INDR 371 LAB 2

Fall 2015, Nov. 2, 2018

These lab exercises constitute part of the course grade. Please save your work at the end of the lab
under a file named name_surname_lab3 under the HOMEWORKS/LAB2 directory. It is expected that
you should complete Exercise 1 and parts of Exercise 2 and Exercise 3 during the lab. The deadline
for Lab-2 Homework submission is Nov. 9 17:00.

1. Semiconductor wafer fabrication is very much error-prone. Assume that a wafer production
facility received an order for a specially designed prototype wafer. The cost of producing
each wafer is estimated to be $20,000. The customer agrees to pay $150,000 for 3 good
wafers, $200,000 for 4 good wafers, and $250,000 for 5 good wafers. Other than 3, 4, or 5
good wafers, all other wafers must be destroyed (i.e. have no value). To obtain the contract,
the wafer fab offers to pay the customer a penalty of $100,000 if at least three good wafers
are not produced. Assume that each wafer is produced independently and the probability
that a wafer is acceptable is 0.65.

This exercise requires setting up several tables and computing different measures (expected
profits, probabilities etc.). This sort of calculation is easily done in a spreadsheet. It is
important to note that your calculations should be general enough to hold when different
inputs (revenues, probabilities etc.) are changed. For this purpose, always define all inputs as
variables whose values can be changed.

a. Let Q be the number of wafers manufactured and X(Q) be the random variable
corresponding to the number of good wafers. Construct a probability table for Q=3,4,
…,10, to compute, P(X(Q)≥3), the probability that there are at least 3 good wafers
when Q wafers are manufactured. Note that this is a Binomial probability. You can
use the BINOMDIST function in Excel to compute Binomial probabilities. Find the
minimum value of Q that ensures that the probability of having at least 3 good wafers
is greater than 0.95.
b. Construct a probability table for Q=1,2,…,10, to compute, P(Q,x), the probability that
there are exactly x good wafers when Q wafers are manufactured. Note that this is a
Binomial probability. You can use the BINOMDIST function in Excel to compute
Binomial probabilities.
c. Construct a profit table for Q=1,2,…,10, to find, Z(Q,x), the profit obtained when Q
wafers are manufactured and x of them are good.
d. Use the probability and profit tables to compute E[Z(Q)], the expected profit when Q
wafers are manufactured. What is the optimal quantity to manufacture to maximize
the expected profit?
e. How does the optimal quantity change when the probability of a wafer being
acceptable changes in the range 0.5 to 0.95? How does the optimal order quantity
change if the production cost per wafer changes in the range from $10,000 to
$40,000?
 f. The wafer company is risk-averse and would also like to avoid the risk of losing
money. In order to compute the probability of losing money when Q wafers are
manufactured, convert the profit table to a binary loss/gain table where L(Q,x), =0 if
Z(Q,x) ≥0 and is 1 otherwise.
g. Using the binary loss/gain table. Compute the probability of losing money for Q=1,2,
…,10.
h. Now assume that consecutive wafer productions are not necessarily independent.
Consider the following model: P(wafer i+1 is good | wafer i is good)=0.9 and P(wafer
i+1 is bad| wafer i is good)=0.1. In addition, P(wafer i+1 is good | wafer i is bad)=0.6,
P(wafer i+1 is bad | wafer i is bad)=0.4. Assume that the first wafer is good with
probability 0.75 and find the expected profit when 4 wafers are scheduled.

2. A firm has received an order for 25 die cast parts made from precious metals. The parts will
sell for $5000 each and it will cost $2500 to produce an individual part. The probability of a
part meeting final inspection equals 0.75. Parts that meet the inspection but are not sold can
be recycled at a value of $1000. Parts that do not meet the inspection have a unit salvage
value of $100. A penalty clause in the contract results in the firm having to pay the customer
$500 per unit short if the number of parts that can be delivered is less than 25. In addition,
if the number that can be delivered is less than 23, there is an additional fixed penalty of
$10,000 (i.e. if 20 parts can be delivered, the total penalty is (25-20)⨯500 +
10,000=$12,500).

a. Generate an Excel chart that shows the expected profit for a given production
quantity.
b. Find the production quantity that maximizes the expected profit
c. Find the probability of losing money when the order quantity in part b. is used.
d. Assume that there is a process improvement that can improve production quality. If
this improvement takes place, the probability of a given part meeting final inspection
will raise to 95%. What is the maximum amount that should be paid for such an
improvement?
e. Find the variance of the profit when the production quantity in part b is used. (Hint:
It may be easier to find the second moment of the profit first).
3. Consider a single machine production system in which thin golden plates are produced for an
electronic circuit. The target plate thickness is 2 ± 0.2millimeters (plates with 1.8-2.2 mm
thickness are acceptable). The historical data shows that the thickness of the golden plates that
have been produced by the current machine (Model A) has a normal distribution with a mean of
2 mm and a standard deviation of 0.2 mm. Assume that each plate production has an
independent success probability to meet the quality requirements and you have a contract to
deliver 20 good items each day to a client. You would like to determine how many items to
produce.
a. Calculate the probability of meeting the golden plate demand if 15, 16, 17 ,…, 40 plates
are produced (Hint: Use BINOM.DIST.RANGE function). Generate an Excel chart to
illustrate your calculations.
b. Find the production quantities that will satisfy the golden plate demand with 90 and 95%
probabilities.
c. Construct a scroll bar to control plate thickness standard deviation for our production to
see how the results of part 3.a and 3.b would change with the increase or decrease in
the standard deviation
d. Now assume that:
 Each plate requires 90 minutes machine time.
 Machines can be operated for 8 hours a day and they require 30 minutes
recalibration to start production at the beginning of each day.
 Due to the failures the availability factor is estimated to be 0.9 for the machines.
Considering the daily production amount you determined in part b for %95 percent
demand satisfaction probability.
e. Assume that Model A machine cost 10.000$ to purchase (we don’t consider the
operating costs). Build a decision support tool (a fancy name for a useful Excel sheet!)
for the facility managers that can calculate the worth of a new machine configuration for
a given:
 Machine time required to produce one plate,
 Standard thickness deviation (assume mean is always the target thickness of 2
mm),
 Machine set up time at the beginning of the day,
 Availability factor.