Beruflich Dokumente
Kultur Dokumente
Supply Chain
Management
Planning, Organizing, and Controlling the Supply Chain
Fifth Edition
Instructors Manual
Ronald H. Ballou
Weatherhead School of Management
Case Western Reserve University
CONTENTS
Preface
Chapter 1
2
3
4
5
6
7
10
11
12
13
14
15
16
iii
1
2
4
9
13
14
17
35
41
48
52
65
84
88
94
121
124
131
134
144
147
148
162
186
190
198
204
208
217
229
230
ii
PREFACE
This instructor's guide provides answers to the more quantitatively oriented problems at
the end of the textbook chapters. If the questions or problems are for discussion or they
involve a substantial amount of individual judgment, they have not been included.
Solutions to the cases and exercises in the text are also included. These generally
require computer assistance for solution.
With the text, you are provided with a collection of software programs, called
LOGWARE, that assist in the solution of the problems, cases, and exercises in the text.
The LOGWARE software along with a users manual is available for downloading from
the Prentice Hall website or this book. The users manual is in Microsoft Word or
Acrobat .pdf formats. This software, along with the users manual, may be freely
reproduced and distributed to your classes without requiring permission from the
copyright holder. This permission is granted as long as the use of the software is for
educational purposes. If you encounter difficulty with the software, direct questions to
Professor Ronald H. Ballou
Weatherhead School of Management
Case Western Reserve University
Cleveland, Ohio 44106
Tel: (216) 368-3808
Fax: (216) 368-6250
E-mail: Ronald.Ballou@CASE.edu
Web site: www.prenhall.com/ballou
iii
CHAPTER 1
BUSINESS LOGISTICS/SUPPLY CHAINA VITAL SUBJECT
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?
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, inventory, 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.
TABLE 2-1
Minimum values.
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.
b
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.
Product
code
08776
12121
10732
11693
10614
12077
07071
10542
06692
09721
14217
11007
Total
Dollar
sales
$71,000
63,000
56,000
51,000
46,000
27,000
22,000
18,000
14,000
10,000
9,000
4,000
$391,000
Cumulative
sales
$ 71,000
134,000
190,000
241,000
287,000
314,000
336,000
336,000
354,000
368,000
378,000
391,000
Cumulative
sales as
% of total
18.2
34.3
48.6
61.6
73.4
80.3
85.9
90.5
94.1
96.7
98.9
100.0
Cumulative
items as
% of total
8.3
16.7
25.0
33.3
41.7
50.0
58.3
66.7
75.0
83.3
91.7
100.0
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 different 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
zip code
11107
42117
74001
59615
a
Catalog
price
$99.95
99.95
99.95
99.95
UPS
zone
2
5
6
8
Transport
costa
$ 7.37
10.46
13.17
18.29
Total
cost
$107.32
110.41
113.12
118.24
Use 27 lb.
(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)
0125
.
YX
0.75 0.25
(1 A) (1 0125
. )X
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:
(1 0125
. )( 0.05)
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
Product
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Cumulative
item proportion, X
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
Cumulative
sales, Y
$ 835,714
1,300,000
1,595,454
1,800,000
1,950,000
2,064,705
2,155,263
2,228,571
2,289,130
2,340,000
2,383,333
2,420,689
2,453,226
2,481,818
2,507,142
2,529,719
2,550,000
2,568,293
2,584,884
2,600,000
Projected
item sales
$ 835,714
464,286
295,454
204,546
150,000
114,706
90,558
73,308
60,559
50,870
43,333
37,356
32,537
28,592
25,324
22,587
20,271
18,293
16,591
15,116
Turnover
ratio
8
8
8
8
6
6
6
6
6
6
4
4
4
4
4
4
4
4
4
4
Total
Average
inventory
value
$104,464
58,036
36,932
25,568
25,000
19,118
15,093
12,218
10,093
8,478
10,833
9,339
8,134
7,148
6,331
5,647
5,068
4,473
4,148
3,779
$380,000
12
This problem involves the application of Equations 3-1 and 3-2. We can develop an 8020 curve based on 30 percent of the items accounting for 70 percent of sales. That is,
X (1 Y ) 0.30(1 0.70)
0.225
0.70 0.30
YX
(1 0.225) X
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:
YA
(1 0.225)( 0.20)
0.5765
0.225 0.20
(1 0.225)( 0.50)
0.8448
0.225 0.50
and 3,000,0000.8448 = 2,534,400. The product group B sales will 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:
B:
1,729,412/9
804,988/5
Total inventory
= 192,157
= 160,988
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)
0156
.
0.65 0.20
YX
and
0.75
(1 0156
. )X
0156
.
X
AxY
0156
. x 0.75
0.288
1 A Y 1 0156
.
0.75
That is, about 29% of the items (0.2885,000 = 1,440 items) produce 75% 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
1 to 1,000 units
1,001 to 2,000 units
>2,000 units
a
Discount
0
1.7%a
3.5%
[(58 - 57)/58][100]=1.7%
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.
X 2 X1
s22
s2
1
N 2 N1
224 185
612 79 2
102 102
39
.
394
36.48 6118
.
2,295 1,342
576 2 3352
56
56
953
10.7
5,924 2,004
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% is 1.48, and for 92%, 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%. Of course, the change in z
is found by taking the difference in z values for 1% 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
Cost of item
Other costs
Profit per item
$5.95
-4.25
-0.30
$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
Finally, the point where the marginal supply cost equals the marginal sales loss is
( y 5)
B
0.10
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%,
and should not allow the out-of-stock level to fall below 1.67 + 5 = 6.67%.
12
CHAPTER 5
ORDER PROCESSING AND
INFORMATION SYSTEMS
All questions in this chapter require individual judgment and response. No answers are
offered.
13
CHAPTER 6
TRANSPORT FUNDAMENTALS
14
The maximum that the power company can pay for coal at its power plant location in
Missouri is dictated by competition. Therefore, the landed cost at the power plant of coal
production costs plus transportation costs cannot exceed $20 per ton. Since western coal
costs $17 per ton at the mine, the maximum worth of transportation is $20 $17 = $3 per
ton. However, if the grade of coal is equal to the coal from the western mines, eastern
coal can be landed in Missouri for $18 per ton. In light of this competitive source,
transportation from the western mines is worth only $18 $17 = $1 per ton.
