Solucionario Ballou

CHAPTER 1
A VITAL SUBJECT
BUSINESS LOGISTICS/SUPPLY CHAIN
12
(a) This problem introduces the student to the evaluation of alternate channels of
production and distribution. To know whether domestic or foreign production is least
expensive, the total of production and distribution costs must be computed from the
source point to the marketplace. Two alternatives are suggested and they can be
compared as follows.
Production at Houston:
Total cost = Production cost at Houston + Transportation and storage costs
= $8/shirt100,000 shirts + $5/cwt. 1,000 cwt.
= $805,000/year
Production at Taiwan:
Total cost = Production cost in Taiwan
+ Transportation and storage costs from Taiwan to Chicago
+ Import duty + Raw material transportation cost from Houston
to Taiwan
= $4/shirt100,000 shirts + $6/cwt. 1,000 cwt. + $0.5/shirt100,000 shirts
+ $2/cwt. 1,000 cwt.
= $458,000/year
Producing in Taiwan would appear to be the least expensive.
(b) Other factors to consider before a final decision is made might be:
(i) How reliable would international transportation be compared with domestic
transportation?
(ii) What is the business climate in Taiwan such that costs might change in favor of
Houston as a production point?
(iii) How likely is it that the needed transportation and storage will be available?
(iv) If the market were to expand, would there be adequate production capacity
available to support the increased demand?
1
CHAPTER 2
LOGISTICS/SUPPLY CHAIN STRATEGY AND PLANNING
13
The purpose of this exercise is to allow the student, in an elementary way, to examine the
tradeoffs between transportation and inventory-related costs when an incentive
transportation rate is offered. Whether the incentive rate should be implemented depends
on the shipment size corresponding to the minimum of the sum of transportation, inven-
tory, and order processing costs. These costs are determined for various shipping
quantities that might be selected to cover the range of shipment sizes implied in the
problem. Table 2-1 gives a summary of the costs to Monarch for various shipment sizes.
From Monarch's point of view, the incentive rate would be beneficial. Shipment sizes
should be approximately doubled so that the 40,000 lb. minimum is achieved. It is
important to note that the individual cost elements are not necessarily at a minimum at
low shipment sizes, whereas order-processing costs are low at high shipment sizes. They
are in cost conflict with each other. Transportation costs are low at high shipment sizes,
but exact costs depend on the minimum volume for which the rate is quoted.
In preparation for a broader planning perspective to be considered later in the text, the
student might be asked what the place of the supplier is in this decision. How does he
affect the decision, and how is he affected by it? This will focus the student's attention
on the broader issues of the physical distribution channel.
2
TABLE 2-1 Evaluation of Alternative Shipment Sizes for the Monarch Electric
Company
Current Proposed
57 motors 114 motors 171 motors 228 motors 285 motors
or or or or or
Type of cost 10,000 lb. 20,000 lb. 30,000 lb. 40,000 lb. 50,000 lb.
Transportation 98,750 58,750 58,750 38,750 38,750
RD = $78,750 = $43,750 = $43,750 = $26,250a = $26,250
Inventory 0.2520057/2 0.25200114/2 0.25200171/2 0.25200228/2 0.25200285/2
carryingb = $1,425a = $2,850 = $4,275 = $5,700 = $7,125
ICQ/2
Order processingc 5,00015/57 5,00015/114 5,00015/171 5,00015/228 5,00015/285
DS/Q = $1,316 = $658 = $439 = $329 = $263a
Handling 0.308,750 0.308,750 0.308,750 0.308,750 0.308,750
HD = $2,625 = $2,625 = $2,625 = $2,625 = $2,625
Total $84,116 $49,883 $51,089 $34,904a $36,263
a
Minimum values.
b
Students should be informed that average inventory can be approximated by one half the shipment size.
c
Demand D has been converted to units per year.
LEGEND
R = transportation rate, $/cwt.
D = annual demand, cwt.
I = inventory carrying cost, %/year.
C = cost of a motor, $/motor.
Q = shipment size in motors, where Q/2 represents the average number of motors maintained in
inventory.
S = order processing costs, $/order.
H = handling costs, $/cwt.
3
CHAPTER 3
THE LOGISTICS/SUPPLY CHAIN PRODUCT
3
The 80-20 principle applies to sales and items where 80 percent of the dollar volume is
generated from 20 percent of the product items. While this ratio rarely holds exactly in
practice, the concept does. We can apply it to these data by ranking the products by
sales, and the percentage that the cumulative sales represent of the total. The following
table shows the calculations.
Cumulative Cumulative
Product Dollar Cumulative sales as items as
code sales sales % of total % of total
08776 $71,000 $ 71,000 18.2 8.3
12121 63,000 134,000 34.3 16.7
10732 56,000 190,000 48.6 25.0
11693 51,000 241,000 61.6 33.3
10614 46,000 287,000 73.4 41.7
12077 27,000 314,000 80.3 50.0
07071 22,000 336,000 85.9 58.3
10542 18,000 336,000 90.5 66.7
06692 14,000 354,000 94.1 75.0
09721 10,000 368,000 96.7 83.3
14217 9,000 378,000 98.9 91.7
11007 4,000 391,000 100.0 100.0
Total $391,000
The 80-20 rule cannot be applied exactly, since the cumulative percent of items does
not break at precisely 20 percent. However, we might decide that only products 08776
and 12121 should be ordered directly from vendors. The important principle derived
from the 80-20 rule is that not every item is of equal importance to the firm, and that dif-
ferent channels of distribution can be used to handle them. The 80-20 rule gives some
rational basis for deciding which products should be shipped directly from vendors and
which are more economically handled through a system of warehouses.
6
(a) Reading the ground transport rates for the appropriate zone as determined by zip code
and the weight of 27 lb. (rounding upward of 26.5 lb.) gives the following total cost
table for the four shipments.
To Catalog UPS Transport Total

zip code price zone costa cost
11101 $99.95 2 $ 8.39 $108.34
42117 99.95 5 10.46 110.41
74001 99.95 6 13.17 113.12
59615 99.95 8 18.29 118.24
a
Use 27 lb.
4
(b) The transport rate structure is reasonably fair, since ground rates generally follow
distance and size of shipment. These are the factors most directly affecting transport
costs. They are not fair in the sense that customers within a zone are all charged the
same rate, regardless of their distance from the shipment origin point. However, all
customers may benefit from lower overall rates due to this simplified zone-rate
structure.
10
(a) This is a delivered pricing scheme where the seller includes the transport charges in
the product price. The seller makes the transport arrangements.
(b) The seller prices the product at the origin, but prepays any freight charges; however,
the buyer owns the goods in transit.
(c) This is a delivered pricing scheme where the freight charges are included in the
product price, however the freight charges are then deducted from the invoice, and
the seller owns the goods in transit.
(d) The seller initially pays the freight charges, but they are then collected from the buyer
by adding them to the invoice. The buyer owns the goods in transit, since the pricing
is f.o.b. origin.
(e) The price is f.o.b. origin. The buyer pays the freight charges and owns the goods in
transit.
Regardless of the price policy, the customer will ultimately pay all costs. If a firm
does not consider outbound freight charges, the design of the distribution system will be
different than if it does. Since pricing policy is an arbitrary decision, it can be argued that
transport charges should be considered in decision making, whether the supplying firm
directly incurs them or not.
11
This shows how Pareto's law (80-20 principle) is useful in estimating inventory levels
when a portion of the product line is to be held in inventory. An empirical function that
approximates the 80-20 curve is used to estimate the level of sales for each product to be
held in inventory. According to Equation 3-2, the constant A is determined as follows:
X (1 Y ) 0.25(1.75)
A= = = 0125
.
YX 0.75 0.25
The 80-20 type curve according to Equation 3-1 is:
(1 + A) (1 + 0125
. )X
Y= =
A+ X 0125
. +X
This formula can be used to estimate the cumulative sales from the cumulative item
proportion. For example, item 1 is 0.05 of the total number of items (20) so that:
5
(1 + 0125
. )( 0.05)
Y= = 0.321
.
0125 + 0.05
Of the $2,600,000 in total annual warehouse sales, item 1 should account for
0.3212,600,000 = $835,714.
By applying this formula to all items, the following inventory investment table can be
developed which shows sales by item. The average inventory investment by item is
found by dividing the turnover ratio into the item sales. The sum of the average
inventory value for each item gives a total projected inventory of $380,000.
Inventory Investment Table

Cumulative Average
Pro- item pro- Cumulative Projected Turnover inventory
duct portion, X sales, Y item sales ratio value
1 0.05 $ 835,714 $ 835,714 8 $104,464
2 0.10 1,300,000 464,286 8 58,036
A
3 0.15 1,595,454 295,454 8 36,932
4 0.20 1,800,000 204,546 8 25,568
5 0.25 1,950,000 150,000 6 25,000
6 0.30 2,064,705 114,706 6 19,118
7 B 0.35 2,155,263 90,558 6 15,093
8 0.40 2,228,571 73,308 6 12,218
9 0.45 2,289,130 60,559 6 10,093
10 0.50 2,340,000 50,870 6 8,478
11 0.55 2,383,333 43,333 4 10,833
12 0.60 2,420,689 37,356 4 9,339
13 0.65 2,453,226 32,537 4 8,134
14 0.70 2,481,818 28,592 4 7,148
15 C 0.75 2,507,142 25,324 4 6,331
16 0.80 2,529,719 22,587 4 5,647
17 0.85 2,550,000 20,271 4 5,068
18 0.90 2,568,293 18,293 4 4,473
19 0.95 2,584,884 16,591 4 4,148
20 1.00 2,600,000 15,116 4 3,779
Total $380,000
12
This problem involves the application of Equations 3-1 and 3-2. We can develop an 80-
20 curve based on 30 percent of the items accounting for 70 percent of sales. That is,
X (1 Y ) 0.30(1 0.70 )
A= = = 0.225
YX 0.70 0.30
Therefore, the sales estimating equation is:
6
(1 + 0.225) X
Y=
0.225 + X
By applying this estimating curve, we can find the sales of A and B items. For
example, 20 percent of the items, or 0.220 = 4 items, will be A items with a cumulative
proportion of sales of:
(1 + 0.225)( 0.20 )
YA = = 0.5765
0.225 + 0.20
and 3,000,0000.5765 = 1,729,412.
The A+B item proportion will be:
(1 + 0.225)( 0.50 )
YA+ B = = 0.8448
0.225 + 0.50
and 3,000,0000.8448 = 2,534,400. The product group B sales will be A+B sales less A
sales, or 2,534,400 1,729,412 = $804,988.
The product group C will be the remaining sales, but these are not of particular
interest in this problem.
The average inventories for A and B products are found by dividing the estimated
sales by the turnover ratio. That is,
A: 1,729,412/9 = 192,157
B: 804,988/5 = 160,988
Total inventory 353,155 cases
The total cubic footage required for this inventory would be 353,1551.5 = 529,732
cu. ft. The total square footage for products A and B is divided by the stacking height.
That is, 529,731/16 = 33,108 sq. ft.
13
This problem is an application of Equations 3-1 and 3-2. We first determine the constant
A. That is,
X (1 Y ) 0.20(1 0.65)
A= = = 0156
.
YX 0.65 0.20
and
(1 + 0156
. )X
0.75 =
0156
. +X
Solving algebraically for X, we have:
7
AxY 0156
. x 0.75
X= = = 0.288
1 + A Y 1 + 0156
. 0.75
That is, about 29 percent of the items (0.2885,000 = 1,440 items) produce 75 percent of
the sales.
14
The price would be the sum of all costs plus an increment for profit to place the
automotive component in the hands of the customer. This would be
25+10+5+8+5+transportation cost, or 53+T. Based on the varying transportation cost,
the following price schedule can be developed.
Quantity Price per unit Discount

1 to 1,000 units 53+5=$58 0
1,001 to 2,000 units 53+4.00=57 1.7%a
>2,000 units 53+3.00=56 3.5%
a
[(58 - 57)/58][100]=1.7%
8
CHAPTER 4
LOGISTICS/SUPPLY CHAIN CUSTOMER SERVICE
6
(a) This company is fortunate to be able to estimate the sales level that can be achieved at
various levels of distribution service. Because of this, the company should seek to
maximize the difference between sales and costs. These differences are summarized
as follows:
Percent of orders delivered

within one day
Contribution to 50 60 70 80 90 95 100
profit -1.8 2.0 3.5 4.0 3.4 2.8 -2.0
The company should strive to make deliveries within one day, 80 percent of the time,
for a maximum contribution to profit.
(b) If a competing company sets its delivery time so that more than 80 percent of the
orders are delivered in one day and all other factors that attract customers are the
same, the company will lose customers to its competitor, as the sales curve will have
shifted downward. Cleanco should adjust its service level once again to the point
where the profit contribution is maximized. Of course, there is no guarantee that the
previous level of profits can be achieved unless the costs of supplying the service can
correspondingly be reduced.
7
(a)
The first task is to determine the increase in sales that can be attributed to the
change in the service policy. To determine if there is a significant change in the
control group, we set up the following hypothesis test:
X 2 X1 224 185 39
z= = = = 394
.
s 2
s 2 2
61 79 2
36.48 + 6118
.
2
+ 1
+
N 2 N1 102 102
Now, referring to the normal distribution table in Appendix A of the text, there is
a significant difference at the 0.01 level in the sales associated with the control group.
That is, some factors other than the service policy alone are causing sales to increase.
Next, we analyze the test group in the same manner.
9
2,295 1,342 953
z= = = 10.7
5762 3352 5,924 + 2,004
+
56 56
This change is also significant at the 0.01 level.

The average increase in sales for the control group is 224/185 = 1.21, or 21
percent. The average sales increase in the test group is 2295/1342 = 1.71, or 71
percent. If we believe that 21percent of the 71percent increase in the test group is due
to factors other than service policy, then 71 21 = 50 percent was the true service
effect. Therefore, for each sales unit, an incremental increase in profit of
(0.4095)(0.50) = $19 can be realized. Since the cost of the service improvement is
$2, the benefit exceeds the cost. The service improvement should be continued.
(b) The use of the before-after-with-control-group experimental design is a methodology

that has been used for some time, especially in marketing research studies. The
outstanding feature of the design is that the use of the control group helps to isolate
the effect of the single service variable. On the other hand, there are a number of
potential problems with the methodology:
The sales distributions may not be normal.

The time that it takes for diffusing the information that a service change has taken
place may distort the results.
The products in the control group may not be mutually exclusive from those in the
test group.
The method only shows the effect of a single step change in service and does not
develop a sales-service relationship.
It may not always be practical to introduce service changes into on-going operations
to test the effect.
8
(a) The optimum service level is set at that point where the change in gross profit equals
the change in cost.
The change in gross profit:
P = Trading margin Sales response rate Annual sales

= 1.000.0015100,000
= $150 per year per 1 percent change in the service level
The change in cost:
C = Annual carrying cost Standard product cost z
10
Demand standard deviation for order cycle
= 0.3010.00400z
Now, set P = C and solve for z.
150 = 1200z
z = 0.125
From the tabulated changes in service level with those changes in z, the service level
should be set between 96-97 percent.
(b) The weakest link in this analysis is estimating the effect that a change in service will
have on revenue. This implies that a sales-service relationship is known.
9
The methodology is essentially the same as that in question 7, except that we are asked to
find X instead of Y. That is,
P = 0.750.001580,000
= 90
and
C = 0.251,000500z
= 1250z
Then,
P= C
90 = 1250z
z = 0.072
From the normal distribution (see Appendix A), the z for an area under the curve of
93 percent is 1.48, and for 92 percent, z is 1.41. Since the difference of 1.48 1.41 =
0.07, we can conclude that the in-stock probability should be set at 92-93 percent. Of
course, the change in z is found by taking the difference in z values for one percent
differences in the area values under the normal distribution curve for a wide range of area
percentages.
10
Apply Taguchis concept of the loss function. First, estimate the loss per item if the
target level of service is not met. We know the profit per item as follows.
11
Sales price $5.95
Cost of item -4.25
Other costs -0.30
Profit per item $1.40
Since one-half of the sales are lost, the opportunity loss per item would be:
Profit per item Sales lost
$1.40 (1/2)(880)
Opportunity loss = = $0.70 /item
880
Current sales
Next, find k in the loss function:
L = k ( y m) 2 out-of-stock % at point where sales are lost

0.70 = k (10 5)2
0.70 = k ( 25) Target %
k = 0.03
Finally, the point where the marginal supply cost equals the marginal sales loss is:
B 0.10
( y 5) = = = 1.67%
2k 2(0.03)
y = 1.67 + 5 = 6.67%
The retailer should not allow the out-of-stock percentage to deviate more than 1.67
percent, and should not allow the out-of-stock level to fall below 1.67 + 5 = 6.67 percent.
12
CHAPTER 7
TRANSPORT DECISIONS
1
Selecting a mode of transportation requires balancing the direct cost of transportation
with the indirect costs of both vendor and buyer inventories plus the in-transit inventory
costs. The differences in transport mode performance affect these inventory levels, and,
therefore, the costs for maintaining them, as well as affect the time that the goods are in
transit. We wish to compare these four cost factors for each mode choice as shown in
Table 7-1 of the manual. The symbols used are:
R = transportation rate, $/unit

D = annual demand, units
C = item value at buyer's inventory, $
C' = item value at vendor's inventory, $
T = time in transit, days
Q = Shipping quantity, units
Rail has the lowest total cost.
TABLE 7-1 An Evaluation of the Transport Alternatives for the Wagner

Company
Cost type Method Rail Piggyback Truck
Transport RD 2550,000 4450,000 8850,000
= $1,250,000 = $2,200,000 = $4,400,000
In-transit 0.2547550,000 0.2545650,000 0.2541250,000
inventorya ICDt/365 (16/365) (10/365) (4/365)
= $260,274 = $156,164 = $564,384
Wagers 0.25475(10,000/2) 0.25456(7,000/2) 0.25412(5,000/2)
inventorya ICQ/2 = $593,750 = $399,000 = $257,500
Electronics 0.25500(10,00/2) 0.25500(7,000/2) 0.25500(5,000/2)
inventory ICQ/2 = $625,000 = $437,500 = $312,500
Total $2,729,024 $3,192,664 $5,534,384
a
C refers to price less transport cost per unit.
2
As in question 1, this problem is one of balancing transport costs with the indirect costs
associated with inventories. However, in this case we must account for the variability in
transit time as it affects the warehouse inventories. We can develop the following
decision table.
17
Service type
Cost type Method A B
129,600 11.809,600
Transport RD = $115,200 =$114,048
0.20509,600 0.20509,600
In-transit ICDt/365 (4/365) (5/365)
inventory = $1,052 = $1,315
Plant ICQ*/2 0.3050(321.8/2) 0.3050(357.8/2)
inventory = $2,684 = $2,684
0.3062(321.3/2) 0.3061.80(321.8/2)
Warehouse ICQ*/2 + 0.306250.5 + 0.3061.8060.6
inventory + ICr = $3,927 = $4,107
Total $122,863 $122,154
Recall that Q* = 2 DS / IC = 2(9,600)(100) / 0.3(50) = 357.8 cwt. for the plant, assuming
the order cost is the same at plant and warehouse. However, for the warehouse, we must
account for safety stock (r) and for the transportation cost in the value of the product.
Therefore,
For A:
Q * = 2 DS / IC = 2(9,600)(100) / 0.3( 62) = 3213

. cwt.
and for z = 1.28 for an area under the normal distribution of 0.90, the safety stock is:
r = zs LT ( d ) = 1.28 1.5 (9,600 / 365) = 50.5 cwt.
For B:
Q * = 2(9,600)(100) / 0.3( 6180

. ) = 3218
. cwt.
and
r = 1.28 1.8 (9,600 / 365) = 60.6 cwt.
Service B appears to be slightly less expensive.
3
The shortest route method can be applied to this problem. The computational table is
shown in Table 7-2. The shortest route is defined by tracing the links from the
destination node. They are shown in Table 7-2 as A D F G for a total distance of
980 miles.
18
TABLE 7-2 Tabulation of Computational Steps for the Shortest Route Method
Applied to Transcontinental Trucking Company Problem
Solved nodes
directly Its closest
connected to connected
unsolved unsolved Total time nth nearest Its minimum Its last
Step nodes node involved node time connectiona
1 A B 186 mi. B 186 mi. AB
A D 276
2 A D 276 D 276 AD*
B C 186+110= 296
3 B C 186+110= 296 C 296 BC
D C 276+ 58= 334
D F 276+300= 576
4 C E 296+241= 537 E 537 CE
C F 296+350= 646
D F 276+300= 576
5 C F 296+350= 646
E G 537+479=1016
D F 276+300= 576 F 576 DF*
6 E G 537+479=1016
F G 576+404= 980 G 980 FG*
a
Asterisk indicates the shortest route
4
In this actual problem, the U.S. Army used the transportation method of linear
programming to solve its allocation problem. The problem can be set up in matrix form
as follows:
Origin Cleve- South San

Destination land Charleston Jose Demand
150 100 800
Letterkenny 150 150 300
325 350 300
Fort Hood 50 50 100
275 325 350
Fort Riley 100 100
375 400 275
Fort Carson 100 100
300 250 450
Fort Benning 100 0 100
Supply 400 150 150
The cell values shown in bold represent the number of personnel carriers to be moved
between origin and destination points for minimum transportation costs of $153,750. An
alternative solution at the same cost would be:
19
Number of
Origin Destination carriers
Cleveland Letterkenny 150
S. Charleston Letterkenny 150
Cleveland Fort Hood 50
San Jose Fort Hood 50
Cleveland Fort Riley 100
San Jose Fort Carson 100
Cleveland Fort Benning 100
The general
approach is to first find the route in ROUTER without regard to the rectilinear distances
of the road network. Because this may produce an infeasible solution, specific travel
distance are added to the database to represent actual distances traveled or to block
infeasible paths from occurring. A reasonable routing plan is shown in Figure 7-1 and
the ROUTER database that generates it is given in Figure 7-2. The total distance for the
route is 9.05 miles and at a speed of 20 miles per hour, the route time is approximately 30
minutes.
20
FIGURE 7-1 ROUTER Solution to the School Bus Routing Exercise
0.5 1.0 1.5 2.0

0
19 17
21 20
0
0.5
22 8 15 16
5 18
9
1.0
3 4 6 7 10 11 12 13 14
1.5
Denotes pickup point
Depot location
FIGURE 7-2 Input Data for ROUTER for School Bus Routing Problem
PARAMETERS AND LABELS
Problem label School Bus Routing Exercise

Grid corner with 0,0 coordinates (NW, SW, SE, or NE) - NW
DEPOT DATA
Depot description - Atlanta Located in zone - 0
Horizontal coordinate 0.14 Vertical coordinate 0.45
Earliest starting time (min) - 0 Latest return time (min) - 9999
Default vehicle speed (miles per hour) - 20
After how many clock hours will overtime begin - 9999
GENERAL DATA
Percent of vehicle in use before allowing pickups - 0
Horizontal scaling factor - 1 Vertical scaling factor - 1
Maximum TIME allowed on a route (hours) - 9999
Maximum DISTANCE allowed on a route (miles) - 9999
LOAD/UNLOAD TIME FORMULA
Fixed time per stop - 0
Variable time per stop by weight - 0 By cube - 0
BREAK TIMES
Duration of 1st break (minutes) - 0 To begin after - 9999
21
Duration of 2nd break (minutes) - 0 To begin after - 9999
Duration of 3rd break (minutes) - 0 To begin after - 9999
Duration of 4th break (minutes) - 0 To begin after - 9999
--STOP DATA
STOP LOAD VOL. LOAD

NO DESCRIPTION TY WGHT CUBE HCRD VCRD ZN TIME BEG1 END1 BEG2 END2
1 Stop 1 D 1 0 0.14 0.80 0 0 0 9999 9999 9999
2 Stop 2 D 1 0 0.14 1.14 0 0 0 9999 9999 9999
3 Stop 3 D 1 0 0.14 1.31 0 0 0 9999 9999 9999
4 Stop 4 D 1 0 0.35 1.31 0 0 0 9999 9999 9999
5 Stop 5/22 D 1 0 0.52 0.61 0 0 0 9999 9999 9999
6 Stop 6 D 1 0 0.58 1.31 0 0 0 9999 9999 9999
7 Stop 7 D 1 0 0.80 1.31 0 0 0 9999 9999 9999
8 Stop 8 D 1 0 1.03 0.61 0 0 0 9999 9999 9999
9 Stop 9 D 1 0 1.03 0.96 0 0 0 9999 9999 9999
10 Stop 10 D 1 0 1.03 1.31 0 0 0 9999 9999 9999
11 Stop 11 D 1 0 1.36 1.31 0 0 0 9999 9999 9999
12 Stop 12 D 1 0 1.48 1.31 0 0 0 9999 9999 9999
13 Stop 13 D 1 0 1.80 1.31 0 0 0 9999 9999 9999
14 Stop 14 D 1 0 1.87 1.31 0 0 0 9999 9999 9999
15 Stop 15 D 1 0 1.84 0.61 0 0 0 9999 9999 9999
16 Stop 16 D 1 0 1.95 0.61 0 0 0 9999 9999 9999
17 Stop 17 D 1 0 1.29 0.10 0 0 0 9999 9999 9999
18 Stop 18 D 1 0 1.26 0.61 0 0 0 9999 9999 9999
19 Stop 19 D 1 0 1.15 0.10 0 0 0 9999 9999 9999
20 Stop 20 D 1 0 0.69 0.23 0 0 0 9999 9999 9999
21 Stop 21 D 1 0 0.14 0.26 0 0 0 9999 9999 9999
VEHICLE DATA
--CAPACITY-- --VEHICLE-- --DRIVER--

OVER
VEHICLE FIXED PER MI FIXED PER HR TIME
NO. DESCRIPTION TP NO WGHT CUBE COST COST COST COST COST
1 Bus 1 1 9999 9999 0 0 0 0 0
SPECIFIED STOP-TO-STOP DISTANCES
STOP STOP STOP STOP DISTANCE

