82

DECISION MAKING Introduction

Thousands of business decisions are made every day and not all will MAKE or BREAK the organization. But each one adds a measure of success (or failure) to the operations, i.e. all decisions have some influence- large or small- on performance. e.g. Arelik Co. managers must decide which of the several potential products to develop. An electronics manufacturer must decide whether to invest in a new process or to stay with a proven one that is producing defective chips at a rate of 1 in 3. Does the ability to make good decisions come naturally or it can be learned? Management scientists hold that education, scientific training, and experience can improve a persons ability to make decisions. The idea of management as a science is founded on its similarity to other sciences as expressed below. Organized principles of knowledge, Use of empirical data, Systematic analysis of data, Repeatable results.

First, the principles and methodology of management, e.g. organization theory, span of control, form an organized or codified body of knowledge. Second, real world data are available for analysis. The business world is essentially a laboratory for the management scientist. Third, an objective and systematic analysis of the data can be made. Fourth, another decision maker could use the same data and arrive at consistent results. Decisions range from simple judgments to complex analyses which can also involve judgment. Judgments typically incorporate basic knowledge, experience and what we often refer to as common sense. They enable us to blend objective and subjective data to arrive at a choice. Fortunately the human brain is capable of selecting and integrating relevant information into a meaningful decision. Quantitative methods of analysis add to the objectivity of such decisions. The appropriateness of a given type of analysis 1. The significance and long-lasting of the decision. e.g. A plant investment may deserve more thorough analysis than a short-term decision to stock Christmas trees for Christmas and New Year eve. 2. The time availability and the cost of analysis. e.g. There must be adequate time to study the complete financial ramifications of a project proposal. 3. The degree of complexity of the decision. Complexity increases when a. many variables are involved, b. the variables are highly interdependent or sequentially related, c. the data describing the variables are incomplete or uncertain. e.g. New factory location decisions are complex, because they involve economic, social, and environmental concerns.

83

Business decision makers have always had to work with incomplete and uncertain data. In some situations a decision maker has (or is assumed to have) complete information about the decision variables; at the other extreme no information available. Managerial decisions are made all along this continuum. Complete Complete Certainty Certainty (All data) Extreme Uncertainty (No data)

Objective Information

Subjective Information

Large Samples

Small Samples

Complete Certainty in decision making requires data on all elements in the population. If such data are not available, large samples lend more certainty than do small ones. Beyond this, subjective information is likely to be better than no data at all. FRAMEWORK FOR DECISION MAKING An analytical and scientific framework for decision implies several systematic steps as explained below. Not all managers follow this formal process- nor all decisions necessitate it. Regardless of the situation, experience and good judgment are always important ingredients in decision making. The steps for scientific decision making are: 1. Define the problem and its parameters (relevant variables), 2. Establish the decision criteria (objectives, to reflect the goals and purpose of the work efforts), 3. Formulation of a model relating the parameters to the criteria, Models can be: a. Verbal (words and descriptions), b. Physical (modified scale. Three dimensional representations of other objects.), c. Schematic (diagrams and charts, graphs and maps), d. Mathematical (equations and numbers), Mathematical models are most useful for understanding complex business problems.

84 4. Generate alternatives by varying the values of the parameters. A decision alternative is a course of action or a strategy that can be chosen by the decision maker. 5. Evaluate the alternatives and select the alternative that best satisfies the criteria. All possibilities must be considered. The results may be undesirable. Outcomes over which the decision maker has little or no control are called States of Nature. 6. Implement the decision and monitor the results. Decision Making is, therefore, the act of selecting a preferred course of action among alternatives. The act of decision making enters into almost all of a managers activities. Managers must reach decisions about objectives and plans for their organizational units. They must decide how to direct, how to organize, how to control. They must not only make many decisions, but also guide subordinates in reaching decisions of their own. Much of managers time is spent in gathering and evaluating information so that he or she will know if a decision is needed and the necessary background information will be available if it is. Businesses and other organizations survive by making and implementing enough of the right decisions; they fail either because they make the right decisions but are unsuccessful in implementing them, or because they make wrong decisions and succeed in implementing them. The success of business and nonprofit organizations hinges on their ability to make good decisions and to implement their decisions well. TYPES OF DECISION MAKING ENVIRONMENTS The types of decisions people make depend on how much knowledge or information they have about the problem scenario. There are three decision making environments. Type 1: Decision Making Under Certainty Decision makers know for sure (i.e. with certainty) the outcome or consequence of every decision alternative. Naturally, they will select the alternative that will result in the best outcome. Linear programming, goal programming, and integer programming are all examples of decision modeling techniques suited for decision making under certainty. Type 2: Decision Making Under Uncertainty The decision maker has no information at all about the various outcomes or states of nature that he or she does not know the probabilities of the various future outcomes. e.g. The probability that Saadet Partisi will control Turkish parliament 20 years from now is not known. It is also impossible to assess the probability for that. Type 3: Decision Making Under Risk The decision maker has some knowledge regarding the probability of occurrence of each outcome or state of nature. e.g. The probability of being dealt a club from a deck of cards is . The probability of rolling a 5 on a die is 1/6. In decision making under risk, the decision maker attempts to identify the alternative that optimizes his or her expected profit or cost.

85 Illustrative Case 1: Ilhandir Lumber Co. lhan Balc is the founder and president of lhandr Lumber Co., a profitable firm located in Gazi Mausa, North Cyprus. lhan identifies his problem as follows; whether to expand his product line by manufacturing and marketing a new product, backyard storage shed. lhan decides his alternatives are to construct: i. ii. iii. a large plant to manufacture the storage sheds, a small plant to manufacture the storage sheds, no plant at all (i.e. he has the option of not developing the new product line).

One of the biggest mistakes that decision makers make is to leave out some important decision alternatives, like no plant at all. Although a particular alternative may seem to be inappropriate or of little value at this time, it may turn out to be the best choice later. lhan determines there are only two possible outcomes: i. ii. The market for storage sheds could be favorable, (meaning that there is a high demand for the product) The market for storage sheds could be unfavorable. (meaning that there is a low demand for the sheds)

A common mistake is to forget about some of the possible outcomes. Optimistic decision makers tend to ignore bad outcomes, whereas pessimistic decision makers tend to ignore a favorable outcome. We have to consider all possibilities. Outcomes over which the decision maker has little or no control are called States of Nature. Ilhans next step is to express the payoff resulting from each possible combination of alternatives and outcomes. Simply, a payoff is the outcome of the decision. Since he wants to maximize his profits in this case, he can use profit to evaluate each consequence. We call such payoffs or profits conditional values. lhan has already evaluated the potential profits associated with the various combinations of alternatives and outcomes. With the favorable market, he thinks a large facility would result in a net profit of 200 000 MU to his firm. (This 200 000 MU is a conditional value because Ilhans receiving the money is conditional upon both his building a large facility and having a good market.) The conditional value if the market is unfavorable would be a 180 000 MU net loss. A small plant would result in a net profit of 100 000 MU in a favorable market, but a net loss of 20 000 MU would occur if the market were unfavorable. Finally doing nothing would result in 0 MU profit in either market. The easiest way to present these values is by constructing a decision table, called Payoff Table. A payoff table is a means of organizing, and illustrating the payoffs from different situations, given various states of nature. Each decision (as alternative) will result in an outcome (or payoff) for each state of nature that will occur in the future.