15
Prior to transport deregulation, it was illegal for a carrier to charge shippers less for the
longer haul than for the shorter haul under similar conditions when the shorter haul was
contained within the longer one. To be fair, the practice probably should be continued.
If competitive conditions do not permit an increase in the rate to Z, then all rates that
exceed $1 per cwt. on a line between X and Z should not exceed $1 per cwt. Therefore,
the rate to Z is blanketed back to Y so that the rate to Y is $1 per cwt. By blanketing the
rate to Z on intervening points, no intervening point is discriminated against in terms of
rates.
16
(a) From text Table 6-4, the item number for place mats is 4745-00. For 2,500 lb., the
classification is 100 since 2,500 lb. is less than the minimum weight of 20,000 lb. for
a truckload shipment. From text Table 6-5, the rate for a shipment 2,000 lb. is
8727/cwt. The shipping charges are $87.27 25 cwt. = $2,181.75.
(b) This is an LTL shipment with a classification of 100, item number 4980-00 in text
Table 6-4. From Table 6-5, the minimum charge is 9351 and the rate for a <500 lb.
shipment is 5401/cwt. Check the charges using the <500 lb. rate and compare it to
the minimum charge. That is,
$54.01 1.5 cwt. = $81.02
Since this is less than the minimum charge of $93.51, pay the minimum charge.
(c) From Table 6-4, the item number is 2055-00 with a classification of 55 for LTL and
37.5 for TL at a minimum weight of 36,000 lb. There are three possibilities that need
to be examined:
(1) Ship LTL at class 55 and 27,000 lb. shipment.
(2) Ship at class 55 and 30,000 lb. rate.
(3) Ship at class 37.5 and 36,000 lb. rate.
14
Lowest cost
3.87 30,000
20,549 lb.
5.65
Since current shipping weight of 24,000 lb. exceeds the break weight, ship as if 30,000
lb. Hence, 3.87 300 = $1,161.00. Now, discount the charges by 40 percent. That is,
$1,161 (1 0.40) = $696.60
21
The question involves evaluating two alternatives. The first is to compute the transport
charges as if there are three separate shipments. The next is to see if a stop-off privilege
offers any cost reduction. The comparison is shown below.
Separate shipments
Loading/unloading
22,000
3,000
15,000
Route
A to D
A to C
B to C
Rate, Stop-off
$/cwt. charge
Charges
$3.20
--$704.00
2.50
--75.00
1.50
--225.00
Total charges $1,004.00
With stop-off
Ship direct to B and split deliver thereafter.
Rate, Stop-off
Loading/unloading Route $/cwt. charge
Charges
25,000
A to B $1.20
$ 300.00
40,000
B to D 2.20
880.00
Stop-off @ C
$25.00
25.00
Stop-off @ D
25.00
25.00
Total charges $1,230.00
Direct shipment
15
16
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:
Cost type
Transport
Method
RD
Rail
2550,000
= $1,250,000
In-transit
0.2547550,000
inventorya
ICDt/365
(16/365)
= $260,274
Wagers
0.25475(10,000/2)
inventorya
ICQ/2
= $593,750
Electronics
0.25500(10,00/2)
inventory
ICQ/2
= $625,000
Total
$2,729,024
a
C refers to price less transport cost per unit.
Piggyback
4450,000
= $2,200,000
0.2545650,000
(10/365)
= $156,164
0.25456(7,000/2)
= $399,000
0.25500(7,000/2)
= $437,500
$3,192,664
Truck
8850,000
= $4,400,000
0.2541250,000
(4/365)
= $56,438
0.25412(5,000/2)
= $257,500
0.25500(5,000/2)
= $312,500
$5,026,438
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.
Service type
17
Cost type
Method
Transport
RD
In-transit
inventory
Plant
inventory
ICDt/365
Warehouse
inventory
Total
ICQ*/2
ICQ*/2
+ ICr
A
129,600
= $115,200
0.20509,600
(4/365)
= $1,052
0.3050(321.8/2)
= $2,684
0.3062(321.3/2)
+ 0.306250.5
= $3,927
$122,863
B
11.809,600
=$114,048
0.20509,600
(5/365)
= $1,315
0.3050(357.8/2)
= $2,684
0.3061.80(321.8/2)
+ 0.3061.8060.6
= $4,107
$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
18
Solved nodes
directly
Its closest
connected to
connected
unsolved
unsolved
Step
nodes
node
1
A
B
A
D
2
A
D
B
C
3
B
C
D
C
D
F
4
C
E
C
F
D
F
5
C
F
E
G
D
F
6
E
G
F
G
a
Asterisk indicates the shortest route
Total time
involved
186 mi.
276
276
186+110= 296
186+110= 296
276+ 58= 334
276+300= 576
296+241= 537
296+350= 646
276+300= 576
296+350= 646
537+479=1016
276+300= 576
537+479=1016
576+404= 980
nth nearest
node
B
Its minimum
time
186 mi.
Its last
connectiona
AB
276
AD*
296
BC
537
CE
576
DF*
980
FG*
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
Destination
Letterkenny
Fort Hood
Fort Riley
Cleveland
150
150
325
50
275
100
375
South
Charleston
100
150
350
San
Jose
800
325
350
Demand
300
300
50
100
400
275
Fort Carson
100
300
100
100
450
250
Fort Benning
100
Supply
400
150
100
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
Origin
Cleveland
S. Charleston
Cleveland
San Jose
Cleveland
San Jose
Cleveland
Destination
Letterkenny
Letterkenny
Fort Hood
Fort Hood
Fort Riley
Fort Carson
Fort Benning
Number of
carriers
150
150
50
50
100
100
100
5
This problem can be used effectively as an in-class exercise. Although the problem
might be solved using a combination of the shortest route method to find the optimum
path between stops and then a traveling salesman method to sequence the stops, it is
intended that students will use their cognitive skills to find a good solution. The class
should be divided into teams and given a limited amount of time to find a solution. They
should be provided with a transparency of the map and asked to draw their solution on it.
The instructor can then show the class each solution with the total distance achieved.
