Past Year Paper Solutions

2009/10(2)

3 Atec is a manufacturer of power supplies. A key component in a power supply is the

PCB board. Atec currently uses two different boards, PCB-1 and PCB-2, which are
sourced from two different suppliers. The unit purchasing prices of PCB-1 and PCB-2
are $25 and $24, respectively. Atec keeps inventory of both boards. The weekly
demands for PCB-1 ( D1 ) and PCB-2 ( D2 ) follow normal distributions as in the
following:
D1 ~ N ( 1 , 1 ) , where 1 500, 1 120
D2 ~ N ( 2 , 2 ) , where 2 550, 2 100
Further analysis reveals that D1 and D2 are negatively correlated with a coefficient of
correlation at 0.2 . The aggregate demand of PCB boards ( D1 D2 ) also follows
a normal distribution:
2 2
D1 D2 ~ N ( 1 2 , 1 2 ) , where 1 2 1 2 , 1 2 1 2 2 1 2

(a) Currently Atec manages the inventories of PCB-1 and PCB-2 independently. For both
types of boards, the fixed ordering cost is $500/order, the replenishment lead time is 4
weeks, and the annual (52 weeks) inventory holding cost rate is 25%. The required
service level (type 1) is 95%, please help Atec to design a (Q, R) inventory control
policy for PCB-1 and PCB-2, respectively (the corresponding z value is 1.645).

(b) Engineers at Atec have recently come up with a universal design PCB-X, which can
perfectly substitute for both PCB-1 and PCB-2. PCB-X can be sourced from a single
supplier at $26.5/unit. Other factors being equal, what will be the ordering quantity
and reorder point for PCB-X under a (Q, R) policy to achieve the same service level at
95%?

(c) Investigate if there is any cost saving for Atec in substituting PCB-1 and PCB-2 with
PCB-X. Based on the result, advise whether Atec should make the substitution.

Solution:

Data given:
1 500, 1 120 ; OiI FEin
ch
550, 2 100 (units/week) Txjoo
fT 256025
2
z
0.2 ,
c1 25, c2 24, c x 26.5 (dollars/unit)
L 4 (weeks), F 500 (dollars/order)
[Link] w
h 25%, =95%, z 1.645 oii T 24
0 25

ML L N1 | P [Link]

WPOL
ge

210
[Link]
 0
t l 69TLZao
f zoo

(a) This is a standard (Q, R) inventory control problem. As the inventories for PCB-1 and
PCB-2 are managed independently, the corresponding ordering quantity and reorder point
can also be solved independently.

For PCB-1:
Demand during replenishment lead time:
D1L ~ N ( 1L , 12L )
1L L 1 4*500 2, 000

1L 4 1 4 *120 240
The ordering quantity can be solved via EOQ model:
2 F (52 1 ) 2*500*52*500
Q1* 2, 040
c1h 25* 25%
The reordering point can be derived as:
R1* 1L z * 1L 2, 000 1.645* 240 2,394
putty
Similarly for PCB-2:

jar
T
if D2 L ~ N ( 2 L , 22L )
L 4*550 2, 200 to
ur
5
2L 2

4 4 *100 200
no
avid
µ 2L 2 r
2 F (52 2 ) 2*500*52*550
Q2* 2,183
c2 h 24* 25%
R2* 2L z* 2L 2, 200 1.645* 200 2,529

(b) With the substitution of PCB-1 and PCB-2, the consolidated demand for PCB-X is:
x 1 2 = 1,050,

x 1 2 1202 1002 2*( 0.2) *120*100 140 (units/week)

Similar approach as in (a) to solve for the ordering quantity and reorder point in (Q, R):
2
DxL ~ N ( xL , xL )
xL L x 4*1, 050 4, 200

xL 4 x 4 *140 280
2 F (52 x ) 2*500*52*1, 050
Q *x 2,871
cx h 26.5* 25%
R*x x z* x 4, 200 1.645* 280 4, 661

(c) The average annual purchasing + inventory related costs are:

Before substitution:

2|P age
 C1 C1ordering C1holding C1 purchase
52 1 Q1*
y
*F R1* 1 L * h * c1 52 1 * c1
Q1* 2
52*500 2040
*500 2394 500* 4 *0.25* 25 52*500* 25
2040 2
a 665, 210
[Link]
52 2 Q2*
C2 *F R2* 2 L * h * c2 52 2 * c2 701, 470
Q2* 2

After substitution:
52 x Q x*
Cx *F Rx* x L * h * c x 52 x * cx 1, 469, 000
Q x* 2

As C x (C1 C 2 ) 1, 469, 000 (665, 210 701, 470) 102, 290 , the total cost will be
higher with the substitution. So, it’s not advised for Atec to substitute PCB-1 and PCB-
1 with PCB-X.

3|P age
2011/12

3. An electronic motor company has a service center in Singapore to provide after-sales

services for two types of motors, Motor A and Motor B. Upon failure, the motors will
be replaced with a spare motor from the inventory that the service center carries. The
weekly demand for spare Motor A follows a normal distribution with mean of 400
units and standard deviation of 50 units. The weekly demand for spare Motor B also
follows a normal distribution but with mean of 500 units and standard deviation of 60
units. The coefficient of correlation between these two demands is 0.6. Inventory of
these two types of motors are both controlled under (Q, R) policy, but they are
managed separately. The lead time of inventory replenishment is 2 weeks and the
required service level (Type-I) is 99% (the corresponding z value is 2.33). The
ordering cost is $100/order. The annual (52 weeks) inventory holding cost rate is 20%.

The engineering department of the company proposes to design a universal Motor C

to substitute Motors A and B. The cost of Motor C is $5.75/unit, which is higher than
that of Motor A at $5/unit and Motor B at $5.5/unit. Motor C has the same failure
property, thus the same demand for spares, as Motors A and B that it substitutes.
Evaluate the impact of the proposed design change upon annual inventory cost
(including both ordering and inventory holding costs) at the service center.

2
Note: for two random numbers following normal distributions, D1 ~ N ( 1 , 1 ) and
2
D2 ~ N ( 2 , 2 ) , their sum also follows a normal distribution.
2 2
D1 D2 ~ N ( 1 2 , 1 2 ) , where 1 2 1 2 , 1 2 1 2 2 1 2 , and is
the coefficient of correlation between the two random variables.

Solution:
The replenishment lead times: l 2 weeks
The ordering cost: c 100 $/order
F
The annual (52 weeks) holding cost rate: h 0.2 .
Service level 99%: z 2.33

The formula for calculation:

C A B E to off or 400
2 2
C A B 2 A B

8 Of 22
2F
c
Ordering quantity: Q *
ph / 52
Ordering cost: 52* cu / Q * a K
*
Average cycle stock: I b Q /2
Safety stock: SS z dl z d l
Inventory holding cost: ( I b SS ) * p * h
E
The calculations are summarized in the table below.