86 States of Nature Unfavorable Market (MU) - 180.000 - 20.000 0

Alternatives Construct a large plant Construct a small plant Do nothing

Favorable Market (MU) 200.000 100.000 0

lhan has to select a decision model and apply it to the data to help to make decision. The type of decision model available for selection depends on the environment in which we are operating and the amount of uncertainty and risk involved. DECISION MAKING UNDER UNCERTAINTY As stated before, this type of decision making environment exists when a manager cannot assess the probabilities of the different outcomes with confidence, or when virtually no probability data is available. The criteria covered to handle such situations are as follows: 1. Maximax Criterion The maximax criterion selects the alternative that maximizes the maximum payoff or consequence over all alternatives, i.e. the decision is selected that will result in the maximum of the maximum payoffs. This is how the criterion derives its name- The maximum of the maxima. The maximax criterion is very optimistic. The decision maker assumes that the most favorable state of nature for each decision alternative will occur. 2. Maximin Criterion The maximin criterion finds the alternative that maximizes the minimum payoff or consequence over all alternatives, i.e. the decision maker selects the decision that will reflect the maximum of the minimum payoffs. The maximin criterion is pessimistic. The decision maker assumes that the minimum payoff will occur. 3. Equally Likely (Equal Likelihood) or Laplace Criterion The equally likely, also called Laplace, criterion finds the decision alternative with the highest average payoff, i.e. the equally likely criterion weights each state of nature equally, thus assuming that the states of nature are equally likely to occur.

4.Criterion of Realism (Hurwicz Criterion) A compromise between the maximax and maximin criteria. The decision maker is neither optimistic (as the maximax criterion assumes) nor totally pessimistic (as the maximin criterion assumes). With the Hurwicz Criterion, the decision payoffs are weighted by a

87 Coefficient of Optimism, a measure of the decision makers optimism. The coefficient of optimism, defined as , is between 0 and 1 (i.e. 0<<1.0). If =1.0, the decision maker is completely optimistic; if =0, the decision maker is completely pessimistic. (Given this definition, 1- is the Coefficient of Pessimism. The advantage of this approach is that it allows the decision maker to build in personal feelings about relative optimism and pessimism. The formula is Criterion of Realism = (maximum payoff for an alternative) +(1-)(minimum payoff for an alternative) 5. Minimax Regret Criterion This criterion is based on opportunity loss. Opportunity Loss, also called Regret, is the difference between the optimal payoff and the actual payoff received. In other words, its the amount lost by not picking the best alternative. Minimax regret finds the alternative that minimizes the maximum opportunity loss within each alternative. Illustrative Case 1 (continued): Lets apply the above mentioned five criteria to the lhandr Lumber Case. Remember that the decision making environment assumes that lhan has no probability information about the two outcomes. 1. Maximax Criterion We first locate the maximum payoff for each alternative and then select the alternative with the maximum number. This criterion decision locates the alternative with the highest possible gain. From the table, we see that Ilhans maximax choice is the first alternative, construct a large plant. The 200 000 MU payoff is the maximum of the maximum payoff for each decision alternative. States of Nature Alternatives Favourable Unfavourable Market (MU) Market (MU) Construct a large 200.000 plant Construct a 100.000 small plant Do nothing 0 Decision : Construct a large plant. 2. Maximin Criteria - 180.000 - 20.000 0

Maximum in the row (MU) 200.000 100.000 0

Maximax!

88 We first locate the minimum payoff for each alternative and then select the alternative with the maximum number. Ilhans maximin choice, do nothing is shown in the table below. The 0 MU payoff is the maximum of the minimum payoffs for each alternative. States of Nature Alternatives Favourable Unfavourable Market (MU) Market (MU) Construct a large 200.000 plant Construct a 100.000 small plant Do nothing 0 Decision: Do nothing - 180.000 - 20.000 0

Maximum in the row (MU) -180.000 -20.000 0

Maximin!

3. Equally Likely (Laplace) Criterion Average payoff is simply the sum of all payoffs for an alternative, divided by the number of outcomes. We then pick the alternative with the maximum average payoff. The Laplace approach assumes that all probabilities of occurrence for the states of nature are equal, i.e. each state of nature is equally likely. The equally likely choice for Ilhandir Lumber is the second alternative, construct a small plant. This strategy, as shown in the table has the maximum average payoff (40 000 MU) over all alternatives.

Alternatives Favourable Market (MU) Construct a large 200.000 plant Construct a 100.000 small plant Do nothing 0 Decision: Construct a small plant

States of Nature Unfavourable Market (MU) - 180.000 - 20.000 0

Row Average (MU) 10.000 40.000 0

Equally Likely!

4. Criterion of Realism (Hurwicz Criterion) If we assume Ilhan sets his coefficient of realism, , to be 0.80, his best decision would be construct a large plant. As seen in the table, this alternative has the highest weighted average payoff : 124 000 = 0.80(200 000) + 0.20(-180 000).

Alternatives Favourable Market (MU)

States of Nature Unfavourable Criterion of Realism Market (MU) or weighted Average

89 ( = 0.8) 124.000 76.000 0 (MU)

Realism!

Construct a large 200.000 plant Construct a 100.000 small plant Do nothing 0 Decision: Construct a large plant

- 180.000 - 20.000 0

5. Minimax regret Criterion To use this criterion, we need to first develop the opportunity loss table. This is done by determining the opportunity loss of not choosing the best alternative for each state of nature. Opportunity loss for any state of nature, or any column, is calculated by subtracting each outcome in the column from the best outcome in the same column. In Ilhandir Lumber Case, the best outcome in the same column. In Ilhandir Lumber Case, the best outcome for a favourable market is 200 000 MU, as a result of first alternative construct a large plant. Likewise, the best outcome for an unfavourable market is 0 MU, as a result of the third alternative, do nothing. We now subtract all payoffs in that column from 0 MU. Table below shows these computations as Ilhans complete opportunity loss table. Once this table has been constructed, we first locate the maximum opportunity loss within each alternative. We then pick the alternative with the minimum number. Ilhans minimax choice is the second alternative, construct a small plant. As shown in table, the regret is 100 000 is the minimum of the maximum regrets over all alternatives. States of Nature Unfavourable Market (MU)

Alternatives Favourable Market (MU) Construct a large 200.000-200.000=0 plant Construct a 200.000-100.000= small plant 100.000 Do nothing 200.000-0=200.000 Decision: Construct a small plant Illustrative Case 2:

Maximum in the row (MU)

0-(- 180.000)=180.000 180.000 0-(- 20.000)=20.000 0-0=0 100.000 200.000

Minimax!

A TV and DVD manufacturing company is considering several alternative methods of expanding its production to accommodate increasing demand for its products. Company planners indicate that there are only four viable options to the company: i. ii. iii. iv. Expand the present plant, Build a new plant, Subcontract out extra production to TV and DVD manufacturers, Do nothing at all.

90

Having identified all the viable alternatives, the decision maker must now list the future events that may occur. Generally, decision makers can identify most future events than can occur; the difficulty is identifying which particular event will occur. These future events (not under the control of the decision maker) are called, as mentioned before, States of Nature. In this case our TV and DVD manufacturer, the greatest uncertainty attaches to future demand for the product. The future events relating to demand are listed as: 1. High demand (resulting from high product acceptance), 2. Moderate demand (resulting from reasonable product acceptance but heavy competitive response), 3. Low demand (resulting from low product acceptance), 4. Failure (no product acceptance). In defining these states of nature, it would be quite usual for the decision maker to attach a money unit (MU) or unit volume to each of the four possible events (outcomes) to define them more accurately. The decision maker constructs a payoff table, which would result from each possible combination of decision alternative and state of nature. The payoff table below shows the extra payoffs expressed in profits earned over next five years.