From the least-distance solutions, the instructor may ask the teams to explain the logic of
their solution process. Finally, the instructor may explore with the class how this and
similar problems might be treated with the aid of a computer.
Although the question asks the student to use cognitive skills to find a good route, a
route can be found with the aid of the ROUTER software in LOGWARE. 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
distances 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.
0.5
1.0
1.5
2.0
0
19
20
21
17
20
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
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
21
--STOP DATA
NO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
STOP
LOAD VOL.
LOAD
DESCRIPTION TY WGHT CUBE HCRD VCRD ZN TIME BEG1 END1
Stop 1
D
1
0
0.14
0.80
0
0
0 9999
Stop 2
D
1
0
0.14
1.14
0
0
0 9999
Stop 3
D
1
0
0.14
1.31
0
0
0 9999
Stop 4
D
1
0
0.35
1.31
0
0
0 9999
Stop 5/22
D
1
0
0.52
0.61
0
0
0 9999
Stop 6
D
1
0
0.58
1.31
0
0
0 9999
Stop 7
D
1
0
0.80
1.31
0
0
0 9999
Stop 8
D
1
0
1.03
0.61
0
0
0 9999
Stop 9
D
1
0
1.03
0.96
0
0
0 9999
Stop 10
D
1
0
1.03
1.31
0
0
0 9999
Stop 11
D
1
0
1.36
1.31
0
0
0 9999
Stop 12
D
1
0
1.48
1.31
0
0
0 9999
Stop 13
D
1
0
1.80
1.31
0
0
0 9999
Stop 14
D
1
0
1.87
1.31
0
0
0 9999
Stop 15
D
1
0
1.84
0.61
0
0
0 9999
Stop 16
D
1
0
1.95
0.61
0
0
0 9999
Stop 17
D
1
0
1.29
0.10
0
0
0 9999
Stop 18
D
1
0
1.26
0.61
0
0
0 9999
Stop 19
D
1
0
1.15
0.10
0
0
0 9999
Stop 20
D
1
0
0.69
0.23
0
0
0 9999
Stop 21
D
1
0
0.14
0.26
0
0
0 9999
BEG2
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
END2
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
VEHICLE DATA
--CAPACITY--
NO.
1
VEHICLE
DESCRIPTION
Bus
TP
1
NO
1
WGHT
9999
CUBE
9999
--VEHICLE-FIXED
COST
0
PER MI
COST
0
--DRIVER-FIXED
COST
0
PER HR
COST
0
OVER
TIME
COST
0
NO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
STOP
NO.
14
14
15
16
18
19
19
19
19
9
9
9
5/22
5/22
5/22
5/22
20
20
20
STOP
DESCRIPTION
Stop 14
Stop 14
Stop 15
Stop 16
Stop 18
Stop 19
Stop 19
Stop 19
Stop 19
Stop 9
Stop 9
Stop 9
Stops 5&22
Stops 5&22
Stops 5&22
Stops 5&22
Stop 20
Stop 20
Stop 20
STOP
NO.
16
15
17
17
9
8
20
5/22
18
20
19
21
1
21
20
9
21
0
5/22
STOP
DISTANCE
DESCRIPTION IN MILES
Stop 16
0.78
Stop 15
0.90
Stop 17
1.06
Stop 17
1.18
Stop 9
0.58
Stop 8
0.76
Stop 20
0.59
Stops5&22
1.14
Stop 18
0.53
Stop 20
1.08
Stop 19
1.11
Stop 21
1.69
Stop 1
0.56
Stop 21
1.05
Stop 20
1.14
Stop 9
0.97
Stop 21
0.84
School
1.03
Stops 5&22
0.55
22
20
21
22
17
0
2
Stop 17
School
Stop 2
0
School
5/22 Stops 5&22
5/22 Stops 5&22
2.43
1.37
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:
Distancea
95.55 mi.
86.45
182.00 mi.
$147.00
54.60
$201.60
Strategy 2 is a mixed strategy involving staying at motels closest to the center of the
stops 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
23
24
TY
D
P
D
D
D
P
D
P
D
D
D
LOAD
WGHT
20
14
18
3
5
4
3
3
3
3
4
VOL.
CUBE
0
0
0
0
0
0
0
0
0
0
0
HCRD VCRD ZN
1147
8197
0
1206
8203
0
1052
7791
0
557
8282
0
527
8341 0
565
8273
0
1031
7954
0
1159
8224
0
1716
7877
0
607
8166
0
527
8351
0
LOAD
TIME
15
45
45
45
45
45
45
45
15
45
45
BEG1
360
360
360
180
360
180
180
180
600
360
360
END1 BEG2
1440
1800
1440
1800
1440
1800
1440
1800
1440 1800
1440
1800
1440
1800
1440
1800
1440
1800
1440
1800
1440
1800
END2
2880
2880
2880
2880
2880
2880
2880
2880
2880
2880
2880
--VEHICLE DATA
-CAPACITY--
NO.
1
2
3
VEHICLE
DESCRIPTION
Truck #1-20
Truck #2-25
Truck #3-30
TP
1
2
3
NO
3
1
1
WGHT
20
25
30
CUBE
9999
9999
9999
--VEHICLE-FIXED
COST
0
0
0
PER MI
COST
1.30
1.30
1.30
--DRIVER-FIXED
COST
0
0
0
PER HR
COST
0
0
0
OVER
TIME
COST
0
0
0
25
Pickup
Pickup
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 3:00AM of day 1
Daytona Beach Deliver 18 pallets
Clearwater
Pickup 14 pallets
Depot
Return time 5:48AM of day 2
Route #2 with 20-pallet truck
Depot
Start time 3:00AM of day 1
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 4:43PM of day 2
26
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
1
1
Load
date
Discharge
date
100
100
XXa
100
C 15
2
69
XX
A 36
Rim restriction
XX
B 39
10
100
100
10
10
10
100
10
XX
XX
XX
100
XX
100
XX
XX
10
1
100
XX
XX
100
10
XX
100
XX
10
100
XX
XX
100
10
100
XX
A 86
XX
100
100
10
100
XX
C 52
XX
100
100
XX
100
100
XX
100
100
1
40
Slack
D 10
Slack
Rim restriction
2
5
XX
10
XX
10
10
1
1
FIGURE 7-5 Transportation Matrix Setup and Solution for the Queens Lines
Tanker Scheduling Problem
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.