1 14 Stop 14 16 Stop 16 0.78
2 14 Stop 14 15 Stop 15 0.90
3 15 Stop 15 17 Stop 17 1.06
4 16 Stop 16 17 Stop 17 1.18
5 18 Stop 18 9 Stop 9 0.58
6 19 Stop 19 8 Stop 8 0.76
7 19 Stop 19 20 Stop 20 0.59
8 19 Stop 19 5/22 Stops5&22 1.14
9 19 Stop 19 18 Stop 18 0.53
10 9 Stop 9 20 Stop 20 1.08
11 9 Stop 9 19 Stop 19 1.11
12 9 Stop 9 21 Stop 21 1.69
13 5/22 Stops 5&22 1 Stop 1 0.56
14 5/22 Stops 5&22 21 Stop 21 1.05
15 5/22 Stops 5&22 20 Stop 20 1.14
16 5/22 Stops 5&22 9 Stop 9 0.97
17 20 Stop 20 21 Stop 21 0.84
22
17 20 Stop 20 21 Stop 21 0.84
18 20 Stop 20 0 School 1.03
19 20 Stop 20 5/22 Stops 5&22 0.55
20 17 Stop 17 0 School 2.43
21 0 School 5/22 Stops 5&22 1.37
22 2 Stop 2 5/22 Stops 5&22 1.03
6
Strategy 1 is to stay at motel M2 and serve the two routes on separate days. Using the
ROUTESEQ module in LOGWARE gives us the sequence of stops and the coordinate
distance. The routes originating at M2 would be:
Route Stop sequence Distancea

1 8,6,1,4,2,3,5,7,9 95.55 mi.
2 10,13,14,17,18,16,12,15,11 86.45
182.00 mi.
a
Includes map scaling factor
The total cost of this strategy would be:
Motel 3 nights @ 49.00 $147.00

Travel 182 miles @ $.30/mi. 54.60
Total $201.60
Strategy 2 is a mixed strategy involving staying at motels closest to the center of the
stop clusters. The route sequences from different motels are:
Route Stop sequence Distance

1 4,2,3,5,7,9,8,6,2 98.50 mi.
2 18,17,13,14,10,11,15,12,16 80.30
178.80 mi.
The total cost of this strategy is:
Motel M1 1st night $ 40.00

M1 2nd night 40.00
rd
M1 3 night 45.00
Travela 214.80 mi. @ 0.30/mi. 64.44
Total $189.44
a
178.80 + 36 = 214.80
Strategy 2 appears to be most economical.
7
(a) Since distances are asymmetrical, we cannot use the geographically based traveling
salesman method in LOGWARE. Rather, we use a similar module in STORM that
allows such asymmetrical matrices, or the problem is small enough to be solved by
inspection. For this problem, the minimal cost stop sequence would be:
23
Bakery!Stop 5!Stop 3!Stop 4!Stop 2!Stop 1!Bakery
with a tour time of 130 minutes.
(b) Loading/unloading times may be added to the travel times to a stop. The problem
may then be solved as in part a.
(c) The travel times between stop 3 and all other nodes are increased by 50 percent. The
remaining times are left unchanged. Optimizing on this matrix shows no change in
the stop sequence. However, the tour time increases to 147.50 minutes.
8
This may be solved by using the ROUTER module in LOGWARE. The screen set up for
this is as follows:
24
FIGURE 7-3 Input Data for ROUTER for Sima Donuts
--PARAMETERS AND LABELS
Problem label - Sima Donuts

Grid corner with 0,0 coordinates (NW, SW, SE, or NE) - NE
DEPOT DATA
Depot description - Atlanta Located in zone - 0
Horizontal coordinate - 2084 Vertical coordinate - 7260
GENERAL DATA
Horizontal scaling factor - 0.363 Vertical scaling factor - 0.363
BREAK TIMES
Duration of 4th break (minutes) - 60 To begin after - 2640
--STOP DATA
STOP LOAD VOL. LOAD
1 Tampa FL D 20 0 1147 8197 0 15 360 1440 1800 2880
2 Clearwater FL P 14 0 1206 8203 0 45 360 1440 1800 2880
3 Daytona Beach F D 18 0 1052 7791 0 45 360 1440 1800 2880
4 Ft Lauderdale FL D 3 0 557 8282 0 45 180 1440 1800 2880
5 N Miami FL D 5 0 527 8341 0 45 360 1440 1800 2880
6 Oakland Park FL P 4 0 565 8273 0 45 180 1440 1800 2880
7 Orlando FL D 3 0 1031 7954 0 45 180 1440 1800 2880
8 St Petersburg FL P 3 0 1159 8224 0 45 180 1440 1800 2880
9 Tallahassee FL D 3 0 1716 7877 0 15 600 1440 1800 2880
10 W Palm Beach F D 3 0 607 8166 0 45 360 1440 1800 2880
11 Puerto Rico D 4 0 527 8351 0 45 360 1440 1800 2880
--VEHICLE DATA
-CAPACITY-- --VEHICLE-- --DRIVER--
OVER
1 Truck #1-20 1 3 20 9999 0 1.30 0 0 0
2 Truck #2-25 2 1 25 9999 0 1.30 0 0 0
3 Truck #3-30 3 1 30 9999 0 1.30 0 0 0
Making a run with ROUTER will give the route design.
25
Pickup
Pickup
FIGURE 7-4 Graphical Display of Route Design for Sima Donuts
The route design involves 3 routes for a total distance of 3,830 miles, a cost of $4,978.71,
and a total time of 100.4 hours. The route details are as follows:
Route #1 with 20-pallet truck
Depot Start time 3:00 a.m. of day 1
Daytona Beach Deliver 18 pallets
Clearwater Pickup 14 pallets
Depot Return time 5:48 a.m. of day 2
Orlando Deliver 3 pallets
W Palm Beach Deliver 3 pallets
Ft Lauderdale Deliver 3 pallets
N Miami Deliver 5 pallets
Miami-Puerto R. Deliver 4 pallets
Depot Return time 4:43 p.m. of day 2
26
Tallahassee Deliver 3 pallets
Tampa Deliver 20 pallets
St Petersburg Pickup 3 pallets
Oakland Park Pickup 4 pallets
Depot Return time 4:03 p.m. of day 2
9
Given sailing times and dates when deliveries are to be made, loadings need to be
accomplished no later than the following dates:
To: A B C D
From: 1 16 40 1
2 69 25 5
The problem can be expressed as a transportation problem of linear programming. There

will be 6 initial states [(1,1), (2,5), (1,16), (2,25), (1,40), and (2,69)] and 6 terminal states
[(D,10), (C,15), (A,36), (B,39), (C,52), and (A,86)]. The linear program is structured as
shown in Figure 7-4.
Using a transportation solution method, we determine one of the optimum solutions.
There are several. The solution is read by starting with the slack on initial loading state
1. This tells us to next select the cell of terminal state 1. In turn, this defines initial state
3 and hence, terminal state 3. And so it goes until we reach the terminal state slack
column. This procedure is repeated until all initial state slacks are exhausted. Our
solution shows two routings. The first is
(1,1)!(D,10)! (1,16)!(A,36)!(2,69)!(A,86).
The second is (2,5)!(C,15)!(2,25)!(B,39)!(1,40)!(C,52). Two ships are needed.
27
FIGURE 7-5 Transportation Matrix Setup and Solution for the Queens Lines
Tanker Scheduling Problem
Loading points and dates

Load 1 2 1 2 1 2
date 1 5 16 25 40 69
Discharge Rim re-
date Slack striction
100 100 1 1 1 1 10
D 10 XXa XX 1 1
100 100 1 1 1 1 10
C 15 XX XX 1 1
100 100 100 100 1 1 10
A 36 XX XX XX XX 1 1
100 100 100 100 1 1 10
B 39 XX XX XX XX 1 1
100 100 100 100 100 1 10
C 52 XX XX XX XX XX 1 1
100 100 100 100 100 100 10
A 86 XX XX XX XX XX XX 1 1
10 10 10 10 10 10 10
Slack 1 1 4 6
Rim re-
striction 1 1 1 1 1 1 6
a
XX inadmissible cells given a high cost
10
This is a problem of freight consolidation brought about by holding orders so they can be
shipped with orders from subsequent periods. The penalty associated with holding the
orders is a lost sales cost.
(i) Orders shipped as received

Weight Rate = Cost
Haysa 10,000 0.0519 = $519.00
Manhattan 14,000 0.0519 = 726.00
Salina 13,000 0.0408 = 530.00
Great Bend 10,000 0.0498 = 498.00
Transportation $2,274.00
Lost sales .00
Total $2,274.00
a
Ship 8,000 lb. as if 10,000 lb.
Average period cost is $2,274.00
28
(ii) Consolidate first period orders with second period orders
Weight Rate = Cost
Hays 16,000 0.0519 = $830.40
a
Manhattan 40,000 0.0222 = 888.00
Salina 26,000 0.0342 = 889.20
Great Bend 10,000 0.0498 = 498.00
Lost sales 1,050.00
Total $4,155.60
a
Ship 28,000 lb. as if 40,000 lb.
The lost sales cost is 1,000 cases $1.05 = $1,050.00 to hold one group of orders for 2
weeks.
Average cost per period is $4,155.60/2 = $2,077.80.
(iii) Hold all orders until the third period.

Weight Rate = Cost
Hays 24,000 0.0426 = $1,022.40
Manhattan 42,000 0.0222 = 932.40
a
Salina 40,000 0.0246 = 984.00
Great Bend 15,000 0.0498 = 747.00
Lost sales 3,150.00
Total $6,835.80
a
Ship as if 40,000 lb.
Lost sales
Hold 1st period orders for 2 periods 1,0001.05.2 = $ 2,100
Hold 2nd period sales for 1 period 1,0001.05 = 1,050
$ 3,150
Average period cost is $6,835.80/3 = $2,278.60
Summary
Ship immediately $ 2,274.00

Hold orders 1 period 2,077.80 Optimum
Hold orders 2 periods 2,278.60
11
Routes are built by placing the trips end-to-end throughout the day from 4 a.m. until 11
p.m., respecting the times that a warehouse can receive a shipment. This is a 19-hour
block of time per day, or there are 95 hours per week per truck in which a truck may
operate. If there were no delivery time restrictions on warehouses and trips could be
29
placed end-to-end for a truck without any slack at the end of the day, the absolute
minimum number of trucks can be found multiplying the number of trips by the route
time, and then dividing the total by the 95 hours allowed per week. That is,
(1) (2) (3)=(1)(2)

Warehouse Number of Total time Total time,
location trips per trip, hr. hr.
Flint 43 1.25 53.75
Alpena 5 10.50 52.50
Saginaw 8 2.25 18.00
Lansing 21 3.75 78.75
Mt. Pleasant 12 5.50 66.00
W. Branch 5 6.00 30.00
Pontiac 43 2.75 118.25
Traverse City 6 10.50 63.00
Petoskey 5 11.75 58.75
Total 539.00
For 539 trip hours, 539/95 = 5.67 rounded to six trucks needed per week. Now, it is
necessary to adjust for the problem constraints. A good schedule can be found by
following a few simple rules that can be developed by examining the data. First, begin
the day with a trip where the driving time to a warehouse is just long enough for the truck
to arrive at the warehouse just after it opens. One-half the driving time should exceed
6:30 4:00 = 2:30, or 2 hr. Trips to Alpena, Traverse City, and Petoskey qualify.
Second, use the short trips at the end of the day to avoid slack time. Third, allocate the
trips to the days using the longest ones first. Make sure that the total trip time for a day
does not exceed 19 hours. For a minimum of six trucks, the following feasible schedule
can be developed by inspection.
Day 1 Day 2 Day 3 Day 4 Day 5

Truck 1 Petoskey 11.75 Petoskey 11.75 Petoskey 11.75 Petoskey 11.75 Petoskey 11.75
W Branch 6.00 W Branch 6.00 W Branch 6.00 5 Flint 6.25 5 Flint 6.25
Flint 1.25 Flint 1.25 Flint 1.25
Total =19.00 hr. Total =19.00 hr. Total =19.00 Total =18.00 hr. Total =18.00 hr.
Truck 2 T. City 10.50 T. City 10.50 T. City 10.50 T. City 10.50 T. City 10.50
2 Lansing 7.50 2 Lansing 7.50 2 Lansing 7.50 2 Lansing 7.50 2 Lansing 7.50
Total =18.00 hr. Total =18.00 hr. Total =18.00 Total =18.00 hr. Total =18.00 hr.
Truck 3 T. City 10.50 Alpena 10.50 Alpena 10.50 Alpena 10.50 Alpena 10.50
2 Lansing 7.50 2 Lansing 7.50 2 Lansing 7.50 2 Lansing 7.50 2 Lansing 7.50
Total =18.00 hr. Total =18.00 hr. Total =18.00 hr. Total =18.00 hr. Total =18.00 hr.
Truck 4 Alpena 10.50 M Pleasant 5.50 M Pleasant 5.50 M Pleasant 5.50 M Pleasant 5.50
Lansing 3.75 4 Pontiac 11.00 4 Pontiac 11.00 4 Pontiac 11.00 4 Pontiac 11.00
3 Flint 3.75 2 Flint 2.50 2 Flint 2.50 2 Flint 2.50 2 Flint 2.50
Total =18.00 hr. Total =19.00 hr. Total =19.00 hr. Total =19.00 hr. Total =19.00
Truck 5 M Pleasant 5.50 M Pleasant 5.50 M Pleasant 5.50 M Pleasant 5.50 M Pleasant 5.50
4 Pontiac 11.00 4 Pontiac 11.00 4 Pontiac 11.00 4 Pontiac 11.00 4 Pontiac 11.00
2 Flint 2.50 2 Flint 2.50 Flint 1.25
Total =19.00 hr. Total =19.00 hr. Total =17.75 Total =16.50 hr. Total = 16.50
Truck 6 M Pleasant 5.50 M Pleasant 5.50 M Pleasant 5.50 W Branch 6.00 W Branch 6.00
4 Pontiac 11.00 3 Pontiac 8.25 6 Saginaw 13.50 2 Saginaw 4.50 10 Flint 12.50
2 Flint 2.50 2 Flint 2.50
Total =19.00 hr. Total =16.25 hr. Total =19.00 hr. Total =10.50 hr. Total =18.50 hr.
30
Although this schedule meets the requirements of the problem, it might be improved by
better balancing the workload across the trucks and the days.
12
(a) A sweep method solution is shown on the following figure. Five trucks are needed
with a total route distance of (30+29+39+44+19.5)10 = 1,615 miles.
20 Route #5
2 3 Load 9 3 1
18 3 4 2
Route #1 5
Load 19 16
4
14 4 3 3 4 Route #4
Load `8
Miles 12
2
x 10 3
10 3
Warehouse 5
8 5
6 4
Route #2 6 Route #3
Load 20 Load 17
4 5 3
2
2
4
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26
Miles x 10
(b) The sweep method is a fast and relatively simple method for finding a solution to
rather complex vehicle routing problems. Solutions can be found graphically without
the aid of a computer. However, there are some limitations. Namely,
The method is heuristic and has an average error of about 10 to 15 percent. This
error is likely to be low if the problem contains many points and the weight of
each point is small relative to the capacity of the vehicle.
The method does not handle timing issues well, such as time windows.
Too many trucks may be used in the route design.
13
This problem may be solved with the aid of ROUTER in LOGWARE. The model input
data may be formatted as shown in Figure 7-6.
31
(a) The solution from ROUTER shows that four routes are needed with a minimum total
distance of 492 miles. The route design is shown graphically in Figure 7-7. A
summary for these routes is given in following partial output report.
Route Run Stop Brk Stem

Route time, time, time, time, time, Start Return No of Route Route
no hr hr hr hr hr time time stops dist,Mi cost,$
1 1.2 1.0 .3 .0 .4 08:59AM 10:12AM 3 29 .00
2 8.9 6.6 1.3 1.0 1.1 08:32AM 05:25PM 19 199 .00
3 6.2 3.7 1.4 1.0 .9 08:42AM 02:54PM 14 112 .00
4 7.5 5.1 1.5 1.0 1.4 08:30AM 04:02PM 12 152 .00
Total 23.8 16.4 4.4 3.0 3.8 48 492 .00
(b) Note that route #1 is short and that a driver and a station wagon would be used for a
route that takes 1.2 hours to complete. By attaching route #1 to route #3, the same
driver and station wagon may be used, and the constraints of the problems are still
met. The refilled station wagon can leave the depot by 3:30-3:45 p.m. and still meet
the customers time windows and return to the depot by 6 p.m.. Thus, only three
drivers and station wagons are actually needed for this problem.
FIGURE 7-6 Input Data for ROUTER for Medic Drugs

--PARAMETERS AND LABELS
Problem label - Medic Drugs

Grid corner with 0,0 coordinates (NW, SW, SE, or NE) - SW
DEPOT DATA
Depot description - Pharmacy Located in zone - 0
Horizontal coordinate - 13.7 Vertical coordinate - 21.2
GENERAL DATA
Horizontal scaling factor - 4.6 Vertical scaling factor - 4.6
BREAK TIMES
Duration of 4th break (minutes) - 0 To begin after 9999
32
--STOP DATA
STOP LOAD VOL. LOAD
1 Covington House D 1 0 23.40 12.90 0 2 540 1020 9999 9999
2 Cuyahoga Falls D 9 0 13.40 13.40 0 18 540 1020 9999 9999
3 Elyria D 1 0 6.30 16.80 0 5 540 1020 9999 9999
46 Westbay D 6 0 8.40 18.00 0 10 630 690 9999 9999
47 Westhaven D 2 0 8.50 18.10 0 5 540 1020 9999 9999
48 Broadfiels Mnr D 6 0 18.20 22.90 0 2 540 1020 9999 9999
--VEHICLE DATA
-CAPACITY-- --VEHICLE-- --DRIVER--
OVER
1 Station wagon 1 50 63 9999 0 0 0 0 0
FIGURE 7-7 Graphical

Solution for Medic Drugs
14
There is no exact answer to this problem nor is one intended. Several approaches might
be taken to this problem. We could apply the savings method or the sweep method to
solve the routing problem for each day of the week, given the current demand patterns.
However, we can see that there is much overlap in the locations of the customers by
delivery day of the week. We might encourage orders to be placed so that deliveries
form tight clusters by working with the sales department and the customers. Perhaps
some incentives could be provided to help discipline the order patterns. The orders
should form a general pattern as shown below. Currently, the volume for Thursday
exceeds the available truck capacity of 45 caskets. Maybe the farthest stops could be
handled by a for-hire service rather than acquiring another truck for such little usage.
33
Monday Tuesday
Friday
Depot
Wednesday Thursday
It appears that the truck capacity is about right, given that some slack capacity is
likely to be needed.
Once the pattern orders are established, either as currently given or as may be revised,
apply principles numbers 1, 3, 4, 5, and 7.
34
CHAPTER 8
FORECASTING SUPPLY CHAIN REQUIREMENTS
4
(a) The answer to this question is aided by using the FORECAST module in
LOGWARE. A sample calculation is shown as carried out by FORECAST. The
results are then summarized from FORECAST output. An example calculation for an
= 0.1 is shown. Other values would be used, ranging from 0.01 to 1.0.
We first calculate a starting forecast by averaging the first four weekly
requirements. That is,
[2,056 + 2,349 + 1,895 + 1,514]/4 = 1,953.50
Now, we back cast this value and start the forecast at time 0. Thus, the forecasts and
the associated errors would be:
Squared
Forecast Error error
F1 = 1953.50
F2 = .1(2056) + .9(1953.50) = 1963.75
F3 = .1(2349) + .9(1963.75) = 2002.28
F4 = .1(1895) + .9(2002.20) = 1991.48
F5 = .1(1514) + .9(1991.48) = 1943.73
F6 = .1(1194) + .9(1943.73) = 1868.76 -749.73 562,095.07
F7 = .1(2268) + .9(1868.76) = 1908.00 399.24 159,392.58
F8 = .1(2653) + .9(1908.00) = 1982.50 745.00 555,025.00
F9 = .1(2039) + .9(1982.50) = 1988.15 56.50 3,192.25
F10 = .1(2399) + .9(1988.15) = 2029.24 410.85 168,797.72
F11 = .1(2508) + .9(2029.24) = 2077.12 478.76 229,211.14
Total squared error 1,677,713.76
The standard error of the forecast is:
Total squared error 1,677,713.76

SF = = = 528.79
N 6
Note: FORECAST does not use N-1 in the denominator.
Repeating this type of analysis, the following table can be developed. The results
from FORECAST are shown.
73
SF
.01 528.72
.05 528.42
.1 528.46
.2 528.89
.5 535.55
1.0 566.07
The that minimizes S F is 0.05.
(b) The forecast errors are computed in part a.
(c) If we assume that the errors are normally distributed around the forecast, we can then
construct a 95 percent confidence band on the forecast. That is, if Y is the actual
volume in period 11, then the range of the forecast (F11 = 2,017.81 for = 0.05) will
be:
Y = F11 + z S F
= 2,017.81 + 1.96528.42
Then,
982.11 Y 3,053.51
All values are in thousands.
5
(a) & (b) The solution to this problem was aided by the use of the exponential smoothing
module in FORECAST. Using the first four week's data to initialize the level/trend
version of the exponential smoothing model and setting and equal to 0.2, the
forecast for next week is F11 = 2,024.47, with a standard error of the forecast of S F =
171.28.
(c) Assuming that the forecast errors are normally distributed around F11, a 95 percent
statistical confidence band can be constructed. The confidence band is:
Y = F11 + z S F
= 2,024.47 + 1.96171.28
where z = 1.96 for 2.5 percent of the area under the two tails of a normal distribution.
The range of the actual weekly volume is expected to be:
1,688.76 Y 2,360.18
74
6
(a) The data may be restated as shown below.
Trend
value,a Seasonal
Sales, S t St t 2
St indexb
27,000 1 27,000 1 41,087 0.66
70,000 2 140,000 4 41,192 1.70
41,000 3 123,000 9 41,298 0.99
13,000 4 52,000 16 41,403 0.32
30,000 5 150,000 25 41,508 0.72
73,000 6 438,000 36 41,613 1.75
48,000 7 336,000 49 41,719 1.15
15,000 8 120,000 64 41,824 0.36
34,000 9 306,000 81 41,929 0.81
82,000 10 820,000 100 42,035 1.95
51,000 11 561,000 121 42,140 1.21
16,000 12 192,000 144 42,245 0.38
500,000 78 3,265,000 650
a
Computed from the linear trend line. For example, for period 1,
S1 = 40,981.6 + 105.31 = 41,087.
b
The ratio of the actual sales S to the trend line value St.
For example, for period 1, the seasonal index is 27,000/41,087 = 0.66.
Given the values from the above table and that t = 78/12 = 6.5, N = 12, and S =
500,000/12 = 41,666, the coefficients in the regression trend line would be:
S t N S t 3,265,000 12 41,666 6.5

b= = = 105.3
t2 N t
2
650 12 6.52
and
a = S b t = 41,666 105.3 6.5 = 40,9816

.
Therefore, the trend value St for any period t would be:
St = 40,981.6 + 105.3t
(b) The seasonal factors are determined by the ratio of the actual sales in a period to the
trend value for that period. For example, the seasonal factor for period 12 (4th
quarter of last year) would be 16,000/42,245 = 0.38. This and the seasonal factors for
all past quarters are shown in the previous table.
75
(c) The forecasts using the seasonal factors from the last 4 quarters are as follows:
Seasonal
t St factors Forecast
13 42,351 0.81 34,304
14 42,456 1.95 82,789
15 42,561 1.21 51,499
16 42,666 0.38 16,213
7
An exponential smoothing model is used to generate a forecast for period 13 (January of
next year). The sales for January through April are used to initialize the model, and an
= 0.2 is used as the smoothing constant. The FORECAST module is used to generate the
forecast. The results are summarized as follows:
Region 1 Region 2 Region 3 Combined

Forecast, F13 219.73 407.04 303.30 938.26
Forecast error, SE 26.89 25.50 17.54 61.41
Note that the sum of the forecasts by region nearly equals the forecast of the combined
usage. However, whether a by-region forecast is better than an overall forecast that is
disaggregated by region depends on the forecast error. The standard error of the forecast
is the best indicator. A comparison of a bottoms-up forecast developed from regional
forecasts to that of a forecast from combined data can be based on the law of variances.
That is, if the usage rates within the regions are independent of each other, the estimate of
the total error can be built from the individual regions and compared to that of the
combined usage data. The total forecast error (variance) from the individual regions S C2
might be estimated as the weighted average of the variances as follows:
F1 2 F2 2 F
S C2 = S E1 + S E2 + 3 S E23
FC FC FC
where
Fi = forecasts of each region
FC = forecast based on combined data
S E2 = variance of the forecast in each region
i
2
S T = total variance of the forecast based on regional data
Therefore,
219.73 407.04 30330

.
S T2 = 26.89 2 + 25.502 + 17.542
930.07 930.07 930.07
= 0.236 72307
. + 0.438 650.25 + 0.326 307.65
= 555.74
76
Then,
S T = 555.74 = 2357
.
Since ST < SC, it appears that a bottom-up, or regional, forecast will have a lower error
than a top-down forecast.
9
(a) See the plot in Figure 8-1. It shows that there is a seasonal component with a very
slight trend to the data as well as some random, or unexplained, variation.
FIGURE 8-1 Plot of time series data for Problem 9

