Decision Theory 1

Decision Theory

12.1. DECISION MAKING UNDER RISK(DMUR)

Decision-making under risk (DMUR) describes a situation in which each strategy results in more
than one outcomes or payoffs and the manager attaches a probability measure to these payoffs. This
model covers the case when the manager projects two or more outcomes for each strategy and
he or she knows, or is willing to assume, the relevant probability distribution of the outcomes.
Thefollowingassumptionsaretobemade:(1)Availabilityofmorethanonestrategies,(2)Theexistence of
more than one states of nature, (3) The relevant outcomes and (4) The probability distribution of
outcomesassociatedwitheachstrategy.Theoptimalstrategyindecisionmakingunderriskisidentified
bythestrategywithhighestexpectedutility(orhighestexpectedvalue).

Problem 12.3.
In a game of head and tail of coins the player A will get Rs. 4/- when a coin is tossed and head
appears; and will lose Rs. 5/- each time when tail appears. Find the optimal strategy of the player.

Solution
Let us apply the expected value criterion before a decision is made. Here the two monetary
outcomes are + Rs. 4/- and – Rs. 5/- and their probabilities are ½ and ½. Hence the expected monetary
value = EMV = u1 p1 + u2 p2 = + 4 × 0.5 + (–5) × 0.5 = – 0.50. This means to say on the average the
player will loose Rs. 0.50 per game.

Problem 12.4.
A marketing manager of an insurance company has kept complete records of the sales effort of
thesalespersonnel.Theserecordscontaindataregardingthenumberofinsurancepoliciessoldandnet
revenues received by the company as a function of four different sales strategies. The manager has
constructed the conditional payoff matrix given below, based on his records. (The state of nature
referstothenumberofpoliciessold).Thenumberwithinthetablerepresentsutilities.Supposeyouare a new
salesperson and that you have access to the original records as well as the payoff matrix. Which
strategy would youfollow?
State of nature N1 N2 N3
Probability 0.2 0.5 0.3
Strategy Utility Utility Utility
S1 (1 call, 0 follow up) 4 6 10
S2 (1 call, one follow up) 6 5 9
S3 (1 call, two follow-ups) 2 10 8
S4 (1 call, threefollow-ups) 10 3 7

Solution
Asthedecisionistobemadeunderrisk,multiplyingtheprobabilityandutilityandsummingthem up give
the expected utility for thestrategy.
 State of Nature N1 N2 N3 Expected utility or expected payoffs
Probability 0.2 0.5 0.3
Strategy Utility Utility Utility
S1 4 6 10 0.2 × 4 + 0.5 × 6 + 0.3 × 10 = 6.8
S2 6 5 9 0.2 × 6 + 0.5 × 5 + 0.3 × 9 = 6.4
S3 2 10 8 0.2 × 2 + 0.5 × 10 + 0.3 × 8 = 7.8
S4 10 3 7 0.2 × 10 + 0.5 × 3 + 0.3 × 7 = 5.6
As the third strategy gives highest expected utility 1 call and 2 follow up yield highest utility.

Problem 12.5.
A company is planning for its sales targets and the strategies to achieve these targets. The data in
terms of three sales targets, their respective utilities, various strategies and appropriate probability
distribution are given in the table given below. What is the optimal strategy?
Sales targets (× lakhs) 50 75 100
Utility 4 7 9
Prob. Prob. Prob.
Strategies
S1 0.6 0.3 0.1
S2 0.2 0.5 0.3
S3 0.5 0.3 0.2

Solution
Expected monetary value of a strategy =  Sales target × Probability
Expected utility of a strategy =  Utility × Probability.