=
=
=
=
=
Cost
$519.00
726.00
530.00
498.00
$2,274.00
.00
$2,274.00
28
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
Hays
24,000 0.0426
Manhattan
42,000 0.0222
40,000 0.0246
Salinaa
Great Bend
15,000 0.0498
Transportation
Lost sales
Total
a
=
=
=
=
=
Cost
$1,022.40
932.40
984.00
747.00
$3,685.80
3,150.00
$6,835.80
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
Hold orders 2 periods
2,278.60
Optimum
11
Routes are built by placing the trips end-to-end throughout the day from 4AM until
11PM, 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
29
operate. If there were no delivery time restrictions on warehouses and trips could be
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,
Warehouse
location
Flint
Alpena
Saginaw
Lansing
Mt. Pleasant
W. Branch
Pontiac
Traverse City
Petoskey
(1)
Number of
trips
43
5
8
21
12
5
43
6
5
(2)
Total time
per trip, hr.
1.25
10.50
2.25
3.75
5.50
6.00
2.75
10.50
11.75
Total
(3)=(1)(2)
Total time,
hr.
53.75
52.50
18.00
78.75
66.00
30.00
118.25
63.00
58.75
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.
Truck 1
Truck 2
Truck 3
Truck 4
Truck 5
Truck 6
Day 1
Petoskey 11.75
W Branch 6.00
Flint 1.25
Total =19.00 hr.
T. City 10.50
2 Lansing 7.50
Total =18.00 hr.
T. City 10.50
2 Lansing 7.50
Total =18.00 hr.
Alpena 10.50
Lansing 3.75
3 Flint 3.75
Total =18.00 hr.
M Pleasant 5.50
4 Pontiac 11.00
2 Flint 2.50
Total =19.00 hr.
M Pleasant 5.50
4 Pontiac 11.00
2 Flint 2.50
Total =19.00 hr.
Day 2
Petoskey 11.75
W Branch 6.00
Flint 1.25
Total =19.00 hr.
T. City 10.50
2 Lansing 7.50
Total =18.00 hr.
Alpena 10.50
2 Lansing 7.50
Total =18.00 hr.
M Pleasant 5.50
4 Pontiac 11.00
2 Flint 2.50
Total =19.00 hr.
M Pleasant 5.50
4 Pontiac 11.00
2 Flint 2.50
Total =19.00 hr.
M Pleasant 5.50
3 Pontiac 8.25
2 Flint 2.50
Total =16.25 hr.
Day 3
Petoskey 11.75
W Branch 6.00
Flint 1.25
Total =19.00
T. City 10.50
2 Lansing 7.50
Total =18.00
Alpena 10.50
2 Lansing 7.50
Total =18.00 hr.
M Pleasant 5.50
4 Pontiac 11.00
2 Flint 2.50
Total =19.00 hr.
M Pleasant 5.50
4 Pontiac 11.00
Flint 1.25
Total =17.75
M Pleasant 5.50
6 Saginaw 13.50
Day 4
Petoskey 11.75
5 Flint 6.25
Day 5
Petoskey 11.75
5 Flint 6.25
Total = 16.50
W Branch 6.00
10 Flint 12.50
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
18
Route #1
Load 19
Route #5
Load 9
2
5
4
3
14
Miles 12
x 10
10
Route #4
Load `8
2
3
Warehouse
5
4
6
4
16
Route #2
Load 20
Route #3
Load 17
2
4
0
0
10 12 14
Miles x 10
16
18
20
22
24
26
(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
Route
no
1
2
3
4
Total
Run Stop
Brk Stem
time, time, time, time, time,
Start Return No of
Route
hr
hr
hr
hr
hr
time
time stops dist,Mi
1.2
1.0
.3
.0
.4 08:59AM 10:12AM
3
29
8.9
6.6
1.3
1.0
1.1 08:32AM 05:25PM
19
199
6.2
3.7
1.4
1.0
.9 08:42AM 02:54PM
14
112
7.5
5.1
1.5
1.0
1.4 08:30AM 04:02PM
12
152
23.8 16.4
4.4
3.0
3.8
48
492
Route
cost,$
.00
.00
.00
.00
.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 3:30-3:45PM and still meet
the customers time windows and return to the depot by 6PM. 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
Earliest starting time (min) - 480 Latest return time (min) - 9999
Default vehicle speed (miles per hour) - 30
After how many clock hours will overtime begin - 168
GENERAL DATA
Percent of vehicle in use before allowing pickups - 0
Horizontal scaling factor - 4.6 Vertical scaling factor - 4.6
Maximum TIME allowed on a route (hours) - 168
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) - 60 To begin after - 720
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
NO DESCRIPTION
1 Covington House
2 Cuyahoga Falls
3 Elyria
TY
D
D
D
LOAD
WGHT
1
9
1
VOL.
CUBE
0
0
0
HCRD VCRD ZN
23.40
12.90
0
13.40
13.40
0
6.30
16.80
0
LOAD
TIME
2
18
5
BEG1
540
540
540
END1 BEG2
1020
9999
1020
9999
1020
9999
END2
9999
9999
9999
32
46
47
48
Westbay
Westhaven
Broadfiels Mnr
D
D
D
6
2
6
0
0
0
8.40
8.50
18.20
18.00
18.10
22.90
0
0
0
10
5
2
630
540
540
690
1020
1020
9999
9999
9999
9999
9999
9999
--VEHICLE DATA
-CAPACITY--
NO.
1
VEHICLE
DESCRIPTION
Station wagon
TP
1
NO
50
WGHT
63
CUBE
9999
--VEHICLE-FIXED
COST
0
--DRIVER--
PER MI
COST
0
FIXED
COST
0
PER HR
COST
0
OVER
TIME
COST
0
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
35
FIGURE 1 Current
Route Design
The questions for management are: How restrictive are the time windows? Can
deliveries be made outside of the account's "open hours", such as by giving the driver the
key to a safe storage area? Can management offer a small incentive to widen the time
window when it otherwise would not be convenient for the account? Relaxing such time
windows is often one of the important sources for cost savings in routing problems.