300
250
Average monthly unit prices
200
150
100
50
0
Jan
Jan
Jan
Jan
Jan
Jly
Jly
Jly
Jly
Jly
Apr
Apr
Apr
Apr
Apr
Oct
Oct
Oct
Oct
Oct
Time, months
(b) A time series model typically will involve only two components: trend and
seasonality. Using 2 years of data should be sufficient to establish an accurate trend
line and the seasonal indices. We can develop the following table for computing a
regression line and seasonal indices.
77
Prices, Time, Trend,a Seasonal
Pt
2
Pt t t Tt indexb St
211 1 211 1 232.4 0.91
210 2 420 4 232.4 0.90
214 3 642 9 232.4 0.92
208 4 832 16 232.4 0.90
276 5 1380 25 232.2 1.19
269 6 1614 36 232.3 1.16
265 7 1855 49 232.3 1.14
253 8 2024 64 232.3 1.09
244 9 2196 81 232.3 1.05
202 10 2020 100 232.3 0.87
221 11 2431 121 232.3 0.95
210 12 2520 144 232.2 0.90
215 13 2795 169 232.3 0.93 0.92
225 14 3150 196 232.2 0.97 0.93
230 15 3450 225 232.3 0.99 0.96
214 16 3424 256 232.3 0.92 0.91
276 17 4692 289 232.2 1.19 1.19
261 18 4698 324 232.2 1.12 1.14
250 19 4750 361 232.2 1.08 1.11
248 20 4960 400 232.2 1.07 1.08
229 21 4809 441 232.2 0.99 1.02
221 22 4862 484 232.2 0.95 0.91
209 23 4807 529 232.2 0.90 0.92
214 24 5136 576 232.2 0.92 0.91
5,575 300 69,678 4,900
a
Computed from the trend regression line. For example, the period 1 trend is T1 =
232.39 - 0.0081 = 232.4.
b
The seasonal index is the ratio of the actual price to the trend for the same period.
For example, the period 1 seasonal index is 211/232 = 0.91.
We also have N = 24, t = 300/24 = 12.5, and P = 5575/24 = 232.29.
Now,
b=
P t N P t = 69,678 24( 232.29)(12.5) = 0.008
t N t 4,900 24(12.5)
2 2 2
and
a = P t b t = 232.29 ( 0.008 )(12.5) = 232.39
Therefore, the trend equation is:
78
Tt = 232.29 0.008 t
Note that the trend is negative for the last two years of data, even though the 5-year
trend would appear to be positive.
Now, computing the trend value Tt for each value of t gives the results as shown
in the previous table. The seasonal index is a result of dividing Pt by Tt for each
period t. The indices are averaged for corresponding periods that are one year apart.
Forecasting into the 5th year shows the potential error in the method. That is, for
January of the 5th year, the forecast is Ft = TtSt-12, or F25 = [232.39 0.00825][0.92]
= 213.6. Repeating for each method, we have:
Actual Forecast Forecast Squared Revised

t price price error error seasonala
25 210 213.6 - 3.6 13.0 0.91
26 223 215.6 7.1 50.4
27 204 222.9 -18.9 357.2
28 244 211.3 32.7 1069.3
29 274 276.3 - 2.3 5.3
30 246 264.6 -18.6 345.9
31 237 257.7 -20.7 428.5
32 267 250.7 16.3 265.7
33 212 236.8 -24.8 615.0
34 211 211.2 - 0.2 0.0
35 188 213.5 -25.5 51.0
36 188 211.2 -23.2 538.2
Total squared error 3,739.5
a
The seasonal index for period 25 is .90. The average of the seasonal index for period 25 12 = 13,
and this period is (0.92 + 0.90)/2 = 0.91.
The standard error of the forecast is S F = 3,739.5 / (12 2) = 19.34 . Now, the forecast
for period 37 would be:
F37 = ( 232.39 0.008 37 )( 0.91) = 21121

.
(c) Using the exponential smoothing module in the FORECAST software, the forecast
for the coming period is F = 201.26, with SF = 17.27. The smoothing constants given
in the problem are the "best" that FORECAST could find.
(d) Each model should be combined according to its ability to forecast accurately. We
can give each a weight in proportion to its forecast error, or standard error of the
forecast (SF). Hence, the following table can be developed:
79
(1) (2) = (1)/36.61 (3)=1/(2) (4)=(3)/4.013
Proportion of Inverse of error
Model type Forecast error total error proportion Model weights
Regression 19.34 0.528 1.894 0.472
Exp. smooth. 17.27 0.472 2.119 0.528
Total 36.61 1.000 4.013 1.000
Therefore, each of the model results is weighted according to the model weights.
The weighted forecast for the upcoming January would be:
(1) (2) (3)=(1)(2)

Weighted
Model type Forecast Model weight proportion
Regression 211.21 0.472 99.69
Exp. smooth. 201.26 0.528 106.27
Weighted forecast 205.96
In a similar fashion, we can weight the forecast error variances to come up with a
weighted forecast error standard deviation SFw. That is,
S Fw = 0.472 19.34 2 + 0.528 17.27 2 = 18.28
A 95 percent confidence band using the combined results might be constructed as:
Y = 205.96 z18.28
where z is 1.96 for 95 percent of the area under the normal distribution.
Y = 205.96 1.9618.28
Hence, we can be 95 percent sure that the actual price Y will be within the following
range:
170.13 Y 241.79
10
The plot of the sales data is shown in Figure 8-2. The plot reveals a high degree of
seasonality with a noticeable downward trend. A level-trend-seasonal model seems
reasonable.
(b) Using the search capability within the FORECAST software, a Level-Trend-Seasonal
form of the exponential smoothing model was found to give the lowest forecast error.
A 14-period initialization and 6 periods to compute error statistics were used. The
respective smoothing constants were = 0.01, = 0.08, and = 0.60. This produced
80
a forecast for the upcoming period (January 2004) of F = 6,327.60 and a standard
error of the forecast of SF = 1,120.81.
FIGURE 8-2 Plot of Time Series Data for Hudson Paper Company
30000
25000
Aggregate sales in 000s
20000
15000
10000
5000
0
Jan
Jan
Jan
Jan
Jan
Jly
Jly
Jly
Jly
Jly
Apr
Apr
Apr
Apr
Apr
Oct
Oct
Oct
Oct
Oct
T im e , m o n th s
(c) Assuming that the forecast errors are normally distributed around the forecast, a 95
percent confidence band on the forecast is given by:
Y = F + zSF
Y = 6,327.60 1.961,120.81
where z = 1.96 for 95 percent of the area under the normal distribution curve.
Therefore, we can be 95 percent sure that the actual sales Y should fall within the
following limits:
4,130.8 Y 8,524.4
11
(a) For A569, the BIAS = 165,698 and the RMSE = 126,567 when using the 3-month
moving average. However, if a level only exponential smoothing model with an =
0.10, the BIAS drops to 9,556 and the RMSE is 118,689. The model fits the data
better and there is a slight improvement in the forecasting accuracy.
For A366, the BIAS = 18,231 and the RMSE = 144,973 when using the 3-month
moving average. A level-trend-seasonal model offers the best fit, but it is suspect
since the data show a high degree of random variability rather than seasonality.
Overall, a simple level-only model is probably better in practice. The model has an
= 0.08, a BIAS = 3,227, and a RMSE = 136,256. This is an improvement over the
3-month moving average.
81
(b) Using the level-only models, the forecast for October for A569 = 193,230 and for
A366 = 603,671.
(c) The 3-sigma (99.7 percent) confidence band on the forecasts would be:
For A569, Y = 193,230 3(118,689), or 0 Y 549,297.
For A366, Y = 603,671 3(136,256), or 194,903 Y 1,012,439.
The actual October usage falls within the 3-sigma confidence bands for each of
these products. The difference of the actual from the forecast for each product is
attributable to the substantial variability in the data, which is characteristic of
purchasing in the steel processing industry.
82
WORLD OIL
Teaching Note
Strategy
The purpose of this case study is to allow students to develop an appropriate forecasting
model for some time series data. Discussion may begin with the nature of this
productone with which most students should be very familiar. Based on the many
available forecasting approaches, students should be encouraged to select several for
consideration. In this note, both exponential smoothing and time series decomposition
are evaluated. Both are appropriate here because (1) they can project from historical time
series data, (2) they can handle seasonality, which appears to be present in the data, (3)
there is enough data to construct and test the models, and (4) the forecast is for a short
period into the future.
Assistance with the computational aspects of this problem is available with the use of
the FORECAST module in the LOGWARE software.
Answers to Questions
(1) Develop a forecasting procedure for this service station. Why did you select your
method?
Both exponential smoothing and time series decomposition forecasting methods are
tested using the FORECAST module in LOGWARE. For exponential smoothing, an
initialization period of one seasonal cycle (52 weeks) plus two weeks are used for a total
of 54 weeks, a minimum requirement in FORECAST. The last 30 weeks of data is used
for computing the error statistics. This number of periods is arbitrary, but seems
reasonably large so as to give stable statistical values. We wish to minimize the forecast
error over time, and FORECAST computes both MAD and RMSE statistics that can be
used to make comparisons among model types. Testing the various exponential
smoothing model types and the time series gives the following statistics.
Smoothing Forecast
constants week 6 of this
Model type MAD BIAS RMSE year
Level only.... .4 37.82 -5.27 67.61 817.35
Level-trend... .2 .5 45.85 7.13 67.80 860.26
Level-seasonal .3 1.0 38.97 11.30 45.71 648.75
Level-trend-
seasonal...... .01 .2 .4 30.27 -6.05 44.17 770.74
TS decomp..... 59.46 37.18 71.85 731.33
The MAD and RMSE statistics show how well the forecast has been able to track
historical fuel usage rates. They are an indication of the accuracy of the forecasting
process in the future on the average. We favor forecasting methods that can minimize
these statistics. In this case, the Level-Trend-Seasonal version of the exponential
smoothing model seems to do this best. Both MAD and RMSE are the lowest for this
model type among the alternatives.
83
Further evidence of the performance of a forecasting method is obtained from a plot
of the forecast against the actual usage rates. This is shown in Figure 1. Note that the
Level-Trend-Seasonal model tracks the usage rates quite well, especially in the more
recent weeks. The modeling process has likely stabilized in the last 30 weeks of the data
and is now tracking quite well.
FIGURE 1 Fit of Level-Trend-Seasonal Exponential Smoothing Model to Fuel

Usage Data on Mondays of the Week
(2) How should the periods of promotions, holidays, or other periods where usage rates
deviate from normal patterns, be handled in the forecast?
If the deviations occur at the same time within the seasonal cycle and with the same
relative intensity, no special procedures are required. The adaptive characteristic of the
exponential smoothing process will automatically incorporate these deviations into the
forecast. However, when the deviations are not regular, as promotions may be timed
irregularly, they may best be handled as outliers in the time series and eliminated from
the time series. The model may be fit without the outliers, and then the effect of them
treated as modifications to the forecast. These modifications can be handled manually.
84
(3) Forecast next Monday's fuel usage and indicate the probable accuracy of the
forecast.
From the Level-Trend-Seasonal exponential smoothing model developed in question 1,

where the smoothing constants are = 0.01, = 0.2, and = 0.4, the forecast for Monday
of week 6 would be 771 gallons. However, this forecast only represents the average fuel
usage.
Determining the accuracy of the forecast requires that the forecast track the mean of
the actual usage, i.e., a bias of 0, and that the forecast errors be normally distributed.
While the BIAS (sum of the forecast errors over the last 30 weeks) is not exactly 0, and
will not likely ever be so, it is low (-6.05), such that we will assume good tracking by the
forecast model. A histogram of the forecast errors can reveal whether they follow the
familiar bell-shaped pattern. Such a histogram is given below. We can conclude that
while the errors are not precisely normally distributed, we cannot reject the idea that they
did not come from a normally distributed population. A goodness-of-fit test could be
used to check this assumption. Although this test is not performed here, it is quite
forgiving, such that the normal distribution of errors assumption is not likely to be
rejected where the data show a reasonably normal distribution pattern. The distribution
here qualifies.
We can now proceed with developing a 95 percent confidence band around the
forecast. The forecast of the actual fuel usage rate Y will be:
F z ($ F ) Y F + z ($ F )
where $ F is the standard error of the forecast. F is the forecast, and z is the number of
standard deviations for 95 percent of the area under a normal distribution. FORECAST
computes the root mean squared error (RMSE) as:
(A
t =1
t Ft )2
RMSE =
N
85
HISTOGRAM FOR FORECAST ERROR OF LAST 30 WEEKS
Class Width = 20.0000 Number of Classes = 10
0% 50% 100%
MID CLASS +----+----+----+----+----+----+----+----+----+----+
< -80.0000 | |
-70.0000 |****** |
-50.0000 |*** |
-30.0000 |******** |
-10.0000 |******** |
10.0000 |***** |
30.0000 |******** |
50.0000 |****** |
70.0000 |* |
90.0000 |* |
110.0000 | |
>= 120.0000 | |
+----+----+----+----+----+----+----+----+----+----+
Since RMSE is uncorrected for degrees of freedom lost, we apply a correction factor
(CF) as a multiplier to RMSE to get the unbiased estimate of the standard error of the
forecast ( F ):
N
CF =
N-n
where n is the number of degrees of freedom lost in the model building process. We
estimate n to be the number of smoothing constants in the model, or three in this case.
Hence,
$ F = RMSE CF
30
= 4417
.
30 3
= 4417
. 1054
.
= 46.56
Now, with z@95% = 1.96 from a normal distribution table, we can be 95 percent confident
that the true 87-octane fuel usage Y on Monday of week 6 will be:
771 1.96(46.56) < Y < 771 + 1.96(46.56)

680 < Y < 862 gallons
86
METRO HOSPITAL
You are the materials manager at Metro Hospital. Approximately one year ago, the
hospital began stocking a new drug (Ziloene) that helps the healing process for wounds
and sutures. It is your responsibility to forecast and order the monthly supply of Ziloene.
The goal is to minimize the combined cost of overstocking and understocking the drug.
Orders are placed and received at the beginning of the month and demand occurs
throughout the month. The following demand and cost data have been compiled.
Costs. If more is ordered than is demanded, a monthly holding cost of $1.00 per case
is incurred. If less is ordered than is demanded, a $2.00 per case lost sales cost is
incurred. The drug has a short shelf life, and any overstocked product at the end of the
month is worthless and no longer available to meet demand.
Demand. The demand for the twelve months of last year was:
Last year's demand
Month 1 2 3 4 5 6 7 8 9 10 11 12
Cases 43 36 24 69 34 75 90 67 59 51 77 50
You believe this demand to be representative of Metro's normal usage pattern.
FIGURE 1 Plot of last year's monthly demand in cases
87
Decision Worksheet
Cases Actual Over @ Short @
Month ordered demand $1/case $2/case Cost, $
1 (13)
2 (14)
3 (15)
4 (16)
5 (17)
6 (18)
7 (19)
8 (20)
9 (21)
10 (22)
11 (23)
12 (24)
Total
Cases Actual Over @ Short @

Month ordered demand $1/case $2/case Cost, $
1 (25)
2 (26)
3 (27)
4 (28)
5 (29)
6 (30)
7 (31)
8 (32)
9 (33)
10 (34)
11 (35)
12 (36)
Total
88
METRO HOSPITAL
Exercise Note
Purpose
Metro Hospital is an in-class exercise designed to illustrate the relationship between good
forecasting and the control of inventory related costs. It shows that accurate forecasting
is a primary factor in minimizing inventory costs. Participants in this exercise use a
variety of methods, often intuition, to forecast demand and to come up with a purchase
quantity. Their performance is measured as over- or understock costs. Using a simple
exponential smoothing forecasting model and an understanding of the standard deviation
of the forecast, an effective purchase plan can be constructed. This process results in
costs that are significantly lower than the majority of the participants are able to achieve
using intuitive methods.
Administration
The descriptive material and the decision worksheet are to be distributed to the class at
the time that the exercise is conducted. To hand out the material ahead of time may take
away much of the drama from the exercise. About one half hour should be scheduled for
running the exercise.
The instructor asks the class to make a decision regarding the size of the order to be
placed in the upcoming period and to record it on the worksheet. The participants are
then informed of the demand for that period from Table 1 after the simulated time of one
month has passed. Given that they now know the actual demand for the period, the
participants are asked to record their costs and then to place an order for the next period.
The pattern is repeated for at least twelve months, a full seasonal cycle. The participants
are asked to sum their costs and to report them to the exercise leader. They are displayed
in a public place, such as a chalkboard, for all to see. Then, the exercise leader
announces his or her cost level that was achieved using a disciplined approach using a
simple forecasting procedure and some basic statistics.
TABLE 1 Actual Demand for Period 13 Through 36

Period 13 14 15 16 17 18 19 20 21 22 23 24
Demand 47 70 55 38 90 24 65 65 23 55 85 66
Period 25 26 27 28 29 30 31 32 33 34 35 36
Demand 53 64 61 63 65 38 80 88 45 70 50 56
Quantitative Analysis
The demand series was generated using a normal distribution with a reasonably high
variance and a very slight upward trend. To illustrate the use of a quantitative approach
to forecasting, an exponential smoothing model was selected, although other methods
such as time series decomposition would also be appropriate. The twelve historical data
points were submitted to the FORECAST module in LOGWARE. A three-month
initialization period and a three-month time period for computing error statistics were
chosen. The smoothing constants for the level, level-trend, and level-trend-seasonal
models were examined. Based on the root mean squared error (RMSE), the best model
89
was the level-trend-seasonal (RMSE = 13.59), but the level model with = 0.19 and
RMSE = 13.89 performed very well and is used here. The model is:
Ft +1 = 0.19 At + 0.81Ft
where
Ft +1 = forecast for next period t + 1

At = actual demand for current period t
Ft = forecast for current period t
LOGWARE gives a forecast value of 58.1 and this is used as the forecast value for
period 13. Applying this simple, level only model to the second year demand as it is
revealed in each period gives the following forecast values:
TABLE 2 Simple Exponential

Smoothing Forecast Values
for the Next Year
Actual
Period demand Forecast
13 47 58.1
14 70 56.0
15 55 58.7
16 38 58.0
17 90 54.2
18 24 61.0
19 65 54.0
20 65 56.1
21 23 57.8
22 55 51.2
23 85 51.9
24 66 54.6
Now we must determine the order quantity. It can be calculated from
Q=F + z(RMSE)
t +1
Recall the RMSE was 13.89 for this model. To be precise, we calculate z by trial and
error. The following order quantity and cost computations can be made for a z value of
0.8 (Table 3).
90
TABLE 3 Purchase Order Quantity and Associated Inventory Costs
Actual Order Units Units
Period demand Forecast quantity over short Cost, $
13 47 58.1 69* 22 22
14 70 56.0 67 3 6
15 55 58.7 70 15 15
16 38 58.0 69 31 31
17 90 54.2 65 25 50
18 24 61.0 72 48 48
19 65 54.0 65 0
20 65 56.1 67 2 2
21 23 57.8 69 46 46
22 55 51.2 62 7 7
23 85 51.9 63 22 44
24 66 54.6 66 0
271
*
Q = 58.1 + 0.8(13.89) = 69.21, or 69
We see from the following graph (Figure 1) that z = 0.8 is optimal.
FIGURE 1 Plot of Total Annual Costs Against the Factor z
310
305
300
295
Cost, $
290
285
280
275
270
265
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Z
91
Figure 2 graphically shows the good purchase pattern of Table 3.
FIGURE 2 Plot of Forecast and Purchase Order Quantity on Product Demand
100
90 Order
quantity
80
70
60
Cases
50 Demand
40 Forecast
30
20
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time period
Summary
The exercise leader should discuss that one of the problems with intuitively forecasting
demand is overreacting to randomness in the demand pattern. This has the effect of
causing extreme over and short costs in inventories. A model for short term forecasting
that is integrated into the purchasing and inventory control process can help to avoid
these extremes and give lower costs. Several forecasting models may perform well, such
as exponential smoothing, a simple moving average, a regression model, or a times series
decomposition model. One of the most practical for inventory control purposes is the
exponential smoothing model. The results from a simple, level only model were
illustrated above using the same information that was available to the participants.
Recognizing that it is less costly to order too much than to order too little, the
purchase quantity should exceed the forecast by some margin. The astute participant will
likely approximate the standard deviation of demand from the range of the demand
values, that is, = (Max - Min)/6. Then, one or two might be used to add a margin of
safety to the forecast and size of the purchase order. This simple approximation
procedure can lead to reasonable results.
92
CHAPTER 9
INVENTORY POLICY DECISIONS
1
The probability of finding all items in stock is the product of the individual probabilities.
That is,
(0.95)(0.93) (0.87) (0.85) (0.94) (0.90) = 0.55
2
(a) The order fill rate is the weighted average of filling the item mix on an order. We can
setup the following table.
(1) (2) (3)=(1)(2)

Frequency Marginal
Order Item mix probabilities of order probability
1 .95.95.95.90.90 = .69 0.20 0.139
2 .95.95.95 = .86 0.15 0.129
3 .95.95.90.90 = .73 0.05 0.037
4 .95.95.95.95.95.90.90 = .62 0.15 0.094
5 .95.95.90.90.90.90 = .59 0.30 0.178
6 .95.95.95.95.95 = .77 0.15 0.116
Order fill rate 0.693
Since 69.3 percent is less than 92 percent, the target order fill rate is not met.
(b) The item service levels that will give an order fill rate of 92 percent must be found by
trial and error. Although there are many combinations of item service levels that can
achieve the desired service level, a service level of 99 percent for items A, B, C, D, E,
and F, and 97 percent to 98 percent for the remaining items would be about right.
The order fill rates can be found as follows.
(1) (2) (3)=(1) (2)

Frequency Marginal
Order Item mix probabilities of order probability
1 (.99)3(.975)2 = .922 0.20 0.184
2 (.99)3 = .970 0.15 0.146
3 (.99)2(.975)2 = .932 0.05 0.047
4 (.99)5(.975)2 = .904 0.15 0.136
5 (.99)2(.975)4 = .886 0.30 0.266
6 (.99)5 = .951 0.15 0.143
Order fill rate 0.922
93
3
This is a problem of push inventory control. The question is one of finding how many of
120,000 sets to allocate to each warehouse. We begin by estimating the total
requirements for each warehouse. That is,
Total requirements = Forecast + zForecast error
From Appendix A, we can find the values for z corresponding to the service level at each
warehouse. Therefore, we have:
(1) (2) (3) (4)=(1)+(2)(3)

Ware- Demand Forecast Values Total require-
house forecast, sets error, sets for z ments, sets
1 10,000 1,000 1.28 11,280
2 15,000 1,200 1.04 16,248
3 35,000 2,000 1.18 37,360
4 25,000 3,000 1.41 29,230
Total 85,000 94,118
We can find the net requirements for each warehouse as the difference between the
total requirements and the quantity on hand. The following table can be constructed:
(1) (2) (3)=(1)(2) (4) (5)=(3)+(4)

Total
Ware- require- On hand Net require- Proration of
house ments quantity ments excess Allocation
1 11,280 700 10,580 3,633 14,213
2 16,248 0 16,248 5,450 21,698
3 37,360 2,500 34,860 12,716 47,576
4 29,230 1,800 27,430 9,083 36,513
94,118 89,118 30,882 120,000
There is 120,000 89,118 = 30,882 sets to be prorated. This is done by assuming that
the demand rate is best expressed by the forecast and proportioning the excess in relation
to each warehouse's forecast to the total forecast quantity. That is, for warehouse 1, the
proration is (10,000/85,000)30,882 = 3,633 sets. Prorations to the other warehouses are
carried out in a similar manner. The allocation to each warehouse is the sum of its net
requirements plus a proration of the excess, as shown in the above table.
4
(a) The reorder point system is defined by the order quantity and the reorder point
quantity. Since the demand is known for sure, the optimum order quantity is:
Q * = 2 DS / IC = 2(3,200)(35) / ( 015
. )(55) = 164.78, or 165 cases
94
The reorder point quantity is:
ROP = d LT = (3,200 / 52) 15

. = 92 units
(b) The total annual relevant cost of this design is:
TC = D S / Q + I C Q * / 2
= (3,200)(35) / 164.78 + ( 015
. )(55)(164.78) / 2
= 679.69 + 679.97
= $1,359.66
(c) The revised reorder point quantity would be:
ROP = (3,200 / 52 ) 3 = 185 units .
The ROP is greater than Q*. It is possible under these circumstances the reorder
quantity may not bring the stock level above the ROP quantity. In deciding whether
the ROP has been reached, we add any quantities on order or in transit to the quantity
on hand as the effective quantity in inventory. Of course, we start with an adequate
in-stock quantity that is at least equal to the ROP quantity.
5
(a) The economic order quantity formula can be used here. That is,
Q * = 2 DS / IC = 2(300)(8,500) / ( 010
. )(8,500) = 77.5, or 78 students
(b) The number of times that the course should be offered is:
N * = D / Q* = 300 / 77.5 = 3.9, or about four times per year
6
This is a single-period inventory control problem. We have:
Revenue = $350/unit
Profit = $350 $250 = $100/unit
Loss = 0.2250 = $50/unit
Therefore,
100
CPn = = 0.667
100 + 50
Developing a table of cumulative frequencies, we have:
95
Cumulative
Quantity Frequency frequency
50 0.10 0.10
55 0.20 0.30
60 0.20 0.50
65 0.30 0.80 Q*
70 0.15 0.95
75 0.05 1.00
1.00
CPn lies between quantities of 60 and 65. We round up and select 65 as the optimal
purchase order size.
7
This question can be treated as a single-order problem. We have:
Revenue = 1 + 0.01 = $1.01/$

Cost/Loss = 0.10(2/365) = $0.00055/$ which is the interest expense for two days
Profit = 1.01 1.00055 = $0.00945/$
and
0.00945
CPn = = 0.945
0.00945 + 0.00055
For an area under the normal curve of 0.945 (see Appendix A), z = 1.60.
The planned number of withdrawals is:
Q* = D + z D = 120 + 1.60(20) = 152.00
The amount of money to stock in the teller machine over two days would be:
Money = Q*75 = 152.0075 = $11,400
8
This is a single-period inventory control problem.
(a) We have:
Profit = 400 320

