Decisions Under Ignorance
Decisions Under Ignorance
Decisions Under Ignorance
IGNORANCE
Decisions Under Ignorance
• When Jane is about to choose a main course, rationality forbids her to choose a3.
• This is because a2 will lead to a better outcome than a3.
Dominance
• The widely accepted dominance principle prescribes that dominated acts must not be
chosen.
Another Example
• You think that having a valid ticket is better than not having one.
• This relation can be represented by 1 and 0.
• it can be clearly seen that a1 dominates a2.
Dominance
• We wish to formalise the idea that one alternative is better than another one.
• Let > be a relation on acts, such that ai > aj if and only if it is more rational to perform
act ai rather than act aj.
• This means that whenever ai > aj is true, then ai should to be chosen (rationally) over aj.
• Furthermore, ai >= aj means that ai is at least as rational as aj.
• And ai ~ aj means that the two acts are equally rational.
Dominance
• The dominance principle comes in two versions; the Weak and the Strong dominance principle.
• Weak Dominance articulates the plausible idea that one alternative is at least as rational as another if all its
outcomes, under every possible state of the world, will be at least as good as that of the other alternative.
• Strong Dominance stipulates that for an alternative to be more rational than another, two conditions must be
fulfilled.
• First, its outcomes under all states must be at least as good as those of the other alternative.
• Secondly, there must also be at least one state under which the outcome is strictly better than that of the
other alternative.
• For example, booking a ticket with Black Jack airlines strongly dominates booking a ticket with Air Mojave.
Dominance
Minimax
Laplace Maximax Maximin Hurwicz
Regret
Criterion Criterion Criterion Criterion
Criterion
Decisions Under Ignorance
If the probabilities of events cannot be determined in a decision problem, such problems are handled within
decision problems in case of uncertainty.
In this type of decision problems; The evaluation of criteria and the choice between alternatives can be
influenced by many different criteria such as the manager's view of events and the policies of the business.
In case of decision making under uncertainty problems;
✓ Laplace Criterion (Eş Olasılık Kriteri)
✓ Optimism (Maximax) Criterion (İyimserlik Kriteri)
✓ Pessimism (Maximin) Criterion (Kötümserlik Kriteri)
✓ Regret (Minimax) Criterion (Pişmanlık Kriteri)
✓ Hurwicz Criterion
are used.
1. Laplace Criterion
Possible Revenues
33,3% 33,3% 33,3%
Factory Size T1 T2 T3
Small F1 80 60 25
Medium F2 20 120 40
Large F3 10 60 180
1. Laplace Criterion
Possible Revenues
33,3% 33,3% 33,3%
Factory Size T1 T2 T3
Small F1 80 60 25
Expected revenues are calculated :
Medium F2 20 120 40
• E(F1)= 80*0,33 + 60*0,33 + 25*0,33 = 55 Large F3 10 60 180
• E(F2)= 20*0,33 + 120*0,33 + 40*0,33 = 60
• E(F3)= 10*0,33 + 60*0,33 + 180*0,33 = 83,33 max. Revenue
• In the case of costs, the option with the minimum expected cost is preferred.
2. Maximax Criterion
This criterion, also known as the criterion of optimism, is used when the decision-maker is optimistic about future.
Maximax implies the maximisation of maximum payoff.
In this criterion, the decision maker is decided in such a way as to provide the most profit among the possible events.
Example: If the problem addressed in the previous example is solved by maximax criterion;
Possible Revenues
Small F1 80 60 25 80
Medium F2 20 120 40 120
Large F3 10 60 180 180
Once the maximum value for each alternative is found, the best of them is preferred.
2. Maximax Criterion
Possible Costs
This criterion, also known as the criterion of pessimism, is used when the decision-
maker is pessimistic about future.
Maximin implies the maximisation of minimum payoff.
In this criterion, the decision maker chooses the best among the worst results.
Example: Possible Revenues
Small F1 80 60 25 25
Medium F2 20 120 40 20
Large F3 10 60 180 10
3. Maximin Criterion
This criterion focuses upon the regret that the decision-maker might have from
selecting a particular course of action.
Regret; is an opportunity (gain) loss that occurs when an option is decided under a
certain condition.
In other words, regret is seen as the difference between the best possible result under
certain conditions and the actual results of the preferred option.