Sales targets (× lakhs) 50 75 100 Expected Monetary Value Expected Utility

Utility 4 7 9
Prob. Prob. Prob.
Strategies.
S1 0.6 0.3 0.1 50×0.6+75×0.3+100×0.1 4 × 0.6 + 7 × 0.3 + 9 × 0.1
= 62.5 = 5.4
S2 0.2 0.5 0.3 50 × 0.2 + 75 × 0.5 + 100 × 4 × 0.2 + 7 × 0.5 + 9 ×
0.3 = 77.5 0.3 = 7.0
S3 0.5 0.3 0.2 50×0.5+75×0.3+100×0.2 4 × 0.5 + 7 × 0.3 + 9 × 0.2
= 67.5 = 5.9

As both expected money value and expected utility of second strategy are higher than the other
two, strategy two is optimal.
12.2. DECISION MAKING UNDERUNCERTAINTY
Decision making under uncertainty is formulated exactly in the same way as decision making
underrisk,onlydifferenceisthatnoprobabilitytoeachstrategyisattached.Letusmakeacomparative
tabletocomparethethree,i.e.decisionmakingundercertainty,risk,anduncertainty.
Decision making under certainty Decision making under risk Decision making under
Uncertainty.
State of Nature State of Nature State of Nature
N N1 N2 N3 N1 N2 N3
Probability p1 p2 p3
Strategy Utility or Payoff Strategy Utility or Payoff Strategy Utility or Payoff.
S1 u1 S1 u 11 u 12 u 13 S1 u 11 u 12 u 13
S2 u2 S2 u 21 u 22 u 23 S2 u 21 u 22 u 23
S3 u3 S3 u 31 u 32 u 33 S3 u 31 u 32 u 33
* One state ofnature * Morethanonestatesofnature *More than one states ofnature.
* Single columnmatrix * Multiple columnmatrix *Multiple column matrix.
* Deterministicoutcomes * Probabilistic outcomes (i.e. * Uncertain outcomes (i.e.
Probabilities are attachedto probabilities are not attachedto
Various states of nature) various states of nature).
* Optimal strategy is the one * Optimal strategy is identified by * Optimal strategy is identified
with the highestutility. the use of expected value using a number of different
criterian. criterian.

In decision making under uncertainty, remember that no probabilities are attached to set of the
statesofnature.Sometimeswemayhaveonlypositiveelementsinthegivenmatrix,indicatingthatthe
company under any circumstances will have profit only. Sometimes, we may have negative elements,
indicatingpotentialloss.Whilesolvingtheproblemofdecisionmakingunderuncertainty,wehavetwo
approaches, the first one is pessimistic approach and the second one is optimistic approach. Let us
examine this by solving aproblem.

Problem 12.6.
The management of XYZ company is considering the use of a newly discovered chemical which,
whenaddedtodetergents,willmakethewashingstet,thuseliminatingthenecessityofaddingsofteners.
Themanagementisconsideringatpresenttime,thesethreealternativestrategies.
S1 = Add the new chemical to the currently marketed detergent DETER and sell it under label
‘NEW IMPROVED DETER’.
S2 = Introduce a brand new detergent under the name of ‘SUPER SOFT’
S3 = Develop a new product and enter the softener market under the name ‘EXTRA WASH’.
Themanagementhasdecidedforthetimebeingthatonlyoneofthethreestrategiesiseconomically
feasible (under given market condition). The marketing research department is requested to develop a
conditional payoff matrix for this problem. After conducting sufficient research, based on personal
interviewsandanticipatingthepossiblereactionofthecompetitors,themarketingresearchdepartment
submits the payoff matrix given below. Select the optimalstrategy.
 State of nature.
Strategies. N1 N2 N3
Utility of Payoffs.
S1 15 12 18
S2 9 14 10
S3 13 4 26

Solution
When no probability is given, depending upon risk, subjective values, experience etc., each
individual may choose different strategies. These are selected depending on the choice criterion. That
is why sometimes the decision making under uncertainty problems are labeled as choice creation
models. Two criterians may be considered here. One is Criterion of Optimism and the other is
Criterion of Pessimism.

12.3. CRITERION OFOPTIMISM

Here we determine the best possible outcome in each strategy, and then identify the best of the
best outcome in order to select the optimal strategy. In the table given below the best of the best is
written in the left hand side margin.
State of nature.
Strategies. N1 N2 N3 Best or Maximum outcome
Utility of Payoffs. (Row maximum)
S1 15 12 18 18
S2 9 14 10 14
S3 13 4 26 26 Maximax.
While applying the criterion of optimism, the idea is to choose the maximum of the maximum
values; the choice process is also known as Maximax.

