the
largest value. The use of expected values distinguishes this approach from the criteria that use only
extreme payoffs. This characteristic makes the approach similar to decision making under risk. A
table illustrates this criterion below.
According to Laplace’s criterion, when the probabilities of conditions are not known, the
probabilities (S1, S2 and S3) are accepted as equal (0.33). No probability has priority to another one.
The weighted value of each business alternative was found by multiplying by all three probabilities
with 0.33 and the added together. Since, the highest value was (-383.67 €), the farmer will choose the
alternative 1:

The decision matrix based on Laplace’s criterion calculation

Source: Pažek
(2008).

The aggregate game criterion results showed that the most profitably alternative compatible with the
assessment of criterion by decision making under uncertainty in agriculture is alternative A1
(production of oil pumpkins on 5 ha arable land and the presumption of 100 % selling of pumpkin
oil):

The summarized results of suggested business alternatives of pumpkin oil production and sales

Source: Pažek (2008).

In the decisions under uncertainty individual decision makers have to choose one of presumed
business alternatives with the extended information about their profitability, outcomes, costs,
financial results, but in the absence of any information about the probabilities of the various states of
nature. The paper presented a decision making process under uncertainty in agriculture. The classical
criterion of Wald’s, Hurwicz’s, Maximax, Savage’s and Laplace’s are assessed and compared in the
case study of pumpkin oil production and selling of pumpkin oil. The assessment was made on the
basis of financial results for individual business alternative evaluation. The results show that
alternative 1 is recommended, where the farmer should prefer the pumpkin oil production on 5 ha
arable land and the total oil production should be sold. We believe that there is a need to place more
emphasis on determining the uncertainty in agriculture, especially in food production and food
processing.
References:
Pažek, K., Rozman, Č. (2008). DECISION MAKING UNDER CONDITIONS OF UNCERTAINTY
IN AGRICULTURE: A CASE STUDY OF OIL CROPS. University of Maribor, Faculty of
Agriculture and Life Sciences, Pivola 10, 2311 Hoče, Slovenia.

Reynolds, B., Schaeffer B. & Sandborn, P. (2008). Decisions Under Uncertainty, ENME 808s
Product & System Cost Analysis, End of Semester Class Project,
http://terpconnect.umd.edu/~sandborn/courses/808S_projects/reynolds.html

Wald, A. (1950): Statistical Decision Functions, Wiley, New York.

Rozman, Č., Pažek, K., Bavec, M., Bavec, F., Turk, J., Majkovič, D. (2006): The Multi-criteria
analysis of spelt food processing alternatives on small organic farms. J. Sustain. Agric. 28(2):159-
179.

Wen, M., Iwamura, K. (2008): Fuzzy facility location-allocation problem under the Hurwicz
criterion. European Journal of Operational Research 184: 627-635.