Alternatives Expand Build Subcontract Do nothing

High 500.000 700.000 300.000 0

States of Nature (Demand) Moderate Low 250.000 300.000 150.000 0 -250.000 -400.000 -10.000 0

Failure -450.000 -800.000 -100.000 0

In the case of making decisions under uncertainty, the decision maker knows which states of nature can happen, but he does not have information which would allow him to specify the probability that these states will happen. If the decision maker wants to use maximax criterion, he first selects the maximum payoff possible for each decision alternative, then chooses the alternative that provides him with the maximum payoff within this group. The alternative within this group which provides the maximum payoff is Build, with an associated extra payoff over five years 700 000 MU.

Alternatives

High

States of Nature Moderate Low

(Demand) Failure

Maximum

91 (MU) Expand Build 500.000 700.000 (MU) 250.000 300.000 (MU) -250.000 -400.000 -10.000 0 (MU) -450.000 -800.000 -100.000 0 in the row (MU) 500.000 Maximax! 700.000 300.000 0

Subcontract 300.000 150.000 Do nothing 0 0 Decision: Build a new plant

If the decision maker tries to maximize his minimum possible payoffs, he begins by first listing the minimum payoff that is possible for each decision alternative. The decision alternative in this criterion is the alternative do nothing at all.

Alternatives

High (MU)

States of Nature Moderate Low (MU) (MU) 250.000 300.000 150.000 0 -250.000 -400.000 -10.000 0

(Demand) Failure (MU) -450.000 -800.000 -100.000 0

Expand 500.000 Build 700.000 Subcontract 300.000 Do nothing 0 Decision: Do Nothing

Maximum in the row (MU) -450.000 -800.000 -100.000 Maximin! 0

For equally likely criterion the choice is to expand the present plant, with an associated extra payoff over five years 12 500 MU.

Alternatives Expand Build Subcontract

High (MU) 500.000 700.000 300.000

States of Nature (Demand) Moderate Low (MU) (MU) 250.000 300.000 150.000 0 -250.000 -400.000 -10.000 0

Failure (MU) -450.000 -800.000 -100.000 0

Row Average (MU) 12.500 -50.000 85.000 0

Equally Likely!

Do nothing 0 Decision: Subcontract

If we assume the decision maker of the company feels fairly optimistic and assigns a value of 0.7 to . Under these conditions, the measure of realism (Hurwicz Criterion) gives to build a new plant as a decision. The advantage of using this criterion of realism is that the decision maker is able to introduce his own personal feelings of relative optimism or pessimism into the decision process.

States of Nature (Demand)

92 Alternatives Expand Build Subcontract High (MU) 500.000 700.000 300.000 Moderate (MU) 250.000 300.000 150.000 0 Low (MU) -250.000 -400.000 -10.000 0 Failure (MU) -450.000 -800.000 -100.000 0 Maximum in the row (MU) .7(500.000)+.3(450.000) = 215 000 .7(700.000)+.3(800.000) = 250 000 .7(300.000)+.3(-100.000) = 180 000 .7(0)+.3(0) = 0
Realism!

Do nothing 0 Decision: Build

The table below shows the regret associated with all combinations of decision alternative and state of nature. These regret values are obtained by subtracting every entry in the payoff table from the largest entry in its column. Applying the minimax regret criterion requires the decision maker to indicate the maximum regret for each decision alternative. He then chooses the minimum of the regret values. In this case, 350 000 MU is the minimum regret value, and this regret is associated with the decision alternative expand. States of Nature (Demand) Moderate Low (MU) (MU) 50.000 0 150.000 300.000 250.000 400.000 10.000 0

Alternatives Expand

High (MU)

Failure (MU) 450.000 800.000 100.000 0

700.000500.000 =200.000 Build 700.000700.000=0 Subcontract 700.000300.000 =400.000 Do nothing 700.000 Decision: Subcontract Example 1:

Maximum in the row (MU) 450.000 800.000 400.000 700.000

Minimax!

The Cyprus Textile Company is contemplating the future of one of its plants located in Gazimagusa, TRNC. Three alternative decisions are being considered: i. Expand the plant and produce light-weight, durable materials for possible sale to the military, a market with little foreign competition. ii. Maintain the status quo at the plant, continuing production of textile goods that are subject to heavy competition, iii. Sell the plant now. If one of the first two alternatives is chosen, the plant will still be sold at the end of the year. The amount of profit that could be earned by selling the plant in a year depends on foreign market conditions, including the status of a trade embargo threat by EU. The following payoff table describes this decision situation.

93

States of Nature Good Foreign Poor Competitive Foreign Relations (MU) Competitive Alternatives Relations (MU) Expand the plant 800.0000 500.000 Maintain the 1.300.000 -150.000 status quo Sell Now 320.000 320.000 Determine the best decision using each of the decision criteria. a. Maximax b. Maximin c. Equally Likely (Laplace) d. Hurwicz (Realism), assume = 0.3 e. Minimax Regret. Solution: a. Maximax Criterion The company would optimistically assume that good competition conditions will prevail in the future, resulting in the following maximum payoffs and decisions.

States of Nature Good Foreign Poor Competitive Foreign Relations (MU) Competitive Alternatives Relations (MU) Expand the plant 800.0000 500.000 Maintain the 1.300.000 -150.000 status quo Sell Now 320.000 320.000 Decision : Maintain the status quo b. Maximin Criterion

Maximum in the row(MU) 800.000 1.300.000 320.000

Maximax!

The company selects the decision that will reflect the maximum of the minimum payoffs. For each decision alternative, the company assumes that the minimum payoff will occur; of these, the maximum is selected. States of Nature Poor Foreign Competitive

Good Foreign Competitive Relations (MU)

Minimum in the row (MU)

94 Alternatives Expand the plant 800.0000 Maintain the 1.300.000 status quo Sell Now 320.000 Decision: Expand the plant c. Equally Likely Criterion The company assigns equal weight to each alternative, and selects the alternative with the maximum of these weighted values. States of Nature Good Foreign Poor Competitive Foreign Relations (MU) Competitive Alternatives Relations (MU) Expand the plant 800.0000 500.000 Maintain the 1.300.000 -150.000 status quo Sell Now 320.000 320.000 Decision: Expand the plant d. Hurwicz (Realism) Criterion For each decision alternative, the maximum payoff is multiplied by , and the minimum payoff is multiplied by 1-. If = 0.3, i.e. the company is slightly optimistic, and 1- = 0.7; the following decision will result. States of Nature Good Foreign Poor Competitive Foreign Relations (MU) Competitive Relations (MU) 500.000 -150.000 320.000 Row Average (MU) 650.000 575.000 320.000
Equally Likely!

Relations (MU) 500.000 -150.000 320.000

500.000 -150.000 320.000

Maximin!

Alternatives

Criterion of Realism or Weighted Average = 0.3 (MU) 590.000 285.000 320.000

Realism!

Expand the plant 800.0000 Maintain the 1.300.000 status quo Sell Now 320.000 Decision: Expand the plant

e. Minimax Regret Criterion

95 The company attempts to avoid regret by selecting the decision alternative that minimizes the maximum regret. The company first selects the maximum payoff under each state of nature; then all other payoffs under the respective states of nature are subtracted from these amounts. These values represent the regret for each decision that would be experienced by the decision maker if a decision were made that resulted in less than the maximum payoff. The maximum regret for each decision must be determined, and the decision corresponding to the minimum of these regret value is selected. States of Nature Good Foreign Poor Competitive Foreign Relations (MU) Competitive Alternatives Relations Expand the plant 1.300.000500.000800.000=500.000 500.000=0 Maintain the 1.300.000500.000-(status quo 1.300.000=0 150.000=650.000 Sell Now 1.300.000500.000320.000=980.000 320.000=180.000
Decision : Expand the plant

Maximum in the row (MU) 500.000 650.000 980.000

Minimax!