Q3. When there are no truck capacity restrictions on a routing problem, the most
efficient route design would be to use one large truck to serve all accounts. Therefore,
we would expect that trucks of larger capacity would reduce the total distance traveled.
A ROUTER optimizing run was made with 600 case capacity trucks to find out. All
other conditions were set at the current design.
36
FIGURE 2
Optimized
Current Route
Design
Compared with
the optimized current design, the potential savings is ($731.31 722.98) 250 =
$2,082.50 per year. Two 600-case trucks would be needed, along with 3 of the smaller
trucks. Although there is a positive savings associated with using larger trucks, the
savings seems quite small and perhaps not worth switching to some larger trucks at this
time. We should note that these savings are a result of comparing full costs, which
include truck depreciation. Although a present value analysis would be appropriate here,
we would need some additional information which might include (1) the proportion of
truck costs allocated to equipment depreciation, (2) the life of the trucks, (3) the
estimated salvage value of the trucks, and (4) and the required rate of return on
investments of this type.
37
Q4. From the incremental costs in the detailed report of the optimized current design,
we can see that account 14 costs $39.05 and account 19 costs $35.04 to make the
deliveries. Since the volume of account 19 is 90 cases and exceeds the 50 case capacity
of an outside transport service, account 19 cannot be considered for alternate delivery. If
it only costs $35.00 for an outside delivery service to handle account 14, then a cost
savings can be realized. Dropping this account and redesigning the routes shows that
only 5 trucks are needed for a route cost of $690.17 and route miles of 286. The total
cost for handling all accounts would be $690.17 + 35.00 = $725.17. Compared with the
optimized current design, this is an annual savings of ($731.31 725.17) 250 =
$1,535.00.
Economically, Roy should use the outside transportation. However, losing direct
control over the deliveries and possible adverse reactions from route salesmen may give
him second thoughts about it.
Q5. Two ROUTER runs were made here in order to determine the effect of a shorter
workday before overtime begins. If no overtime is allowed, then six trucks are required
for a total daily routing cost of $754.77. If some overtime is permitted, then the route
cost can be reduced to $733.28 with 6 trucks required. Compared with the optimized
current design, Roy Fowler would seem to be putting himself at a disadvantage by not
wanting to pay overtime.
Q6. Moving the warehouse to a more central location would have the appeal of being
closer to all stops and would result in shorter routes. Testing this shows that total route
distance can be reduced to 305 miles with a daily cost of $716.91. However, in order to
meet the time window and other constraints, an extra truck is needed. There is a potential
annual savings of ($731.31 716.91) 250 = $3,600.00. Since it costs $15,000 to make
38
the move, a simple return on investment would be (3,600/15,000) 100 = 24%. At this
ROI, the move should be seriously considered.
Q7. Routing and scheduling beer trucks from a central depot is quite similar to
delivery problems found in manufacturing, retail, and service industries. Several
examples are listed below.
Summary
Roy Fowler should consider optimizing his route design. This can be done at no
investment to him, and he can gain about a 7% reduction in operating costs on the
average. A substantial benefit can be realized by widening the time windows. He should
especially see if those time windows that cause routes to overlap themselves can be
widened. The potential for doing this is up to $14,418 per year. He should explore the
possibility and reasonableness of serving accounts 14 and 19 by an outside transport
service. Finally, he should consider relaxing his rigid policy of not wanting to pay
overtime, especially if the union is successful in negotiating a 7-hour workday.
There seems to be little opportunity for reducing cost by increasing truck capacity or
relocating the warehouse. Although there are operating cost reductions available, they do
not seem sufficient to justify a change in truck size unless Roy can make good use of the
trucks in other ways.
39
40
This problem is complex, considering that it is quite dynamic. Although the patient pick
up list appears to be fixed for a particular day, changes take place in the form of
cancellations and occasional additions. Patient appointment times are met a high
percentage of the time, but not always. Patients to be returned to their origins may find
alternate means of transportation so that 100 percent of the patients picked up may not
have to be returned. For simplicity, it will be assumed that the patient pick up list is fixed
for a particular day and appointment times are rigid. In addition, the returns are not
directly considered in designing the routes. Rather, they are expected to be seeded into
the pick up routes. This need not be a serious limitation as long as van capacity (number
of pick-up patients to the number of available seats) is not highly utilized on a pickup
route and van utilization for pick ups drops as appointments are concentrated in the
morning hours and return times are shifted toward the afternoon hours, as shown in
Figure 1.
41
Appointment times
Return times
A typical days patient list can be submitted to ROUTER. Coordinate points for zip
code areas are scaled from the map in Figure 2 of the case study. Appointment times are
represented as time windows with a 2-hour gap for pick up. A large number of 15passenger and 6-passenger vans are assumed available to start the ROUTER model.
Since ROUTER is a basic vehicle routing and scheduling model, it simply assigns
patients to vans and gives a sequence to pick them up. It assumes that a single van leaves
and returns to MMC only once in a day. Analysis must account for this, since vans make
several out and back trips per day.
It is not possible to accurately recreate the current routing patterns for the vans from
the data given. However, optimized routing and scheduling can be found through the
application of ROUTER. First, the typical daily appointment list is solved with
ROUTER using a 2-hour time window where the ending window time is the appointment
time. An average speed of 27.5 miles per hour, variable costs of $0.11 per mile1, and a
large number of 15-passenger and 6-passenger vans are used. A complete database is
shown in the appendix of this note.
Using the appointment list for the typical day as representative with all 56 patients,
ROUTER shows that 6 vans are needed, where one is a 15-passenger type (route 1). This
is routing where no patient is transported by the contract transport service. The routing
represents a total of 329 miles driven at a variable cost of 329 mi. $0.11/mi. = $36.19.
The routes are placed end-to-end to cover the full day and then ranked, first by the
number of patients transported throughout the day on a van and second by the total miles
driven, as shown in Table 1. A reasonable procedure for allocating routes to vans is to
assign routes with the largest number of patients first to the largest vans. This is because
1
Per mile variable cost is determined as repair cost/30,000 mi. per yr. = $1,000/30,000 = $0.03 per mi. plus
fuel cost at $1.00 per gallon/average miles per gallon = 1/13 = $0.08 per mi. for a total of 0.08 + 0.03 =
$0.11 per mile.