RA I l 8004 2.33 27 4|P age

 b
Motor A B C
Sooo
y [Link]
Cost pA 5 pB 5.5 pC 5.75
Mean A 400 B 500 C 900 [Link]
Demand Standard
deviation A 50 B 60 C 98.5 f gooff
2
I
Ordering Quantity Q *A 2040 QB* 2174 QC* 2853
Cost 1020 1196 1640 I E
Average Q

I
1020 1087 1427
cycle stock
Inventory Safety stock 165 198 325
ai
Holding
1185 1413 2014
cost
Total inventory cost 2205 2610 3654

The results show that by having a more expensive but more ‘common’ motor C, the
ii
total inventory actually goes down as the increased commonality reduces the level of
I
inventory.

7853 is
His Co

27 o Y
f't

5|P age
2013/14

1 A Singapore-based baby food manufacturer has found a high correlation between its
annual sales and the number of births nationally in the preceding year. Table 1 shows
the sales of the baby food and birth figures of Singapore in the past 5 years (year 5 is
the most recent year).
Table 1

Year
1 2 3 4 5
Sales (in $ 1000) 99 92 85 72 68 y
Birth (in thousands) 35 34 31 29 28

(a)
Singapore next year (assuming one step forecast).
(b) Develop a linear regression model to forecast the sales of the baby food for next year.

Solution:
y bo t b i k
(a)
Exponential Smoothing: te I 5 t 39th tutor
Ft At (1 ) Ft 1
ft Ft
1

Results:

Period Actual ES 0.25

1 35 -
2 34 isan 35.00
3 31 34.75
4 29 33.81
5 28 32.61
6 31.46
d
Asi ca t I Fs
Forecast of birth next year is: 31,460 fo
(b)
Assume a linear relationship between annual sales and the number of births nationally
in the preceding year:

Birth (x) Sales (y)

35.0 92
34.0 85
31.0 72
29.0 68

6|P age
 Linear regression:

Sums
x(k) 129
x(k)^2 4183
y(k) 317
y(k)^2 25497
x(k)y(k) 10314

K K K K
xk2 yk xk x k yk
k 1 k 1 k 1 k 1 4183*317 129*10314
b0 2
49.4
K K 4* 4183 1292
K xk2 xk
k 1 k 1
K K K
K x k yk xk yk
k 1 k 1 k 1 4*10314 129*317
b1 2
4.0
K
2
K 4* 4183 1292
K x k xk
k 1 k 1

Sales forecast for next year:

00
y b0 b1 * 28 49.4 4* 28 62.6(thousand )

2(a) SGR is a Singapore-based retailer that sells a particular brand of scooters, which are
produced by a manufacturer in China. The scooters are shipped to SGR through a
third party logistics company. For each order, the logistics company charges a fixed
fee of $1,000 per order plus a variable fee of $10 per scooter (for example, if the SGR
orders 500 units, the shipping cost charged by the logistics company would be
1000+10×500 = 6000). The demand for the scooters is stable at 750 units per week.
The cost of the scooter is $50 per unit. The retail price that SGR sells the scooter is
$99 per unit. The annual (52 weeks) inventory holding cost rate for SGR is 20%.

Determine the optimal order size for SGR to minimize the total cost (including
shipping and inventory holding cost).

Solution:

D 750 Demand units/week

V
f 1, 000 Shipping fixed cost $ / shipment v
r 10 Shipping variable cost $ / unit
c 50 Product cost (in dollars/unit)
v
h 20% Inventory holding cost rate (in percentage/year) v
Q Ordering quantity (units)
7|P age
Transportation cost per order:
f Q r
Annual transportation cost:
52 D
CT f Q r
Q

Annual inventory holding cost at the warehouse:

Q
CI hc
2

Total annual logistics cost:

52 D Q
C f Q r hc
Q 2
52 Df Q
hc 52 D r
Q 2

The optimal ordering quantity (also the optimal shipping size) is:
dC 52 Df hc
0
dQ Q2 2
2*52 Df 2*52*750*1000
Q* 2, 793
hc 20% *50

2(b) Kevin’s retail shop sells a particular type of industrial tape. The demand for the tape is
fairly stable at 600 units per year. Kevin purchases the tape from a distributor. The
distributor offers discounts based on the size of an order. More specifically, for orders
of less than 500 units, the price is 0.3 dollars per unit; for orders of 500 or more but
fewer than 1000, the price is 0.29 dollars per unit; for orders of 1000 or more, the
price is 0.28 dollars per unit. There is also a fixed cost of $8 for placing an order. The
annual inventory holding cost rate is 20%.

Help Kevin to determine the optimal order size to minimize the total cost (including
purchasing, ordering, and inventory holding cost).

Solution:
D 600 [Link] Year
D 600 Demand units/year
F 8 fixed ordering cost $ / order
h 20% Inventory holding cost rate (in percentage/year) 500 o 3 dollar

[Link] u
Sool Luv our
The unit cost of the product is:
028 do'itarki
0.30 for 0 Q 500 woo
c(Q) 0.29 for 500 Q 1000
0.28 for 1000 Q f 8 1order
In o 2
year
8|P age
 8
Q FIE fEo O 2 10
3

aw
For each unit cost, the EOQ quantities are:
2*600*8
=400 for 0 Q 500, Realizable
0.2*0.30

Q * 2 DF 2*600*8
=406 for 500 Q 1000, Not realizable
Q or2 029

hc 0.2*0.29
2*600*8
=414 for 1000 Q, Not realizable
0.2*0.28
500
The candidates for optimal order quantities are: 400, 500, and 1000. cow
The cost function is:
D Q
C D*c *F * h*c
Q 2
600*0.30+600*8/400+400/2*20%*0.30=204 for Q 400
600*0.29+600*8/500+500/2*20%*0.29=198.1 for Q 500
600*0.28+600*8/1000+1000/2*20%*0.28=200.8 for Q 1000

Thus, the optimal order size is: 500

2(c) HomeAppliance sells a particular type of vacuum cleaner. The unit cost of a vacuum
cleaner is 100 dollars. The monthly demand for the vacuum cleaners follows a normal
distribution with mean 800 units and standard deviation of 100 units. HomeAppliance
purchases the vacuum cleaners from a manufacturer in China. The lead time for order
delivery is 2 weeks. The fixed cost of placing an order is 500 dollars. The annual
inventory holding cost rate for HomeAppliance is 25%. HomeAppliance uses a base
stock policy to manage the inventory of the vacuum cleaners and aims to achieve a
service level (type-I) of 99% (the z-value of a standard normal distribution
corresponding to 0.99 is 2.33). Due to scheduling constraint, HomeAppliance can only
place orders on a monthly or quarterly basis. In other words, orders can only be placed
at the beginning of a month or a quarter.