12.4. CRITERION OFPESSIMISM

When criterion of pessimism is applied to solve the problem under uncertainty, first determine
worst possible outcome in each strategy (row minimums), and select the best of the worst outcome
inordertoselecttheoptimalstrategy.Theworstoutcomesareshowninthelefthandsidemargin.
State of nature.
Strategies. N1 N2 N3 Worst or minimum out come
Utility of Payoffs. (Row minimums)
S1 15 1 18 12 Maximin
2
S2 9 1 10 9
4
S3 13 4 26 4
Best among the worst outcome is 12, hence the manager selects the first strategy. Maximin
assumescompletepessimism.Maximaxassumescompleteoptimism.Toestablishadegreeofoptimism
or pessimism, the manager may attach some weights to the best and the worst outcomes in order to
reflect in degree of optimism or pessimism. Let us assume that manager attaches a coefficient of
optimism of 0.6 and then obviously the coefficient of pessimism is 0.4. The matrix shown below
showshowtoselectthebeststrategywhenweightsaregiven.
Strategy. Best or maximum Payoffs Worst or minimum Payoffs Weighted Payoffs.
Weights. 0.6 0.4
S1 18 12 0.6 × 18 + 0.4 × 12 = 15.6
S2 14 9 0.6 × 14 + 0.4 × 9 = 12.0
S3 26 4 0.6 × 26 + 0.4 × 4 = 17.2
Maximum.

12.5. CRITERION OFREGRET

In this case, we have to determine the regret matrix or opportunity loss matrix. To find the
opportunity loss matrix (column opportunity loss matrix), subtract all the elements of a column from
the highest element of that column. The obtained matrix is known as regret matrix. While selecting the
best strategy, we have to select such a strategy, whose opportunity loss is zero, i.e. zero regret. If we
selectanyotherstrategy,thentheregretistheelementatthatstrategy.Forthematrixgiveninproblem
12.6 the regret matrix is
State of nature.
Strategies.
N1 N2 N3
Utility of Payoffs.
S1 0 2 8
S2 6 0 16
S3 2 10 0
Rule for getting the regret matrix: In each column, identify the highest element and
then subtract all the individual elements of that column, cell by cell, from the highest element
toobtainthecorrespondingcolumnoftheregretmatrix.
To select the optimal strategy we first determine the maximum regret that the decision maker can
experience for each strategy and then identify the maximum of the maximum regret values. This is
shown in the table below:
State of nature.
Strategies. N1 N2 N3
Regret or Opportunity loss. Maximum regret.
S1 0 2 8 8 minimax.
S2 6 0 16 16
S3 2 10 0 10
Select the minimum of the maximum regret (Minimax regret). The choice process can
be known as minimax regret. Suppose two strategies have same minimax element, then the
manager needs additional factors that influe nce hisselection.
12.6. EQUAL PROBABILITYCRITERION
Aswedonothaveanyobjectiveevidenceofaprobabilitydistributionforthestatesofnature,one can use
subjective criterion. Not only this, as there is no objective evidence, we can assign equal probabilities
to each of the state of nature. This subjective assumption of equal probabilities is known
asLaplacecriterion,orcriterionofinsufficientreasoninmanagementliterature.
Once equal probabilities are attached to each state of nature, we revert to decision making under
risk and hence can use the expected value criterion as shown in the table below:
State of nature
N1 N2 N3 Expected monetary value (EMV)
Probabilities 1/3 1/3 1/3
Strategy Utility or Payoffs.
S1 15 12 18 15 × 1/3 + 12 × 1/3 + 18 × 1/3 =15 Maximum
S2 9 14 10 9 × 1/3 + 14 × 1/3 + 10 × 1/3 = 11
S3 13 4 26 13 × /3 + 4 × 1/3 + 26 × 1/3 = 14 1/3
As S1 is having highest EMV it is the optimal strategy.