Loss = 320 300
Then,
96
400 320
CPn = = 0.80
( 400 320) + (320 300)
We now need to find the sales that correspond to a cumulative frequency of 0.80. In the
following table:
Cumulative
Sales Frequency frequency
500 0.2 0.2
750 0.2 0.4
1,000 0.3 0.7
1,250 0.2 0.9 Q*
1,500 0.1 1.0
1.0
Q* lies between 1,000 and 1,200 in the cumulative frequency table. We choose to
roundup to Q* = 1,250 units.
(b) Carrying the excess inventory to next year,
80
CPn = = 0.556
80 + ( 0.2 320)
where the loss is the cost of holding a unit until the next year. The Q* now lies between
750 and 1,000 units. We choose 1,000 units. Holding the excess units means a potential
loss of 0.2320 = $64/unit, whereas discounting the excess units represents a loss of only
320 300 = $20/unit. Therefore, Cabot will need fewer units if they are held over in
inventory.
9
(a) The optimum order quantity is:
Q * = 2 DS / IC = 2(1,250)(52)( 40) /(0.3)(56) = 556 cases
and the reorder point quantity is:
ROP = d LT + z sd'
where
sd' = sd LT = 475 2.5 = 751
and z P=0.80 = 0.84.
97
Now,
ROP = (1,250)( 2.5) + ( 0.84 )( 751) = 3,756 cases
Policy: When the amount of inventory on hand plus any quantities on order or in
transit falls below ROP, reorder an amount Q*.
(b) For the periodic review system, we first estimate the order review time:
T * = Q * / d = 556 / 1,250 = 0.44 weeks
The max level is:
M * = d (T * + LT ) + z sd'
where sd' now is:
sd' = sd T * + LT = 475 0.44 + 2.5 = 814 cases
Hence,
M * = 1,250( 0.44 + 2.5) + 0.84(814 ) = 4,359 cases
Policy: Find the amount of stock on hand every 0.44 weeks and place a reorder for
the amount equal to the difference between the quantity on hand and the max level
(M*) of 4,359 cases.
(c) The total annual relevant cost for these policies is:
TC = DS / Q + ICQ / 2 + ICzsd' + kDs d' E( z ) / Q
For the reorder point system:
TCQ = 1250(52)(40)/556 + .3(56)(556)/2
+ .3(56)(.84)(751) + 10(1250)(52)(751)(.1120)/556
= 4,676.26 + 4,670.40 + 10,598.11 + 98,332.37
= $118,277.14
For the periodic review system:
TCP = 1250(52)(40)/556 + .3(56)(556)/2
98
+ .3(56)(.84)(814) + 10(1250)(52)(814)(.1120)/556
= 4,676.26 + 4,670.40 + 11,487.17 + 106,581.29
= $127,415.12
(d) The actual service level achieved is given by:
sd' E( z )
SL = 1
Q
For the reorder point system:
751( 01120
. )
SLQ = 1 = 1 015
.
556
or demand is met 85 percent of the time.
For the periodic review system:
814( 01120
. )
SLP = 1 = 1 016
.
556
or demand is met 84 percent of the time.
(e) This requires an iterative approach as follows:
Compute Q = 2 DS / IC
Compute P = 1 QIC / Dk , then z, then E (z)
Compute Q = 2 D( S + ksd' E( z ) ) / IC
Go back and stop when there is no change

in either P or Q
After the initial value of Q = 556.3, the process can be summarized in tabular form.
99
Step Q P z E(z)
1 778.4 0.9856 2.19 0.0050
2 860.0 0.9799 2.06 0.0072
3 889.9 0.9778 2.01 0.0083
4 899.6 0.9777 2.00 0.0085
5 902.8 0.9767 1.99 0.0087
6 902.8 0.9767 1.99 0.0087
Now, for P = 0.9767, z = 1.99
ROP = 1,250(2.5) + 1.99(751) = 4,620 cases
and the total relevant cost is:
TCQ = DS / Q + ICQ / 2 + ICzsd' + kDsd' E( z ) / Q

= 65,000( 4 ) / 902.8 + 0.3(56)(902.8 ) / 2
+0.3(56 )(199
. )( 751)
+10( 65,000 )( 751)( 0.0087 ) / 902.8
= $40,275
This is considerably less than the $118,277.14 for the preset P at 0.80.
If you solve this problem using INPOL, you will get a slightly different answer. That
is, Q* = 858. This simply is because z is carried to two significant digits rather than
the four significant digits used in the above calculations.
10
Refer to the solution of problem 10-9 for the general approach.
(a) Q* = 556.3 cases
and
ROP = d LT + z LT sd2 + d 2 sLT

2
= 1,250( 2.5) + 0.84 2.5 4752 + 1,250 2 0.52

= 3125
, + 0.84(977.08 )
= 3,946 cases
(b) An approximation for T* = Q*/d, or
T* = 556/1,250 = 0.44 weeks
and approximating sd' as
100
sd' = (T * + LT )sd2 + d 2 sLT
2
= ( 0.44 + 2.5)( 4752 ) + 1,250 2 ( 0.5) 2

= 1,027 cases
So,
Max = d (T * + LT ) + z sd'
= 1,250( 0.44 + 2.5) + 0.84(1,027)
= 4,537 cases
(c) According to INPOL,
TC = 4,686 + 4,686 + 128,195 + 13,862 = $151,429

Q
TC = 4,686 + 4,686 + 134,751 + 14,571 = $158,694

P
(d) According to INPOL,
SL = 80.28 percent
Q
SL = 79.27 percent
P
(e) According to INPOL,
Q* = 930 cases, ROP = 5,128 cases,
TC = $49,532, SLQ = 99.22 percent

Q
T* = 0.76 weeks, MAX = 6,257 cases
TC = $52,894, SLP = 99.18 percent

P
11
(a) The production run quantity is:
2 DS p 2(100 )( 250 )( 250 ) 300

Q *p = = = 1,000 units
IC pd 0.25( 75) 300 10
(b) The production run cycle is:
Q *p = 1,000 / 300 = 333

. days
101
(c) The number of production runs is:
D / Q *p = 100( 250 ) / 1,000 = 25 runs per year
12
(a) The order quantity is:
Q * = 2 DS / IC = 2( 2,000)( 250 )(100

. ) / ( 0.30 )(35) = 309 valves
and the reorder point quantity is:
ROP = d LT + z sd LT
but sd = 0 . Therefore,
ROP = ( 2,000 / 8 )(1) = 250 valves
(b) Boxes are set up that contain 309 valves - the optimum order quantity. When an
order arrives from a supplier, 250 valves are set aside in a separate box and are
treated as the backup stock. The residual 309 250 = 59 valves are used on the
production line. When the 59 valves at the production line are used up, the backup
box containing 250 valves is brought to the production line and the empty box is sent
to the supplier refilling. One hour later when the order arrives, there will be zero
valves remaining at the production line. Then, 250 valves are set aside and 59 are
sent to the production line. The cycle is then repeated.
This problem approach is similar to that of the KANBAN system. Lead times are
very short so that lead times are virtually certain. Demand is certain, since it is fixed
by the production schedule. Boxes or cards are used to assure movement of the most
economic quantity. KANBAN is essentially classic economic reorder point inventory
control under certainty.
13
(a) The economical quantity of cars to be called for at a time is found by the economic
order quantity formula:
Q * = 2 DS / IC = 2( 40 )(52 )(500 ) / ( 0.25)(90,000)(30 ) / 2,000 = 78.5, or 79 cars
(b) This is the reorder point quantity:
ROP = d LT + z sd LT
where z = 1.28 from Appendix A for an area under the curve equal to 0.90.
Therefore,
102
ROP = 40(1) + 128
. (333
. ) 1
= 44.3 cars, or 44.3(90,000 / 2,000)
= 1,994 tons of soda ash
14
(a) This is a reorder point design under conditions of uncertainty for both demand and
lead-time. We assume that the probability of an out of stock is given. Therefore, the
order quantity is:
Q * = 2 DS / IC = 2(50)(365)(50) / ( 0.30)( 45) = 367.7 units
and
ROP = d LT + z sd'
where
z = 1.04 (see Appendix A) for the area under the curve equal to 0.85 and
sd' = sd2 LT + d 2 sLT

2
= 152 ( 7 ) + (50 2 )( 2 2 ) = 107.6 units
Therefore,
ROP = 50( 7) + 104

. (107.6 ) = 4619
. units
(b) This is the periodic review system design under uncertainty. The complexity requires
us to make some approximations here. The time interval for review of the stock level
is:
T * = Q * / d = 367.7 / 50 = 7.35 days
The MAX level is:
MAX = d (T * + LT ) + z sd'
where z = 1.04 and sd' is approximated as:
sd' = (T * + LT )( sd2 ) + d 2 ( sLT

2
)
= ( 7.35 + 7)(152 ) + 50 2 ( 2 2 )
= 115.0 units
Therefore,
103
MAX = 50(7.35 + 7) + 1.04(115.0)
= 837.1 units
(c) Since the service level is specified, the probability is not set at the optimum level.
Knowing the out-of-stock cost allows us to find the most appropriate service level.
Since this is an iterative process, we use INPOL to carry out the calculations.
The optimized service level yields a reorder point design of
Q* = 410 units and ROP = 571 units
and the total relevant cost drops from $12,642 in part a to $8,489. The demand in
stock in part a was 97.74 percent, and it now increases to 99.81 percent.
15
(a) Find the common review time:
T * = 2( O + s I ) / I Ci Di
= 2(100 + 0) / [( 0.3 / 52 )( 2 25 2,000 + 1 90 500 )]
= 2.5 weeks
Then,
M *A = d A (T * + LT ) + z A sd A T * + LT
where z = 1.282 for P = 0.90

A
M *A = 2,000( 2.5 + 15
. ) + 1282
. (100) 2.5 + 15
. = 8,256 units
and
M B* = 500( 2.5 + 15
. ) + 0.842( 70) 2.5 + 15
. = 2,118 units
where z = 0.842 for P = 0.80.

B
The control system works as follows: the stock levels of both items are reviewed
every 2.5 weeks. The reorder size for A is the difference between the amount on hand
and 8,256 units. The reorder size for B is the difference between the amount on hand
and 2,118 units.
(b) The average amount in inventory is expected to be:
104
AIL = d T * / 2 + z sd T * + LT
For A:
AIL A = 2,000( 2.5) / 2 + 128

. (100 ) 2.5 + 15
. = 2,756 units
For B:
AILB = 500( 2.5) / 2 + 0.842( 70 ) 2.5 + 15

. = 743 units
(c) The service level is given by:
SL = 1 sd' E( z ) / d T *
For A:
SLA = 1 100 2.5 + 15

. ( 0.0475) / 2,000( 2.5) = 0.998
For B:
SLB = 1 70 2.5 + 15
. ( 01120
. ) / 500( 2.5) = 0.987
(d) We set T* = 4 and cycle through the previous calculations. Thus, we have:
M *A = 11,301 units M B* = 2,888 units
AIL = 4,301
A AIL = 1,138
B
SL = 0.999
A SL =0 .991
B
16
This problem is one of comparing the combined cost of transportation and in-transit
inventory. In tabular form, we have the following annual costs:
Cost type Formula Rail Truck

Transportation RD 6(40,000)(1.25) 11(40,000)(1.25)
= $300,000 = $550,000
In-transit ICDT/365 0.25( 250 )( 40,000 )( 21) 0.25( 250 )( 40,00 )( 7 )
inventory 365 365
= $143,836 = $47,945
Total $443,836 $597,945
You should select rail.

17
105
The two transport options from the consolidation point are diagrammed in Figure 9-1.
Whether to choose one mode over the other depends more than transportation costs alone.
Because the transport modes differ in the time in transit, the cost of the money tied up in
the goods while in transit must be considered in the choice decision. This in-transit
ICDt
inventory cost is estimated from . The following design matrix can be developed.
365
Cost type Method Air Ocean

Transportation RD $180,800 $98,800
In-transit inventory ICDt/365 3,447* 34,467
Total $184,247 $133,267
*
ICDt/365 = 0.17(185)(20,000)(2)/365= 3,447
Ocean appears to be the lowest cost option even when a substantial in-transit inventory
cost is included. The ocean option assumes that the trucking cost to move the goods from
the consolidation point to the Port of Baltimore is included in the ocean carrier rate.
Multiple sourcing points FIGURE 9-1 The

Consolidation Operation for a
Hydraulic Equipment
Manufacturer
Consolidation Baltimore
point
20 days
2 days
Sao
Paolo
18
The demand pattern is definitely lumpy, since s = 327 > d = 169. To develop the min-
d
max system of inventory control, we first find Q*. That is,
Q * = 2 DS / IC = 2(169)(12)(10) / 0.20( 0.96 + 0.048) = 448.5 units
The ROP is
ROP = d LT + z sd' + ED
where
z = 1.04 from Appendix A,
106
ED = 8 unitsthe average daily demand rate,
and
sd' = sd2 LT + d 2 sLT

2
= 327 2 ( 4) + 169 2 ( 0.8 2 )

= 667.8 units
So,
ROP = 169(4) + 1.04(667.8) + 8
= 1,378.5 units
The max level is:
M* = ROP + Q* ED
= 1,378.5 + 448.5 8
= 1,819 units
19
(a) The basic relationship is:
IT = I i n
We know that I = $5,000,000. If there are 10 warehouses, the amount of inventory in

T
a single one would be:
I1 = I T / 10 = 5,000,000 / 3162
. = 1,581139
,
The inventory in all 10 warehouses would be $1,581,13910 = $15,811,390.
(b) The inventory in a single warehouse would be:
IT = 1,000,000 9 = 3,000,000
In each of three warehouses, we would have:
I = 3,000,000 / 3 = $1,732,051
107
and in all three warehouses, we would have $1,732,0513 = $5,196,152.
20
(a) The turnover ratio is the annual demand (throughput) divided by the average
inventory level. These ratios for each warehouse and for the total system are shown
in the table below.
Annual Average
Ware- warehouse inventory Turnover
house thruput level ratio
21 2,586,217 504,355 5.13
24 4,230,491 796,669 5.31 Avg. = 5.59
20 6,403,349 1,009,402 6.34
13 6,812,207 1,241,921 5.49
2 16,174,988 2,196,364 7.36
11 16,483,970 1,991,016 8.28
4 17,102,486 2,085,246 8.20
1 21,136,032 2,217,790 9.53
23 22,617,380 3,001,390 7.54
9 24,745,328 2,641,138 9.37
18 25,832,337 3,599,421 7.18
12 26,368,290 2,719,330 9.70
15 28,356,369 4,166,288 6.81
14 28,368,270 3,473,799 8.17
6 40,884,400 5,293,539 7.72
7 43,105,917 6,542,079 6.59
22 44,503,623 2,580,183 17.25
8 47,136,632 5,722,640 8.24
17 47,412,142 5,412,573 8.76
16 48,697,015 5,449,058 8.94
10 57,789,509 6,403,076 9.03
19 75,266,622 7,523,846 10.00
3 78,559,012 9,510,027 8.26 Avg. = 8.66
5 88,226,672 11,443,489 7.71
818,799,258 97,524,639 8.40
The overall turnover ratio is 8.40. Ranking the warehouses by throughput and
averaging turnover ratios for the top three and the bottom three warehouses shows
that the lowest volume warehouses have a lower turnover ratio (5.59) than the highest
volume warehouses (8.66). There are several reasons why this may be so:
The larger warehouses contain the higher-volume items such as the A items in the
line. These may carry less safety stock compared with the sales volume.
Conversely, the low-volume warehouses may have more dead stock in them.
108
There may be start-up (fixed) stock in the warehouses, needed to open them, that
becomes less dominant with greater throughput.
(b) A plot of the inventory-throughput data is shown in Figure 9-2. A linear regression
line is also shown fitted to the data. The equation for this line is:
Inventory = 200,168 + 0.1132Throughput
FIGURE 9-2 Plot of Inventory and Warehouse Thruput for California Fruit
Growers Association
12
10
Average inventory level, $ (Millions)
E s t im a t in g lin e
6
0
0 20 40 60 80 100
A n n u a l w a r e h o u s e t h r u p u t , $ ( M il lio n s )
(c) The total throughput for the three warehouses is:
Warehouse Throughput
1 $21,136,032
12 26,368,290
23 22,617,380
Total $70,121,702
Using this total volume and reading the inventory level from Fig. 9-2 or using the
regression equation, we have:
Inventory = 200,168 + .01132(70,121,702)
= $8,137,945
109
(d) Warehouse 5 has a throughput of $88,226,672. Splitting this throughput by 30
percent and 70 percent, we have:
0.3088,226,672 = 26,468,002
0.7088,226,672 = 61,758,670
88,226,672
Estimating the inventory for each of the new warehouses using the regression
equation, we have:
Inventory = 200,168 + 0.113226,468,002 = $3,196,346
and
Inventory = 200,168 + 0 .113261,758,670 = $7,191,249
for at total inventory in the two warehouses of $10,387,595
21
The order quantity for each item when there is no restriction on inventory investment is:
Q * = 2 DS / IC
We first find the unrestricted order quantities.
Q A* = 2(51,000)(10) / 0.25(17
. ) = 1,527 units
QB* = 2( 25,000)(10) / 0.25(325
. ) = 784 units
QC* = 2(9,000)(10) / 0.25( 2.50) = 537 units
The total inventory investment for these items is:
IV = C A ( Q A / 2) + CB ( QB / 2) + CC ( QC / 2)
= 175
. (1,527 / 2) + 3.25( 784 / 2) + 2.50(537 / 2)
= $3,28138 .
Since the total investment limit is exceeded, we need to revise the order
quantities. For each product:
Q * = 2 DS / [C( I + )]
110
For product A:
. ( 0.25 + )]
Q *A = 2(51,000)(10) / [175
For product B:
QB* = 2( 25,000)(10) / [3.25( 0.25 + )]
For product C:
QC* = 2(9,000)(10) / [2.50( 0.25 + )]
Now, the investment limit must be respected so that:
3,000 = C A ( Q A / 2 ) + CB ( QB / 2 ) + CC ( QC / 2 )
Expanding we have:
3,000 = 175
. . ( 0.25 + )]
2(51,000)(10) / [175
+325
. 2( 25,000)(10) / [3.25( 0.25 + )]
+2.50 2(9,000)(10) / [2.50( 0.25 + )]
We now need to find an value by trial and error that will satisfy this equation. We
can set up a table of trial values.
Investment in
Total
inventory
A B C value, $
0.03 1,262.44 1,204.53 633.87 3,100.84
0.04 1,240.48 1,183.58 622.84 3,046.90
0.045 1,229.92 1,173.51 617.54 3,020.97
0.049 1,221.67 1,165.63 613.40 3,000.70
0.05 1,219.63 1,163.69 612.37 2,995.69
0.10 1,129.16 1,077.36 566.95 2,773.47
When the term I+ is the same for all products, as in this case, may be found
directly from Equation 10-30.
We can substitute the value for = 0.049 into the equation for Q* and solve.
Hence, we have:
111
Q A* = 2(51000)(10) / [175
. ( 0.25 + 0.049)] = 1,396 units
QB* = 2( 25,000)(10) / [3.25( 0.25 + 0.049)] = 717 units
QC* = 2(9,000)(10) / [2.50( 0.25 + 0.049)] = 491 units
Checking:
1.75(1,396)/2 + 3.25(717)/2 + 2.50(491)/2 = $3,000
22
We first check to see whether truck capacity will be exceeded. Since three items are to
be placed on the truck at the same time, the items are jointly ordered. The interval for
ordering follows Equation 9-23, or:
2( O + S i ) 2( 60 + 0)
T* = =
I Ci Di 0.25[50(100)(52) + 30(300)(52) + 25( 200)(52)]
120
= = 0.022 years, or 1.144 weeks
0.25(988,000)
Now, from
DT
i
i
*
wi Truck capacity
[100(70) + 300(60) + 200(25)][1.144] = 34,320 lb.
The truck capacity of 30,000 lb. has been exceeded, and the order quantity or the order
interval must be reduced. Given the revised Equation 9-31, the increment to add to I can
be found. That is,
2O
= 2
I
Truck capacity
C D
Di wi i i
2( 60)
= 2
0.25
30,000
(50(10)(52) + 30(30)(52) + 25( 20)(52))
[100(52)( 70) + 300(52)( 60) + 200(52)(10)]
120
2
0.25
30,000
(988,000)
2,340,000
= 0.73895 0.25 = 0.48895
Revise T*, the order interval by:
112
2( O + S i ) 2( 60 + 0)
T* = =
( I + ) Ci Di ( 0.25 + 0.48895)[50(100)(52) + 30(300)(52) + 25( 200)(52)]
120
= = 0.01282 years, or 0.6667 weeks
0.73895(988,000 )
Once again, we check that the truck capacity has not been exceeded.
[100(70) + 300(60) + 200(25)][0.66667] = 30,000 lb.
Therefore, place an order every 4.7, or approximately five days.
23
The average inventory for each item is given by:
Q*
AIL = + z sd'
2
2 DS
where sd' = sd LT and Q* is found by Q * = . z@ 95% = 1.65 from the normal
IC
distribution in Appendix A. The results of these computations can be tabulated.
A B C D E
sd' 7.75 15.49 19.36 11.62 27.11
Q* 188.38 238.28 421.23 361.98 565.14
AIL 106.98 144.70 242.56 200.16 327.30
Summing the AIL for each product gives a total inventory of 1,022 cases.
24
The peak quantity of an item to appear on a shelf can be approximated as the order
quantity plus safety stock, or
Q + z sd' 250 boxes
where z@93% = 1.48 from Appendix A and sd' = sd LT = 19 1 = 19 boxes. The

economic order quantity is
2 DS 2(123 52 )(125
. )
Q* = = = 255.42 boxes
IC 019
. (129
. )
Checking to see if the shelf space limit will be exceeded by this order quantity
113
255.42 + 1.48(19) = 283.54 boxes
The quantity is greater than the 250 allowed. Subtracting the safety stock from the limit
gives 250 28 = 222 boxes. The order quantity should be limited to this amount.
25
The plot of average inventory to period facility throughput (shipments) gives an overall
indication of how the company is managing collectively its inventory for all stocked
items. We can see that the relationship is linear with a zero intercept. This suggests that
the company is establishing its inventory levels directly to the level of demand
(throughput). An inventory policy, such as stocking to a number of weeks of demand,
may be in effect.
Overall, the inventory policy seems to be well executed in that the regression line fits
the point for each warehouse quite well. The terminal with an inventory level of $6,000
seems to be an outlier and it should be investigated. If its high turnover ratio were
brought in line with the other terminals, an inventory reduction from $6,000 to $4,000 on
the average could be achieved.
The stock-to-demand inventory policy should be challenged. An appropriate
inventory policy should show some economies of scale, i.e., the inventory turnover ratio
should increase as terminal throughput increases. Whereas the current policy is of the
form I = 0.012 D , a better policy would be I = kD 0.7 , where D represents terminal
throughput and I is the average inventory level. The coefficient 0.012 for the current
policy is found as the ratio of 6,000/500,000 = 0.0.12 for the last data point in the plot.
The k value for the improved policy needs to be estimated. From the cluster of the lowest
throughput facilities, the average inventory level is approximately $2,000 with an average
throughput of about $180,000. Therefore, from
I = kD 0.7
2,000 = k (180,000) 0.7
2,000 = k ( 4,771.894)
2,000
k=
4,771.894
k = 0.419
Reading values from the plot, the following table can be developed showing the
inventory reduction that might be expected from revised inventory policy. (Note: If the
inventory-throughput values cannot be adequately read from the plot, the values in the
following table may be provided to the students.)
Actual Estimated inventory, $ Revised inventory, $

Terminal Inventory, $ Shipments, $ I = 0.012 D I = 0.419 D 0.7
1 2,000 150,000 1,800 1,760
2 1,950 195,000 2,340 2,115
3 2,000 200,000 2,400 2,152
4 2,050 200,000 2,400 2,152
114
5 3,900 320,000 3,840 2,991
6 6,000 330,000 3,960 3,056
7 4,500 390,000 4,680 3,435
8 4,300 410,000 4,920 3,558
9 5,500 500,000 6,000 4,088
Totals 32,200 2,695,000 32,340 25,307
Revising the inventory control policy has the potential of reducing inventory from the
32,340 25,307
linear policy by x100 = 21.7% .
32,340
26
We can use the decision curves of Figure 9-23 in the text answer this question since it
applies to a fill rate of 95 percent and an = 0.7. First, determine K for an inventory
throughput curve for the item, which is
D1 (117 x12) 0.3

K= = = 1.466
TO 6
Next,
tD10.7 12(117 x12) 0.3

X = = = 0.90
ICK 0.20( 400)(1.466)
and with z 1.96 from Appendix A
zs LT 1.96(15) 2
Y = a
= = 0.18
KD (1.466)(117 x12) 0.7
The demand ratio r is 42/177 = 0.36. The intersection of r and X lies below the curve Y
(use curve Y = 0.25), so do not cross fill.
27
Regular stock
For two warehouses, estimate the regular stock for the three products.
115
2dS
Q
RS = = IC
2 2
2(3,000)( 25)
0.02(15)
Product A RS A1 = = 354 units
2
2(5,000)( 25)
0.02(15)
RS A2 = = 457 units
2
2(8,000)( 25)
0.02(30)
RS B1 = = 408 units
Product B 2
2(9,500)( 25)
0.02(30)
RS B 2 = = 445 units
2
2(12,500)( 25)
0.02( 25)
Product C RSC1 = = 559 units
2
2(15,000)( 25)
0.02( 25)
RSC 2 = = 612 units
2
Regular system inventory for two warehouses is RS2W = 354 + 457 + 408 + 445 + 559 +
612 = 2,835.
Regular stock for a central warehouse
2(8,000)( 25)
0.02(15)
RS A = = 577 units
2
2(17,500)( 25)
0.02(30)
RS B = = 604 units
2
2( 27,500)( 25)
0.02( 25)
RSC = = 829 units
2
Total central warehouse regular stock is RS1W =577 + 604 + 828 = 2,009 units.
116
Safety Stock
Product A
SS = zsd LT
SS A1 = 1.65(500) 0.75 = 714 units
SS A2 = 1.65(700) 0.75 = 1,000 units
where z@0.95 = 1.65 from Appendix A
Product B
SS B1 = 1.65( 250) 0.75 = 357 units
SS B 2 = 1.65(335) 0.75 = 479 units
Product C
SS = zsd LT
SS C1 = 1.65(3,500) 0.75 = 5,001 units
SS C 2 = 1.65(2,500) 0.75 = 3,572 units
System safety stock is SS2W = 714 + 1,000 + 357 + 479 + 5,001 + 3,572 = 11,123 units
For each product, the estimated standard deviation of demand on the central warehouse
is:
s A = s12 + s22 = 500 2 + 700 2 = 860 units

s B = 250 2 + 3352 = 418 units
s B = 3,500 2 + 2,500 2 = 4,301 units
The safety stock is:

