SS / 08-10 Date : 14/11/2008 Operations Research Decision Theory Decision situations that involve multiple stages- called sequential

decision problems, they are characterized by a sequence of decisions in which following each decision , a chance event occurs which in turn influences the next decision. In analyzing multiple stage decision situations we have to evaluate the decision proceeding in a backward manner by evaluating the best course of action at the later stages to decide the best action at the earlier stages. For this purpose the decision tree or the decision flow diagram as it is sometimes called , is a very effective device. A decision tree is a graphic representation of the sequence of action event combinations available to the decision maker. It depicts in a systematic manner all possible sequences of decisions and consequences. Each such sequence is shown by a distinct path through the tree. A decision tree enables the decision maker to see the various elements of his problem in proper perspective and in a systematic manner. It may be mentioned that the criterion on the basis of which the decisions are made in the decision tree approach is generally the expectation principle. So we may choose the alternative that maximizes the expected profit , or the alternative that minimizes the expected cost .. and so on. Example : An oil company has recently acquired rights in a certain area to conduct surveys and test drillings to lead to lifting oil if it is found in commercially exploitable quantities. The area is considered to have good potential for finding oil in commercial quantities. At the outset , the company has the choice to conduct further geological tests or to carry out a drilling programme immediately. On the known conditions , the company estimates that there is a 70:30 chance of further tests showing a success. Whether the tests show the possibility of ultimate success or not or even if no tests are undertaken at all, the company could still pursue its drilling programme or alternatively consider selling its rights to drill in the area. Thereafter , however if it carries out the drilling programme, the likelihood of final success or failure is considered dependent on the foregoing stages. Thus : if successful tests have been carried out , the expectation of success in drilling is given as 80:20 if the tests indicate failure , the expectation of success in drilling is given as 20:80

if no tests have been carried out at all the possible outcomes and the net present value of each is as follows:

Outcome Value ( Rs Million) Success : With prior tests Without prior tests Failure : With prior tests Without prior tests Sale of exploitation rights: Prior tests show success Prior tests show failure Without prior tests

Net Present 100 120 -50 -40 65 15 45

a) draw up a decision ( probability ) tree diagram to represent the above information ; and b) evaluate the tree in order to advise the management of the company on its best course of action. The decision tree is given as under

22 Decision node Chance node

Failure 0.45 sell 1 2 0 Succ ess 0.55 drill

6 5 Failure 0.2

Success 0.8

1 0 0

+ve 0.7

1
sell -ve 0.3

1
4 5

drill

Success 0.2

10 0

sell 1 5

Failure 0.8 -50

A square represents a decision node or decision fork at which the decision maker has to take a decision, while a circle represents a chance node or chance fork at which events (i.e the states of nature) are branched out. At decision node 3 , there are branches , representing three alternatives of drilling, testing for oil and selling of rights of which the decision maker has to select one. Now if company decides to drill, there are two possible events it may get oil or not, which are shown as branches emanating from circle, whose probabilities are 0.55 and 0.45 respectively . A profit of Rs 120 million would result in case oil is obtained and a loss of Rs 40 million if it is not . The second alternative at the decision node 3 is to go for a test, which may give positive or negative results, with probabilities 0.7 and 0.3 respectively . In case a positive result is indicated a choice has to be made as to whether to sell the rights for Rs 65 million or to drill , which is likely to succeed and

fail with chances 80 :20 . Therefore a decision node number 2 is shown by a square . Similarly for a negative indication there are two options- to sell for Rs 65 million or to drill, which has a 20 % chance of success . Thus decision node 1 is indicated here. The third alternative is to sell the rights for Rs 45 million. The method of solution : as mentioned earlier , the decisions have to be evaluated in a backward manner by evaluating the best course of action at the later stages so as to decide on the best course of action on the earlier stages. Thus a solution to the problem is obtained by working backwards from right to left through the tree. This is called as rollback technique . In this technique we assume that we have reached various decision nodes on the tree and then decide on the optimal act conditional on having reached the node being analysed. Thus on reaching any decision node, we evaluate each of the alternative courses of action available there and select the most appropriate one. We begin with the right most decision node and after having analysed it , we move backward and analyze the preceeding decision node in a similar manner. The process is continued till the first , the leftmost decision node is analysed. The decision at the first node represents the best initial decision. We begin with the evaluation of decision node 1 in the first instance. Note that this point is reached after a negative is indicated by the test. The expected values of each of the two alternatives , to drill and to sell , are obtained below . Alternative 1 drill Outcome Success Failure Pro b 0.2 0.8 Condition al Value 100 Expected Value

20 (40) Total (20) 2 sell 1.0 15 15 The expected value of the drilling option is a loss of Rs 20 million while if the site is sold, we can get Rs 15 million. On the basis of the criterion of maximization of the expected profit , our decision would be to sell the site, which is conditional upon the negation indicated by the test. Now evaluating the decision node 2. for this node the alternatives together with their expected values are shown below. Alternative 1 2 Sell drill Outcome success Pro b 1.0 0.8 Condition al Value 65 100 Expected Value 65 80

(50) (10) Total 70 The two alternative courses of action , namely selling and drilling have expected values equal to Rs 65 million and Rs 70 million , respectively . Obviously therefore provided that a positive result is indicated by the test, the best course would be to go in for oil drilling. Now at this stage we have two conditional decisions sell the site if a negative is obtained on the test and drill in case the test indicates a positive . At each node , the branches on which we have not to move, representing the options ruled out , have been shown cancelled. Next we move to the decision node where a decision has to be taken whether to drill at the outset, to undertake a test or to sell the rights outright. The alternatives along with their expected values are shown . Also the expected values associated with the different chance nodes and the decision nodes are indicated in which the decision tree is given is reproduced. Alternative 1 drill drill Outcome success failure Pro b 0.5 5 0.4 5 0.7 0.3 1.0 Condition al Value 120 (40) Total Total 70 15 Total 45 Expected Value 66 (18) 48 70 49.0 4.5 53.5 45

failure

0.2

2 3

test sell

positive negative

The expected value of the alternative of carrying out a test is Rs 53.5 million , which is the highest of the three. Therefore it is better to test before drilling . This is the initial decision . The overall decision can now be stated as : The test be carried out . If it proves negative , the rights should be sold to give a return of Rs 15 million . To proceed with drilling , if that happens , would expectedly lead to a loss. However if the test proves positive , the drilling should be undertaken.