12.7. DECISIONMAKINGUNDERCONFLICTANDCOMPETITION
In the problems discussed above, we have assumed that the manager has a finite set of strategies
andhehastoidentifytheoptimalstrategydependingontheconditionofcompletecertaintytocomplete
uncertainty. In all the models, the assumptions made are (1) Various possible future environments that
the decision maker will face can be enumerated in a finite set of states of nature and (2) The complete
payoff matrix is known. Now, let us consider that two rationale competitors or opponents are required
to select optimal strategies, given a series of assumptions, including: (1) The strategies of each party
areknowntobothopponents,(2)Bothopponentschoosetheirstrategiessimultaneously,(3)thelossof one
party equals exactly to gain of the other party, (4) Decision conditions remain the same, and (5) It is a
repetitive decision making problem (refer to Gametheory).
Two opponents are considered as two players, and we adopt the convention that a positive
payoff will mean a gain to the row player A or maximizing player, and a loss to the column player
B or minimizing player. (Refer to 2 person zero sum game).
Consider the matrix given: maximin identifies outcome for player A and Minimax identifies the
optimal strategy outcome for player B. This is because each player can adopt the policy, which is best
to him. A wants to maximize his minimum outcomes and B wants to minimize his maximum loses.
Player B
B1 B2 B3 B4 Row minimum
A1 8 12 7 3 3
Player A A2 9 14 10 16 9
A3 7 4 26 5 4
Column maximum 9 14 26 16
A selects the second strategy as it guaranties him a minimum of 9 units of money and B chooses
strategy 2 as it assures him a minimum loss of 9 units of money. This type of games is known as pure
strategy game.The alement where minimax point and maximin point are same known as saddle
point.
12.8. HURWICZ CRITERION (CRITERION OFREALISM)
This is also known as weighted average criterion, it is a compromise between the maximax and
maximindecisionscriteria.Ittakesbothofthemintoaccountbyassigningthemweightsinaccordance with the
degree of optimism or pessimism. The alternative that maximizes the sum of these weighted payoffs is
thenselected.

Problem 12.7.
The following matrix gives the payoff of different strategies (alternatives) A, B, and C against
conditions (events) W, X, Y and Z. Identify the decision taken under the following approaches:
(i) Pessimistic, (ii) Optimistic, (iii) Equal probability, (iv) Regret, (v) Hurwicz criterion. The decision
maker’s degree of optimism () being0.7.
Events

Solution (for i, ii, and iii)

 W X Y Z
Rs. Rs. Rs. Rs.
A 4000 –100 6000 18000
B 20000 5000 400 0
C 20000 15000 – 2000 1000

W X Y Z Maximum regret. Rs.

Regret Rs. RegretRs. RegretRs Regret Rs —
A 16000 15100 0 0 16000
B 0 10000 5600 18000 18000
C 0 0 8000 17000 17000

Pessimistic Optimistic
Equal Probability value
Maximin value Maximax value
A – Rs. 100/- Rs. 18000 Rs. ¼ (4000 – 100 + 6000 + 18000) = Rs. 6975/-
B Rs. 0/- Rs. 20000 Rs. ¼ (20000 + 5000 + 400 + 0) = Rs. 6350/-
C – Rs. 2000 Rs. 20000 Rs. ¼ (20000 + 15000 – 2000 + 1000) = Rs. 8,500/-

Under pessimistic approach, B is the optimal strategy, under optimistic approach B or C are
optimal strategies, and under equal probability approach C is the optimal strategy.
(iv) Giventablerepresentstheregretsforeveryeventandforeachalternativecalculatedby:
= ith regret = (maximum payoff – ith payoff) for the jth event.
As strategy A shows minimal of the maximum possible regrets, it is selected as the optimal
strat egy.
(v) For a given payoff matrix the minimum and the maximum payoffs for each alternative is
given in the tablebelow:
 Alternative Maximum payoff Minimum payoff. Payoff = × maximum payoff + (1–)
Rs Rs minimum payoff , where = 0.7 (Rs)
A 18000 – 100 0.7 × 18000 – 0.3 × 100 = 12570
B 20000 0 0.7 × 20000 + 0.3 × 0 = 14000
C 20000 –2000 0.7 × 20000 – 0.3 × 2000 = 13400
Under Hurwicz rule, alternative B is the optimal strategy as it gives highest payoff.