SS = zs LT
SS A = 1.65(860) .75 = 1,229 units
SS B = 1.65(418) .75 = 597 units
SSC = 1.65(4,301) .75 = 6,146 units
Total safety stock in the central warehouse SS1W = 1,229 + 597 + 6,146 = 7,972 units.
Total inventory with two warehouses RS2W + SS2W = 2,835 + 11,123 = 13,958 units and
for a central warehouse RS1W + SS1W = 2,009 + 7,972 = 9,981 units. Centralizing
inventories reduces them by 13,958 9,981 = 3,977 units.
28
The solution to this multi-echelon inventory control problem is approached by using the
base-stock control system method. The idea is that inventory at any echelon is to plan its
inventory position plus the inventory from all downstream echelons.
117
First, compute the average inventory levels for each customer. This requires finding
Q and the safety stock. Q is found from the EOQ formula.
For customer 1
2( 425 x12)(50)
Q1 = = 270 units
0.2(35)
Q1 270
AIL1 = + zs d1 LT3 = + 1.65(65) 0.5 = 211 units
2 2
where z@0.95 =1.65 from Appendix A
For customer 2
2(333 x12)(50)
Q2 = = 239 units
0.2(35)
Q2 239
AIL2 = + zs d 2 LT3 = + 1.65(52) 0.5 = 180 units
2 2
For customer 3
2( 276 x12)(50)
Q3 = = 218 units
0.2(35)
Q3 218
AIL3 = + zsd 3 LT3 = + 1.65(43) 0.5 = 159 units
2 2
Total customer echelon inventory is AILC = 211 + 180 + 159 = 550 units
For the distributors echelon
Q D = 2,000 units
QD 2,000
AILD = + zsd D LTD = + 1.28(94) 1.0 = 1,120 units
2 2
where z@0.90 =1.28 from Appendix A
The expected inventory that the distributor will hold is the distributor echelon inventory
less the combined inventory for the customers, or 1,120 - 550 = 570 units.
118
COMPLETE HARDWARE SUPPLY, INC.
Teaching Note
Strategy
Complete Hardware Supply is an exercise involving the control of inventoried items
collectively. Data for a random sample of 30 items from the company's total of 500 items
held in inventory are given. The objective is to manage the total dollar value allowed to
be held as inventory. Several alternatives can be considered for changing inventory
levels, some of which require an investment other than in inventory.
The number of items that must be analyzed and the multiple scenarios that are to be
examined can be computationally time consuming. It is strongly suggested that students
use the INPOL module within LOGWARE to aid analysis. The current database has
been prepared and is available in the LOGWARE software.
The Base Case

We begin with the current data optimized as a reorder point design. The optimum order
quantities and associated inventory levels are found. The base case costs are shown as
follows:
Fixed order quantity policy

Purchase cost $556,912
Transport costa 0
Carrying cost 4,425
Order processing cost 4,425
Out-of-stock cost 0
Safety stock cost 2,529
Total cost $568,291
Total investment $27,801
aIncluded in the purchase cost
We note that optimizing the current design shows that investment of $27,801 exceeds
the allowed investment level of $18,000. Ways need to be explored to reduce this.
Transmit Orders More Rapidly

Instead of mailing orders to vendors, Tim O'Hare can buy a facsimile machine and
transmit orders electronically. This scenario can be tested by reducing the lead times in
the base case by two days, or (2/5) = 0.40 weeks and increasing order processing costs by
two dollars, and then optimizing again. INPOL shows that there will be a slight increase
in operating costs from $568,291 to $568,640, an incremental increase of $349.
Projecting this to all 500 items, we have 349(500/30) = $5,817. Since both operating cost
and inventory investment level increase, there is no economic incentive to implement this
change.
Faster Transportation
Suggesting that vendors who are located some distance (>600 miles) from the warehouse
use premium transportation is a possible way of reducing lead times, and therefore safety
119
stock levels. Of course, the increase in transportation cost for those affected vendors is
likely to lead to a price increase to cover these costs. This scenario is tested by reducing
the lead-time in weeks to 2.2 for those vendors over 600 miles from the warehouse. For
these same vendors, a five percent price increase is made.
Compared with the base case, there is little change in the inventory investment
($27,801 vs. $27,746); however, operating costs increase. The total costs now are
$585,490 compared with the base case of $568,291, an increase of $27,199. The major
portion ($17,159) of this comes from the increase in price. We conclude that this is not a
good option for Tim.
Reduce Forecast Error

Reducing the forecast error involves reducing the standard deviation of the forecast error.
Testing this option requires taking 70 percent of the base-case forecast error standard
deviations and optimizing the design once again.
These changes have a positive impact on operating costs and inventory investment.
Operating cost now is $567,529 and inventory investment is $24,739. This is a saving in
operating costs of $762 per year. For all 500, we can project the savings to be
762(500/30) = $12,700. Based on a simple return on investment, we have:
12,700
ROI = 0.25, or 25% / year
50,000
This would appear to be attractive since carrying costs are 25 percent per year and the
company's return on investment probably makes up about 80 percent of this value.
Reduce Customer Service

At this point, we have only accepted the idea of reducing the forecast error. However,
inventory investment remains too high. We can now try to reduce it by reducing the
service levels. This is tested by dropping the service index from its current 0.98 level to a
level where inventory investment approximates $18,000. This is done, assuming the
forecast software will be purchased and the forecast error reduced by 30 percent. By trial
and error, the service index is found to be 0.54, which gives an investment level of
$18,028. The revised service level compared with the base case is summarized below for
the 30 items.
120
Base Base
Item case Revised Item case Revised
1 99.88% 96.26% 16 99.98% 99.56%
2 99.92 98.02 17 99.90 97.57
3 99.96 98.54 18 99.95 97.81
4 99.98 99.15 19 99.89 95.96
5 99.98 99.45 20 99.97 98.15
6 99.96 98.60 21 99.69 89.53
7 99.97 98.84 22 99.97 98.96
8 99.96 98.61 23 99.97 98.96
9 99.92 97.29 24 99.96 97.58
10 99.98 99.26 25 99.92 99.33
11 99.99 99.70 26 99.97 96.68
12 99.99 99.43 27 99.93 97.45
13 99.92 97.30 28 99.89 98.78
14 99.98 99.14 29 99.97 96.92
15 99.96 98.84 30 99.91 96.78
Notice how little the service level changes, even with a substantial reduction in the
service index.
Conclusions
Tim can make a good economic argument for purchasing software that will reduce the
forecast error. The only questions here are whether the software can truly produce at
least the error reduction noted and whether a 25 percent return on investment is adequate
for the risks involved.
Arguing to accept a service reduction in order to lower the investment level is a little
less obvious since we do not know the effect that service levels have on sales. However,
Tim may point out that the service levels need to be changed so little that it is unlikely
that customers will detect the change. He might also raise the question as to whether
customer service levels were too high initially, and suggest that customers be surveyed as
to the service levels that they do need.
121
AMERICAN LIGHTING PRODUCTS
Teaching Note
Strategy
American Lighting Products is a manufacturer of fluorescent lamps in various sizes for
industrial and consumer use. As frequently happens in business, top management has
requested that inventories be reduced across the board, but it does not want to sacrifice
customer service. Sue Smith and Bryan White have been asked to eliminate 20 percent
of the finished goods inventory. Their plan is to reduce the number of stocking locations
and, thereby, eliminate the amount of inventory needed. Of course, they must recognize
that with fewer stocking points, transportation costs are likely to increase and customer
delivery times may increase as well. On the other hand, facility fixed cost may be
reduced.
The purpose of this case is to allow students to examine inventory policy and
planning through aggregate inventory management procedures. They also can see the
connection between location and inventory levels.
(1) Evaluate the companys current inventory management procedures.
The companys procedures for controlling inventory levels are at the heart of whether
inventory reductions are likely to be achieved through inventory consolidation. The
company appears to be using some form of reorder point control for the entire system
inventory, but it is modified by the need to produce in production lot sizes. It is not clear
how the reorder point is established. If it is based on economic order quantity principles,
then the effect of the principles becomes distorted by the need to produce to a lot size that
is different from the economic order quantity. Therefore, average inventory levels in a
warehouse will not be related to the square root of the warehouses throughput (demand),
i.e., throughput raised to the 0.5 power.1 Rather, the throughput will be raised to a higher
exponent between 0.5 and 1.0.
The above ideas can be verified by plotting the data given in Table 1 of the case and
then fitting a curve of the form I = TP . Note: The curve can be found from standard
linear regression techniques when the equation is converted to a linear form through a
logarithmic transformation, i.e., lnI = ln + lnTP. The results are shown in Figure 1.
The inventory curve is I = 2.99TP 0.816 with r = 0.86, where I and TP are in lamps. The
projected inventory reduction can be calculated by using this formula.
From the plot of the inventory data, we can see that there is substantial variation
about the fitted inventory curve. There is not a consistent turnover ratio between the
warehouses. This probably results from the centralized control policy. On the other
hand, improved control may be achieved by using a pull procedure at each MDC. The
data available in the case do not let us explore this issue.
1
Based on the economic order quantity formula, the average inventory level (AIL) for an item held in
inventory can be estimated as AIL = Q / 2 = 2 DS / IC / 2 . Collecting all constants into K, we have
AIL=K(D)0.5, where D is demand, or throughput.
122
FIGURE 1 Plot of
MDC average
inventory vs. annual
throughput.
(2) Should establishing the LOC be pursued?
One of the ideas proposed in the case is to consolidate all Consumer product line items
into one large order center (LOC). Evaluating the impact of the LOC on inventory
reduction requires that an assumption be made as to how much demand and associated
inventory of the total belongs to Consumer products. Table 2 of the case gives the order
and back order breakdown by sales channel. Using this data, total consumer demand is
312,211 line items, or 33.4 percent of the total line items. The assumption is that the
same percentage applies to total demand. Hence, Consumer demand is
33.4%169,023,000 = 56,453,682 lamps. From the inventory-throughput curve, we can
estimate the amount of inventory needed at the single LOC. That is, I =
2.997(56,453,682)0.816 = 6,339,684 lamps. If Consumer products account for 33.4% of
total inventory, then there are 33.4%23,093,500 = 7,713,229 lamps in Consumer
inventory. The reduction that can be projected is 7,713,229 6,339,684 = 1,373,545
lamps for a reduction of
1,373,545
Reduction = 100 = 17.8%
7,713,229
in Consumer inventory levels, but only a 6 percent reduction in overall inventory levels.
The 20 percent reduction goal is not achieved. Other alternatives need to be explored.
(3) Does reducing the number of stocking locations have the potential for reducing
system inventories by 20 percent? Is there enough information available to make a good
inventory reduction decision?
The second alternative proposed in the case is to reduce the number of MDCs from eight
to a smaller number. In order to evaluate this proposal, it needs to be determined which
MDCs will be consolidated and the associated total demand flowing through the
consolidated facilities. The inventory-throughput relationship can then be used to
estimate the resulting inventory levels. For example, if the Seattle and Los Angeles
MDCs are combined, the consolidated demand would be 4,922,000 + 21,470,000 =
26,392,000 lamps. The combined inventory is projected to be I = 2.997(26,392,000)0.816 =
123
3,408,852 lamps, compared with the inventory for the two locations of 4,626,333, as
shown in Table 1. This yields a 26.3 percent reduction from current levels.
Table 1 shows other possible MDC consolidations and the resulting inventory
reductions that can be projected.
TABLE 1 Inventory Reduction for Selected MDC Combinations, in Lamps

Combined Combined Inventory
MDC combination demand inventory reduction
Seattle/Los Angeles 26,392,000 3,408,852 1,217,481
Kansas City/Dallas 29,194,000 3,701,403 50,181
Chicago/Ravenna 49,174,000 5,664,257 -557,590
Atlanta/Dallas 39,314,000 4,718,862 1,224,721
Kansas City/Chicago 39,271,000 4,714,650 -933,900
Ravenna/Hagerstown 64,046,000 7,027,231 1,715,607
K City/Dallas/Chicago 52,515,000 5,976,377 -36,377
Ravenna/Htown/Chicago 87,367,000 7,508,054 3,423,196
Atlanta/Dallas/K City 55,264,000 5,242,351 2,293,566
From the MDC combinations in Table 1, proximity to each other is a primary

consideration in order to not increase transportation costs or jeopardize delivery service
any more than necessary. Several options can be identified that yield a 20 percent
inventory reduction. These are:
Inventory Total
reduction, inventory
Option MDC combinations lamps reduction
1 LA/Seattle 1,217,481
Ravenna/Htown/Chicago 3,423,196
Total reduction 4,640,677 20.1%
Kansas City/Hagerstown 1,224,721
Ravenna/Hagerstown 1,715,602
Ravenna/Hagerstown 1,715,602
Atlanta/Dallas/K City 2,293,566
Options 1 and 3 achieve the 20 percent reduction goal, although other MDC
combinations not evaluated may also do so. The maximum reduction would be achieved
with one MDC. The total inventory would be I = 2.997(169,023,000)0.816 = 15,512,812
lamps, for a system reduction of 32.8 percent. However, we must recognize that as the
number of warehouses is decreased, outbound transportation costs will increase. Inbound
transportation costs to the combined MDC will remain about the same, since
124
replenishment shipments are already in truckload quantities. Some difference in cost will
result from differences in the length of the hauls to the warehouses. On the other hand,
outbound costs may substantially increase, since the combined MDC locations are likely
to be more removed from customers then they are at present. Outbound transportation
rates will be higher, as they are likely to be for shipments of less-than-truckload
quantities. If the sum of the inbound and outbound transportation cost increases is
greater than the inventory carrying cost reduction, then the decision to reduce inventories
must be questioned.
Calculating all transportation cost changes is not possible, since the case study does
not provide sufficient data on outbound transportation rates. However, they should be
determined before and after consolidation to assess the tradeoff between inventory
reduction and transportation costs increases. On the other hand, inbound transportation
costs can be found, as shown below for option 1, where the consolidation points are Los
Angeles and Hagerstown.
Annual Combined
TL rate, demand, Transport annual Transport
Location $/TL lamps cost, $ demand, lamps cost, $
Seattle 1800 4,922,000 253,131a
Los Angeles 1800 21,470,000 1,104,171 26,392,000 1,357,302
Ravenna 250 25,853,000 184,664
Hagerstown 475 38,193,000 518,334 87,367,000 1,185,695
Chicago 350 23,321,000 233,210
Total 113,759,000 2,293,510 113,759,000 2,542,997
a
(4,922,000/35,000)1800 = 253,131
There will be a net increase in inbound transportation costs of $2,542,997 2,293,510 =

$249,487 for option 1.
In addition, the annual fixed costs for the MDCs will be less, since the total space
needed in the consolidated facilities should be less than that for the existing facilities.
Again, the case study does not estimate the fixed costs for existing or potential locations.
We do know that taking them into account would favor consolidation.
In summary, the costs associated with option 1, that just meets the 20 percent
inventory reduction goal, would be:
Cost type Cost savings, $

Inventory carrying cost reduction 0.200.8824,640,677 = 818,615
Warehouse cost 0.104,640,677 = 464,068
Warehouse fixed cost Unknown, but may be included in warehouse cost
Outbound transportation cost Unknowndata not given
Inbound transportation cost (249,487)
Although Sue and Bryan could report a substantial savings in inventory related costs,
they should be encouraged to include fixed costs and transportation costs so as to report
the true benefits of the inventory reduction plan.
125
(4) How might customer service be affected by the proposed inventory reduction?
The general effect of inventory consolidation is to reduce the number of stocking points
and make them more remote from customers. That is, the delivery distance will be
increased if inventory consolidation is implemented. Therefore, delivery customer
service may be jeopardized and must be considered before deciding to consolidate
inventories.
From Table 3 of the case, it can be seen that customer lead times remain constant for
a variety of locations with the exception of Kansas City. Since consolidation points will
be selected among the existing locations, outbound lead times will remain unaffected.
Customer service due to location should be constant, at least for a moderate degree of
consolidation.
Customer service due to stock availability will be affected if safety stock levels are
reduced after consolidation. Although the inventory-throughput relationship projects
adequate safety stock to maintain the current first-time delivery levels, it does not account
for any increase in lead times that may occur between the current system of MDCs and
the consolidated ones. By comparing the weighted inbound lead times for the existing
distribution system and option 1, as shown in Table 2, the average inbound lead-time is
slightly reduced through consolidation. Lead-time variability is usually related to
average lead-time. This should have a favorable affect on inventory levels since
uncertainty is reduced. First-time deliveries should not be adversely affected by
consolidation, according to option 1.
TABLE 2 A Comparison of Inbound Lead Times for the Existing Distribution

System and a Consolidated Distribution System (Option 1)
(a) Current Distribution System
Inbound Weighted
lead time, lead time,
Master Distribution Center Shipments days days
Atlanta 26,070,000 2 0.308
Chicago 23,321,000 1 0.138
Dallas 13,244,000 3 0.235
Hagerstown 38,193,000 1 0.226
Kansas City 15,950,000 2 0.094
Los Angeles 21,470,000 5 0.635
Ravenna 25,853,000 1 0.153
Seattle 4,922,000 6 0.175
Total 169,023,000 1.964
126
(b) Consolidation Option 1
Inbound Weighted
lead time, lead time,
Master Distribution Centera Shipments days days
Atlanta 26,070,000 2 0.308
Dallas 13,244,000 3 0.235
Htown/Ravenna/Chicago 87,367,000 1 0.517
Kansas City 15,950,000 2 0.094
Los Angeles/Seattle 26,392,000 5 0.781
Total 169,023,000 1.935
a
Consolidation is assumed to take place at the MDC with the largest number of current shipments.
127
AMERICAN RED CROSS: BLOOD SERVICES
Teaching Note
Strategy
The American Red Cross Blood Services has a mission to provide the highest quality
blood components at the lowest possible cost. High quality blood products are provided
to regional hospitals, but managing the inventory to meet demand as it occurs is a
difficult problem. Blood is considered a precious product, especially by those who give it
voluntarily. So, managing this perishable product carefully is a foremost concern.
Blood is a vital product to those in need of it for emergencies and a precious product
to those requiring it for elective surgery and other treatments. The goal is to always have
what is needed but never so much that this perishable product has to outdated. Managing
the blood inventory is quite difficult because (1) forecasting demand is not particularly
accurate, (2) the planning horizon for collections can be up to a year long with uncertain
yields, (3) the life of blood products ranges from 42 days to as short as five days, (4) once
scheduled, blood donors are never turned away except for medical reasons, and (5) there
is a limited opportunity to sell blood outside of the local region if too much is on hand.
Overall, this situation has many characteristics of a supply driven inventory
management problem, which requires inventory management techniques different from
those for typical consumer products.
The intended purpose of this case study is for students to examine an inventory
situation where there is limited control over the amount of the product flowing into
inventory. This supply-driven inventory situation is likely to be quite different from that
discussed on the introductory level. Students are encouraged to consider the various
elements that affect inventory levels of individual products and how they interact. These
elements are (1) demand forecasting, (2) collections, (3) decision rules for creating blood
derivatives, (4) product prices, and (5) inventory policy. It is expected that students will
be able to make general suggestions for improvement.
Questions
(1) Describe the inventory management problem facing blood services at the American
Red Cross.
One of the major problems facing the American Red Cross (ARC) is that the availability
of blood is supply-driven, meaning that quantities of blood received for processing to
meet demand in the short term are unknown, yet they must be placed in inventory if
demand is less than the collected quantities. Blood availability is a function of number of
factors that cannot be well-controlled by the regional blood center in the short run,
causing wide variability in supply. The usage of blood at hospital blood banks, which
creates the demand on ARCs blood inventories, is also uncertain and varies from day to
day and between hospital facilities.
The yield of blood at the point of collection is random and does not necessarily give
the product mix needed to meet demand. Different blood types can only be known by a
probability distribution as to the percentage of the blood types that exist in the general
population. In the short term, the demand for blood types may differ from the collected
128
quantities, resulting in a potential for under- and over-stocking, since blood is drawn
from all qualified donors as they arrive at collection sites.
Forecasting demand for blood products will likely be reasonably accurate for a base
load. Surgery loads on hospitals are scheduled in advance so that blood needs will be
known with a fair degree of certainty, although each operation will not typically use the
full amount of blood allocated to it. However, emergency blood needs are not well
predicted, and they can cause spikes in demand and unplanned draws on inventory. A
problem is establishing how much accuracy is needed for good inventory management.
Inventory policy for managing inventory levels is a mixed strategy of product pricing,
derivative product selection for processing at the time of collection, conversion to other
products later in the product life cycle, product sell off, emergency supply (call for
blood), discount pricing, and stocking rules for hospitals. Although there are many
avenues to controlling inventory levels, shortages and outdating cannot always be
avoided. It is not clear that these procedures lead to an optimal control of inventory
levels.
Competition from local independent blood banks that sell selected blood products at
low prices makes it difficult for ARC to cover costs. ARC provides a wider range of
products, but it has difficulty-differentiating price among derivative products so that it
might compete effectively. Given pressures for hospitals to increase efficiency, they will
shop around for the lowest-priced blood products. ARC is having difficulty maintaining
its position as the dominant supplier of blood products in the region, which results in the
greater uncertainty in managing inventory levels.
In summary, blood is a precious product given by volunteers for the benefit of others.
Donors have the right to expect that their contribution will be handled responsibly. To
ARC, this means managing the blood supply so that recipients receive a high-quality
product at the lowest possible price. To achieve this goal, ARC manages the blood
supply through four inter-connected elements: (1) estimating the blood product needs
over time, (2) planning the collection of whole blood, (3) deciding which derivative
products and their amounts should be created from whole blood, and (4) controlling the
inventory levels to avoid outdating. The volunteer nature of the blood giving and donor
attitudes surrounding it, long planning lead times and the associated uncertainties, rising
competition among some products from local blood banks, and the uncertainties of blood
needs all make blood supply management a unique inventory management problem.
(2) Evaluate the current inventory management practices in light of ARCs mission.
Performance of blood management can be evaluated on two levels: customer service and
cost. Tables 8 and 9 of the case show that in March standards were not quite met overall.
Within specific product types, there was up to an eight percent deficit. Both order fill
rate and item fill rate were less than 100 percent for most products. There would seem to
be some room for improvement, especially in managing the variation among product
types.
From a cost standpoint, it is not known how efficiently the blood supply is managed
since no costs are reported. In addition, the revenue that the blood products generate is
not known. We would like to know how prices of the various products are set so that
revenues might be maximized, considering competition among some of the product line.
129
We do expect that demand is price elastic, since hospitals do shop around for blood
products that are available from local, commercial, and community blood banks. On the
other hand, ARC is the sole regional supplier of certain products such as platelets.
Setting product fill-rate standards at various levels can influence costs. We do not
know this effect.
Setting inventory levels by a number of days of inventory rule of thumb is simple
but not as effective as planning inventory levels based on the uncertainties that occur in
demand forecasts and supply lead times. The number-of-days of inventory rule does tend
to lead to too much inventory or to too many out-of-stock situations.
The plan for evaluation, if enough data were available, would be to establish a base
case of cost and service. This, then, would provide a basis for evaluating the effect of
change in the supply procedures.
(3) Can you suggest any changes in ARCs inventory planning and control practices that
might lead to cost reduction or service improvement?
Suggestions for improvement in blood supply management stem from a basic

understanding of the nature of the demand-supply relationship. When supply is uncertain
and all supply must be taken that is available, there is the possibility that significant
excess inventory will occur. The goal is to manage the demand in the short run to
reduce inventory levels when overstocking occurs, rather than focusing on managing
supply. Several approaches for doing this are:
Aggressively price selected products that are in excess supply and are nearing their
expiration dates, e.g. run a sale or offer price discounts.
Sell off excess supply to secondary demand sources or other regions of the ARC.
Temporarily adjust return rules for hospitals.
Bring demand more in line with supply by converting products into derivative ones
that have excess demand, e.g., reprocess whole blood into plasma.
Encourage hospitals to buy certain products in excess supply for a more favorable
status in buying other products that are in short supply, such as phersis platelets and
rare whole blood types.
Try to create excess demand for all products, especially those items that are
available from local blood banks, through promotion of ARCs distinct advantages,
such as quality, high service levels, and a wide range of blood derivative products.
Offer two-for-one sales, such that if a hospital buys one blood product, it may
receive another at a favorable price.
Pool the risk of uncertain demand by maintaining a central inventory for all
hospitals, or managing the inventories at all hospitals, as well at ARC, collectively.
Provide quick deliveries or transfers among inventory locations.
ARC should attempt to be the premier provider of blood products and leverage the
advantage. This will allow it to maintain a degree of control over the demand for blood.
Effectively controlling demand in turn allows it to control its costs and avoid product
outdating.
130
(4) Is pricing policy an appropriate mechanism to control inventory levels? If so, how
should price be determined?
From the previous discussion, it can be seen that price plays a role in controlling demand.
Since there appears a relationship between demand and price for some products,
especially among those products offered by local blood banks that compete with ARC
blood products, price may be an effective weapon to meet competition. Rather than
setting price based on the cost of production, ARC might consider raising the price on
products for which it is the sole provider, such as platelets, and then meeting the price of
competitors on whole blood. Although ARC strives to be a nonprofit organization, the
increased volume that an effective pricing strategy promotes would allow more of the
fixed costs to be covered. This may lead to lower overall average prices for ARCs
products.
Blood could also be priced as a function of its freshness at two or more levels.
Although blood that has been donated within 42 days legally can be utilized, the quality
of blood does not remain the same for the entire 42-day period. A chemical compound
found in blood, called 2,3-DPG, decreases with the age of the stored blood, and is
believed to be important in oxygen delivery. For this reason, certain procedures such as
heart transplants and neonatal procedures require that blood be fresh, usually donated
within 10 days or less. Thus, a simple pricing policy could be to charge a higher price for
blood that is less than 10 days old, and a lower price for blood that is between 10 and 42
days old. Price differences here are based on product quality.
131
CHAPTER 10
PURCHASING AND SUPPLY SCHEDULING DECISIONS
1
(a) The following requirements schedules will lead to the proper timing and quantities for
the purchase orders.
Desk style A
Week
1 2 3 4 5 6 7 8
Sales forecast 150 150 200 200 150 200 200 150
Receipts 200 300 300 300 300
Qty on hand 0 50 200 0 100 250 50 150 0
Releases to prod. 300 300 300 300
Desk style B
Week
1 2 3 4 5 6 7 8
Sales forecast 60 60 60 80 80 100 80 60
Receipts 100 100 100 100 100
Qty on hand 80 20 60 0 20 40 40 60 0
Releases to prod. 100 100 100 100 100
Desk style C
Week
1 2 3 4 5 6 7 8
Sales forecast 100 120 100 80 80 60 60 80
Receipts 100 100 100 100 100
Qty on hand 200 100 80 80 0 20 60 0 60
Releases to prod. 100 100 100 100 100
Summing the releases for these three desk release schedules gives a production
requirements schedule for desks in general and sheets of plywood in particular. That
is,
Week
1 2 3 4 5 6 7 8
Desk requirement 500 100 400 500 200 400 100 0
Plywood sheetsa 1500 300 1200 1500 600 1200 300 0
a
Desk requirements times 3
Now, find the purchase order releases for the plywood sheets.
Week
1 2 3 4 5 6 7 8
Sales forecast 1500 300 1200 1500 600 1200 300 0
Receipts 600 1000 1000 1000 1000
Qty on hand 2400 900 1200 1000 500 900 700 400 400
Releases to prod. 1000 1000 1000 1000
132
Therefore, purchase orders should be placed in weeks 1, 2, 3, and 4 for 1000 sheets
each.
(b) Using Equation 10-2 in the text, the probability of not having the plywood sheets at
the time needed would be:
Pc 5
Pr = 1 = 1 = 0.02
Cc + Pc . +5
01
From Appendix A, z@1-.02 = 2.05. Therefore, the lead-time should be:
T * = LT + z sLT = 14 + 2.05( 2) = 181