The decision to expand the plant was designated most often by four of the five decision criteria. The decision to sell was never indicated by any criterion. This is because the payoffs for expansion are always better than the payoffs for selling. The alternative of selling is said to be dominated by the alternative of expanding. The dominated decision alternatives can be removed from the payoff table to reduce the complexity of the decision analysis. Example 2: Ali Caliskan and Ayse Mutlu have known each other since high school. Two years ago they entered the same university and today they are taking undergraduate courses in the Faculty of Business. Both hope to graduate with degrees in finance. In an attempt to make extra money and to use some of the knowledge gained from their business courses, Ali and Ayse have decided to look into the possibility of starting a small company that would provide word processing services to students who needed term papers or other reports prepared in a professional manner. Using a systems approach, Ali and Ayse have identified thre strategies. Strategy 1 is to invest in a fairly expensive microcomputer system with high-quality laser printer. In a favourable market, they should be able to obtain a net profit of 10 000 MU over the next two years. If the market is unfavourable, they can lose 8 000 MU. Strategy 2 is to purchase a less expensive system. With a favourable market, they could get a return during the next two years of 8 000 MU. With this unfavourable market, they would incur a loss of 4 000 MU. Their final strategy, Strategy 3, is to do nothing. Ali is basically a risk taker, while Ayse tries to avoid risk.

96 a. What type of decision procedure should Ali use? What would Alis decision be? b. What type of decision maker is Ayse? What decision would Ayse use? c. If Ali and Ayse were indifferent to risk, what type of decision approach should they use? What would you recommend if this were the case? Solution: The problem is one of decision making under uncertainty. Before answering the specific questions, a decision table should be developed showing the alternatives, states of nature, and related consequences.
States of Nature Maximum in Minimum in Row Favourable Unfavourable the row the row Average Market (MU) Market (MU) (MU) (MU) (MU) _____ 10 000 -8 000 10 000 Max! -8 000 1 000 8 000 -4 000 8 000 -4 000 2 000 Max! 0 0 0 0 Max! 0 _____

Alternatives Strategy 1 Strategy 2 Strategy 3

a. Since Ali is a risk taker, he should use the maximax decision criteria. This approach selects the row that has the highest or maximum value. The 10 000 MU value, which is the maximum value from the table, is in row 1. Thus Alis decision is to select Strategy 1, which is an optimistic decision approach. b. Ayse should use the maximin criteria, because she wants to avoid risk. The minimum or worst outcome for each row, or strategy, is identified. The maximum of these values is selected. Thus, Ayse would select Strategy 3, which reflects a pessimistic approach. c. If Ali and Ayse are indifferent to risk, they could use the equally likely approach. This approach selects the alternative that maximizes the row averages. Thus, using the equally likely approach, the decision is to select Strategy 2, which maximizes the row averages. DECISION MAKING UNDER RISK In many real-world situations, it is common for the decision maker to have some idea about the probabilities of occurrence of the different outcomes or states of nature. These probabilities may be based on the decision makers personal opinions about future events, or on data obtained from market surveys, expert opinions, and so on. As noted previously, when the probability of occurrence of each state of nature can be assessed, the problem environment is called Decision Making under Risk. The most widely used decision making criteria under risk is Expected Value. The expected value (or expected monetary value) is a weighted average of decision outcomes in which each future state of nature is assigned a probability of occurrence.

n EV(x)= P(x).xi i=1

97

Where

xi= outcome i
P(x)= probability of outcome i

Illustrative Case 1 (Continued): If you recall the Ilhans case and assume he believes the probability of a favourable market is the same as the probability of an unfavourable market. That is, each state of nature has a 0.50 probability. Which alternative would give him the greatest EV? To determine this, Ilhan has computed the EV for each alternative, as shown in the table below. The largest EV of $40 000 results from the second alternative, construct a small plant". State of Nature Favourabl Unfavourable EV Computed (MU) Alternatives e Market (MU) Market (MU) Construct a 200 000 -180 000 0.5(200.000)+0.5(180.000)=10.000 large plant Construct a 100 000 -20 000 0.5(100.000)+0.5(-20.000)=40.000 small plant Do nothing 0 0 0.5(0)+0.5(0)=0 Probabilities 0.5 0.5 Decision: Construct a small plant

Maximum!

Observe that the EV represents the long run average payoff, although the actual payoff from a decision will be one of the payoffs listed in the decision table. EV is, however, an acceptable criterion for many business decisions since companies make similar decisions on are repeated basis over time. Example 1 (Continued): Assume that it is now possible for the Cyprus Textile Co to estimate a probability of 0.70 that good foreign competitive conditions will exist and a probability of 0.30 that poor foreign conditions will exist in the future. Determine the best decision using expected value. Solution: The expected value for each decision alternative is computed as follows:

Alternatives

Good Foreign Competitive Conditions

States of Nature Poor Foreign Competitive Conditions

EV Computed (MU)

98 (MU) 800.000 1.300.000 320.000 (MU) 500.000 -150.000 320.000 0.3

Expand the plant Maintain the status quo Sell now

0.7(800.000)+0.3(500.000)= 710.000 0.7(1.300.000)+0.5(20.000) Maximum! =865.000 0.7(320.000)+0.3(320.000)= 320.000

Probabilities 0.7 Decision: Maintain status quo

Expected Value of Perfect Information Occasionally additional information is available, or can be purchased, regarding future events, enabling the decision maker to make a better decision. e.g. A company could hire an economic forecaster to determine more accurately the economic conditions that will occur in the future. However, it would be foolish to pay more for this information than it stands to gain in extra profit from having the information. The information has some maximum value that is the limit of what the decision maker would be willing to spend. To compute the expected value of perfect information, first look at the decisions under each state of nature. If information that assured us which state of nature was going to occur (i.e. perfect information) could be obtained, the best decision for that state of nature could be selected. For example in the textile company example (Example1), if the company executives knew for sure that good competitive conditions would prevail, they would maintain the status quo. If they knew for sure that poor competitive conditions would occur, then they expand. Example 1 (Continued): The probabilities of each state of nature (i.e. 0.70 and 0.30) indicate that good competitive conditions will prevail 70% of the time and poor competitive conditions will prevail 30% of the time (if this decision situation is repeated many times). In other words, even though perfect information enables the investor to make the right decision, each state of nature will occur only a certain portion of the time. Thus, each of the decision outcomes obtained using perfect information must be weighted by its respective probability; 0.70(1.300.000 MU)+0.30(500.000 MU)= 1.060.000 MU 1.060.000 MU is the Expected Value of the decision given perfect information. The expected value of perfect information is the maximum amount that would be paid to gain information that would result in a decision better than the one made without perfect information. Recall that the EV decision without perfect information was to maintain the status quo and the EV= 865 000 MU. EVPI= Expected Value Given Perfect Information Expected Value without Perfect Information