42
the ambulance service charges on a per patient basis, and these routes would be most
costly to subcontract. If the number of patients is tied for a van, rank first the van with
the shortest total route distance.
TABLE 1 Route Design for Representative Patient Appointments with
No Transport Subcontracting
Van
no.
1
1
1
1
1
2
2
3
3
4
4
5
6
1
Start from
MMC
6:07AM
6:52AM
8:55AM
11:04AM
11:41AM
7:24AM
10:59AM
7:41AM
10:45AM
7:37AM
11:52AM
7:24AM
7:43AM
Return to
MMC
6:28AM
8:28AM
10:30AM
11:31AM
1:27PM
9:05AM
11:57AM
9:16AM
12:27PM
9:12AM
1:34PM
8:44AM
9:03AM
Patients on
route
1
81
4
1
5
5
5
5
5
5
5
4
3
56
Route distance
7 mi.
22
33
9
35
32
13
29
33
30
33
25
28
329 mi.
Route time
0.3 hr.
1.6
1.6
0.4
1.8
1.7
1.0
1.6
1.7
1.6
1.7
1.3
1.6
17.6 hr.
$ 5,750 1
28,500 2
6,000
141,000
8,686 3
$189,936
The question now is whether substituting contract carriage for some of the routes will
significantly lower vehicle and driver costs while only slightly increasing variable costs.
Since the routes are ranked in Table 1, a reasonable rule for dropping routes would be to
start from the bottom of the route list in Table 1 and work upward. This will determine
the number of vans needed and the patients to be transported by each mode. These
results are shown in Table 2. An unlimited number of subcontracting trips is assumed.
43
Vans
6
5
4
3
2
1
0
1
Driver
$141,000
117,500
94,000
70,500
47,000
23,500
0
Vehicle
$34,250
29,500
20,000
15,250
10,500
5,750
0
Variable
(gas +
repair)
$14,686
12,946
11,286
8,623
5,986
3,798
0
Subcontracting
$
0
6,235 1
14,549
35,333
56,117
76,901
116,390
Total
$189,936
166,181
139,835
129,706
119,603
109,949
116,390
Annual subcontracting cost is $8.66 per round trip 240 days per year 3 patients = $6,235.20
(2) How many subcontracted trips should MMC negotiate and at what price?
One 15-passenger van appears to be the optimal number, but this requires 37 patients
20 days per month = 740 subcontracting trips per month. This is more than the 500
allowed. If the 500-trip limit (500/20 days per mo. = 25 patients per day) is to be
respected, approximately one additional van will be needed. On the other hand, the cost
of the additional van is the annual cost of a van + the annual drivers salary + the gas and
repair cost = $19,000/4 + 23,500 + 1,1882 = $29,438/yr. to transport 10 patients per day.
At this rate, MMC could afford to pay $29,438/240/10 = $12.27 per round trip (find 10
patients for van 2 in Table 1). Possibly the subcontractor would be willing to offer this
additional transport for a price between the $8.66 and $12.27 per trip. It would be in
Macs interest to do this. Renegotiating all trips at the 750-trip level is another
possibility. MMC can afford to pay a subcontracting cost of up to $189,936 - 31,0483 =
$158,888/yr. This would require transporting 37 patients per day at a maximal average
trip cost of $158,888/240/37 = $17.89. Since $17.89 is the optimal worth of
subcontracting, MMC probably can negotiate a rate much less than this while still
encouraging the subcontractor to provide service at the higher level.
(3) MMC is considering using all 6-passenger vans. Would this be a good decision?
Developing daily routes and assigning them to 6-passenger vans is the same as for Table
1. This initial allocation without subcontracting is shown in Table 3. Considering the
subcontracting trip limit of 500 patients per month (25 patients per day), two vans are
needed in the MMC fleet. This number of vans is determined by counting the patients
from the bottom of Table 3 until approximately 25 patients are found, considering
increments of full vans and noting the number of vans remaining at the top of the list.
(Assume 27, an extra 2 patients can be accommodated by the subcontractor at no
additional cost.) The variable cost for two vans is shown below.
Two
6-passenger
44
vans
Annual cost of 2 drivers
Vehicle mileage + repair cost
Subcontracting cost
Total
1
2
47,000
5,9601
56,1172
$109,077
47,000
5,986
56,117
$109,103
In contrast, the 15-passenger and 6-passenger van configuration compared with the two
6-passenger van configuration saves $109,103 109,077 = $26 per year. The initial
outlay is less for the smaller van by $4,000. From purely an economic standpoint, using
just 6-passenger vans would appear to be the best choice, however there may be customer
service reasons for having the larger van available to meet peak demand needs.
TABLE 3 Route Design for 6-Passenger Vans Only
Van
no.
1
1
1
1
1
2
2
2
3
3
4
4
5
6
Start from
MMC
6:07AM
6:52AM
7:37AM
11:04AM
11:52AM
7:22AM
8:55AM
10:59AM
7:41AM
10:45AM
7:24AM
11:41AM
7:24AM
7:43AM
Return to
MMC
6:28AM
7:37AM
9:12AM
11:31AM
1:34PM
8:19AM
10:30AM
11:57AM
9:16AM
12:27PM
9:05AM
1:27PM
8:44AM
9:03AM
Patients on
route
1
3
5
1
5
5
4
5
5
5
5
5
4
3
56
Route distance
7 mi.
12
30
9
33
13
33
13
29
33
32
35
25
28
332 mi.
Route time
0.3 hr.
0.8
1.6
0.4
1.7
1.0
1.6
1.0
1.6
1.7
1.7
1.8
1.3
1.3
17.8 hr.