. days
Another week should be added to the current lead-time of 2 weeks.
2
(a) Using Equation 10-2, the probability of not having the item when needed for
production is:
Pc 150
Pr = 1 = 1 = 0.0001
Cc + Pc ( 0.2 35 / 365) + 150
The time to place an order ahead of need is:

T * = LT + z s LT = 14 + 3.6( 4) = 28 days
where z@1-.0001 = 3.6 from Appendix A.
(b) Use part period cost balancing. The unit carrying cost is (0.2/52)35 = 0.134. Then,
(Q=250) Week 4
0.134[500 + 200]/2 = 46.9
(Q=1350) Weeks 4 + 5
0.134[(1350 + 1050)/2 + (1050 + 200)/2] = 244.6
The carrying cost closest to the order cost of $50 is Q = 250. Order this amount.
3
Using the requirements planning procedure, we can develop a schedule of material flows
through the network over the next 10 weeks.
Whse 1 1 2 3 4 5 6 7 8 9 10
Requirements 1200 1200 1200 1200 1200 1200 1200 1200 1200 1200
Schd receipts 7500 7500
On-hand qty 1700 500 6800 5600 4400 3200 2000 800 7100 5900 4700
Releases 7500 7500
133
Whse 1 1 2 3 4 5 6 7 8 9 10
Requirements 1200 1200 1200 1200 1200 1200 1200 1200 1200 1200
On-hand qty 1700 500 6800 5600 4400 3200 2000 800 7100 5900 4700
Releases 7500 7500
Whse 2 1 2 3 4 5 6 7 8 9 10
Requirements 2300 2300 2300 2300 2300 2300 2300 2300 2300 2300
Schd receipts 7500 7500 7500
On-hand qty 3300 1000 6200 3900 1600 6800 4500 2200 7400 5100 2800
Releases 7500 7500 7500
Whse 3 1 2 3 4 5 6 7 8 9 10
Requirements 2700 2700 2700 2700 2700 2700 2700 2700 2700 2700
Schd receipts 7500 7500 7500 7500
On-hand qty 3400 700 5500 2800 100 4900 2200 7000 4300 1600 6400
Releases 7500 7500 7500 7500
Regnl whse A 1 2 3 4 5 6 7 8 9 10
Requirements 22500 0 0 15000 0 15000 7500 0 7500 0
On-hand qty
52300 29800 29800 29800 14800 14800 14800 7300 7300 1300 1300
Releases to
plant 15000 15000
Whse 4 1 2 3 4 5 6 7 8 9 10
Requirements 4100 4100 4100 4100 4100 4100 4100 4100 4100 4100
Schd receipts 7500 7500 7500 7500 7500
On-hand qty 5700 1600 5000 900 4300 200 3600 7000 2900 6300 2200
Releases 7500 7500 7500 7500 7500
Whse 5 1 2 3 4 5 6 7 8 9 10
Requirements 1700 1700 1700 1700 1700 1700 1700 1700 1700 1700
On-hand qty 2300 600 6400 4700 3000 1300 7100 5400 3700 2000 300
Releases 7500 7500
Whse 6 1 2
3 4 5 6 7 8 9 10
Requirements 900
900 900 900 900 900 900 900 900 900
On-hand qty 1200 300 6900 6000 5100 4200 3300 2400 1500 600 7200
Releases 7500 7500
Regnl whse B 1 2 3 4 5 6 7 8 9 10
Requirements 22500 0 7500 0 15000 7500 0 7500 7500 0
Schd receipts 15000 15000 15000
On-hand qty
31700 9200 24200 16700 16700 1700 9200 1700 9200 9200 9200
Releases to
plant 15000 15000
134
Plant 1 2 34 5 6 7 8 9 10
Requirements 0 0 030000 0 0 30000 0 0 0
On-hand qty 0 0 0 0 10000 10000 10000 0 0 0 0
Releases-matls 40000 20000
Summing the releases to the plant shows that the plant should place 30,000 cases into
production in weeks 4 and 7.
Because demand is shown to be constant, the average inventory must be one-half the
order quantity. For the six field warehouses and a shipping quantity of 7500, the average
long run inventory would be (7500/2)6 = 22,500 cases. For the regional warehouses,
the average inventory would be (15,000/2)2 = 15,000 cases. For the plant, the average
inventory would be 20,000/2 = 10,000 cases. The total system average inventory would
be 22,500 + 15,000 + 10,000 = 47,500 cases.
4
(a) The leverage principle shows the relative change that must be made in cost, price, or
sales volume to affect a given change in the profit level. Usually it is used in
reference to the cost of goods sold to show the impact that small changes in the cost
of goods will have on profits and the important role that purchasing plays in the
profitability of the firm. The following simple profit and loss statements will show
how much change is needed in various activities to increase profits by 10 percent.
Sales Price L&S OH COG

Current (+4%) (1%) (-3%) (-6%) (-2%)
Sales $55.0 $57.2 $55.5 $55.0 $55.0 $55.0
Cost of goods 27.5 28.6 27.5 27.5 27.5 27.0
Labor & salaries 15.0 15.6 15.0 14.5 15.0 15.0
Overhead 8.0 8.0 8.0 8.0 7.5 8.0
Profit $ 4.5 $ 5.0 $ 5.0 $ 5.0 $ 5.0 $ 5.0
Due to the magnitude of cost of goods sold, it requires less than a two percent
change in COG to increase profits by 10 percent.
(b) The current ROA as:
Profit margin = (4.5/55)100 = 8.2 percent

Investment turnover = 55/20 = 2.75
ROA = 2.758.2 = 22.6 percent
Reducing cost of goods by 7 percent will increase profits to 55 27.50.93 15 8

= $6.43 and the profit margin now is 6.43100/55 = 11.7 percent. Inventory at 20
percent of total assets is $4 million. If the cost of goods is reduced by 7 percent,
inventory value will decline to $40.93 = $3.72. Total assets will be 3.72 + 16 =
$19.72 million. The investment turnover is 55/19.72 = 2.789. The ROA now will be
11.72.789 = 32.63 percent.
135
5
(a) A mixed purchasing strategy will generally be beneficial when prices show a definite
seasonality, they are predictable, and inventory costs associated with forward buying
are not excessive. In the problem, we should consider forward buying in the first half
of the year and hand-to-mouth buying in the last half. To test the various strategies,
compare (1) hand-to-mouth buying, (2) forward buying every 2 months, (3) forward
buying every 3 months, and (4) forward buying for the first 6 months. The results are
summarized in Table 10-1.
The inventory for the hand-to-mouth buying strategy can be approximated as
50,000/2 = 25,000. The carrying cost would be 0.304.9825,000 = $37,350 per
year.
The carrying cost for the two month forward buying strategy is:
0.304.88[(0.5100,000/2) + (0.550,000/2)] = $54,900
For the 3-month forward buying strategy:
0.34.56[(0.5300,000/2) + (0.550,000/2)] = $119,700
From the total costs in Table 10-1, the best strategy is to forward buy the first six-
month's requirements in January and hand-to-mouth buy for the last six months.
(b) Some possible disadvantages are:
Prices may fall rather than rise in the first six months
There may not be adequate storage space to accommodate such a large purchase.
The materials may be perishable and not easily stored.
Uncertainties in the requirements and carrying costs may void the strategy.
136
TABLE 10-1 A Comparison of Various Forward Buying Strategies with Hand-to-Mouth Buying
Hand-to-mouth buy 2-month forward buy 3-month forward buy 6-month forward buy
Price, Quantity, Price, Quantity, Price, Quantity, Price, Quantity,
$/unit units Total $/unit units Total $/unit units Total $/unit units Total
Jan 4.00 50,000 $200,000 4.00 100,000 $400,000 4.00 150,000 $600,000 4.00 300,000 $1,200,000
Feb 4.30 50,000 215,000
Mar 4.70 50,000 235,000 4.70 100,000 470,000
Apr 5.00 50,000 250,000 5.00 150,000 750,000
May 5.25 50,000 262,000 5.25 100,000 525,000
Jun 5.75 50,000 287,500
Jly 6.00 50,000 300,000 6.00 50,000 300,000 6.00 50,000 300,000 6.00 50,000 300,000
Aug 5.60 50,000 280,000 5.60 50,000 280,000 5.60 50,000 280,000 5.60 50,000 280,000
Sep 5.40 50,000 270,000 5.40 50,000 270,000 5.40 50,000 270,000 5.40 50,000 270,000
Oct 5.00 50,000 250,000 5.00 50,000 250,000 5.00 50,000 250,000 5.00 50,000 250,000
Nov 4.50 50,000 225,000 4.50 50,000 225,000 4.50 50,000 225,000 4.50 50,000 225,000
Dec 4.25 50,000 212,000 4.25 50,000 212,000 4.25 50,000 212,500 4.25 50,000 212,500
Subtotals $2,987,500 $2,932,500 $2,887,500 $2,737,500
Inventory costs 37,350 54,900 72,150 119,700
Totals $3,024,850 $2,987,400 $2,959,650 $2,857,200
Average price/unit $4.98 $4.88 $4.81 $4.56
137
6
(a) On the average, a total expenditure of 1.1025,000 = $27,500 should be made for
copper each month.
(b) For the next 4 months, the dollar averaging purchases would be:
(1) (2) (3)=(1)(2) (4)=(2)/2

Price, No. of Total Average
Month $/lb. lb. cost,$ inventory, lb.
1 1.32 20,833 27,500 10,417
2 1.05 26,190 27,500 13,095
3 1.10 25,000 27,500 12,500
4 0.95 28,947 27,500 14,474
100,970 $110,000 12,622a
a
50,486/4 = 12,622
The average per-lb. cost would be $110,000/100,970 = $1.089. The inventory

carrying cost over 4 months would be 0.201.089(4/12) 12,622 = $916.
If hand-to-mouth were used, we would have:
(1) (2) (3)=(1)(2) (4)=(2)/2

Price,No. of Total Average
Month $/lb. lb. cost,$ inventory, lb.
1 1.32 25,000 33,000 12,500
2 1.05 25,000 26,250 12,500
3 1.10 25,000 27,500 12,500
4 0.95 25,000 23,750 12,500a
100,000 $110,500 12,500
a
50,000/4 = 12,500
The average per-lb. cost would be $110,500/100,000 = $1.105. The inventory

carrying cost over 4 months would be 0.201.105(4/12) 12,500 = $921.
If 100,000 lbs. of copper were purchased, the two strategies can be compared as
follows.
Purchase Inventory Total

Strategy cost cost cost
Dollar averaging $108,900 + 916 = $109,816
Hand-to-mouth 110,500 + 921 = 111,421
Dollar averaging buying would be preferred.
7
For an inclusive quantity discount price incentive plan, we first compute the economic
order quantities for each range of price. Using
Q * = 2 DS / IC
we compute
138
Q1* = 2(500)(15) / ( 0.20)( 49.95) = 38.75 cases
Q2* = 2(500)(15) / ( 0.20)( 44.95) = 40.85 cases
Since Q2* is outside of the second price bracket, Q1* is the only relevant quantity. Now
we check the total cost at Q1* and at the minimum quantities within the price break. We
solve:
TCi = Pi D + DS / Qi + ICi Qi / 2
At Q = 38.75
TC = 49.95500 + 50015/38.75 + 0.249.9538.75/2

= $25,362
At Q = 50
TC = 44.95500 + 50015/50 + 0.244.9550/2

= $22,850
At Q = 80
TC = 39.95500 + 50015/80 + 0.239.9580/2

= $20,388
Floor polish should be purchased in quantities of 80 cases.
8
This noninclusive price discount problem requires solving the following relevant total
cost equation for various order quantities until the minimum cost is found.
TCi = Pi D + DS / Qi + ICi Qi / 2
The computations can be shown in the table below given that D = 1,400, S = 75, and I =
0.25.
139
Q Price P D +DS/Q +ICQ/2 = Total cost
20 795 1,113,000.00 5,250.00 1,987.50 $1,120,237.50
50 795 1,113,000.00 2,100.00 4,968.75 1,120,068.75
100 795 1,113,000.00 1,050.00 9,937.50 1,123,987.50
200 795 1,113,000.00 525.00 19,875.00 1,133,400.00
300 200795+100750 1,092,000.00 350.00 29,250.00 1,121,600.00
300
400 200795+200750 1,081,500.00 262.50 38,625.00 1,120,387.50
400
500 200795+200750 1,068,200.00 210.00 47,687.50 1,116,097.50
+100725
500
550 200795+200750 1,063,363.64 190.91 52,218.75
1,115,773.27
+150725
550
600 200795+200750 1,059,333.33 175.00 56,750.00 1,116,258.33
+200725
600
The optimal purchase quantity is 550 motors.
9
(a) This problem is a good application of the transportation method of linear
programming. We begin by determining the costs for the current sourcing
arrangement.
Source Destination Price Transport Volume Cost

Dayton Cincinnati 3.40 0.05 5,000 $17,250
Dayton Baltimore 3.40 0.15 1,000 3,550
Kansas City Dallas 3.45 0.08 2,500 8,825
Minneapolis Los Angeles 3.25 0.24 1,200 4,188
Total $33,813
To optimize, we establish the following transportation cost matrix and solve it using
any appropriate method, such as the TRANLP module in LOGWARE.
Cincin- Los
nati Dallas Angeles Baltimore Capacity
3.40 3.44 3.49 3.46
Minneapolis 1200 1200
3.55 3.53 3.65 3.63
Kansas City 4800
3.45 3.52 3.67 3.55
Dayton 5000 2500 1000 9999
Requirements 5000 2500 1200 1000
The total cost for this solution is $33,788, or a savings of $25 over the current sourcing.
140
(b) Because Minneapolis is at capacity, this supplier should be examined further. If
unlimited capacity were available at Minneapolis, all requirements would be met by
this supplier for a total cost of $33,248, or a savings of $565 for this material.
(c) The above analysis does indicate that too many suppliers are being used. Only two
are needed if Minneapolis continues to supply at the current level. If Minneapolis can
be expanded, it becomes the only supplier. Of course, whether the company would
risk a single supplier for this material must be left unanswered.
10
(a) The deal-buying equation (Equation 10-5) can be applied to this problem. First, find
the optimal order quantity before the discount.
2 DS 2(120,000 )( 40 )
Q* = = = 566 units
IC 0.30(100 )
Next, find the adjusted order quantity after the discount has been applied.
dD pQ * 10(120,000) 100(566)
Q$ = + = + = 42,700 units
( p d ) I p d (100 5)( 0.30) (100 5)
A large order size of 42,700 units should be placed.
(b) The time that an order of this size will be held before it is depleted is given by:
Q$ 42,700
= = 0.356 years, or 18.5 weeks
D 120,000
141
INDUSTRIAL DISTRIBUTORS, INC.
Teaching Note
Strategy
The purpose of the Industrial Distributors case study is to illustrate the computation of
purchase quantities under inclusive and noninclusive price discounts and transport rate-
weight breaks. The INPOL module of LOGWARE is helpful in conducting the analysis.
As a teaching strategy, it may be worthwhile to begin any class discussion with the cost
tradeoffs that are present in such a problem as this. This will help to establish the nature
of the total cost equation that needs to be solved in this problem.
(1) What size of replenishment orders, to the nearest 50 units, should Walter place, given
the manufacturer's noninclusive price policy?
When price discounts are offered, purchase quantities are not simply determined by a
single formula. Due to discontinuities in the total cost curve as a function of order
quantity, the optimal order quantity is found by computing total costs for different
quantity values. In this case of both price and transport rate breaks plus warehousing
costs that can be affected by the order size, the following annual total cost formula is to
be solved.
SD ICQ
TC = PD + RD + + + W ( Q - 300)
Q 2
where
TC = total cost for quantity Q, $

PD = purchase cost for price P, $
RD = transport costs at rate R, $
SD/Q = ordering cost at quantity Q, $
ICQ/2 = carrying cost at quantity Q, $
W(Q-300) = public warehousing cost if Q is greater than 300 units, $
W = public warehousing rate, $ per unit per year
D = annual demand, units
P = price for orders of size Q, $ per unit
R = transport per unit for shipments of size Q, $ per unit
S = order processing cost, $ per order
I = annual carrying cost, %
C = product value, $ per unit
Q = size of purchase order, units
Under noninclusive price discounts, price is an average, determined by the number of

units in each break. For example, if 250 units are to be ordered, the average price per
unit would be computed as:
142
(100 $700) + (100 $680) + (50 $670)
P250 = = $686.00
250
A table of annual costs can now be developed, as shown in Table 1. To the nearest 50
units, the optimal purchase quantity should be 250 units.
(2) If the manufacturer's pricing policy were one where the prices in each quantity break
included all units purchased, should Walter change his replenishment order size?
The average price per unit is more easily determined in this case than the previous one.
Since all units are included in the price break back to the first unit, the average price is
simply the price associated with a given purchase quantity.
Finding the optimal purchase quantity is simply a matter of determining the total cost
for the quantities, found by the economic order quantity formula, assuming these
quantities are feasible, and for the quantities at the transport rate-weight break. The
comparison is made among the total costs of these alternatives. These costs are shown in
Table 2.
The order quantities, as determined by the economic order quantity formula for the
base price of $700, would be:
2 DS 2(1500 )( 25)
Q* = = = 18.8, or 19 units
IC 0.3( 700 + 7.2 )
where C is the $700 price per unit at Baltimore plus the $45 transport cost from
Baltimore, as determined by an LTL shipment (19 units 250 lb. = 4,750 lb.) at $18
2.5 cwt. = $45 per unit. The Q values for the other prices in the schedule lie outside the
feasible range of the price used to compute Q.
The optimal strategy is to purchase 201 units per order, which is one unit into the last
price break. Yes, Walter should alter his buying strategy.
TABLE 1 Annual Costs by Quantity Purchased for Noninclusive Price Discounts

Average Purchase Transport Ordering Carrying Warehouse
Quantity price cost cost cost cost cost Total cost
19 $700.00 $1,050,000 $67,500 $2,049 $2,045 $0 $1,121,594
50 700.00 1,050,000 67,500 750 5,588 0 1,123,838
100 700.00 1,050,000 67,500 371 11,287 0 1,129,158
150 693.33 1,039,995 67,500 250 16,613 0 1,124,363
160 692.50 1,038,750 45,000 234 17,340 0 1,101,324
200 690.00 1,035,000 45,000 187 21,707 0 1,101,818
250 686.00 1,029,000 45,000 150 26,850 0 1,101,000Opt.
300 683.33 1,063,286 45,000 125 32,100 0 1,102,225
400 680.00 1,020,000 45,000 94 42,600 1,000 1,108,694
a
EOQ at a price of ($700 + 45) per unit.
b
First price break.
c
Transport rate break.
d
Second price break.
TABLE 2 Annual Costs by Quantity Purchased for Inclusive Price Discounts
143
Average Purchase Transport Ordering Carrying Warehouse
Quantity price cost cost cost cost cost Total cost
19 $700.00 $1,050,000 $67,500 $2,049 $2,045 $0 $1,121,594
19 680.00 Infeasible
19 670.00 Infeasible
101 680.00 1,020,000 67,500 371 10,984 0 1,098,855
160 680.00 1,020,000 45,000 234 17,040 0 1,082,274
201 670.00 1,005,000 45,000 187 21,105 0 1,032,732Opt.
a
Feasible EOQ at a price of ($700 + 45) per unit.
b
Infeasible EOQ at a price of ($680 + 45) per unit.
c
Infeasible EOQ at a price of ($670 + 30) per unit.
d
First price break.
e
Transport rate break.
f
Second price break.
144
CHAPTER 12
STORAGE AND HANDLING DECISIONS
2
Various alternatives are evaluated in Tables 12-1 to 12-4. The annual costs of each
alternative are plotted in Figure 12-1. The best economic choice is to use all public
warehousing.
146
TABLE 12-1 Costs for a Pure Public Warehouse Strategy
Privately-operated Rented
Ware- Space
house require- Private Monthly Monthly Rented Monthly Monthly
thruput, ments, allo- fixed cost variable allo- storage handling Monthly
Month lb.a sq. ft.b cation cost cation cost cost total cost
Jan. 1,000,000 62,500 0% $0 $0 100% $30,000c $50,000d $80,000
Feb. 800,000 50,000 0 0 0 100 24,000 40,000 64,000
Mar. 600,000 37,500 0 0 0 100 18,000 30,000 48,000
Apr. 400,000 25,000 0 0 0 100 12,000 20,000 32,000
May 200,000 12,500 0 0 0 100 6,000 10,000 16,000
June 50,000 3,125 0 0 0 100 1,500 2,500 4,000
July 250,000 15,625 0 0 0 100 7,500 12,500 20,000
Aug. 450,000 28,125 0 0 0 100 13,500 22,500 36,000
Sept. 600,000 37,500 0 0 0 100 18,000 30,000 48,000
Oct. 700,000 43,750 0 0 0 100 21,000 35,000 56,000
Nov. 800,000 50,000 0 0 0 100 24,000 40,000 64,000
Dec. 900,000 56,250 0 0 0 100 27,000 45,000 72,000
Totals 6,750,000 421,875 $0 $0 $202,500 $337,500 $540,000
a
Thruput (lb.) = Sales ($)/($5/lb.)
b
Space requirements (sq. ft.) = Thruput (lb.) 1/0.40 0.1/10(5) = Thruput 0.6250
c
Given a turnover ratio of 2 and 100% of the demand through the rented warehouse, then 1,000,000 1.00/2 0.06 = $30,000
d
1,000,000 1.00 0.05 = $50,000
147
TABLE 12-2 Costs for a Mixed Warehouse Strategy Using a 10,000 Square Foot Privately-
Operated Warehouse
Ware- Space
Jan. 1,000,000 62,500 16% $9,792d $3,200e 84% $25,200f $42,000g $80,192
Feb. 800,000 50,000 20 9,792 3,200 80 19,200 32,000 64,192
Mar. 600,000 37,500 27 9,792 3,200 73 13,140 21,900 48,032
Apr. 400,000 25,000 40 9,792 3,200 60 7,200 12,000 32,192
May 200,000 12,500 80 9,792 3,200 20 1,200 2,000 16,192
June 50,000 3,125 100 9,792 1,000 0 0 0 10,792
July 250,000 15,625 64 9,792 3,200 36 2,700 4,500 20,192
Aug. 450,000 28,125 36 9,792 3,200 64 8,640 14,400 36,032
Sept. 600,000 37,500 27 9,792 3,200 73 13,140 21,900 48,032
Oct. 700,000 43,750 23 9,792 3,200 77 16,170 26,950 56,112
Nov. 800,000 50,000 20 9,792 3,200 80 19,200 32,000 64,192
Dec. 900,000 56,250 18 9,792 3,200 82 22,140 36,900 72,032
Totals 6,750,000 421,875 $117,504 $36,200 $147,930 $246,550 $548,184
a
b
c
10,000/62,500 = 0.16
d
(3510,000/20) + 1010,000/12 = $9,792 per month
e
1,000,0000.160.02 = $3,200
f
d
1,000,000 0.84 0.05 = $42,000
148
TABLE 12-3 Costs for a Mixed Warehouse Size Strategy Using a 30,000 Square Foot Privately-
Operated Warehouse
Ware- Space
Jan. 1,000,000 62,500 48%c $29,375d $9,600e 52% $15,600f $26,900g $80,575
Feb. 800,000 50,000 60 29,375 9,600 40 9,600 16,000 64,575
Mar. 600,000 37,500 80 29,375 9,600 20 3,600 6,000 48,575
Apr. 400,000 25,000 100 29,375 8,000 0 0 0 37,375
May 200,000 12,500 100 29,375 4,000 0 0 0 33,375
June 50,000 3,125 100 29,375 1,000 0 0 0 30,375
July 250,000 15,625 100 29,375 5,000 0 0 0 34,375
Aug. 450,000 28,125 100 29,375 9,000 0 0 0 38,375
Sept. 600,000 37,500 80 29,375 9,600 20 3,600 6,000 48,575
Oct. 700,000 43,750 69 29,375 9,600 31 6,510 13,020 58,505
Nov. 800,000 50,000 60 29,375 9,600 40 9,600 16,000 64,575
Dec. 900,000 56,250 53 29,375 9,600 47 12,690 21,150 72,815
Totals 6,750,000 421,875 $352,500 $94,200 $61,200 $104,170 $612,070
a
b
c
30,000/62,500 = 0.48
d
(3530,000/20) + 1030,000/12 = $29,375 per month
e
1,000,0000.480.02 = $9,600
f
d
1,000,000 0.52 0.05 = $26,000
149
TABLE 12-3 Costs for a Mixed Warehouse Size Strategy Using a 40,000 Square Foot Privately-
Operated Warehouse
Ware- Space
Jan. 1,000,000 62,500 64%c $39,167 $12,800e 36% $10,800f $18,000g $80,767
Feb. 800,000 50,000 80 39,167 12,800 20 4,800 8,000 64,767
Mar. 600,000 37,500 100 39,167 12,800 0 0 0 51,167
Apr. 400,000 25,000 100 39,167 8,000 0 0 0 47,167
May 200,000 12,500 100 39,167 4,000 0 0 0 43,167
June 50,000 3,125 100 39,167 1,000 0 0 0 40,167
July 250,000 15,625 100 39,167 5,000 0 0 0 44,167
Aug. 450,000 28,125 100 39,167 9,000 0 0 0 48,167
Sept. 600,000 37,500 100 39,167 12,000 0 0 0 51,167
Oct. 700,000 43,750 91 39,167 12,800 0 1,890 3,150 57,007
Nov. 800,000 50,000 80 39,167 12,800 20 4,800 8,000 64,767
Dec. 900,000 56,250 71 39,167 12,800 29 7,830 13,050 72,847
Totals 6,750,000 421,875 $470,004 $115,000 $30,120 $50,200 $665,324
a
b
c
40,000/62,500 = 0.64
d
(3540,000/20) + 1040,000/12 = $39,167 per month
e
1,000,0000.640.02 = $12,800
f
d
1,000,000 0.36 0.05 = $18,000
150
FIGURE 12-1
670 Total Annual
650 Costs for a
Combined
Total cost, $000s
630 Warehouse
610 Size Using
Private and
590
Public
570 Warehouse
550
Space
530
0 10,000 30,000 40,0 00
Private warehouse space, sq. ft.
3
The annual cost of public warehousing is:
Handling $ 600,000
Storage 300,000
Total $ 900,000
The costs of private warehousing are:
Annual operating $ 250,000