Assuming there are 4 weeks in a month, 3 months in a quarter, and 4 quarters in a

year, determine if HomeAppliance should replenish the inventory of the vacuum
cleaner monthly or quarterly. Show the steps of your calculations to justify your
answer.

Solution:

9|P age
 D ~ N ( , ) Monthly demand distribution
800 Mean demand (units/month)
100 Standard deviation of demand (units/month)
L 2 Order replenishment lead time (weeks)
V
F 500 Fixed ordering cost (in dollars/order)
v
c 100 Product cost (dollars/unit)
v
h 25% Inventory holding cost rate (in percentage/year) v
Type-1 service level: = 99% , the corresponding z 2.33 v
2
Demand in an order cycle (T+L): DT L ~ N( T L , T L )
If monthly: T = 1month = 4 weeks
12
T L (T L) (2 4) * *100 122.5
48
AM
Annual inventory related costs: you you
C1 C1ordering C1holding 12* F T
z* T L *h*c
2
800
12*500 2.33*122.5 *0.25*100 23,135.6
2

If quarterly: T = 1 quarter = 3 months = 12 weeks

12
T L (T L) (12 2) * *100 187.1
48

Annual inventory related costs: I3

C2 C 2 ordering C 2 holding 4* F T
z* T L *h*c
2
3*800
4*500 2.33*187.1 *0.25*100 42,898.6
2

Thus, monthly replenishment of inventory should be selected.

10 | P a g e
2014/15

2. A warehouse keeps inventory of a special type of tires for cars. The unit cost of each
tire is 100 dollars. The monthly demand for the tier follows a normal distribution with
mean 1000 units and standard deviation of 100 units. Currently, the warehouse uses a
base stock policy for inventory management. It replenishes its inventory every 12
weeks. The lead time for order delivery is 4 weeks. The cost of placing an order is 500
dollars, regardless of the size of order. The annual inventory holding cost rate is 20%.
(Note: 1 year = 12 months = 52 weeks; the z-value of a standard normal distribution
corresponding to 0.99 is 2.33).

(a) Under the base stock policy, what is the order-up-to level required to achieve a
service level (Type 1) of 99%? And, what is the annual total inventory related
cost (including ordering cost and inventory holding cost) under the base stock
policy?
(b) Help the warehouse to design a (Q, R) policy and determine the annual inventory
related costs (including ordering cost and inventory holding cost).

Solutions:
D ~ N ( , ) Monthly demand distribution
1000 Mean demand (units/month)
100 Standard deviation of demand (units/month)
T 12 Ordering intervals (weeks)
L 4 Order replenishment lead time (weeks)
F 500 Fixed ordering cost (in dollars/order)
r
c 100 Product cost (dollars/unit)
h 20% Inventory holding cost rate (in percentage/year)
l
(a)
2
Demand in an order cycle (T+L): DT L ~ N( T L , T L )
12
T L (T L) (12 4) *1000* 3692
52
12
T L (T L) (12 4) * *100 192
52

Type-1 service level: = 99% , the corresponding z 2.33

Finding S so that service level can be met:
S z T L T L 3692 2.33*192 4139

Annual inventory related costs:

52
C1 C1ordering C1holding *F T
z* T L * h*c
T 2
52 12*12 / 52*1000
*500 2.33*192 *0.2*100
12 2
38,807
11 | P a g e
(b)
Determine the (Q, R) Policy:
Ordering quantity:
2 F *12 2*500*1000*12
Q* 775
hc 20% *100

Demand during replenishment lead time:

DL ~ N ( L , L2 )
L L 4*(12 / 52) *1000 923

L L 4*12 / 52 *100 96

The reordering point can be derived as:

R L z * L 923 2.33*96 1147

Annual total inventory related cost:

C 2 C 2 ordering C 2 holding
12 Q*
*F z* L *h*c
Q* 2
12*1000 775
*500 2.33*96 *0.2*100
775 2
19,966

12 | P a g e
2015/16

Table 1 shows an electronics retailer’s sales of air conditioners (in thousand units) and the
average daily temperature in the past 4 years (year 4 is the most recent year).

Table 1

Year
1 2 3 4
Sales (’000 )
f
70
30
83
31
92
32
86
31.5 Y
I
Based on the information provided, answer the following questions:

(a) Based on the time series data of sales in the past 4 years, use a single exponential

next year.

(b)
the sales of air conditioners and average daily temperature are linearly related, develop
a causal model to forecast the retailer’s sales of air conditioners for next year.

Solutions:
(a)
Exponential Smoothing:
Ft At (1 ) Ft 1
fzn [Link] t's
70
ft Ft 25 83 t fl 025 73.25
1
Results:

Period Actual ES 0.25 N 32 5

1 70 -
2 83 70.00
3 92 73.25
4 86 77.94
5 79.95 4183 Heal
Forecast of sales for the next year is: 79,950

(b)
Linear regression:

Sums
x(k) 124.5
x(k)^2 3877.25
y(k) 331
y(k)^2 27649
x(k)y(k) 10326
13 | P a g e
 K K K K
xk2 yk xk x k yk
k 1 k 1 k 1 k 1 3877.25*331 124.5*10326
b0 2
253.4
K K 4*3877.25 124.52
K xk2 xk
k 1 k 1
K K K
K x k yk xk yk
k 1 k 1 k 1 4*10326 124.5*331
b1 2
10.8
K K 4*3877.25 124.52
K xk2 xk
k 1 k 1

Sales forecast for next year:

y b0 b1 *32.5 253.4 10.8*32.5 97, 600

2. A contract manufacturer produces a special type of valves for the oil and gas industry.
The demand for the valves is fairly stable at 5000 units per month. The manufacturer
can produce the valve at a rate of 1000 units per day (assuming there are 12 months in
a year, and 30 days in a month). The setup cost for each production run is estimated to
be $300. The annual inventory holding cost rate is 20%. The value of each valve is
$65. Based on the information provided above, answer the questions below:

(a) Determine the size of each production run for the manufacturer to minimize annual
inventory holding and setup costs.

(b) Suppose each production run has to be batches of 400. In other words, the size of a
production run needs to be multiples of 400, for example 400, 800, 1200… etc. Based
on the new information, determine the optimal size of each production run.