45
STOP
DESCRIPTION---- TY
Baker
P
Boyd
P
Carver
P
Ivey
P
Rashed
P
Walsh
P
Johnson
P
Burgess
P
Delgado
P
Fairrow
P
Middlebrooks
P
Suech
P
Lawson
P
Reed
P
Bongiovanni
P
Miller
P
Talley
P
Williams
P
Dumas
P
Taylor
P
Barker
P
Lhota
P
Manco
P
Webb
P
Wilson
P
Arrington
P
Staunton
P
Wall
P
Williams
P
Caruso
P
West
P
Amaro
P
Brown
P
Ciesicki
P
Pinkevich
P
Staufer
P
Winterich
P
Brown
P
Ball
P
Lanza
P
Mayernik
P
Suech
P
Heffner
P
Jarrell
P
LOAD VOLUME
WGHT
CUBE
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
46
45
46
47
48
49
50
51
52
53
54
55
56
Piatak
Swaysland
Baer
Wills
Fauber
Mullins
Pack
Westerfield
Lisiewski
McPherson
Mykytuk
Gutschmidt
VEHICLE
NO. DESCRIPTION---1 15-pasngr vehcl
2 6-pasngr vehcl
P
P
P
P
P
P
P
P
P
P
P
P
TP
1
2
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
7.00
6.10
3.80
4.70
5.20
7.20
6.20
5.90
5.00
4.20
6.00
5.50
3.50
6.70
4.90
6.10
6.80
6.30
5.00
6.30
5.50
6.20
6.50
5.90
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
480
480
540
540
660
660
660
675
690
720
750
780
640
640
700
700
780
780
780
795
810
840
870
900
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
9999
47
It should be obvious to Anita that the current distribution system may not be performing
at optimum. Allowing carriers to decide the routes to use when they are being paid on a
mileage basis is like asking a fox to watch the chicken coop. She really cannot expect
that carriers will be motivated to seek out optimal routing patterns. Therefore, she should
determine the best routes between regional and field warehouses and insist that carriers
invoice according to the mileages along these specified routes.
Using the map provided in Figure 1 of the case study and the current assignments of
field warehouses to regional warehouses, she can develop a database for the ROUTE
module in LOGWARE. The database is shown in the Supplement to this note. Running
ROUTE will give the optimal routes from which she can develop transportation costs, as
shown in Table 1 below. Compared with the current cost level, a savings of
$652,274 630,140 = $22,134 per year
can be realized. This savings does not require any investment; however, it will be
magnified by the growth in demand in the next five years.
48
Field
warehouse
Los Angeles
Phoenix
Salt Lake City
San Francisco
Portland
Butte
Seattle
Optimal
route
miles
219
588
815
183
293
676
467
3,241
Average
number of
tripsa
366.7
200.0
116.7
280.0
143.3
16.7
186.7
Transport
cost, $
104,399b
152,880
123,644
66,612
54,583
14,676
113,346
630,140
Field warehouse
Los Angeles
Phoenix
Salt Lake City
San Francisco
Portland
Butte
Seattle
If served
from Burns
806 mi.
973
536*
555
293*
676*
467*
If served
from Fresno
219 mi.*
588*
815
183*
757
1,120
925
This shows that Salt Lake City would be better served out of Burns rather than Fresno.
Burns currently is near its capacity limit, so to assign Salt Lake City's volume to it
would require expansion. In the short term, 35,000/8 = 4,375 cwt. of inventory capacity
is needed. However, (43,000 + 5,000 + 56,000)/8 = 13,000 cwt. of the 15,000 cwt. of
available capacity is currently being used. At minimum, an additional increment of
capacity is required at a cost of $300,000. Reassignment of Salt Lake City to Burns
would save 815 536 = 279 miles per trip. On 116.7 trips per year, the annual savings
would be 1.30 279 116.7 = $42,327. The simple return on investment (ROI) would
be:
ROI
$42,327
100 141%
.
$300,000
49
If the anticipated growth in demand is realized, the number trips to Salt Lake City would
increase to 56,000/300 = 186.7. The projected savings in the fifth year would be 1.30
279 186.7 = $67,716. The average annual savings would be (42,327 + 67,716)/2 =
$55,022. The average annual ROI is:
ROI
$55,022
100 18.3%
$300,000
Anita must now compare this return to other worthy investments in the firm to see if this
opportunity is worth the risk.
(3) Is there any merit to consolidating the regional warehousing operation at Reno, NV?
If Reno were to replace the Burns and Fresno warehouses, an initial cost of $2,000,000
would be incurred to establish the new location and shut down the existing warehouses.
Against this cost would be a savings of 40 percent of the inventory in the two regional
warehouses. That is, the total inventory is 393,000/8 = 49,125 cwt. now and 528,000/8 =
66,000 cwt. in five years. The inventory cost savings now would be 0.40 0.35 60
49,125 = $412,650 and in five years 0.40 0.35 60 66,000 = $554,400. However,
transportation costs will increase compared with the two-warehouse distribution system.
If Reno is used, the transportation costs would be:
Optimal
Field warehouse
route
served from Reno
miles
Los Angeles
472
Phoenix
732
Salt Lake City
520
San Francisco
228
Portland
542
Butte
823
Seattle
716
Totals
4,033
a
472 336.7 1.30 = $225,007
Current
no. of
trips
366.7
200.0
116.7
280.0
143.3
16.7
186.7
Current
transport
cost, $
225,007a
190,320
78,889
82,992
100,969
17,867
173,780
869,824
5th-year
no. of
trips
440.0
280.0
186.7
350.0
190.0
50.0
263.3
5th-year
transport
costs, $
269,984
266,448
126,209
103,740
133,874
53,495
245,080
1,198,830
We cannot make a fair comparison with the current system design since there is
inadequate capacity at Burns to handle the growth in volume. Two additional units of
capacity will be needed for a total of $600,000. Then, the net investment attributable to
Reno is 2,000,000 600,000 = $1,400,000. The revised transportation cost for the fifth
year is as follows.
50
Regional
warehouse
Fresno
Fresno
Burns
Fresno
Burns
Burns
Burns
Totals
Field
warehouse
Los Angeles
Phoenix
Salt Lake City
San Francisco
Portland
Butte
Seattle
Optimal
route
miles
219
588
536
183
293
676
467
2,962
5th-year
number of
trips
440.0
280.0
186.7
350.0
190.0
50.0
263.3
Transport
cost, $
125,268
214,032
130,092
83,265
72,371
43,940
159,849
828,817
Now,
Transport costs
Transport costs
Net increase
Less inventory
savings
Net cost savings
Current year
$587,813
869,824
$282,011
5th year
$828,817
1,198,830
$370,013
(412,650)
$130,639
(554,400)
$184,387
Burns/Fresno
Reno
The average annual savings for Reno now is (130,639 + 184,387)/2 = $157,513. The
relevant return on investment is
ROI
$157,513
100 1125%
.