Annual lease payment 3150,000 = 450,000
Other fixed (one time) 400,000
The savings in operating costs of lease vs. public warehousing is:
Savings = $900,000 250,000 = $650,000/yr.
151
TABLE 12-5 Ten-Year Cash Flow Stream for Public vs. Leased Warehouse Comparison
Savings
Depre- Savings less
Savings: Pre-tax ciation less depre- After-tax Discount Dis-
Lease vs. net cash sche- depre- Taxes ciation Savings net cash factor counted
Year public flow dule ciation (35%) & tax less tax flow 1/(1+i)j cash flow
0 $0 (3,050)a $0 0 0 0 0 ($3,050) ($3,050)
1 650 650 57b 593 208 385 442c 442 0.9009 398
2 650 650 57 593 208 385 442 442 0.8116 359
3 650 650 57 593 208 385 442 442 0.7312 323
4 650 650 57 593 208 385 442 442 0.6587 291
5 650 650 57 593 208 385 442 442 0.5935 262
6 650 650 57 593 208 385 442 442 0.5346 236
7 650 650 58 592 207 385 443 443 0.4817 213
8 650 650 0 650 228 422 442 442 0.4339 183
9 650 650 0 650 228 422 442 442 0.3909 165
10 650 650 0 650 228 422 442 442 0.3522 149
$6,500 $3,450 $400 $6,100 $2,139 $3,961 $4,361 $1,311 NPV = ($471)
a
Capitalization lease plus initial cash outlay, i.e., $2,650,154 + 400,000 = $3,050,154
b
Depreciation charge for each of seven years is 1/7 = 0.1429 such that 400,0000.1429 = $57,143
c
Add back depreciation, i.e., 385 + 57 = $442
152
Capitalizing the lease over ten years, we have:
(1 + 011
. )10 1
PV = 450,000 = $2,650,154
. (1 + 011
011 . )10
The initial investment in $000s then is:
Initial investment = $2,650 + 400 = $3,050
The ten-year cash flow stream is shown in Table 12-5. Since the savings are expressed to
favor leasing and the net present value is negative, choose public warehousing.
4
Given:
k = $210/sq ft.
S = 100,000 sq. ft.
C = $0.01/ft.10,000 = $100/ft.
The width is:
C + 8k
W* = S
2C + 8 k
100 + 8( 210)
= 100,000
2(100) + 8( 210)
= 308 ft.
The length is:
L* = S / W * = 100,000 / 308 = 325 ft.
5
Space layout according to text Figure 12-4(a) can be determined by the application of
Equations 12-8 and 12-9. These equations specify the best number of shelf spaces and
the best number of double racks, respectively. Equations 12-10 and 12-11 give the length
and width of the building.
The optimal number of shelf spaces would be:
153
1 dCh + 2aCs + 2C p K ( w + a ) L
m1* =
L 2( dCh + C p ) 2h
1 400,000( 0.001) + 2(10)( 0.50) + 2(3.00) 50,000(8 + 10)( 4)

=
4 2( 400,000)( 0.001) + 300
.
2( 4)

= 120.48, or 121
The optimal number of double racks would be:
1 2( dCh + C p ) K ( w + a ) L
n1* =
w + a dCh + 2aCs + 2C p 2h
1 2[( 400,000 )( 0.001) + 3.00] 50,000(8 + 10 )( 4 )

=
8 + 10 400,000( 0.001) + 2(10 )( 0.50 ) + 2(3.00 )
2( 4)

= 52
The warehouse length would be:
u1 = n1* ( w + a ) = 52(8 + 10) = 936 ft.
and the width would be:
v1 = 2a + m1* L = 2(10) + 121( 4) = 504 ft.
6
According to Equation 12-17, the number of truck doors can be estimated by:
DH
N=
CS
Therefore,
N = (7512,000) 3/(312,000) 8 = 9.37, or 10 doors
7
Summarizing the given information as follows:
Three Five Seven

type 1 type 2 type 3
units units units
Initial investment $60,000 $50,000 $35,000
Useful life 10 yr. 10 yr. 10 yr.
Salvage value @15%
of initial cost $ 9,000 $ 7,500 $ 5,250
154
Annual operating
expenses $ 6,000 $12,500 $21,000
Return on investment
before tax 20% 20% 20%
An initial solution to this problem can be found through a discounted cash flow
analysis. Three alternatives are to be evaluated.
(1 + 0.2)10 1 1
PV1 = 60,000 + 6,000 10
9,000 10
0.2(1 + 0.2) (1 + 0.2)
= 60,000 + 6,000( 4.2) 9,000( 016
. )
= $83,760
PV2 = 50,000 + 12,500( 4.2) 7,500( 016

. )
= 50,000 + 52,500 1,200
= $101,300
PV3 = 35,000 + 21,000( 4.2) 5,250( 016

. )
= 35,000 + 88,200 840
= $122,360
The low present value of the Type 1 truck indicates that from among these three
alternatives, this would be the best buy.
8
Given:
Initial cost of equipment = $4,000

Operating costs 500 + 40(t - 1)2 - 30(t - 1)
Salvage value Sn = I(1 t/7)
Rate of return on investment = 20%
Replacement is expected to be with equipment of like kind
The best replacement year can be found by comparing the equivalent annual cost of a
sequence of similar equipment replaced every n years. The equivalent annual cost is:
[ ][ ]
AC n = I + j=1 (C j /(1 + i ) j ) ( S n /(1 + i ) n ) i (1 + i ) n /(1 + i ) n 1
n
Solving this equation for different years is facilitated if the equation is set up in tabular
form, as shown in Table 12-6.
The equipment should be replaced at the end of the third year of service although a 5-
year replacement cycle is also attractive.
155
TABLE 12-6 Equivalent Annual Cost Computations for Problem 8
(1) (2) (3) (4) (5)=(1+2-3)(4)
Operating Salvage Factor
Year, Initial costs, value, i(1 + i ) n Equivalent annual
n investment, I C j / (1+ i) j
S n / (1 + i ) n
(1 + i ) 1
n cost, ACn
1 $4,000 $416 $2,857 1.20 $1,871
2 4,000 770a 1,984 0.65 1,821
b
3 4,000 1,117 1,323 0.47 1,783
4 4,000 1,488 827 0.39 1,818
5 4,000 1,898 459 0.33 1,795
6 4,000 2,350 191 0.30 1,848
7 4,000 2,841 0 0.28 1,915
a
$416 + [500 + 40(2-1)2 30(2-1)]/(1+0.20)2 = $770
b
$770 + [500 + 40(3-1)2 30 (3-1)]/1+0.20)3 = $1,117
9
(a1) Layout by popularity involves locating the more frequently ordered items closest to
the outbound dock. Based on the average number of daily orders on which the item
appears, the items closest to the outbound dock would be ranked as follows:
B,I,E,A,F,H,J,C,G,D
The storage space might then be:
Inbound
D J, C, G
H, F, A H, J
A, E E
B B, I, E
Outbound
(a2) Layout by cube places the smallest items nearest the outbound dock. Using the
individual item size, the ranking would be as follows: A,E,I,C,J,H,G,B,F
The layout of items in the storage bays would be:
156
Inbound
F, D B
H, D H, G, B
D, J, C E, I, C
E, A E
Outbound
(a3) The cube-per-order index is created by ratioing the average required cubic footage
of a product to the average number of daily orders on which the item is requested.
Hence, this index is found as follows:
(1) (2) (3)=(1)/(2)

Space
required Daily CPO
Product cu. ft. orders index
A 5,000a 56 89
B 30,000 103 291
C 15,000 27 556
D 17,000 15 1,133
E 55,000 84 655
F 11,000 55 200
G 7,000 26 269
H 28,000 45 622
I 13,000 94 138
J 9,000 35 257
a
500 sq. ft. stacked 10 ft. high
Locating the products with the lowest index values nearest to the outbound dock
results in the following ranking and layout: A,I,F,J,G,B,C,H,E,D
157
Inbound
D E
E, F E
H, C B
F, I, A F, J, G, B
Outbound
(b) All of the above methods assume (1) that the product is moved to the storage
locations in large unit loads, but retrieved from the storage locations in relatively
small quantities and (2) that only one product is retrieved during an out-and-back trip.
Therefore, these methods do not truly apply to the situation of multiple picks on the
same trip. However, they may be used with some degree of approximation if the
products can be aggregated as one and grouped together or zoned in the same section
of the warehouse.
10
This extra challenging problem requires some knowledge of linear programming. It may
be formulated as follows:
Let Xij represent the amount per 1,000 units of product j stored in location i. Let Cij
be the handling time associated with storage bay i and product j. Gj is the capacity of
a bay for product j and Rj is the number of units of product j required to be stored.
The linear programming statement is:
Objective function
Zmin = .90X11 + .75X12 + .90X13 + .80X21 + .65X22 + .95X23
+ .60X31 + .70X32 + .65X33 + .70X41 + .55X42 + .45X43
+ .50X51 + .50X52 + .45X53 + .40X61 + .45X62 + .35X63
Subject to:
Capacity restrictions on bays
158
20X11 + 33.3X12 + 16.7X13 100
20X21 + 33.3X22 + 16.7X23 100
20X31 + 33.3X32 + 16.7X33 100
20X41 + 33.3X42 + 16.7X43 100
20X51 + 33.3X52 + 16.7X53 100
20X61 + 33.3X62 + 16.7X63 100
and storage requirements restrictions on products
X11 + X21 + X31 + X41 + X51 + X61 11
X12 + X22 + X32 + X42 + X52 + X62 4
X13 + X23 + X33 + X43 + X53 + X63 12
Solving the linear programming problem by means of any standard transportation code of
linear programming, such as LNPROG in LOGWARE, yields:
X12 = 1.610 The total minimum

X21 = 1.020 handling time is
X22 = 2.390 138.68 hours
X31 = 5.000
X43 = 5.988
X51 = 4.980
X53 = 0.024
X63 = 5.988
where Xs are in thousands of units. That is, product 1 should be stored in bays 3, 4, and 5
in quantities of 1,020, 5,000, and 4,980, respectively. Product 2 should be stored in bays
1 and 2 in quantities of 1,610 and 2,390, respectively. Product 3 should be stored in bays
4, 5, and 6 in quantities of 5,988, 24, and 5,988, respectively. Graphically, this is:
Bay 1 2 3 4 5 6 Require
Product -ments
1 1,020 5,000 4,980 11,000
2 1,610 2,390 4,000
3 5,988 24 5,988 12,000
% of bay
capacity 53.7 100.0 100.0 100.0 100.0 100.0
159
CHAPTER 14
THE LOGISTICS PLANNING PROCESS
3
The MILES module within the LOGWARE software is used to solve this problem. It
computes distance based on the great circle distance formula using longitude and latitude.
(a) The estimated road distance is 1,380 miles.
(b) The estimated road distance is 830 miles.
(c) Since both latitudes are in the same hemisphere, no adjustments need to be made.
The estimated distance is 244 miles, or 2441.61 = 393 km.
(d) In this case, one point is east and the other west of the Greenwich line. Therefore, we
need to set a sign convention. Let's set west longitudes as + and east longitudes as .
Thus, 2.20o E longitude is entered into MILES as 2.20 o. The estimated distance is
250 miles, or 2501.61 = 402.5 km.
4
Suppose that a certain linear grid coordinate system has been overlaid on a map of the
United States. The grid numbers are calibrated in miles, and there is a road circuity
factor of 1.21. Find the expected road distances between the following pairs of points:
Equation 14-1 in the text is used to approximate distances from linear coordinates.
The K factor in the equation is set at 1.21.
(a) Lansing, MI to Lubbock, TX
Location X Coordinate Y Coordinate

a. From Lansing, MI 924.3 1675.2
To Lubbock, TX 1488.6 2579.4
b. From El Paso, TX 1696.3 2769.3
To Atlanta, GA 624.9 2318.7
c. From Boston, MA 374.7 1326.6
To Los Angeles, CA 2365.4 2763.9
d. From Seattle, WA 2668.8 1900.8
To Portland, OR 2674.2 2039.7
D = 121
. (924.3 1,488.6 ) 2 + (1,675.2 2,579.4 ) 2 = 1,290 miles
(b) El Paso, TX to Atlanta, GA
D = 121
. (1,696.3 624.9)2 + ( 2,769.3 2,318.7)2 = 1,406 miles
(c) Boston, MA to Los Angeles, CA
202
D = 121
. ( 374.7 2,365.4)2 (1,326.6 2,7639
. )2 = 2,971 miles
(d) Seattle, WA to Portland, OR
D = 121
. ( 2,668.8 2,674.2)2 + (1,900.8 2,039.7)2 = 168 miles
5
The plot of the truck class rates is shown in Figure 14-1. The rates show a high degree of
linearity. A linear regression was found with aid of the MULREG module in
LOGWARE. The rate equation was determined to be:
R = 5.1745 + 0.0041D
The standard error of the estimate SE is 0.9766.

The coefficient of determination r2 is 0.928.
The best single estimate of the rate at 500 miles is:
R = 5.1745 + 0.0041500
= $7.23/cwt.
Assuming the error around the regression line is normally distributed, a 95 percent
confidence band would give a range for the actual rate. That is,
Y = R 1.96SE
= 7.23 1.914
where 1.96 is the normal deviate for the normal distribution representing 95 percent of
the area in a two-tailed distribution. The range of the estimate is:
$5.32/cwt. Y $9.14/cwt.
The r2 value of 0.928 indicates that a linear rate equation explains about 93 percent of
the variation in the data with distance. Such a simple relationship seems to represent the
rates quite well.
203
FIGURE 14-1 Plot of Truck Class Rates
20
18
16
14
Class rate, $/cwt.
12
10
8
Estimating line
6
4
2
0
0 500 1000 1500 2000 2500 3000 3500
Distance, miles
6
A plot of the average inventory level versus warehouse throughput is shown in Figure 14-
2. The multiple regression software in LOGWARE was used to test two equation forms.
The first was of the form
I = aTP b
and the other was of the form
I = a + bTP
Both forms showed high r2 values, with the exponential form being slightly higher at
0.9406. It was selected as the equation form to use. This equation was:
I = 0.704 TP 0.83
where TP and I are both expressed in thousands of dollars. We can now estimate that for
an annual warehouse throughput of $50,000,000, the average inventory would be:
I = 0.704 50,000 0.83

= 5,593939
. , or $5,593,939
This type of relationship is very useful in network planning, especially warehouse

location, to estimate how inventory levels will change when sales are reallocated to a
varying number of warehouses.
204
FIGURE 14-2 Plot of Inventory Levels and Warehouse Throughput for California
Fruit Growers Association
Average inventory level, $(Millions)

12
10
4
Estimating line
2
0
0 20 40 60 80 100
Annual warehouse thruput, $(Millions)
205
USEMORE SOAP COMPANY
Teaching Note
The purpose of this case study is to provide students with the opportunity to evaluate and
design a large-scale production-distribution network using real data and cost
relationships. To assist in the substantial amount of computational effort in this problem,
an interactive computer program (WARELOCA) is available in the LOGWARE
collection of software modules.
Major Issues
The text of the case suggests a number of questions that are critical to production-
distribution network design. These reduce to three major issues, namely:
(1) Should plant capacity be added and, if so, when and where?
(2) How many warehouses are optimal and where should they be located?
(3) Should the current customer service level be retained?
Although no change can be made in the network without potentially affecting other
variables, the attempt here will be to treat these questions sequentially to converge on a
good network design.
Numerous computer runs were made to provide the basic information needed in the
analysis. The more meaningful runs are summarized in Appendix A to this note. Tables
1 and 2 compare selected runs for both the current-year and the future-year time periods.
This information is used throughout the analysis of the major issues.
The Plant Expansion Issue

An attempt to meet 5-year growth goals using current plant capacity will cause the
system having a total capacity of 1,630,000 cwt. to be out of capacity in 1.7 years. That
is,
5th-year demand 1,908,606 cwt.

Current demand 1,477,026
Net increase 431,580 cwt.
Therefore, the average annual growth rate is 431,580/5 = 86,316 cwt. So, in (1,630,000
1,477,026)/86,316 = 1.7 years all available capacity will be depleted.
If no expansion of plant capacity occurs, then 1,908,606 1,630,000 = 278,606 cwt.
will potentially be lost by the 5th year. Sales are $100 million on 1.477 million cwt. in
volume for a product value of $67.7/cwt. With a profit margin of 20 percent, the profit
per cwt. would be 20%$67.7/cwt, or $20/1.477, = $13. Thus, 278,60613 = $3.16
million are lost in sales. The weighted profit loss over the five-year period would be:
2/5 (0) + ( 3/5) ((0 + 3.6))/2) = $1.08m/yr.
206
TABLE 1 Current-Year Comparison of Network Alternatives ($000s)
Improved Optimum Optimum Maximum
Bench- bench- number number Relaxed Relaxed oppor-
Cost type mark mark of whses of whses service (1) service (2) tunity
Production $30,762 $30,678 $30,673 $30,675 $30,678 $30,673 $30,386

Warehouse operations 1,578 1,468 1,608 1,572 1,296 1,420 1,529
Order processing 369 354 370 358 349 354 358
Inventory carrying 457 431 508 490 390 445 500
Transportation
Inbound 2,050 1,802 1,976 1,860 1,249 1,178 1,178
Outbound 6,896 6,991 6,310 6,365 7,238 6,698 6,458
Total costs $42,112 $41,725 $41,447 $41,321 $41,201 $41,043 $40,409
Customer
service:
300 mi. 93% 93% 98% 92% 75% 88% 81%
600 mi. 98% 98% 100% 100% 98% 100% 94%
No. of stocking
points 22 21 31 30 19 26 40
No. of plants 4 4 4 4 4 4 6
Savings vs.
benchmark $0 $387 $665 $791 $911 $1,069 $1,703
Savings vs. improved

benchmark $0 $0 $278 $404 $524 $ 682 $1,316
Comments: Service 600 mi con- 600 mi con- Unlimited

to match straint on straint on service,
bench- current opt no. of whses, and
mark warehouses warehouses plant cap.
207
TABLE 2 Future-Year Comparison of Alternatives ($000s)
Add plant Memphis Memphis
No plant Add plant @ Memphis and opt no. and opt no.
Cost type expansion @ Memphis & Chicago of whses of whses
Production $33,965 $39,517 $39,548 $39,524 $39,522

Warehouse operations 1,496 1,842 1,847 2,028 1,976
Order processing 393 462 454 470 460
Inventory carrying 431 505 497 591 573
Transportation
Inbound 1,647 2,350 2,000 2,614 2,426
Outbound 7,230 9,030 9,036 8,117 8,222
Total costs $45,164 $53,705 $53,382 $53,342 $53,179
Customer
service:
300 mi 98% 94% 95% 98% 92%
600 mi 99% 98% 98% 100% 100%
No. of stocking
points 20 21 20 31 30
No. of plants 4 5 6 5 5
Comments: Not all High Service

demand service held at
met level benchmark
208
Based on a simple rate of return on investment, capturing this profit potential would yield
1.08/4 = 27 percent annually on a $4,000,000 investment for expansion. The return
would increase to 90 percent per year with the full loss in the 5th year. The potential
seems great enough to justify one unit of expansion (1,000,000 cwt.). Two units of
expansion probably cannot be justified, since adequate capacity would be available from
the first capacity unit to meet demand requirements. The only benefit would be from the
network design improvement. The savings would be about $323,000 per year in the fifth
year (see Table 2) comparing one additional plant with two additional plants and keeping
the current number of warehouses. The simple return on investment using fifth-year
savings would only amount to about 8 percent (323,000100/4,000,000 = 8.1%).
The next question is: Where should the expansion take place at an existing plant
or at one of the two proposed locations? From a test of expanding any of the four
existing plants or the two proposed plant locations (runs 10 through 16 in Appendix A of
this note), it would appear that Memphis would be the lowest cost site in the 5th year
with Chicago next at only an additional cost of $76,000 per year (compare runs 14 and 15
in Appendix A). Adding a plant at a new location rather than expanding an existing plant
site saves a minimum of $281,000 annually (compare runs 11 and 14 in Appendix A of
this note), which results from placing plant capacity closer to warehouses.
Selecting Warehouses
A simple test on the number of warehouses in the network shows that transportation costs
are dropping more rapidly than inventory related costs are increasing (see Figure 1). This
means that 40 active warehouses will have the lowest total cost. However, some of these
warehouses will have low throughput. In order to maintain a minimum replenishment
frequency and shipment size, a minimum throughput needs to be met. Approximately a
truckload every two weeks, or 10,400 cwt. of throughput per year, is the minimum
activity needed to open a warehouse. Therefore, any warehouse showing less than this
throughput will be eliminated from consideration.
Under various assumptions about plants and their capacities, demand growth, and
service levels, 30 to 31 warehouses seem most economical with no deterioration on
service over the benchmark network. The following table shows selected results.
209
Percent
Type of Plant of demand Total No. of
run Year capacities 300 mi. cost whses
Benchmark Current Current 93 $42,112 22

Improved
benchmark Current Current 93 41,725 21
Improved Current
benchmark 5th yr. + Memphis 94 53,705 31
Current
yr. whses Current 92 41,321 30
5th yr. 5th Current
whses year + Memphis 92 53,179 30
Note that this conclusion about the number of warehouses depends on the previous
conclusion that a Memphis plant should be added by the fifth year. The number of
warehouses should be increased from the present 22 in both the current year and the fifth
year.
42 100
99
41.8 98
Service (right scale)
97
% of demand < 300 mi.

Total cost, $(000,000s)
41.6 96
95
41.4 94
Cost (left scale)
93
41.2 92
Practical design
91
41 90
22 26 30 31 36 40
Number of warehouses
FIGURE 1 Cost and Customer Service Profiles for Alternative Network Designs
More detailed economic analysis shows that if the plants are held at current
throughput levels, a savings realized from 30 warehouses would be $41,725,000
41,321,000 = $404,000 (see previous table). If current plant capacities are used and the
Memphis plant is on-stream in year five, the savings of the added warehouse would be:
$53,705,000 53,179,000 = $526,000
210
On the average, there can be savings of approximately ($404,000 + 526,000)/2 =
$465,000 per year by increasing the number of warehouses to 30 from the current 22.
Since these are public warehouses, little or no investment would be required to
implement the change.
Although the number of warehouses remains relatively unchanged from the current
year to the 5th year, there is some shifting among the particular warehouses in the mix.
The 30 warehouses in the current year should be numbers:
1,2,3,4,5,7,8,11,13,14,15,16,17,18,19,20,21,25,28,31,32,33,34,35,36,37,38,40,44,45
providing that the loading on the current plants is allowed up to the limits of their current
capacity. When the Memphis plant is brought on-stream by the end of the second year,
the warehouse mix should begin to evolve to numbers:
1,2,3,4,5,7,8,11,13,14,15,17,18,19,20,21,25,28,29,31,32,34,35,36,37,38,40,44,45,47
As the Memphis plant is bought on stream, the Memphis public warehouse is closed
and the volume is shifted to the Memphis plant as a warehouse. In addition, the
Richmond, VA warehouse is closed and the Las Vegas, NV warehouse is opened. The
number of warehouses remains at 30.
Both in the base year and in the future year, the throughputs in the plants serving as
warehouses are within acceptable limits as the following summary shows.
Plant Current- Future-

as a Thruput year year
warehouse limits solution solution
Covington 450,000 cwt. 254,471 cwt. 306,478 cwt.