99

EVPI= 1.060.000MU- 865.000 MU = 195.000 MU The EVPI, 195 000 MU, is the maximum amount that the investor would pay to purchase perfect information from some other source, such as an economic forecaster. Of course, Perfect Information is rare and is usually unobtainable. Typically, the decision maker would be willing to pay some smaller amount, depending on how accurate (i.e. close to perfection) the information believed to be. Illustrative Case 1 (Continued): Ilhan Bal has been approached by Scientific Marketing, Inc., a market research firm, with a proposal to help him make the right decision regarding the plant. Scientific claims that its analysis will tell Ilhan with certainty whether the market will be favourable for his propped product. In other words, it will change Ilhan's problem environment from one of decision making under risk to one of decision making under certainty. This information could prevent Ilhan making an expensive mistake. Scientific would charge Ilhandirs Lumber 65 000 MU for the information. What should Ilhan do? Is the information worth 65 000 MU? Although such perfect information is usually not available in practice, determining its value can be very useful. It places an upper bound on what we should be willing to spend on any information, perfect or otherwise. EVPI = EVwPI - Max EV From the table and with the assumption that probability of each state of nature is 0.5, we obtain: EVwPI = 0.5 (200 000 MU) + 0.50 (0 MU) = 100 000 MU That is, if we had perfect information, we would expect an average payoff of 100 000 if the decision could be repeated many times. The maximum EV without perfect information is 40.000 MU. Hence; EVPI = EVwPI - EV = 100 000 MU - 40 000 MU = 60 000 MU Thus, the most John should pay for the perfect information is 60 000 MU. This, of course, is again based on the assumption that the probability of each state of nature is 0.50. Example 3: Ayse Mutlu is considering the possibility of opening a small dress shop on Salamis Road, a few blocks from the University. She has located a good mall that attracts students. Her options are to open a small shop, a medium-sized shop, or no shop at all. The market for a dress shop can be good, average, or bad. The probabilities for these three possibilities are 0.2 for a good market, 0.5 for an average market and 0.3 for a bad market. The net profit or loss for the medium-sized and small shops for the various market conditions is given in the following table. Building no shop at all yields no loss or no gain. What do you recommend?

100

Alternative Small shop Medium-sized shop No shop Solution:

States of Nature Good Market(MU) Average Market(MU) 75.000 25.000 100.000 35.000 0 0

Bad Market(MU) -40.000 -60.000 0

Since the decision-making environment is risk (probabilities are known), it is appropriate to EV criterion. The problem can be solved by developing a payoff table that contains all alternatives, states of nature, and probability values. The EV for each alternative is also computed, as in the following table:

Alternative

Small shop Medium-sized shop No shop 0 0 Prob. 0.20 0.50 Decision: Build the medium sized shop

Good Market(MU) 75.000 100.000

States of Nature Average Market(MU) 25.000 35.000

Bad Market(MU) -40.000 -60.000 0 0.30

EV (MU) 15.500 19.500 0

Maximum!

EV (Small Shop) = 0.2(75.000MU) + 0.5(25.000MU) + 0.3(-40.000MU) = 15.500MU EV (Medium Size Shop) =0.2(100.000MU) + 0.5(35.000MU) + 0.3(-60.000MU) =19.500MU EV (No Shop) = 0.2(0MU) + 0.5(0MU) + 0.3(0MU) = 0MU As can be seen, the best decision is to build the medium-sized shop. The EV for this alternative is the maximum, 19500 MU. MARGINAL ANALYSIS WITH A LARGE NUMBER OF ALTERNATIVES AND STATES OF NATURE So far, we have considered cases in which there are only a few alternatives and states of nature. What happens when we have a large number of alternatives? e.g. a large restaurant might be able to stock from 0 to 100 cartons of doughnuts. Furthermore, demand could also range from 0 to 100 cartons per day. In this case we have to analyze 101 possible alternatives and states of nature. This would require a very large decision table. When we can identify a marginal profit and loss, it is possible to use marginal analysis to obtain the best decision without using a large decision table. Marginal Analysis is a decision making approach that can help select the optimal inventory level. It involves two new terms: marginal profit and marginal loss. e.g. suppose a daily paper costs 0.19 MU and can be sold for 0.35 MU per issue. But if a paper is not sold at

101 the end of the day, it is completely worthless (a 0 MU salvage value). In this case, the marginal profit (MP) is the profit made by selling each additional paper, namely 0.16 MU/issue. (= 0.35MU/issue 0.19 MU/issue). The marginal loss (ML), is the loss caused by stocking, but not selling, each additional paper it would be 0.19 MU for every paper remaining at the end of the day. When there are a manageable number of alternatives and states of nature and we know the probabilities for each state of nature, marginal analysis with discrete distributions can be used. When there are a very large number of possible alternatives and states of nature and the probability distributions of the states of nature can be described with a normal distribution, marginal analysis with the normal distribution is appropriate. Marginal Analysis with Discrete Distributions Find the best inventory level to stock is not difficult when we follow the marginal analysis procedure. Given any inventory level, we would only add another unit to our inventory level if its expected marginal profit equals or exceeds it expected marginal loss. This relationship is expressed symbolically as follows. First, we get P = Probability that demand will be greater than or equal to a given supply (or the probability of selling at least one additional unit). 1-P = Probability of demand will be less than supply. The expected MP is then found by multiplying the probability that a given unit will be sold by the marginal profit, P (MP). Similarly, the expected ML is the probability of not selling the unit multiplied by the ML, or (1-P)(ML). The optimal decision rule is to stock the additional unit, if P(MP) (1-P)(ML) P(MP) ML - P(ML) P(MP) + P(ML) ML P __ML__ MP+ML In other words, as long as the probability of selling one more unit (P) _ML__ , we MP+ML would stock the additional unit. Steps of Marginal Analysis with Discrete Distributions 1. Determine the value of ML/MP+ML for the problem. 2. Construct a probability table and add a cumulative probability column. 3. Keep ordering inventory as long as the probability (P) of selling at least one additional unit is greater than ML/ML+MP Illustrative Case 3:

102 Pleasure Caf is a popular Izmir dining spot on the corner of the Lozan Meydani. Its specialty is coffee and doughnuts; it buys the doughnuts fresh daily from a large industrial bakery. The Caf pays 4 MU for each carton (containing two dozen doughnuts) delivered each morning. Any cartons not sold at the end of the day are thrown away, for they would not be fresh enough to meet the cafs standards. If a carton of doughnuts is sold, the total revenue is 6 MU. Hence, the marginal profit per carton of doughnuts is MP = 6 MU/carton 4 MU/carton = 2 MU/carton The marginal loss is ML = 4 MU/carton 0 MU/carton = 4 MU/carton From the past sales, the cafs manager estimates that the daily demand will follow the probability distribution shown in the following table. The manager then follows three steps to find the optimal number of cartons of doughnuts to order each day. Daily Sales Probability (Cartons of of Demand will be at Doughnuts) this level 4 0.05 5 0.15 6 0.15 7 0.20 8 0.25 9 0.10 10 0.10 Step1 Determine the value of __ML__ ML+MP for the decision rule P __ML__ = _4 _ = 0.66 ML+MP 4+2 P 0.66 Step2 add a new column to the table to reflect the table to reflect the prob. That doughnuts sale will be at each level on greater.

Daily Sales Probability of Demand Probability that (Cartons of will be at this level Demand will be at this

103 Doughnuts) 4 5 6 7 8 9 10 level or greater 1.00 0.95 0.80

CP (0.66)

0.05 0.15 0.15 0.20 0.25 0.10 0.10

0.65 0.45 0.20 0.10

Step3 Keep ordering additional cartons as long as the probability of selling at least one additional carton is greater than P, which is the indifference probability. If Pleasure Caf orders 6 cartons, MP will still be greater than ML. P at 6 cartons (0.80) 0.66 The optimal decision is the order 6 cartons each day. Example 4: Ali Gezer is the manager of Izmir Cheese, which produces cheese spreads and other cheese related products. E-Z spread cheese is a product that has always been popular, especially by children. The probability of sales, in cases, is as follows: Demand In Cases 10 11 12 13 14 Probability 0.2 0.3 0.2 0.2 0.1