Solutions:
(a)
Demand rate: D = 5000 (units/month) = 60,000 units/year
4Out for two
Production rate: R = 1000*30*12 = 360,000 units/year
Fixed setup cost: F = 300 (dollars/setup)
Unit product value (in this case, cost): c = 65 (dollars/unit)
Inventory carrying cost rate h = 20% /year

This is an EOQ problem with a constant production rate.

h 1 D h 1 60000 *0.2 0.1667
R 360000

The optimal size of the production run:

2 DF 2*60000*300
Q* 1823
hc 16.67% *65

(b)
The two eligible production sizes that are closest to the EOQ are: 1600 and 2000.
For Q=1600,
14 | P a g e
 D Q 60000 1600
C (Q ) *F *h *c *300 *16.67% *65 19918
Q 2 1600 2
For Q=2000,
D Q 60000 2000
C (Q ) *F *h *c *300 *16.67% *65 19836
Q 2 2000 2

So 2000 will be the optimal production size.

O
3. A machine tool company sells two kinds of electronic motors with code names:
Motor-A and Motor-B, respectively. The demands for both motors are quite stable but
at different rates. Table 2 shows the costs and monthly demands of these two motors.
For both motors, the fixed ordering cost is 5000 dollars/order. The annual inventory tooooly
holding cost rate is 20%. Sooo pun
Df a ooo dollars for
Table 2

Cost Demand rate C e i bb hilarity

(dollars/motor) (motors/month)
Motor-A 160 5000 f o 2
Motor-B 150 8000

Based on the information provided above, answer the questions below:

(a) Suppose the company manages the inventory of these two motors independently,
use Economic Order Quantity model to determine the optimal ordering frequency
for Motor-A and Motor-B, respectively (Note: you can express ordering frequency
in terms of the number of orders per year).

(b) Suppose the company decides to synchronize the ordering of these two motors. In
a synchronized order, the shop places a combined order for both kinds of motors
instead of placing 2 separate orders for each kind of motor. In other words, the
company orders the two motors in a bundle. Assuming all data remain unchanged,
determine the optimal frequency of the synchronized ordering.

Solutions:
(a)
[Link] [Link]
F 5000 dollars/order
h 20% /year
Q
DA 5000 units/month 2 160
cA 160 dollars/unit
DB 8000 units/month
cB 150 dollars/unit
E
For Motor-A: Q A*
2 DA F 2*(5000*12) *5000
4330 Q Q
hc A 20% *160
15 | P a g e
Ordering Frequency:
DA 5000*12
fA 13.86 times/year
Q *A 4330

For Motor-B:
2 DB F 2*(8000*12) *5000
Qs* 5657
hcB 20% *150

Ordering Frequency:
DB 8000*12
fA 16.97 times/year
QB* 5657

(b) With synchronized ordering, there is a single ordering frequency.

Assuming the time interval between orders is Tc. It needs to satisfy the following
relationship (demand = supply)
Q A QB
Tc
DA DB
Under synchronized ordering, the annual inventory holding cost for Motor-A would
be:
QA DA * Tc
* h * cA * h * cA
2 2

The annual inventory holding cost for Motor-B would be:

QB DB * Tc
* h * cB * h * cB
2 2
1
The annual ordering cost would be: F
Tc

The total inventory related cost is a function of Tc:

DA * Tc DB * Tc 1
C (Tc ) * h * cA * h * cB F
2 2 Tc
DA * c A DB * cB 1
* h * Tc F
2 Tc

Minimizing the above function, the optimal ordering frequency would be:
F 5000
Tc*
DA * c A DB * cB 5000*12*160 8000*12*150
*h *0.2
2 2
0.04564 years

Optimal ordering Frequency: 1/0.04564 = 21.91 times/year.

16 | P a g e
2016/17

O 1
smoothing parameter in exponential smoothing (ES). Suppose that MA with m = 2
-
step forecast, which yields the following table.

Table 1

Period t Demand At MA (m=2) ES ( )

1 116 - -
2 111 - 116.00
3 119 113.50 115.00
4 106 115.00 115.80
5 110 112.50 113.84
6 118 108.00 113.07
7 119 114.00 114.06
8 100 118.50 115.05
9 105 109.50 112.04
10 112 102.50 110.63
11 104 1 108.50 110.90
12 A B

(a) Complete the empty cells A and B in the table, and write down your steps of
calculation.

(b) With the data available, which forecasting method is more accurate based on Mean
Absolute Deviation (MAD)?

Solutions:

The forecasting and the absolute errors are summarized in the table below:

Period t Demand At MA (m=2) Absolute ES ( ) Absolute

error error
1 116 - -
2 111 - 116.00 5.00
3 119 113.50 5.50 115.00 4.00
4 106 115.00 9.00 115.80 9.80
5 110 112.50 2.50 113.84 3.84
6 118 108.00 10.00 113.07 4.93
7 119 114.00 5.00 114.06 4.94
8 100 118.50 18.50 115.05 15.05
9 105 109.50 4.50 112.04 7.04
10 112 102.50 9.50 110.63 1.37
11 104 108.50 4.50 110.90 6.90
12 A=108 7.67 B=109.52 6.29
17 | P a g e
 (a)
its on III sicilian
A: moving average (112+104)/2 = 108 no tz
B: exponential smoothing 104 * 0.2 + 110.90 * (1-0.2) = 109.52

(b)

MAD for the moving average forecast: 7.67

MAD for the exponential smoothing forecast: 6.29 < 7.67

Thus, forecast by the exponential smoothing method is more accurate.

2 An airline has conducted a market survey to determine the optimal price of flight
tickets from Singapore to a popular tourist place in Indonesia. Table 2 summarizes the
monthly demand (y) for the flight tickets at different price points (x).

Table 2

Price (xk) $100 $150 $200 $250 $300

Demand (yk) 1250 900 750 500 200

Additional data analysis based on Table 1 is given below:

5 5 5 5
xk 1000 , yk 3, 600 , xk2 225, 000 , yk2 3, 225, 000 ,
k 1 k 1 k 1 k 1
K
x k yk 595, 000
k 1

s
(a) Based on the information provided, help the company develop a linear regression
model to predict the demand of the flight tickets with respect to price.

(b) Based on the linear regression model derived in (a), determine the price for the flight
ticket so that monthly revenue can be maximized (note: revenue is equal to price times
the number of flight tickets sold).

Solution:
(a)
Linear regression model: y b0 b1 x , where x: price; y: demand

K K K K
xk2 yk xk x k yk
k 1 k 1 k 1 k 1 225, 000*3, 600 1, 000*595, 000
b0 2 2
1, 720
K
2
K
5* 225, 000 1, 000
K x k xk
k 1 k 1
K K K
K x k yk xk yk
k 1 k 1 k 1 5*595, 000 1, 000*3, 600
b1 2 2
5
K K
5* 225, 000 1, 000
K xk2 xk
k 1 k 1

18 | P a g e
 y b0 b1 x 5 x 1720

(b) Revenue: R( x ) (b0 b1 x ) x ( 5 x 1720) x

First order condition:
R ( x) 10 x 1720 0, x* 172 dollars/unit

Dee
Second order condition (optional): R ( x ) 10 ,
So the optimal solution is a maximum.

r Sio
3 c of the coming end-
A stationery shop is considering stocking calendars in anticipation
of-the-year sales season. Each calendar costs $2.8 for the shop to purchase, and the
shop sells it for $8.0 per unit. The shop anticipates 5 possible scenarios for the sales
season, spanning from the Worst to the Great. The discrete probability for each
scenario to take place and the corresponding demand for the calendar is summarized
in Table 3. For example, there is a 35% probability for the sales season to be Average,
and if it happens, the demand for the calendar would be 800 units.
D
Table 3
Scenario Worst Poor Average Good Great
Demand 300 500 800 1200 1800
Probability 10% 20% 35% 25% 10%

(a) Suppose unsold calendars at the sales season are scraped with no salvage value, how
many calendars should the shop order?