Problem 12.8.
A newspaper boy has the following probabilities of selling a magazine. Cost of the copy is
Rs.0.30andsalepriceisRs.50.Hecannotreturnunsoldcopies.Howmanycopiescanheorder?
No. of copies sold Probability
10 0.10
11 0.15
12 0.20
13 0.25
15 0.30
Total 1.00

Solution
The sales magnitude of newspaper boy is 10, 11, 12, 13, 14 papers. There is no reason for him to
buy less than 10 or more than 14 copies. The table below shows conditional profit table, i.e. the profit
resulting from any possible combination of supply and demand. For example, even if the demand on
some day is 13 copies, he can sell only 10 and hence his conditional profit is 200 paise. When stocks
11 copies,hisprofitis220paiseondayswhenbuyersrequest 11,12,13,and14copies.Butontheday when he
has 11 copies, and the buyers buy only 10 copies, his profit is 170 paisa, because one copy is
unsold.Hencepayoff=20×copiessold–30×copiesunsold.Henceconditionalprofittableis:
Conditional Profit Table (paisa)
Possible Stock Action

Possible demand Probability 10 copies 11 copies 12 copies 13 copies 14 copies

(number of copies)
10 0.10 200 170 140 110 80
11 0.15 200 220 190 160 130
12 0.20 200 220 240 210 180
13 0.25 200 220 240 260 230
14 0.30 200 220 240 260 280
 Expected Profit Table
Expected Profit from Stocking in Paisa

Possible demand Probability 10 copies 11 copies 12 copies 13 copies 14 copies

10 0.10 20 17 14 11 8
11 0.15 30 33 28.5 24 19.5
12 0.20 40 44 48 42 36
13 0.25 50 55 60 65 57.5
14 0.30 60 66 72 78 84
Total Expected Profit in Paisa 200 215 222.5 220 205

The newsboy must therefore order 12 copies to earn the highest possible average daily profit
of222.5paise.Henceoptimalstockis12papers.Thisstockingwillmaximizethetotalprofitsover
aperiodoftime.Ofcoursethereisnoguaranteethathewillmakeaprofitof222.5paisetomorrow.
However,ifhestocks12copieseachdayundertheconditiongiven,hewillhaveaverageprofitof
222.5 paisa per day. This is the best be can do because the choice of any of the other four possible
stock actions will result in a lower daily profit.
(Note: The same problem may be solved by Expected Opportunity Loss concept as shown
below)
EOL(ExpectedOpportunityLoss)canbecomputedbymultiplyingtheprobabilityofeachof state
of nature with the appropriate loss value and adding the resultingproducts.
For example: 0.10 × 0 + 0.15 × 20 + 0.20 × 40 + 0.25 × 60 + 0.30 × 80 = 0 + 3 + 15 + 24 = 50 paise.

Conditional Loss Table in Paisa

Possible Stock Action (Alternative)

Possible demand
Probability 10 copies 11 copies 12 copies 13 copies 14 copies
Number of copies(event)
10 0.10 0 30 60 90 120
11 0.15 20 0 30 60 90
12 0.20 40 20 0 30 60
13 0.25 60 40 20 0 30
14 0.30 80 60 40 20 0
 If the newspaper boy stocks 12 papers, his expected loss is less.
Possible demand
Probability 10 copies 11 copies 12 copies 13 copies 14 copies
Number of copies (event)
10 0.10 0 3 6 9 12
11 0.15 3 0 4.5 9 13.5
12 0.20 8 4 0 6 12
13 0.25 15 10 5 0 7.5
14 0.30 24 18 12 6 0
EOL (Paisa) 50 35 27.5 30 45