Scribd
Upload
24 views

Exercises On Decision Trees and Payoff Tables: Costs, N

The document discusses exercises involving decision trees and payoff tables. It includes questions about identifying the best decision alternative based on expected values for different probabilities of states of nature. It also includes a question about calculating expected travel time using a payoff table for two road options, and questions about how perfect and imperfect forecast information would impact the expected travel time. Finally, it discusses a scenario where a TV retailer must decide how many TVs to order based on uncertain customer demand, including constructing a payoff table and decision tree to determine the expected profit and value of perfect information.

Uploaded by

Viktor
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
24 views

Exercises On Decision Trees and Payoff Tables: Costs, N

The document discusses exercises involving decision trees and payoff tables. It includes questions about identifying the best decision alternative based on expected values for different probabilities of states of nature. It also includes a question about calculating expected travel time using a payoff table for two road options, and questions about how perfect and imperfect forecast information would impact the expected travel time. Finally, it discusses a scenario where a TV retailer must decide how many TVs to order based on uncertain customer demand, including constructing a payoff table and decision tree to determine the expected profit and value of perfect information.

Uploaded by

Viktor
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Exercises on Decision Trees and Payoff Tables

Questions 1, 2 and 3 are based on the following table. In the following table of
costs, n1 and n2 are the states of nature, and a1, a2 and a3 are your decision
alternatives:

n1 n2
a1 5 3
a2 1 4
a3 2 k

Q.1. Suppose k = 6. If n1 occurs with probability p, If n1 occurs with probability p, for


what values of p will a3 be the best decision alternative based on expected
values, for what values of p will a1 be the best and for what values of p will a2
be the best?

Q.2. Suppose n1 occurs with probability p. For what all values of k is a3 always better
than a1 (based on expected values)?

Q.3. What is the EVPI?

Questions 4, 5 and 6 are based on the following:

You and your friend agree to form a carpool for travelling from the campus to
the city and back, whenever required, especially on holidays. Your friend prefers
to use the “Bypass Road”. Though you prefer to use the “AB Road”, you agree
that you should take the “Bypass Road” if traffic on “AB Road” was jammed. The
following payoff table provides the one-way time estimate in minutes for
travelling to or from the city, from or to the campus.

State of Nature
No traffic jam Traffic Jam on
on AB Road (s1) AB Road (s2)
Decision Alternative
Bypass Road 50 50
AB Road 35 65

Based on the experience of both of you with traffic conditions on these roads,
both of you agree on a 0.10 probability that “AB Road” would be jammed. You
also agree that weather affects the traffic conditions on “AB Road”, and agree
upon the following estimates on conditional probabilities:
P(C | s1) = 0.75, P(O | s1) = 0.20, P(R | s1) = 0.05,

P(C | s2) = 0.15, P(O | s2) = 0.25, P(R | s2) = 0.6,

where C = clear sky, O = overcast, R = rain.

Q.4. Which decision alternative would you choose? What is the expected travel time?

Q.5. You are considering whether or not to buy the report of PQR’s weather forecast
whose outcome will be either “high chance of rain” or “low chance of rain”. If
PQR forecasts “high chance of rain”, you believe that there is a 70% chance of
traffic jam on AB road. If PQR forecasts “low chance of rain”, you believe that
there is a 20% chance of traffic jam on AB road. What is the expected time you
can save if you buy PQR’s weather forecast report?

Q.6. What is the expected time you can save if you have perfect information?

Q.7. TV-Town must decide how many, if any, new Sanapony 50 inch television set to
order for next month. The sets costs TV-Town $ 1850 each and sell for $2450
each. Because Sanapony is coming out with a new line of big screen television
sets in a month, any sets not sold during the month will have to be marked down
to 50% of the normal retail price to be sold at the TV-Town clearance centre. TV-
Town estimates that if it does not have enough television sets on hand to satisfy
demand, it will suffer a goodwill loss of $ 150 for each customer who cannot get
a set. TV-Town management feels that the maximum customer demand over the
next month will be for three big screen sets. Note that the states of the nature
correspond to the demand and the decision alternatives correspond to the
number of sets that TV-Town plans to order.

a. Write down TV-Town’s payoff table.

Now, suppose the manager of TV-Town estimates the probabilities of possible


customer demands for the next month as follows
P (Demand = 0 sets) = 0.20
P (Demand = 1sets) = 0.30
P (Demand = 2 sets) = 0.30
P (Demand = 3 sets) = 0.20

b. Draw the decision tree. How many Sanapony big screen television sets
should TV-Town order? What is TV-Town’s expected profit?

c. What is the maximum amount TV-Town should pay, for perfect


information.
d. The manager of TV-Town is considering conducting a telephone survey of
30 randomly selected customers. The survey will determine whether at
least one of the 30 is likely to buy a big screen set within the next month.
What is the maximum amount TV-Town should pay, for the telephone
survey?

You might also like