(b) The shop recently signed a contract with a paper-recycling company, who is willing to
buy any unsold calendars from the shop at a certain salvage value. Determine the
minimum salvage value that would make the shop order 1200 units of the calendar.

Solution:
(a) The critical ratio is:
r c 8 2.8
Cr 0.65
r s 8 0

So, the optimal quantity to order is to cover 65% of demand.

The different order quantity and the accumulative probability to cover demand:

Demand 300 500 800 1200 1800

Probability 10% 20% 35% 25% 10%
Order quantity 300 500 800 1200 1800
Accumulative probability 10% 10%+20% 65% 90% 100%
So the optimal quantity to order is: 800
(b)
Order quantity of 1200 would cover 90% of the demand.

fo'twt 35 t 25
The critical ratio is: Esu
r c 8 2.8
Cr 0.9

FIE
r s 8 s
o s s
19 | P a g e
 Solving the equation above:
s = 2.22

If the salvage value is 2.22 dollar/unit or above, the shop would order 1200.

4 An industrial engineer was hired by a retailer to improve supply chain management. In

his first project, the engineer was assigned to work on the inventory management of 2
products, codenamed SKU-1 and SKU-2. The inventories of these 2 products are
currently managed separately. After extensive data collection and some preliminary
data analysis, the engineer has come up with Table 4 below.

Shipping cost
Product Demand Purchasing cost
Fixed cost Variable cost
code (units/month) (dollars/unit)
(dollars/order) (dollars/unit)
SKU-1 50 10 100 0.8
SKU-2 40 15 100 1.0

In the table, demand represents the quantity of products that the retailer sells to
consumers. Purchasing cost is how much the retailer pays its supplier for the product.
The shipping cost is how much the retailer pays a logistics company for transporting
the products from the supplier to the retailer. Furthermore, the retailer’s annual
inventory carrying cost is 20%. And, the inventory carrying cost is based on the total
landed cost, which in this case includes the purchasing cost and the variable shipping
cost. For time units, assume 1 year = 12 months = 52 weeks.

(a) Determine the optimal order quantity for SKU-1 and the corresponding total annual
cost (including purchasing, shipping, and inventory holding costs).

(b) The logistics company recently makes an offer: if the retailer orders SKU-1 and SKU-
2 in a combined order, the logistics company only charges the fixed cost of shipping
once, instead of twice as in two separate orders. To take advantage of the offer, the
retailer needs to order SKU-1 and SKU-2 as a bundle at fixed time intervals.
Determine the optimal time intervals between such bundled orders.

Solution:
(a)
The relevant information for SKU-1:
D1 50 Demand mean units/month
F 100 Shipping fixed cost $ / shipment
r1 0.8 Shipping variable cost $ / unit
c1 10 Product cost (in dollars/unit)
h 20% Inventory holding cost rate (in percentage/year)

Transportation cost per order: f1 Q1 c1

Annual transportation cost:

20 | P a g e
 12 D1
C1T F Q1 r1
Q1
Q1
Annual inventory holding cost at the warehouse: C1I h c1 r1
2
Total annual cost:
12 D1 Q1
C1 F Q1 r1 h c1 r1 12 D1c1
Q1 2
12 D1F Q1
12 D1r1 h c1 r1 12 D1c1
Q1 2
12 D1F Q1
h c1 r1 12 D1 c1 r1
Q1 2

The optimal ordering quantity is:

dC1 12 D1F h c1 r1
0
dQ1 Q12 2
2*12 D1F 2*12*50*100
Q1* 236
h c1 r1 20% * 10 0.8

(b)
With synchronized ordering, there is a single ordering frequency.
Assuming the time interval between orders is Tc.
It needs to satisfy the following relationship (demand = supply)
Q1 Q2
Tc (months)
D1 D2
Under synchronized ordering, the annual inventory holding cost for SKU-1 would be:
Q1 D1Tc
h c1 r1 h c1 r1
2 2

Similarly, the annual inventory holding cost for SKU-2 would be:
Q2 D2Tc
h c2 r2 h c2 r2
2 2
12
The annual ordering cost would be: F
Tc

The total inventory related cost is a function of Tc:

D1Tc D2Tc 12
C (Tc ) h c1 r1 12 D1c1 h c2 r2 12 D2 c2 F
2 2 Tc
D1 c1 r1 D2 c2 r2 12
hTc F 12 D1c1 D2 c2
2 Tc

Minimizing the above function, the optimal ordering frequency would be:

21 | P a g e
 12 F 100
Tc*
D1 c1 r1 D2 c2 r2 50*(10 0.8) 40*(15 1)
*h *0.2
2 2
3.2 months

5 A technology company based in Singapore has developed a new medical device. The
company sets the price of the device at 480 dollars/unit. The demand for the device is
estimated in Table 5 (for example, there is a 25% probability for the demand to be 300
units).
Table 5
Quantity Probability
300 25%
500 40%
600 35%

The company is negotiating with a supplier for the production of the device. The fixed
cost for the supplier to set up the production line is 20,000 dollars. The variable cost
of production is 90 dollars/unit. At the end of the product’s lifecycle, any unsold
device is useless to the company.

(a) If the supplier charges $180/unit to produce the medical device for the company
and refuse to buy back any unsold devices, how many units should the company
order? And, what is the expected profit for the company and its supplier?

(b) The supplier has developed a technology to recycle the medical device for
materials, the value of which is 30 dollars/unit to the supplier. If the supplier still
charges $180/unit to produce the medical device but agrees to buy back any unsold
devices from the company at $20/unit, how many units should the company order?
And, what is the expected profit for the company and its supplier?