A case of E-Z spread cheese sells for 100 MU and has a cost of 75 MU. Any cheese that is not sold by the end of the week is sold to a local food processor for 50 MU. Ali never sells cheese that is more than a week old. How many cases of E-Z Spread Cheese should Ali produce each week to maximize average profit? Solution: Demand 10 11 12 13 14 Prob. Of Demand Cum. Prob. 0.2 1.0 0.3 0.8 0.2 0.5 CP(0.5) 0.2 0.3 0.1 0.1 1.0 ML= Cost/ case-Salvage Value/case= 75 MU/case-50 MU/case= 25 MU/case MP= Price/ case-Cost /case= 100 MU/case-75 MU/case= 25 MU/case

104 P __ML__ = __25_ ML+MP 25+25 = 0.5

Ali should produce 12 cases of E-Z spread each week because it has a cumulative probability (0.5) = critical probability (0.5). Marginal Analysis with the Normal Distribution When the product demand or sales follow a normal distribution, which is a common business situation, marginal analysis with the normal distribution can be applied. First we need to find four values. 1. The average or mean sales for the product, 2. The standard deviation of sales, 3. The marginal profit for the product, MP 4. The marginal loss for the product, ML Once these quantities are known, the process of finding the best stocking policy is somewhat similar to marginal analysis with discrete distributions. We let I opt=optimal stocking level. Step of Marginal Analysis with Normal Distribution: 1. Determine the value of P __ML__ ML+MP 2. Find the probability (1-P), which denotes the service level 3. Find Iopt using the relationship Z = Iopt - To solve for the optimal stocking policy: Iopt = + Z

Illustrative Case 4 a) Demand for copies of the Cyprus Tribune newspaper at Necos Newstand is normally distributed and has averaged 50 papers per day, with a standard deviation of 10 papers. With a ML of 0.04 MU and a MP of 0.06 MU, what daily stocking policy should Neco follow? Step 1: Neco should stock Tribunes until the probability of having a demand at a given level or greater is at least P __ML ML+MP = ___0.04 = 0.4 0.04+0.06

Z=0.25 Step 2: 1-P = 1.00 - 0.40 = 0.60. 0.60 is the probability of being serviced, i.e service level probability. 0.10 0.4 Step 3: 0.50 0.60
=50 Iop.

Mean daily sales

105 We look at the normal distribution for 0.50 or 0.10 (0.50+0.10), since the normal table has cumulative areas under the curve either between the left side and any point or from mean value and end point in order to get the corresponding z value. Z = 0.25 standard deviation from the mean.

Hence, Iopt = + Z = 0.50 = 0.25(10) = 52.5 or 52 newspapers Thus, Neco should order 52 Cyprus Tribune. b) This same procedure can be used when P is greater than 0.50. Lets say Necos Newstand also stocks the Cyprus Times and its ML is 0.08 MU and MP is 0.02 MU. The daily sales have averaged 100 Cyprus Tribunes, with a standard deviation of 10 papers. The optimal stocking policy is as follows: P __ML ML+MP Z=0.84 0.30 SL=0.2 0.80
=100

= __0.08 = 0.8 0.08+0.02

0.50

SL = 1.00 - 0.80 = 0.20 From normal curve table we find Z for an area under the curve of 0.30 is -0.84. Therefore Z = -0.84 standard deviation from the mean for an area Iopt = 100 0.84(10) = 91.6 or 91 papers So Neco should order 91 papers every day Example 5: The marginal loss on Cyprus Reds, a brand of grape fruits from Guzelyurt, TRNC, is 35 MU/case. The marginal profit is 15 MU per case. During the past year, the mean sales of Cyrpus Reds in cases were 45.000 cases, and the standard deviation was 4450. How many cases of Cyrpus Reds should be brought to market? Assume that sales follow a normal distribution.

Z=-0.52 Solution:

106 0.50 0.70 =45000

SL=0.30 P __ML ML+MP = __35 = 0.70 35+15

0.20

SL = 1.00 - 0.70 = 0.30 From normal curve table we find Z =-0.52 Z = -0.52 standard deviation from the mean for an area Iopt = 45.000 0.52(4450) = 42 686 cases Result: Bring 42.686 cases to market. Example 6: Ali Durmaz, a college senior, sells cherry tomatoes every spring. Ali buys them for 9 MU a crate and resells them for 16 MU per crate. If a crate is not sold on the first selling day, it is worth 3 MU as salvage. Alis examination of past sales records indicates that demand for this particular item is normally distributed, with a mean of 120 crates daily and a standard deviation of 38 crates. What should Alis stock be? Solution: Ali can use equation P __ML___ ML+MP to calculate the minimum required probability of selling at least an additional unit to justify his stocking of that unit with data he has available. ML= Cost/ unit - Salvage Value/unit = 9 MU/crate - 3 MU/ crate = 6 MU /crate MP= Price/ unit- Cost/unit = 16 MU/crate - 9 MU/ crate = 7MU /crate P __ML = __6 = 0.462 ML+MP 6+7 This value is the critical probability or stock-out risk, Ali must be 0.462 sure of selling at least an additional unit before it would pay him to stock that unit. Z=0.1 0.038 0.462
Iopt

0.50

0.538

=120 crates/daily =38 crates/daily Example 7:

The shaded area is 1-P in other words that is the Service Level. For 0.538, we see that the appropriate z value is 0.1. This means that point Iopt is 0.1 standard deviation to the right of the mean. Thus Iopt is found as follows: Iopt = + Z Iopt = 120+0.1(38) = 124 crates The optimum stock for Ali to order is 124 crates.

107

The Magusa Livestock Company receives order for an average of 6000 dozen quail eggs a week. The standard deviation of weekly orders is 425 dozen. The eggs cost 7 MU/dozen and are resold for 10 MU/dozen. If the eggs are not shipped within a week, their fertility is impaired and Maguss cannot sell them as first quality; they can however be sold for 1MU/dozen. Calculate Magusas optimum weekly order of eggs. Solution: ML = Cost/ dozen- Salvage Value/dozen = 7 MU/dozen - 1 MU/ dozen = 6 MU /dozen MP = Price/ dozen- Cost/dozen = 10 MU/dozen - 7 MU/ dozen = 3MU /dozen P __ML = __6 ML+MP 6+3 = 0.67 Service Level = 1-P = 1.00 - 0.67 = 0.33. From normal distribution we find Z= -0.44.

Z=-0.44 0.17 0.33 0.67

0.50
.

Iopt =6000 dozen =425 dozen

Iopt = + Z = 6000+(-0.44)(425) = 5.813 dozens It is recommended to order 5.813 dozen of quail eggs each week. DECISION TREES In the preceding examples, each problem involved the selection of a course of action at a specified point in time. Although any problem that can be presented in a decision table can also be graphically illustrated, in a decision tree. In addition to this, sometimes a problem may involve a series of decisions, each dependent on the one preceding it. In these instances a single-stage approach of the type previously demonstrated must be abandoned in favour of one that is appropriate for multistage problems. One relatively simple device for examining successive alternatives is the decision tree. A tree diagram is a graphic construction that represents the various alternatives in a decision problem as if they were branches of a tree. Each main branch of the tree represents an alternative course of action. The decision tree is simply a probability tree with the payoffs for the chosen alternatives identified at the end of each branch sequence. It is therefore similar to the payoff tables described and demonstrated earlier. However, the decision tree is not normally used when opportunity losses are involved. Instead, the basic decision criteria are those of maximizing expected gain and/or minimizing expected losses.

