Write short note on decision-making under uncertainty. Or Explain the decision-making under uncertainty. Or Discuss the modern approaches to decision making under uncertainty.Ripunjay Tiwari
Ans. When a decision involves conditions about which the manager has no information, either about the outcome or the relative chances for any single outcome, he is said to be operating under conditions of uncertainty. Because the manager does not have any information on which he can develop any analysis, the best he can do is to be aware that he has no opportunity of predicting the events. Under these conditions, a number of different decision criteria have been proposed as possible bases for decision-making. These are as follows-
(i) Maximizing the maximum possible payoff – the maximax criterion (optimistic).
(ii) Maximizing the minimum possible payoff- the maximin criterion (pessimistic).
(iii) Minimizing the maximum possible regret to the decision maker the minimax criterion (regret).
(iv) Assuming equally likely probabilities for the occurrence of each possible state of nature -the insufficient criterion (insufficient reasoning).
In the first step of decision-making under uncertainty, a conditional value pay-off table is constructed. The next step is the selection and application of one of the decision criteria. This can be done in the following way –
(i) Maximax Criterion– This decision criterion is applied by the most optimist decision maker when he thinks optimistically about the happening of events affecting a decision. If this philosophy is followed, the manager will choose that alternative under which it is possible to receive the most favorable pay-off.
However, it is dangerous to use this criterion because it ignores possible losses and gains of making or not making a profit. A pay-off table is given in table 3.1.
Referring to table 3.1, the manager using the maximax criterion will list · the most favorable pay-off for each alternative as follows –
Centralization Rs. 30 crores
Decentralisation Rs. 35 crores
When maximum monetary pay-off is the objective, the decision maker will decentralize. But, he is not sure that the given pay-off materialize because the nature of demand is completely unknown.
(ii) Maximin Criterion – This criterion is adopted by the most pessimistic decision maker. The manager believes that worst possible may take place. This pessimism causes the selection of that alternative which maximizes the least favorable pay-off. In table 3.2, the minimum pay-off for each alternative is as follows –
Centralization Rs. 15 crores
Decentralization Rs. 10 crores
The decision would be to centralize the distribution accordingly because it maximizes the minimum pay-off.
(iii) Minimax Criterion– Minimax criterion leads to the minimization of regret. The managerial regret is defined as the pay-off for each alternative under every state of nature of competitive action subtracted from the most favorable pay-off which is possible with the happening of the particular event. When manager chooses an alternative and when a state of nature takes place which does not result in the most favorable pay-off, regret takes place. The manager is regretful that the alternative chosen did not lead to the best payoff. Because the manager does not know the probability of occurring a particular event, he will select an alternative that minimizes his regret. Referring to table 3.1, the pay-off table using minimax criterion can be formed as shown in table 3.2.
Regret table reveals that in the case of high demand, the manager has a regret of Rs. 15 crores if he selects the alternative of centralized distribution. In the same way, in the case of low demand, he has a regret of Rs. 20 crores if he selects the alternative of decentralized distribution. He will select centralization because he is interested in minimizing his regret.
(iv) Insufficient Reason Criterion – Three preceding decision ‘Criteria assume that without any previous experience, it is not possible or worthwhile to allocate any probability to the states of situation. In this case also, probability can be allocated though there is no criterion for allocating the probability. The situation is referred to as insufficient reason criterion or La Place criterion, or Bay’s postulate. In this criterion, equal probability is allocated to each event because the probability cannot be allocated on any basis. The basis is that the probability of happening an event is unknown, every event should be deal as equal. Applying this criterion in table 3.1, following pay-offs will be expected from two alternative actions –