Solutions:
(a)
Notations:
wholesale price: w=180;
retail price: r = 480;
production cost: c = 90;
demand: d;
order quantity: q.
fixed cost: f = 20,000

The supplier’s expected profit:

m q( w c) f
The retailer’s expected profit:
n

r min(q, di ) r q w pi
i 1

For q 300
22 | P a g e
 r q *(r w) q *(480 180) 300q

For 300 q 500

r 300* r * 25% q * r *75% q * w
36000 180* q

For 500 q 600

r 300* r * 25% 500* r * 40% q * r *35% q * w
132, 000 12* q

For 800 q
r 300* r * 25% 500* r * 40% 600* r *35% q * w
232800 180* q

The maximum expected profit is when q 500 ,

m q ( w c) f 500*(180 90) 20, 000 25, 000
r 36000 180*500 126, 000

(b)
buyback price to the company: sc = 20;
salvage value to the supplier: ss = 30;

The company’s expected profit:

n

r min(q, di ) r max(q, di ) di s q w pi
i 1

For q 300
r q *(r w) q *(480 180) 300q

For 300 q 500

r 300* r (q 300) * sc * 25% q * r *75% q * w
300*(r sc ) * 25% q * sc * 25% r *75% w
34500 185q

For 500 q 600

r 300* r (q 300) * sc * 25% 500* r (q 500) * sc * 40%
q * r *35% q * w
300* 25% 500* 40% *(r sc ) q sc *65% r *35% w
126500 q

For 600 q

23 | P a g e
 r 300* r (q 300) * sc * 25%
500* r (q 500) * sc * 40%
600* r (q 600) * sc *35% q * w
[300* 25% 500* 40% 600*35%]*(r sc ) q *( sc w)
223100 160q

The maximum expected profit is when q 600 ,

r 126500 q 127100

The supplier’s expected profit:

n

m q( w c) f ( ss sc ) max(q, di ) di pi
i 1

600*(180 90) 20000 (600 300) * 25% (600 500) * 40% *(30 20)
35150

24 | P a g e
2017/18

3 A warehouse handles parcel deliveries for online retailers. The amount of parcels

O
received each day by the warehouse is a random variable with the following discrete
probability distribution (Table 3). For example, there is a 2.5% probability for the
number of parcels in a day to be 1500.

Table 3

Parcels 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000
Probability 2.5% 5% 10% 20% 22.5% 15% 10% 7.5% 5% 2.5%

r
The warehouse employs a team of workers to deliver the parcels. Each worker can
deliver 500 parcels per day, and costs the warehouse $300 to hire a worker for a day.
In case the amount of parcels exceeds the capacity of the workers, the warehouse
would outsource delivery of the surplus parcels to a 3rd party logistics company, which
charges $5 per parcel for the delivery service. Based on the information provided
above, answer the questions below:
WorkerId
C
(a) What is the underage cost for the warehouse to NOT have enough workers?
(5 marks)
(b) What is the overage cost for the warehouse to have too many workers?
(5 marks)
(c) How many workers should the warehouse hire?
(10 marks)

Solutions
a) First, we convert the demand in terms of parcels to demand in terms of workers.
25 (5 points)
g 545
Demand for workers 3 4 5 6 7 8 9 10 11 12
Probability 2.5% 5% 10% 20% 22.5% 15% 10% 7.5% 5% 2.5%

Having made the transformation, it becomes easy to determine underage cost:

If we had excess parcels, we would have had to call in the 3rd party logistics company,
and it would have cost $5 / parcel * 500 parcels/worker = $2500.

Normally, it would have cost the warehouse only $300, if we have sufficient workers.
So the underage cost is $2500 - $300 = $2200.

Note: It’s also right to leave the demand in terms of number of parcels:
$5-$0.60=$4.40. That means you convert the overage cost in terms of workers to
parcels (i.e.: $300/500 = $0.60).

b) What is the overage cost of having too many workers?

(5 points)
$300 per delivery worker, as an extra worker will be paid this amount without
delivering any parcels. Note that there is no salvage value to have too many delivery
workers.
25 | P a g e
 c) How many workers should the warehouse hire?
(10 points)
First, obtain the critical ratio, which is:
2200/ (2200+300) = 0.88.

Then, obtain the cumulative distribution of demand:

Demand for workers 3 4 5 6 7 8 9 10 11 12

Cumulative Probability 2.5% 7.5% 17.5% 37.5% 60% 75% 85% 92.5% 97.5% 100%

Critical ratio is between 9 and 10, so the warehouse need to hire10 workers.

4 A furniture store sells a particular brand of chairs in Singapore. The store buys the
chairs from China via Alibaba, an e-commerce platform. The cost of the chair is $50
per unit, and the store sells it for $100 per unit. The demand for the chair at the
furniture store is quite stable at 100 units per month. The annual inventory holding
cost rate for the furniture store is 24%. Besides purchasing cost and inventory holding
cost, the only other cost that is significant to the furniture store is shipping cost. Note:
the shipping cost is charged separately and is NOT part of the purchasing cost.

(a) Suppose Alibaba charges a fixed fee of $800 for shipping, what will be the optimal
order size for the furniture store?
(10 marks)

Solution:

Demand rate: D = 100 (units/month) = 1200 units/year

(Fixed) shipping cost: F = 800 (dollars/order)
Unit product cost c = 50 (dollars/unit)
Inventory carrying cost rate h= 24% /year

This is a standard EOQ problem.

2 DF 2*100*12*800
Q* 400
hc 24% *50

(b) Suppose Alibaba has a tiered fee structure for shipping, i.e. for any purchase order
below 30000 dollars, the shipping fee is 800 dollars per order; for any purchase order
between 30000 dollars and 50000 dollars, the shipping fee is 500 dollars per order; for
purchase orders that is 50000 dollars or above, the shipping fee will be waived
completely, what will be the optimal order size for the furniture store under this new
fee structure?
(10 marks)
Solution:

The minimum order size to qualify for $500/order shipping cost: 30000/50 = 600;
The minimum order size to qualify for waiver of shipping cost: 50000/50 = 1000.
26 | P a g e
 Thus, the ordering cost is a function of order size as the following:
800, if Q 600 r
F si
400, if 600 Q 1000 r
0, if Q 1000
r [Link]
For Q<600, optimal Q is 400 as in part (a);
The resulted total cost is:
* *
C total 2 C shipping C holding
D Q2*
*F *h*c
Q1* 2
100*12 400
*800 * 24% *50
400 2
4800
For 600<= Q < 1000, F = 500 dollars/order,
The corresponding EOQ quantity would be:
2 DF 2*100*12*500
Q2* 316 , which is less than 600.
hc 24% *50
Thus, the optimal quantity to order is Q2* 600

The resulted total cost is:

* *
C total 2 C shipping C holding
D Q2*
*F *h*c
Q1* 2
100*12 600
*500 *0.24*50
600 2
4600

For Q > 1000, F = 0 dollars/order, only inventory holding cost is relevant.

To minimize inventory holding cost, Q is to be minimized. Q3* 1000
The resulted total cost is:
* Q3* 1000
C total 3 C holding *h*c *0.24*50 6000
2 2
Compare the total costs under different ranges, Q2* 600 gives the lowest cost, thus is
the optimal quantity to order.