4. Minimax (Regret) Criterion
To implement this method, the decision maker must first prepare the regret table,
which turns the payments table into opportunity costs.
MR example; Durumlar
Şeçenekler 500 Çekim 750 Çekim 1000 Çekim
Yeni bir MR Ünitesi Satın Alma -15.000 200.000 300.000
Yeni iki MR Ünitesi Satın Alma -150.000 100.000 725.000
Dışarıdan Temin 15.000 22.500 40.000
In case of 500 shots (1st column), the maximum revenue is 15.000 TL.
A manager who chooses the outsourcing (dışarıdan temin) option will not have any
regrets when the demand amount is 500 shots.
Fırsat Kayıpları/Pişmanlıklar Durumlar
Şeçenekler 500 Çekim 750 Çekim 1000 Çekim
Yeni bir MR Ünitesi Satın Alma
Yeni iki MR Ünitesi Satın Alma
Dışarıdan Temin 0
4. Minimax (Regret) Criterion
However, if the decision maker had decided to buy a new MR unit, s/he would have
regretted for loss of 15,000 - (- 15,000) = 30,000 TL.
Durumlar
Şeçenekler 500 Çekim 750 Çekim 1000 Çekim
Yeni bir MR Ünitesi Satın Alma -15.000 200.000 300.000
Yeni iki MR Ünitesi Satın Alma -150.000 100.000 725.000
Dışarıdan Temin 15.000 22.500 40.000
However, if the decision maker had decided to buy two new MR unit, s/he would have
regretted for loss of 15.000-(-150.000)=165.000TL.
Durumlar
Şeçenekler 500 Çekim 750 Çekim 1000 Çekim
Yeni bir MR Ünitesi Satın Alma -15.000 200.000 300.000
Yeni iki MR Ünitesi Satın Alma -150.000 100.000 725.000
Dışarıdan Temin 15.000 22.500 40.000
For other columns (for 750 shots and 1000 shots) the same calculations are made and
then, the below table of opportunity losses is obtained.
Durumlar
Şeçenekler 500 Çekim 750 Çekim 1000 Çekim
Yeni bir MR Ünitesi Satın Alma -15.000 200.000 300.000
Yeni iki MR Ünitesi Satın Alma -150.000 100.000 725.000
Dışarıdan Temin 15.000 22.500 40.000
After the decision maker has created the regret table, the maximum regret values are
determined for each alternative.
And then an alternative with a minimum regret value is preferred.
The decision maker will have minimal regrets by choosing this way.
The maximax and the maximin criteria assumes that the decision-maker is either optimistic
or pessimistic.
Decision makers may remain undecided between optimism and pessimism because of their
past experience.
A more realistic approach would be to take into account the degree or index of
optimism or pessimism of the decision-maker in the process of decision-making.
The Hurwitz criterion is based on the weighting of optimistic and pessimistic results.
The weight value (α) takes a value between 0 and 1. (0≤α≤1)
The decision maker first gives a weight to the optimistic result (α).
In this case, (1-α) would naturally be the weight of the pessimistic result.
5. Hurwicz Criterion
The decision maker firstly determines the most optimistic and pessimistic results for
each decision alternative (for every row).
Then, calculations are made using the determined α and (1- α) values:
α.(max. value in the row)+ (1- α).(min. value in the row)
Durumlar
Şeçenekler 500 Çekim 750 Çekim 1000 Çekim
Yeni bir MR Ünitesi Satın Alma -15.000 200.000 300.000
Yeni iki MR Ünitesi Satın Alma -150.000 100.000 725.000
Dışarıdan Temin 15.000 22.500 40.000
5. Hurwicz Criterion
In our example; It was assumed that the decision maker chose the weight α = 0.5 for
the optimistic result.
Durumlar
In this case; (1- α)=0,5 Şeçenekler 500 Çekim 750 Çekim 1000 Çekim
Yeni bir MR Ünitesi Satın Alma -15.000 200.000 300.000
Yeni iki MR Ünitesi Satın Alma -150.000 100.000 725.000
Dışarıdan Temin 15.000 22.500 40.000
• Buying 1 MR unit=0,5x(300.000)+0,5x(-15.000)=142.500
• Buying 2 MR units=0,5x(725.000)+0,5x(-150.000)=287.500
• Outsourcing =0,5x(40.000)+0,5x(15.000)=27.500
5. Hurwicz Criterion