New York 380,000 302,043 380,523
Arlington 140,000 66,592 66,161
Long Beach 180,000 95,943 117,288
Customer Service
Currently, a high proportion of demand (93 percent) is located within 300 miles of a
stocking point. Since the service distance may be up to 600 miles and still meet the
company's service policy, should the service level be reduced somewhat to effect a cost
saving? For example, using the improved benchmark as the base case (run 2), 93 percent
of the demand is within 300 miles and 97.5 percent is within 600 miles. If a 600-mile
constraint is applied to the current network configuration (run 23), 75 percent of the
demand is within 300 miles and 98 percent is still within 600 miles. The total costs are
reduced from $41,725,000 to $41,201,000, or a savings of $524,000 per year. In
addition, if the number of warehouses in the network is optimized, the costs can be
reduced by another $158,000 per year (run 23 vs. run 22). However, $278,000 of the
total $524,000 + 158,000 = $628,000 can be realized without a service change. This
leaves approximately $404,000 that can be saved by a relaxed service restriction.
The question now becomes one of whether the higher costs associated with the more
restrictive service level are justified. Since there is no sales-service relationship for this
problem, we can only estimate the worth of the service. That is, can enough sales be
211
generated to cover the higher service level? If physical distribution costs for the com-
pany are 15 percent of sales, which is probably a conservative estimate, then 1/0.15 =
$6.70 in sales must be generated for each dollar that is added to distribution costs.
Therefore, to cover $404,000 in cost would require
$404,000 $6.70
= 38,124 cwt.
$0.71 / lb.100lb./cwt.
increase in sales. In terms of overall demand, this would be 38,124100/1,477,026 = 2.5

percent.
But not all customers would experience a higher service level. Comparing the
demand centers for 299,818 cwt. of demand shows a reduction in warehouse to customer
miles. Thus, moving from a minimum cost network to one with a high service level,
where the percent of demand less than 300 miles increases from 75 percent to 93 percent,
requires that the 38,124 cwt. increase in demand occur in the 299,818 cwt. of demand
affected by the change. This would be a 13 percent increase.
The products are not highly differentiated from others in the marketplace so that
service plays an important role in selling these products. Whether a 93 75 = 18
percentage points increase in service can result in a 2.5 percent increase in overall sales
cannot be judged by the distribution department alone. The sales department must play
an important part in indicating whether the additional sales are possible. If they are not
likely to be realized, there is no incentive for a network other than the minimum cost one.
If this information is not available from sales, the conclusion is likely to be to
maintain the status quo as represented by the benchmark. That is, one-day service is
most likely to guide the design.
Overall Analysis and Summary

The recommended design would involve an immediate increase in the number of
warehouses from 22 to 30. In addition, there should be an immediate reallocation of
demand among the existing plants. No reduction in the customer service level seems
justified at this time. Therefore, a total cost reduction of $42,112,000 41,321,000 =
$791,000 per year seems immediately achievable (run 1 vs. run 18). By the end of the
2nd year, the Memphis plant should be brought on stream and the network should begin
to evolve from the current design (run 24) to that for the fifth year (run 25). The addition
of a plant is justified from the high rate of return realized from the profit potential of
being able to continue meeting the growth in demand.
For the current year, a breakdown of the service and the cost changes show the
following:
212
Current-
Bench- year Change from
Cost type mark design benchmark
Production $30,762 $30,675 $ -87 -0.3%

Whse operations 1,578 1,572 - 6 -0.4
Order processing 369 358 -11 -3.0
Inventory carrying 457 490 +33 +7.2
Transportation
Inbound 2,050 1,860 -190 -9.3
Outbound 6,896 6,365 -531 -7.7
Total costs ($000s) $42,112 $41,320 $-792 -1.9%
By the fifth year, total distribution costs should be $53,179,000, or

$53,179,000/1,908,606 = $27.86, compared with the current-year cost of
42,112,463/1,477,026 = $28.51 per cwt. If current year costs are projected to the fifth
year demand level, the 5th-year production/distribution costs might be 28.511,908,606 =
$54,414,357, or a savings of $54,414,357 53,179,000 = $1,235,357 per year. Of
course, these savings can only be realized through the addition of capacity at Memphis
for $4,000,000. If this capacity is useful for at least 15 years, the amortization of
$4,000,000/15 = $267,000 per year would yield a net savings of $532,000 per year.
Overall, the design change appears to be justified.
213
APPENDIX A Listing of Selected Computer Runs
Service Percent of
Run.Run No of Plant Demand con- No of Total demand within
no. description plants capacity level straint whses costs 300 600 Comments
1 Benchmark 4 - Current - mi 22 $42,112 93% 98% Current network design

2 Improved benchmark 4 Current Current 300 21 41,725 93 98 No investment required
3 No serv constraint 4 Current Current 9000 18 40,896 71 89
4 Max opportunity 6 Current Current 9000 40 40,409 81 94 Added plants at 1m cwt
5 Future yr-imp bmk 4 Crnt+1m 5th yr. 300 21 53,777 93 98 Plant cap + 1m cwt
6 Test 27 whses 4 Current Current 300 26 41,744 95 100
10 Exp Covington 4 Current 5th yr. 300 21 54,145 94 98 Covington cap + 1m cwt
11 Exp New York 4 Current 5th yr. 300 21 53,986 93 98 New York + 1m cwt
12 Exp Arlington 4 Current 5th yr. 300 21 54,709 94 98 Arlington cap + 1m cwt
13 Exp Long Beach 4 Current 5th yr. 300 21 55,251 94 98 Long Beach cap + 1m cwt
14 Add Memphis 5 Current 5th yr. 300 21 53,705 94 98 Add Memphis at 1m cwt
15 Add Chicago 5 Current 5th yr. 300 20 53,781 94 98 Add Chicago at 1m cwt
16 Add Mem & Chi 6 Current 5th yr. 300 20 53,382 95 98 Add Chi & Mem at 1m cwt
ea
17 No plant expansion 4 Current 5th yr. 300 20 45,164 98 100 Only 85.4% of demnd
served
18 Optimum whses 4 Current Current 300 31 41,447 98 100 Plants at current
capacity
19 Optimum whses 4 See cmt Current 300 30 41,563 97 100 Plants at current thruput
20 Optimum whses 5 Current 5th yr. 300 31 53,342 98 100 Memphis at 1m cwt
21 Test cust service 4 Current Current 600 31 40,996 80 100 Whses at opt no = 31
22 Test cust service 4 Current Current 600 26 41,043 88 100 Whses from opt no = 31
23 Test cust service 4 Current Current 600 19 41,201 75 98 Whses from current 22
24 Optimum whses 4 Current Current 375 30 41,321 92 100 Service level at bmk
25 Optimum whses 5 See cmt 5th yr. 375 30 53,179 92 100 Serv at bmk/Mem @ 1m cwt
214
ESSEN USA
Teaching Note
Strategy
Essen USA is concerned with entire supply channel performance. The supply channel
consists of four echelons ranging from factory to customers. The purpose of this case
study is for the student to manipulate the supply channel variables using a channel
simulator in order to improve individual member and system-wide performance. The
channel variables include forecasting methods, inventory policies, transportation services,
production lot sizes, order processing costs, and stock availability levels. Students should
seek to optimize channel performance, although it is not expected that the optimum
actually can be found or verified. However, improving performance over existing levels
is achievable.
The SCSIM module of LOGWARE is used to simulate the demand and product flows
throughout the multi-echelon supply chain. SCSIM is an ordinary Monte Carlo day-to-
day type of simulator. Using a simulator for performance improvement requires thinking
of it in terms of as an experimental methodology. That is, a single run of the simulator is
a particular event sequence generated from random numbers. Changing the seed number
in the simulator causes a different set of random numbers to be generated and possibly
another outcome from the same input data. A simulation run with a specified seed
number should be viewed as a single statistical observation and multiple outcomes from
various seed numbers should be treated as a statistical sample and analyzed accordingly,
i.e., comparing means and standard deviations.
Each simulation is run for a period of 11 years with results taken from years 2
through 11. The first year is not used since it can show unstable results due to startup
conditions. The results appear to reach steady state by the second year, and the results for
the 10 years thereafter are averaged to give a reasonable representation of channel
performance for a given run. The database used to represent the current performance of
the channel, as derived from the case study, is summarized in the Appendix A of this note
and a typical run report is shown in Appendix B.
This case provides students with the opportunity to observe the operation of a multi-
echelon supply channel and to assess the impact of changing key operating variables on
individual members as well as on channel-wide performance. The effect on cost and
customer service as well as sales, inventory, and back order levels of demand patterns,
demand forecasting methods, inventory control methods, transportation performance,
production lot sizing, order processing procedures, and item fill rates can be observed in
both graphical and report forms. Most importantly, students can see the effects of supply
chain decisions rather than project the results statistically.
Questions
1. What can you say about the logistics performance throughout the supply channel for
Essen and its customers?
215
General observations
It is recognized that Essen must deal with demand that has significant seasonal peaks
at gift-giving times of the year as shown in Figure 1. Compared with a smooth demand
pattern, this can cause increasing demand variability upstream from the customers, as
illustrated in Figure 2. This bull whip effect is partly a result of the demand for an
upstream member being derived from the order size and pattern of its immediate
downstream channel member. Forecast accuracy, lead-time uncertainty, and inventory
control method also affect demand variability and the resulting cost of that variability.
Figure 1 Typical Demand Pattern for Essen Over the Period of One Year
Essen
Ware-
Retailer warehouse
house Factory
Distri Factory
warehouse
-butor
Retailer
Retailer
Figure 2 Increasing Demand Variability of Upstream Channel Members for a

Four-Year Period
216
Benchmark
Running the simulator (SCSIM) with a seed number of 123456 and simulated period
of 11 years with results taken from the last 10 years, the channel generates average
annual sales of $109.5 million for a net average annual system profit contribution of
$24.4 million, as shown in Table 1. The question arises as to whether channel
performance can be improved and profits increased. At least two observations can be
made that suggest there is room for improvement. First, the inventory levels for both the
retailers warehouse and Essens warehouse are quite high compared with the Retailer
(see Figure 3). It is possible that Retailer inventories are too low. However, the
inventory turnover ratio is about seven for the Essens warehouse (see Table 1). This is
not particularly high for a food product that might have a turnover at least in the range of
10 to 12. The turnover for the retailers warehouse appears more in line with industry
norms of about 13 (see Table 1).
Retailer
warehouse Essen
warehouse
Retailer
Figure 3 Inventory Levels for Four Years Using Benchmark Data
Second, the backorders at the Retailer level do not seem to recover well from the
seasonal spike in demand. Correspondingly, the Retailer inventory turnover is 81 (see
Table 1), which is quite high. The low percentage of demand filled on request (<50%)
suggests that inadequate inventory is being maintained to meet reasonable fill rates.
Third, customer service levels are also low for the retailers warehouse and Essens
warehouse. Backorder occurrences are high for both channel members. Although
inventory levels are adequate most of the time, seasonal demand rippling through the
supply chain causes a significant number of back orders before inventory can be
replenished.
The observation is that there is an opportunity to improve channel performance,
especially in terms of customer service. A major concern is how to mange the seasonal
demand pattern that is causing the cyclical behavior throughout the echelons of the
217
channel. Current performance of the channel members is summarized in Table 1 for four
simulation runs using different seed numbers.
Table 1 Average Annual Performance of Channel Members and the System at

Benchmark
Channel member Run 1 Run 2 Run 3 Run 4 Average
Essens factory
Total cost $73,105,904 $72,967,088 $72,616,192 $72,477,376 $72,791,640
Units produced 37,918 37,846 37,664 37,592 37,755
Cost per unit $1,928 $1,928 $1,928 $1,928 $1,928
Essens
warehouse
TO ratio 6.52 6.59 6.61 6.58 6.58
Fill rate <50% <50% <50% <50% <50%
Cost $5,578,291 $5,549,202 $5,521,236 $5,540,447 $5,547,294
Cost per unit $147.02 $146.50 $146.32 $146.65 $146.62
Retailers
warehouse
TO ratio 12.95 13.01 12.96 12.89 12.95
Fill rate <50% <50% <50% <50% <50%
Cost $3,873,236 $3,895,406 $3,853,890 $3,912,699 $3,883,808
Cost per unit $101.94 $102.15 $102.14 $103.07 $102.33
Retailer
TO ratio 81.02 80.38 80.45 80.60 80.61
Fill rate <50% 53.02% 54.09% <50% <50%
Units sold 38,017 37,983 37,774 37,804 37,895
Cost $2,884,527 $2,768,697 $2,939,464 $2,981,607 $2,893,574
Cost per unit $76.19 $72.89 $77.82 $78.87 $76.44
System
Profit $24,426,593 $24,591,053 $24,235,498 $24,340,563 $24,398,426
Profit as % of sales 22.23% 22.40% 22.20% 22.28% 22.28%
Seed Number 123456 444444 555555 666666
2. What steps would you suggest taking to improve logistics performance throughout the
channel? Do any of the changes involve Essen? If so, does the company directly
realize any cost and/or operating performance improvements?
A number of actions can be taken to lower costs and improve customer service.
Improving the forecast, shortening the lead times, changing the inventory control policy,
and changing production lot sizes are all variables that can be altered for possible
performance improvement. The interactions among these variables and the large number
of variable combinations preclude finding the optimal set. However, they can be
explored in a systematic way to find improvement. The primary focus of this analysis
will be to increase the fill rates at the risk of increasing costs. Ultimately, revenues,
through improved customer service, may be preserved or increased to more than
compensate for reduced profits.
218
Retailer Level
Start with the retailer because of the proximity to the customer. Fill rates need to be
improved, probably in the 95-99% range as specified in the database. Inventory turns can
be guided by the industry average of 12 turns per year. Where the two cannot be jointly
met, service will prevail.
Clearly, putting additional inventory at the retail point will improve customer service.
Using the companys current inventory policy of stocking to demand, the target level can
be raised without changing the review time. Exploring different target levels shows 14
days to offer about 35 turns and a 99+% fill rate. Because of the high cost of a back
order, total costs at the retail level drop significantly.
Altering the forecasting method and the settings associated with the method yield
little opportunity for improvement. Using an exponential smoothing model with a high
smoothing constant to better follow the seasonal changes in demand results in increased
costs. Lowering the smoothing constant to 0.1 did not offer improvement either.
Altering the number of periods in the moving average model did not improve costs and
only degraded performance. Shortening the review time in the stock-to-demand reorder
policy did have a positive effect on fill rate, but resulted in high costs and lower
inventory turns. The tradeoff did not seem beneficial, given the fill rate and turnover
targets.
Retail Warehouse Level

Determining an improved policy at this level is difficult because a 95% fill rate and
10 to 12 inventory turns is an illusive goal. Using service as the primary target, a stock-
to-demand control policy is used with a review period of 7 days and a target of 25 days of
inventory. The forecasting method is moving average with a period of 7 days. The
performance achieved at this channel level is about nine inventory turns per year and a
97% fill rate.
Essens Warehouse Level

The performance at Essens warehouse level seems to mimic that at the retail
warehouse level except that there is more demand variability. Again, an inventory
turnover ratio in the target range cannot be achieved while maintaining a high fill rate
level. Trying to achieve high service levels with high levels of inventory is difficult,
probably due to the extensive demand variability that filters back to this member of the
channel. Multiple simulation runs show that a high fill rate cannot consistently be
achieved even when on the average inventory levels are high. However, average
performance shows an 82% fill rate and 1.5 inventory turns per year based on a 7-day
moving average forecasting model and a stock-to-demand inventory control policy with a
review time of 7 days and an inventory target of 25 days.
Essens Factory
The concern with the factory level in the channel is whether product should be
manufactured in a larger lot size, but with slightly higher production time variability.
The reduced costs seem to out weigh the negative effects of increased variability.
Producing in the larger lot size is favored.
219
Overall
Using the objective of improving customer service, it is not surprising that supply
channel costs increase as shown from the reduced profit in Table 2 compared with Table
2. The average fill rate has increased for all members of the channel, but the cost effects
are spread disproportionately among the members. Even with a higher fill rate, the
retailer benefits from a substantial reduction in the cost per unit sold. On the other hand,
the cost for handling a unit of the product at Essens warehouse is substantially increased.
Essen should take advantage of the cost reduction from producing in the larger batch size,
but this does not offset the higher cost at the companys warehouse. As an upstream
channel member, Essen undoubtedly suffers from variability in demand, which it cannot
entirely control.
The retailer benefits from the action to increase fill rates across the channel.
However, Essen is put at a disadvantage and may take a counter action to improve its cost
position. Essen may simply lower its inventory level by reducing the reorder target
quantity from 25 to 10 days. This reduces Essens per-unit warehouse cost, but it also
increases the costs for the retailer. The reduced inventory level at the Essen warehouse
causes lower fill rates for the downstream retailer. Unless the retailer can find an
incentive to reward Essen for its good service, it will be difficult for Essen to provide the
level of service that the retailer would like and that is economically beneficial to Essen.
Table 2 Average Annual Performance of Channel Members and the System as

Revised
Channel member Run 1 Run 2 Run 3 Run 4 Average
Essens factory
Total cost $68,638,738 $74,761,258 $78,418,188 $75,400,666 $74,304,713
Units produced 35,950 39,286 41,268 39,622 39,032
Cost per unit $1,909 $1,903 $1,900 $1,903 $1,904
Essens
warehouse
TO ratio 1.47 1.42 1.51 1.46 1.47
Fill rate 90.65% 100% 63.16% 72.92% 81.68%
Cost $12,060,444 $12,449,170 $11,895,743 $12,178,581 $12,145,984
Cost per unit $317.51 $326.28 $311.87 $319.61 $318.82
Retailers
warehouse
TO ratio 9.24 9.02 9.14 9.28 9.17
Fill rate 96.64% 96.29% 97.60% 98.09% 97.16%
Cost $4,018,143 $4,080,160 $4,043,973 $3,992,791 $4,033,767
Cost per unit $105.70 $107.09 $106.55 $105.51 $106.21
Retailer
TO ratio 35.44 35.36 34.99 35.14 35.23
Fill rate 99.52% 99.53% 99.70% 99.40% 99.54%
Units sold 38,017 37,936 37,979 37,730 37,916
Cost $727,378 $717,363 $709,772 $748,198 $725,677
Cost per unit $19.13 $18.91 $18.69 $19.75 $19.12
System
Profit $24,423,849 $17,627,666 $14,690,765 $17,184,464 $18,481,686
Profit as % of sales 22.23% 16.08% 13.38% 15.69% 16.85%
Seed Number 123456 111111 222222 333333
220
3. Would shipping by airfreight from Germany be a benefit to channel performance? To
Essen?
No. Selling candies to end customers at $2,890 per thousand lb. using airfreight shipping
results in obvious loss to Essen. The cost of shipping by air is $1,833 per thousand lb.,
plus $1,000 material cost and $850 production cost results in a total cost much higher
than selling price. There is no point to using airfreight. Running the simulation with the
higher freight rate but lower variability confirms that channel profits would be negative.
4. Is there a benefit to producing in the larger 20,000-pound batch size?
Yes. This was tested in question 2. From Tables 1 and 2, it can be seen that production
costs drop from $1,928 per unit to $1,904 per unit. The overall channel cost reduction
overshadows the negative effects of greater length and variability in production time.
Appendix A Simulation Database for Essen USA Under Current Conditions
Title: ESSEN USA
Initialization
123456 Seed value
11 Length of simulation, years
2890 Annual price, $/unit
Customer demand pattern

Generate daily demand
100 Average daily demand, units
15 Standard deviation of daily demand, units
1 Annual demand growth increment, %
Monthly seasonal indices
Month Index Month Index Month Index Month Index
1 0.25 4 0.75 7 0.75 10 0.75
2 1.25 5 0.75 8 0.75 11 1.50
3 1.25 6 0.75 9 0.75 12 2.50
Retailer/Level 1
Product item data
2220 Item value in inventory, $/unit
1 Customer order filling cost, $/unit
35 Purchase order processing cost, $/order
25 Inventory carrying cost, %/year
1 Average customer order fill time, days
0 Customer order fill time standard deviation, days
221
98 In-stock probability, %
670 Back order cost, $/unit
Forecasting method
Moving average
7 Number of periods
Reorder policy
Stock-to-demand control method
10 Target days of inventory
7 Review time in days
Distributor/Level 2
Product item data
20 Retailer order filling cost, $/unit
2 Average retailer order fill time, days
0.2 Retailer order fill time standard deviation, days
Forecasting method
Moving average
30 Number of periods
Reorder policy
Warehouse/Level 3
Product item data
15 Distributor order filling cost, $/unit
3 Average distributor order filling time, days
0.3 Distributor order fill time, days
Forecasting method
Moving average
360 Number of periods
Reorder policy
222
Factory/Source
Product item data
850 Production cost, $/unit
10 Minimum production lot size, units
10 Warehouse order filling cost, $/unit
8 Average production time, days
2 Production time standard deviation, days
1000 Purchase cost, $/unit
Transportation
Transport between Distributor and Retailer
25 Transport cost, $/unit
1 Average time in-transit, days
0 Transit time standard deviation, days
Transport between Warehouse and Distributor
Transport between Factory and Warehouse
Appendix B Benchmark Simulation Results with Seed Number 123456 and

Simulation Length of 11 Years
SUPPLY CHANNEL REPORT FOR SIMULATED YEARS 2 TO 11

Yearly Simulated
average period FINANCIAL PERFORMANCE
$109,868,552 $1,098,685,520 Revenue
37,918,000 379,180,000 Cost of purchased goods
71,950,552 719,505,520 Gross margin
32,230,300 322,303,000 Production cost
Transportation costs:
949,025 9,490,250 Distributor to retailer
2,656,010 26,560,100 Warehouse to distributor
2,957,604 29,576,040 Factory to warehouse
Sales order handling cost for:

38,017 380,168 Customer orders
759,890 7,598,900 Retailer orders
569,145 5,691,450 Distributor orders
Order processing cost for:

1,638 16,380 Orders to distributor
683 6,825 Orders to warehouses
675 6,750 Orders to factory
Inventory costs
223
260,414 2,604,145 Retailer
1,627,909 16,279,092 Distributor
1,988,984 19,889,837 Warehouse
Back order costs

2,584,458 25,844,580 Retailer
535,730 5,357,300 Distributor
363,478 3,634,775 Warehouse
$24,426,593 $244,265,928 Net profit contribution
Appendix C Simulation Database for Essen USA as Revised for Service

Improvement
Title: ESSEN USA
Initialization
123456 Seed value
11 Length of simulation, years
2890 Annual price, $/unit
Customer demand pattern
Generate daily demand
100 Average daily demand, units
15 Standard deviation of daily demand, units
1 Annual demand growth increment, %
Monthly seasonal indices
Month Index Month Index Month Index Month Index
1 0.25 4 0.75 7 0.75 10 0.75
2 1.25 5 0.75 8 0.75 11 1.50
3 1.25 6 0.75 9 0.75 12 2.50
Retailer/Level 1
Product item data
1 Customer order filling cost, $/unit
1 Average customer order fill time, days
0 Customer order fill time standard deviation, days
Forecasting method
Moving average
7 Number of periods
Reorder policy
224
Distributor/Level 2
Product item data
20 Retailer order filling cost, $/unit
2 Average retailer order fill time, days
0.2 Retailer order fill time standard deviation, days
Forecasting method
Moving average
7 Number of periods
Reorder policy
Warehouse/Level 3
Product item data
15 Distributor order filling cost, $/unit
3 Average distributor order filling time, days
0.3 Distributor order fill time standard deviation, days
Forecasting method
Moving average
7 Number of periods
Reorder policy
Factory/Source
Product item data
825 Production cost, $/unit
20 Minimum production lot size, units
10 Warehouse order filling cost, $/unit
10 Average production time, days
2.1 Production time standard deviation, days
1000 Purchase cost, $/unit
225
Transportation
Transport between Distributor and Retailer
Transport between Warehouse and Distributor
Transport between Factory and Warehouse
226

Solucionario Ballou

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Solucionario Ballou

Transféré par

Droits d'auteur :

Formats disponibles

CHAPTER 1

Producing in Taiwan would appear to be the least expensive.

To Catalog UPS Transport Total

Inventory Investment Table

Therefore, the sales estimating equation is:

and 3,000,0000.5765 = 1,729,412.

The A+B item proportion will be:

Solving algebraically for X, we have:

Quantity Price per unit Discount

Percent of orders delivered

This change is also significant at the 0.01 level.

(b) The use of the before-after-with-control-group experimental design is a methodology

The sales distributions may not be normal.

The change in gross profit:

P = Trading margin Sales response rate Annual sales

The change in cost:

C = Annual carrying cost Standard product cost z

Now, set P = C and solve for z.

Next, find k in the loss function:

L = k ( y m) 2 out-of-stock % at point where sales are lost

R = transportation rate, $/unit

Rail has the lowest total cost.

TABLE 7-1 An Evaluation of the Transport Alternatives for the Wagner

Q * = 2 DS / IC = 2(9,600)(100) / 0.3( 62) = 3213

r = zs LT ( d ) = 1.28 1.5 (9,600 / 365) = 50.5 cwt.

Q * = 2(9,600)(100) / 0.3( 6180

r = 1.28 1.8 (9,600 / 365) = 60.6 cwt.

Service B appears to be slightly less expensive.

Origin Cleve- South San

Supply 400 150 150

0.5 1.0 1.5 2.0

Problem label School Bus Routing Exercise

STOP LOAD VOL. LOAD

--CAPACITY-- --VEHICLE-- --DRIVER--

SPECIFIED STOP-TO-STOP DISTANCES

STOP STOP STOP STOP DISTANCE

Route Stop sequence Distancea

The total cost of this strategy would be:

Motel 3 nights @ 49.00 $147.00

Route Stop sequence Distance

The total cost of this strategy is:

Motel M1 1st night $ 40.00

Strategy 2 appears to be most economical.

with a tour time of 130 minutes.

Problem label - Sima Donuts

Making a run with ROUTER will give the route design.

FIGURE 7-4 Graphical Display of Route Design for Sima Donuts

The problem can be expressed as a transportation problem of linear programming. There

Loading points and dates

(i) Orders shipped as received

Average period cost is $2,274.00

Average cost per period is $4,155.60/2 = $2,077.80.

(iii) Hold all orders until the third period.

Average period cost is $6,835.80/3 = $2,278.60

Ship immediately $ 2,274.00

(1) (2) (3)=(1)(2)

Day 1 Day 2 Day 3 Day 4 Day 5

Route Run Stop Brk Stem

FIGURE 7-6 Input Data for ROUTER for Medic Drugs

Problem label - Medic Drugs

FIGURE 7-7 Graphical

[2,056 + 2,349 + 1,895 + 1,514]/4 = 1,953.50

The standard error of the forecast is:

Total squared error 1,677,713.76