5 A fashion retailer in Singapore outsources the production of a special kind of

handbags to a supplier in Vietnam. The retailer sets the retail price of the handbags at
$500/unit. At the end of the sales season, the retailer can salvage unsold handbags at
$50/unit. For the forthcoming sales season, the retailer anticipates 3 sales scenarios,
namely bad, average, and good. The sales volume and probability of each scenario is
given in Table 4.

27 | P a g e
 Table 4

Scenario Sales volume Probability

Bad 300 30%
Average 500 45%
Good 800 25%

For the supplier, the fixed cost of production setup is $20000 and the variable cost of
production is $100/unit.

(a) What is the global optimal order quantity for the overall supply chain? And, what’s
the maximum total expected supply chain profit?
(7 marks)

(b) If the wholesale price that the supplier charges the retailer is $200/unit, what should
be the retailer’s order quantity? And, what is the expected profit for the retailer?
(8 marks)

(c) Suppose the only variable that the supplier and the retailer can negotiate in a supply
contract is the wholesale price, what is the range of wholesale price that both the
supplier and the retailer will be profitable and maximum expected profit for the
supply chain can be achieved?
(5 marks)

Solution:
retail price: r = 500
salvage value: s = 50
production variable cost: c = 100
demand: di (d1=300, d2=500, d3=800)
probability: pi (p1=30%, p2=45%, p3=25%)
fixed cost: f = 20,000
order quantity: q

(a) For overall supply chain,

n

s {min(q, di ) r max(q, di ) di s q c f }pi

i 1
For q 300
r q *(r c) f 400q 20000

For 300 q 500

3

s d1r (q d1 ) s qc f p1 (qr qc f ) pi
i 2

20500 265q
For 500 q 800

28 | P a g e
 2

s di r (q di ) s qc f pi (qr qc f ) p3
i 1

121750 62.5q
For 800 q
3

s di r (q di ) s qc f pi
i 1

211750 50q

The global optimal order quantity would be q = 800

The maximum expected profit is: 121750 + 62.5*800 = 171,750

(b)
The retailer’s expected profit:
n

r min(q, di ) r max(q, di ) di s q w pi
i 1

For q 300
r q *(r w) q *(500 w)

For 300 q 500

r 300* r (q 300) * s *30% q * r *70% q * w
300*(r s ) *30% q * s *30% r *70% q*w
40500 (365 w) * q
For 500 q 800
r 300* r (q 300) * s *30% 500* r (q 500) * s * 45%
q * r * 25% q * w
300*30% 500*50% *(r s ) q * s *75% q * r * 25% q * w
141750 (162.5 w) * q
For 800 q
r 300* r (q 300) * s *30% 500* r (q 500) * s * 45%
800* r (q 800) * s * 25% q * w
[300*30% 500* 45% 800* 25%]*(r s ) q *( s w)
231750 (50 w) * q

Thus, when w = 200,

300q, when q 300
40500 165q, when 300 q 500
r
141750 37.5q, when 500 q 800
231750 150q, when q 800

The optimal order quantity by the retailer would be q = 500

The maximum expected profit for the retailer is: 123000
The supplier’s profit: m q ( w c) f = 500*(200 - 100) – 20000 = 30000
29 | P a g e
(c)
Based on part (b), for the retailer to order 800 instead of 500, we need to have
162.5 w 0
And, to ensure rationality and the supplier be profitable, we need to have:
m q ( w c) f 800( w 100) 20000 0
50 w 0

Thus, by aggregating all the conditions, the wholesale price needs to be:
162.5 w 125

30 | P a g e
2018/19

1 Table 1 shows the sales of electric cars (in thousand units) and the average gasoline
price in the past 3 years in an automobile market (year 3 is the most recent year).

Table 1

Year

Sales (’000 )
1
50
2
80
3
100the
Gasoline price ($/gallon) 2.0 2.5 3.0

Based on the information provided, answer the following questions:Be

(a) Based on the time series data of sales in the past 3 years, use a single exponential
5 to forecast the sales of electric cars in year 4.

(b) It’s estimated that the average gasoline price in year 4 will be 3.5 $/gallon. Suppose
the sales of electric cars and average gasoline price are linearly related, develop a
causal model to forecast the sales of electric cars in year 4.

Solution:
period Ac ual ES
(a) 50 O
Exponential Smoothing: so
Ft At (1 ) Ft 1 O 5150 t
2
fo
ft Ft 3 100
1
Results: Ez HA [Link] a

Period Actual ES 0.5

xeo
1 50 - dC.loo3tlt4l.fOY0FX8DtLos
2 80 50.000
3 100 65.000
4 82.500
o 5 10011 0.5 65

Forecast of sales for the next year is: 82,500

(b)
Linear regression:
[Link] 50

Sums
x(k) 7.50
x(k)^2 19.25
y(k) 230
y(k)^2 18900
x(k)y(k) 600.00

31 | P a g e
 K K K K
xk2 yk xk x k yk
k 1 k 1 k 1 k 1 19.25* 230 7.5*600
b0 2
48.333
K K 3*19.25 7.52
K xk2 xk
k 1 k 1
K K K
K x k yk xk yk
k 1 k 1 k 1 3*600 7.5* 230
b1 2
50
K
2
K 3*19.25 7.52
K x k xk
k 1 k 1

Sales forecast for next year:

y b0 b1 *3.5 48.333 50*3.5 126.667 thousand

2 An electronics company manufactures a device in Singapore and sells it to South East

Asian countries. The unit cost of the device is $10/unit. The demand for the device is
fairly steady at a rate of 500 units/month. On the production side, the setup cost for
each production run is $2000. The rate of production is 2000 units/month. The
inventory holding cost rate is 20%. Answer the following questions.

(a) Determine the optimal size of each production run and the annual inventory related
cost (including ordering cost and inventory holding cost).

(b) Suppose the time interval between production runs has to be multiples of 3 months,
meaning that it could only be 3 months, 6 months, …, etc., determine the optimal
size of each production run with the additional information.

Solution:
(a)
Demand rate: D = 500 (units/month) = 6,000 units/year y Q n c
Production rate: R = 2000 (units/month) = 24,000 units/year
Fixed setup cost: F = 2000 (dollars/setup) v
v
Unit product value (in this case, cost): c = 10 (dollars/unit)
Inventory carrying cost rate h = 20% /year h l Pak
v

fi ojo
This is an EOQ problem with a constant production rate. 2
h 1 D h 1 6000 * 20% 15%
R 24000

The optimal size of the production run:

tooo
2 DF 2*6000* 2000 500
Q* 4000
hc 15% *10

(b)
8northS
At the EOQ quantity, the time interval between production runs is:
Q * 4000 hunt
T *

D 500
8 months f or to

32 | P a g e
 The two eligible time intervals that are closest to the EOQ results are: 6 and 9 months.
For T=6 month, Q=3000,
D Q i 6000 3000
C (Q ) *F *h *c * 2000 *15% *10 6250
Q 2 3000 2
For T=9 month, Q=4500,
D Q 6000 4500
C (Q ) *F *h *c * 2000 *15% *10 6042
Q 2 4500 2

So T=9 month, and Q=4500 will be the optimal production size.