108

All decision trees are similar in that they contain decision points or nodes and state of nature points or nodes: A decision node from which one of several alternatives may be chosen O A state-of-nature node out of which one state of nature will occur. Analyzing problems with decision trees involves five steps: 1. Define the problem 2. Structure or draw the decisions tree. 3. Assign probabilities to the states of nature 4. Estimate payoffs for each possible combination of alternatives and states of nature 5. Solve the problem by computing EV s for each state of nature node. This is done by working backward, i.e. starting at the right of the tree and working back to decision nodes on the left. Also, at each decision node, the alternative with the best EV is selected. Illustrative Case I (Continued): Lets take another look at the Ilhandir Lumber Company case. You may recall that Ilhan was trying to decide whether to expand his penetration by building a new plant to produce storage shed. A simple decision tree to represent Ilhans decision is shown in Figure1. Note that the tree represents the decision and outcomes in sequential order. First, Ilhan decides whether to construct a large plant, a small plant or no plant. Then, once that decision is made, the possible states of nature as outcomes (favourable or unfavourable market) will occur. A State of Nature Node A Decision Node 1 Favorable Market Unfavorable Market Favorable Market
2

Unfavorable Market

Fig. 1: Ihans decision tree. A completed and solved decision tree for Ilhandir Lumber is presented in Fig. 2. Note that the payoffs are placed at the right side of each of the trees branches. The probabilities are placed in parentheses next to each state of nature. The EVs for each state-of-nature node are then calculated and placed by their respective nodes. The EV of the first node is 10 000 MU. This represents the branch from the decision node to construct a large plant. The EV for

109 node 2, to construct a small plant, is 40 000 MU. Building no plant or doing nothing has, of course, a payoff of 0 MU. The branch leaving the decision node leading to the state-of-nature node with the highest EV should be chosen. In Ilhandirs case, a small plant should be built.
EV=10,000MU Max EV=40,000MU

Favorable Market (0.5) 1 Unfavorable Market (0.5) Favorable Market (0.5)

2

200,000 MU -180,000 MU 100,000 MU -20,000 MU

Unfavorable Market (0.5)

EV=40000MU

0 MU Fig. 2: Completed and Solved Decision Tree for Ilhandir Lumber Co. When a sequence of decisions needs to be made, decision trees are much powerful tools than decision tables. All possible outcomes and alternatives must be included in their logical sequence. This is one of the strengths of using decision trees in making decisions. The user is forced to make decisions in a logical, sequential manner. Illustrative Case 1 (Continued): Lets say that Ilhan has two decisions to make, with the second decision dependent on the outcome of the first. Ilhan has the option of conducting his own marketing research survey at a cost of 10 000 MU. The information form his survey could help him decide whether to construct a large plant, a small plant, or not to build at all. Ilhan recognizes that such a market survey will not provide him with perfect information, but it may help quite a bit nevertheless.

Fig.3: Larger Decision Tree with Payoffs and Probabilities for Ilhandir Lumber (below).

110 Payoffs Second Decision Point

EV=106,400 MU

First Decision Point

Favorable Market (0.78) 190,000

Unfavorable Market (0.22) -190,000 Favorable Market (0.78) 90,000 EV=63,600 MU

3
EV=106,400 MU

Unfavorable Market (0.22)

-30,000 -10,000

1
EV=49,200 MU EV=-87,400 MU

Favorable Market (0.27) Unfavorable Market (0.73) Favorable Market (0.27)

190,000 -190,000 90,000 -30,000 -10,000

EV=2,400 MU

EV=2,400 MU

Unfavorable Market (0.73)

EV=10,000 MU

Favorable Market (0.50)

200,000

Unfavorable Market (0.50) -180,000 Favorable Market (0.50) 100,000

7

EV=40,000 MU

EV=40,000 MU

Unfavorable Market (0.50) -20,000 0

111

112

Examining the tree, we see that Ilhans first decision point is whether to conduct the 10 000 MU market survey. If he chooses not to do the study (the lower part of the tree), he can either construct a large plant, a small plant or no plant. This is Ilhans second decision point. The market will either be favourable (0.50 probability) or unfavourable (also 0.50 probability) if he builds. The upper part of the figure reflects the decision to conduct the market survey. State of nature, node 1, has two branches. There is a 45% chance that the survey results will indicate a favourable market for storage sheds. We also note that the probability is 0.55 that the survey results will be negative. (Here we assume that Ilhans experience provides the probabilities and accept them as reasonable). The rest of the possibilities of the probabilities shown in parentheses are all conditional probabilities e.g. 0.78 is the probability of a favourable market for the sheds given a favourable result from the market survey. Of course, you would expect to find a high probability of a favourable market given that the research indicated that the market was good. Dont forget, though, there is a chance that Ilhans 10 000 MU market survey didnt result in perfect or even reliable information. Any market research study is subject to error. In this case, there is a 22% chance that the market for sheds will be unfavourable given that the survey results are positive. We note that there is 27% chance that the market for sheds be favourable given that Ilhans survey results are negative. The probability is much higher, 0.73, that the market will actually be unfavourable given that the survey was negative. Finally, when we look to the payoff column in Fig.3, we see that 10 000 MU, cost of marketing study, had to be subtracted from each of the top 10 tree branches. With all probabilities and payoffs specified, we can start calculating the expected value at each state of nature node. We begin at the end or right side of the decision tree and work back toward the origin. When we finish the best deciding will be known. 1. Given favourable survey results, EVnode 2 = 0.78(190.000) + 0.22(-190.000) = 106.400 MU MAXIMUM! EVnode 3 = 0.78(90.000) + 0.22(-30.000) = 63.600 MU EVno plant =10.000 MU If the survey results are favourable ` a large plant` should be built. 2. Given a negative survey results, EVnode 4 = 0.27(190.000) + 0.73(-190.000) = -87.400 MU EVnode 5 = 0.27(90.000) + 0.73(-30.000) = 2.400 MU MAXIMUM! EVno plant =-10.000 MU Thus, a given a negative survey results, Ilhan should build a small plant with an expected value of 2.400 MU. 3. Continuing on the part of the tree and moving backward, we compute EV of conducting the market survey.

113 EVnode 1 = 0.45(106.400) + 0.55(2.4000) = 49.200 MU 4. If the market survey is not conducted. EVnode 6 = 0.50(200.000) + 0.50(-180.000) = 10.000 MU EVnode 7 = 0.50(100.000) + 0.50(-20.000) = 40.000 MU EVno plant = 0 MU Thus, building `a small plant` is the best choice, given that the marketing research is not performed. 5. We move back to the first decision node and choose the expected value of conducting a survey is 49 200 MU, versus an EV of 40 000 MU for not conducting the study. The best choice is to seek marketing information. If the survey results are favourable, Ilhan should construct a large plant; but if the research is negative, Ilhan should construct a small plant. Expected Value of Sample Information With the market survey, the decision maker for example knows what will be his best decision under positive or negative survey results. But the decision maker must also realize that conducting a market research is not free. He must know that the actual value of doing a survey is. One way of measuring the value of market information is to compute the expected value of sample information (EVSI) EVSI = (EV with sample information + cost) (EV without sample information)

Ilhans new decision tree is represented in Fig.3. Lets take a careful look at this more complex tree. Note that all possible outcomes and alternatives are included in their logical sequence. This is one of the strengths of using decision trees in making decisions. Illustrative Case I (Continued): With the market research Ilhan knows that his best decision will be to build a large plant if the survey is favourable, or a small plant if the survey results are negative. But as mentioned above, Ilhan also realizes that conducting the market research is not free. He would like to know what the actual value of doing such a market research is. In Ilhans case, his EV would be 59260 MU if he hadnt already subtracted the 10 000 MU study cost from each payoff. But EV of not gathering sample information is 40 000 MU. Thus, EVSI = (49200 MU +10 000 MU) 40 000 MU = 19 200 MU This means that Ilhan could have paid up to 19 200 MU for a market research and still come out ahead. Since it costs only 10 00 MU, the survey is indeed worthwhile. Example 8:

114 Bora Deniz has enjoyed sailing small boats since he was 10 years old, when his father started sailing with him. Today, Bora is considering the possibility of starting a company to produce small boats for the recreational market. Unlike other mass-produced sailboats, however, these boats will be made specially for children between he ages of 10 and 15. The boats will be of the highest quality and extremely stable and the sail size will be reduced to prevent problems of capsizing. Because of the expense involved in developing the initial molds and acquiring the necessary equipment to produce fibreglass sailboats for young people children, Bora has decided to conduct a pilot study to make sure that the market for the sailboats will be adequate. He estimates that the pilot study will cost him 10 000 MU. Furthermore, the pilot study can be either successful or not successful. His basic decision is whether to build a large manufacturing facility, a small manufacturing facility, or no facility at all. With a favourable market, Bora can expect to make 90 000 MU from the large facility or 60 000 MU from the smaller facility. If the market is unfavourable, however Bora estimates that he would lose 30 000 MU with a large facility, and he would lose only 20 000 MU with the small facility. Bora estimates that the probability of a favourable market given a successful pilot study is 0.8. The probability of an unfavourable market given an unsuccessful pilot study result is estimated to be 0.9. Bora feels that there is a 50 50 chance that the pilot study will be successful. Of course, Bora could bypass the pilot study and simply make the decision as to whether to build a large plant, small plant, or no facility at all. Without doing any testing in a pilot study, he estimates that the probability of a successful market is 0.6. What do you recommend? Solution: Before Bora starts to solve this problem, he should develop a decision tree that shows all alternatives, states of nature, probability values, and economic consequences. This decision tree is shown in Fig.4

Fg.4: Boras Decision Tree, Listing Alternatives, States of Nature, Probability Values, and Financial Outcomes (below).

115 Payoffs (MU) Favorable Market (0.8) 2 Unfavorable Market (0.2) Favorable Market (0.8)
3

80,000 -40,000 50,000

Unfavorable Market (0.2)

-30,000 -10,000

1 Favorable Market (0.1) 4 Unfavorable Market (0.9) Favorable Market (0.1)

5

80,000 -40,000 50,000 -30,000 -10,000

Unfavorable Market (0.9)

Favorable Market (0.6) 6 Unfavorable Market (0.4) Favorable Market (0.6)

7

90,000 -30,000 60,000

Unfavorable Market (0.4) -20,000 0

116

115 Payoffs (MU)

EV=56,000 MU

Favorable Market (0.8) Unfavorable Market (0.2) Favorable Market (0.8) Unfavorable Market (0.2)

80,000 -40,000 50,000 -30,000 -10,000

2
EV=34,000 MU

3
EV=23,000 MU EV=56,000 MU

1
EV=-28,000 MU

Favorable Market (0.1) Unfavorable Market (0.9) Favorable Market (0.1) Unfavorable Market (0.9)

80,000 -40,000 50,000 -30,000 -10,000

4
EV=42,000 MU EV=-10,000 MU EV=-22,000 MU

EV=42,000 MU

Favorable Market (0.6) Unfavorable Market (0.4) Favorable Market (0.6)

90,000 -30,000 60,000

EV=42,000 MU

EV=28,000 MU

Unfavorable Market (0.4) -20,000 0

116

117 Once the decision tree has been developed, Bora can solve the problem by computing EV s starting at the endpoints of the decision tree. The final solution is shown on the revised decision tree, Fig. 5. The optimal solution is to not conduct the study but to construct the large plant directly. The expected value (EV) is 42 000 MU Calculations: 1) EVnode 2 = 0.8(80.000) + 0.2(-40.000) = 56.000MU EVnode 3 = 0.8(50.000) + 0.2(-30.000) = 34.000MU EVno shop = 10.000MU EVnode 2 = 56.000 MU is selected for the decision node, that is to build large shop. 2) EVnode 4 = 0.1(80.000) + 0.9(-40.000) = -28.000MU EVnode 5 = 0.1(90.000) + 0.73(-30.000) = -22.000MU EVno shop = -10.000MU EVno shop= -10.000MU is selected for decision node, that is no shop. 3) EVnode 1= 0.5(56.000) + 0.5(-10.000) = 23.000MU 4) EVnode 6 = 0.6(90.000) + 0.4(-30.000) = 42.000MU EVnode 7 = 0.6(60.000) + 0.4(-20.000) = 28.000MU EVno shop = 0 MU The maximum payoff (42.000MU) is chosen, i.e to construct large shop 5) We compare EV= 23.000 MU of conducting study with EV= 42.000 MU of not conducting study and the maximum EV= 42.000 MU is chosen. As a result, Bora should not conduct any study but to construct a large shop directly. Example 9: Mehmet Kaplan is thinking about opening a bike shop in his home town, Gazimagusa. Mehmet loves to take his own bike on 50 km trips to Karpaz with his friends, but he believes that any small business should be started only if there is a good chance of making a profit. Mehmet can open a small shop, a large shop, or no shop at all. Because there will be a five year lease on the building that Mehmet is thinking about using, he wants to make sure that he makes the correct decision. Mehmet is also thinking about hiring his marketing professor, Dr. Mustafa Tumer, to conduct a marketing research study. If the study is conducted, the results could be either favourable or unfavourable. Mehmet Kaplan has done some analysis about the profitability of the bike shop. If Mehmet builds a large bike shop, he will earn 60 000 MU if the market is favourable, but he will loose 40 000 MU if the market is unfavourable. The small shop will return a 30 000 MU profit in a favourable market and a 10 000 MU loss in an unfavourable market. At the present time, he believes that there is a 50 50 chance that the market will be favourable. His marketing professor will charge him 5 000 MU for the marketing research. It is estimated that there is a 0.6 probability that the survey will be favourable. Furthermore, there is a 0.9 probability that the market will be favourable given a favourable outcome from the study.

118 However, Dr. Tumer warned Mehmet that there is only a probability of 0.12 of a favourable market if the marketing research results are not favourable. Mehmet is confused. What should he do? Solution: EVnode 2 = 0.9(55.000) + 0.1(-45.000) = 45.000 MU EVnode 3 = 0.9(25.000) + 0.1(-15.000) = 21.000 MU EVnode 4 = 0.12(55.000) + 0.88(-15.000) = -10.000 MU EVnode 6 = 0.5(60.000) + 0.5(-40.000) = 10.000 MU EVnode 7 = 0.5(30.000) + 0.5(-10.000) = 10.000 MU
EVnode 1

= 0.6(45.000) + 0.4(-5.000) = 25.000 MU

Since EV (market survey) > EV (no survey), Mehmet should conduct the survey. Since EV (large shop | favourable survey) > EV (small shop | favourable survey), and EV (no shop | favourable survey), Mehmet should build a large shop if the survey is favourable. If the survey is unfavourable, he should build nothing since EV (no shop | unfavourable survey) > EV (large shop | unfavourable survey) and EV (small shop | unfavourable survey).

119

Payoffs (MU)
EV=45,000 MU

Favorable Market (0.9) Unfavorable Market (0.1) Favorable Market (0.9) Unfavorable Market (0.1)

55,000 -45,000 25,000 -15,000 -5,000

2
EV=21,000 MU

3
EV=25,000 MU EV=45,000 MU

1
EV=-33,000 MU

Favorable Market (0.12)

55,000

Unfavorable Market (0.88) -45,000 Favorable Market (0.12) 25,000 EV=-10,200 MU

5
EV=-5,000 MU

Unfavorable Market (0.88)

-15,000 -5,000

EV=10,000 MU

Favorable Market (0.5) Unfavorable Market (0.5) Favorable Market (0.5)

60,000 -40,000 30,000

EV=10,000 MU

EV=10,000 MU

Unfavorable Market (0.5) -10,000 0

72