3 A company based in Singapore is in the business of designing and selling scooters.

The production of the scooters is outsourced to a supplier in China. The unit cost of a
scooter is 500 dollars. The monthly demand for the scooter follows a normal
distribution with mean value of 500 units and standard deviation of 100 units. The
company currently manages its inventory using a base stock policy, and replenishes its
inventory every 6 weeks. The lead time of order delivery from its supplier is 3 weeks.
The cost of placing an order is 3000 dollars. The annual inventory holding cost rate is
20%. Note: conversions between time units are 1 year = 12 months = 48 weeks; the z
value corresponding to 95% is 1.645.

(a) Under the base stock policy, what is the order-up-to level required to achieve a service
level (Type 1) of 95%? And, what is the annual inventory related costs (including
ordering cost and inventory holding cost)?

(b) Help the company to design a (Q, R) policy to achieve a service level (Type 1) of
95%, and calculate the inventory related costs (including ordering cost and
inventory holding cost).

Solution
(a)
2
D ~ N( , ) Monthly demand distribution
500 Mean demand (units/month)
100 Standard deviation of demand (units/month)
T 6 Ordering intervals (weeks)
L 3 Order replenishment lead time (weeks)
F 3000 Fixed ordering cost (in dollars/order)
c 500 Product cost (dollars/unit)
h 20% Inventory holding cost rate (in percentage/year)

2
Demand distribution in (T+L): DT L~ N ( T L, T L )
1
T L (T L) (6 3) *500* 1125
4
6 3
T L (T L) *100 150
4

33 | P a g e
Finding S so that service level can be met:
S z T L T L 1125 1.645*150 1372

Annual inventory related costs:

C1 C1ordering C1holding
48
*F T
z* T L *h*c
T 2
48 6 / 4*500
*3000 1.645*150 * 20% *500
6 2
86175

(b)
Ordering quantity:
2 F *12 2*3000*500*12
Q* 600
hc 20% *500

Demand during replenishment lead time:

DL ~ N ( L , L2 )
L L 3 / 4*500 375

L L 3 / 4 *100 87

The reordering point can be derived as:

R L z * L 375 1.645*87 517

Annual total inventory related cost:

C 2 C 2 ordering C 2 holding
12 Q*
*F z* L *h*c
Q* 2
12*500 600
*3000 1.645*87 * 20% *500
600 2
74246

4 A start-up company has developed a new type of medical device. The company is
negotiating a contract with a distributor for sales of the device. For the company, the
fixed cost of production is $250,000 and the variable cost of production is $250/unit.
The distributor is able to sell the medical device to hospitals at a price of $800/unit.
The demand for the medical device from hospitals is uncertain, with quantities and
corresponding probabilities shown in the Table 2 below. As the medical device is very
specialized, the value of any leftovers robots is negligible, i.e. can be assumed to be 0.

34 | P a g e
 Table 2

Demand (units) Probability

300 20%
500 50%
800 30%

The company and the distributor have agreed upon a wholesale price of $500/unit.
The manufacturer is willing to buy back any leftover devices from the distributor at
$100/unit. Based on the information above, answer the following questions.

(a) How many medical devices should the distributor order from the company?

(b) What are the expected profits for the company and distributor, respectively?

Solution:
(a)
Retail price: r = 800
production variable cost: c = 250
demand: di (d1=300, d2=500, d3=800)
probability: pi (p1=20%, p2=50%, p3=30%)
fixed cost: f = 50,000
wholesale price: w = 500
order quantity: q

With buyback price at: sbb 100 , the distributor’s profit function is:
3

r min(q, di ) r max(q, di ) di sbb q w pi

i 1
For q 300 ,
r q *(r w) 300q

For 300 q 500

r 300* r (q 300) * sbb * 20% q * r *80% q * w
300*(r sbb ) * 20% q * sbb * 20% r *80% w
42000 160* q

For 500 q 800

r 300* r (q 300) * sbb * 20% 500* r (q 500) * sbb *50%
q * r *30% q * w
217, 000 190q

For 800 q
r 3, 000* r (q 300) * sbb * 20% 500* r (q 500) * sbb *50%
800* r (q 800) * sbb *30% q * w
385, 000 400* q
35 | P a g e
 The optimal order quantity by the distributor would be q = 500
(b)
The maximum expected profit for the distributor is:
r 42000 160* q 42000 160*500 122, 000

The expected profit function for the company is:

3

m q( w c) f max(q, di ) di sbb pi
i 1

When q = 500, there will be leftovers when demand turns out to be 300 units, hence
1

m q( w c) f q di sbb pi
i 1

q( w c) f sbb q d1 pi
500*(500 250) 50, 000 100*(500 300) * 20%
35, 000

5 Figure 1 shows the battery architecture of Tesla Model S. The battery pack consists of
16 battery modules, each battery module consists of 444 Lithium ion battery cells, and
each battery cell is measured by 18mm×65mm (diameter × height) in dimensions,
hence called 18650 batteries. The production of the battery pack is highly automated.
The battery cells are slotted into a case structure in an optimized array to make a
battery module. The battery modules are then fitted into a battery pack that is sealed
with high-strength steel panels. Each battery module has its own sensors for
monitoring temperature, and battery modules are separated from each other with heat
resistant materials in a battery pack that also serves as the floor of the car.

Figure 1

Discuss the characteristics of Tesla’s battery architecture, and the implications on

supply chain management in terms of key technological and economic advantages.

36 | P a g e
Solution (reference):

The key characteristic of Telsa’s battery architecture is that it’s modular. Standard
and generic battery cells are combined to make battery modules, which are further
combined to make battery packs.

On the technological fronts, some of the key implications include:

Better safety, as battery modules are partitioned and monitored individually for
temperature, thus preventing thermal runaway;
Easier assembly, as generic battery cells are assembled into modules, which
are further assembled to make battery packs;
Lower centre of gravity, better stability for the car, as the battery is put as the
floor of the vehicle.

On the economic fronts, some of the key implications:

Lower cost due to economies of scale, as battery cells are generic and standard
that can be mass produced;
Higher product variety, as battery modules could be flexibly combined to
achieve different driving ranges without incurring significantly higher cost of
battery design;
Lower maintenance cost, as modules can be repaired or replaced separately.

The End

37 | P a g e