$1,400,000
This is probably not a sufficient return to justify the warehouse at Reno. Rather, if Orion
wishes to serve the increasing demandthere is no requirement to do sothen the better
strategy would be to expand Burns by 20,000 cwt. and serve Salt Lake City from this
location.
An interesting question to pose to students is: What does it mean to only serve
demand up to the limits of capacity? Orion could serve only the more profitable demand
and avoid the risks of expansion.
51
R & T WHOLESALERS
Teaching Note4
The objective of this assignment is to minimize the total monthly delivery costs for R&T
Wholesalers, a company distributing general products throughout India. The focus is on
one warehouse acting as a truck depot that delivers merchandise to retailers located in
surrounding towns. Delivery expenses are minimized through the optimal utilization of
the trucks, crews, and related expenses. The constraints and other considerations listed
below were used when designing the solution methodology and identifying the optimal
solution.
Constraints
Operating Schedule
- Trucks make deliveries every day of the week except Saturday or Sunday
- Normal operation is for trucks to be loaded overnight and leave from the
warehouse in the morning
- Trucks make deliveries within towns from 9 a.m. to 6 p.m.
- Earliest start time for trucks is 12 a.m. in the morning of delivery
- Trucks returned to the depot require 2 hours for reloading and subsequent sameday delivery
Visits Per Month
- Every town has a known number of visits per month
Truck Capacity
- T407 trucks have a capacity of Rs500,000
- T310 trucks have a capacity of Rs350,000
Truck Operating Costs
- Trucks operate at an average speed of 40 km/hr
- T407 trucks have an operating cost of Rs13,500
- T310 trucks have an operating cost of Rs7,000
- T407 trucks have a running cost of Rs5 per kilometer
- T310 trucks have a running cost of Rs3 per kilometer
- Each truck has a crew of two, a driver and a helper
- The driver is paid Rs2,200 per month
- The helper is paid Rs1,400 per month
- Each crewmember receives Rs60 per day for meals and other expenses while on
the road
Working Schedule
- Flexibly planned breaks for crewmembers are at approximately 6 a.m., 12 p.m.
and 6 p.m.
- Breakfast and lunchtime breaks are 30 minutes each and dinner is 60 minutes
The solution to this exercise is provided by Nutthapol Dussadeenoad, Inderjot Gandhi, Earle Keith, Lisa
Kuta, Jan Shahan, and
Piyanuch Vichitakul who were students in the MBA program of the Weatherhead School of Management.
52
1. Route 43 cities
altogether
Cost = Rs116,678
Cost = Rs135,254
Cost = Rs109,092
Cost = Rs78,328
Cost = Rs65,780
Satisfy?
Y
STOP
53
54
55
fixed cost
allowance
Week 1&3
Week 2&4
Running cost
Common Carrier
Total
of
of
of
truck 2 (T310)
driver
helper
4
2
2
people
people
people
Week 1 & 3
Week 2 & 4
To Guntur
To Rajahmundry
=
=
=
x
x
x
=
=
=
=
2 weeks
2 weeks
2 weeks
x
x
x
3,146 km
2,414 km
35
76
km
km
2
2
2
5 days
5 days
4 days
x
x
x
x
x
x
7,000
2,200
1,400
Rs 60
Rs 60
Rs 60
=
=
=
14,000
4,400
2,800
=
=
=
2,400
1,200
960
x
x
3 Rs/km
3 Rs/km
x
x
2 weeks
2 weeks
=
=
18,876
14,484
x
x
15 Rs/km
15 Rs/km
x
x
4 weeks
4 weeks
=
=
2,100
4,560
65,780
56
Truck #
DAY 1
II
DAY 2
II
DAY 3
II
DAY 4
II
Route
1 Depot
Podili
Kondulur
Tanguturu
Depot
25 Depot
Jangareddygudem
Kakinada
Depot
Start Time
Crew #
Driver I, Helper I
7:04 PM
5.46 AM
405
339
5:45 PM
6:39 AM
7 Depot
Tadikonda
Tenali
Chirala
Vuyyuru
Depot
7:59 AM
8 Depot
Sattenapalle
Ongole
Depot
Route
Miles
4:27 AM
5 Depot
Chilakalurupet
Narasaraopet
Macheria
Depot
13 Depot
Kaikalur
Bhimavaram
Tadepallegudem
Depot
End Time
Driver I, Helper I
8:18 PM
306
Driver II, Helper II
8:47 PM
232
6:47 AM
Driver I, Helper I
7:07 PM
253
7:01 AM
8:12 PM
19 Depot
Narasapur
Mandapeta
Depot
5:19 AM
11 Depot
Pamarru
Machilipatnam
Palakolu
Depot
7:19 AM
307
Driver I, Helper I
7:03 PM
349
Driver II, Helper II
5:19 PM
240
57
DAY 5
24 Depot
Eluru
Chintalapuidi
Depot
II
2 Depot
Kani Giri
Bestavaipetta
Giddalur
Markapur
Depot
6:56 AM
Driver I, Helper I
3:11 PM
190
3:44 AM
11:21 PM
525
End Time
Route
Miles
Guntur
Rajahmundry
Outsource
WEEK 2 & 4
DAY 1
Truck #
Route
Start Time
3 Depot
Ongole
Depot
5:01 AM
II
DAY 2
II
Driver I, Helper I
3:28 PM
18 Depot
Jaggayyapeta
Eluru
Depot
6:34 AM
13 Depot
Vuyyuru
Machilipatnam
Gudivada
Depot
7:41 AM
19 Depot
Hunuman Junction
Chirala
Bapatia
Repalie
Depot
8:26 AM
Crew#
278
Driver II, Helper II
5:00 PM
276
Driver I, Helper I
3:30 PM
153
Driver II, Helper II
7:49 PM
175
58
DAY 3
II
DAY 4
II
DAY 5
II
Outsource
14 Depot
Tenali
Palakolu
Depot
7:35 AM
24 Depot
Tanuku
Nidadvole
Kovvur
Depot
5:12 AM
16 Depot
Nuzvid
Tadepallegudem
Bhimavaram
Depot
7:27 AM
29 Depot
Piduguralia
Addanki
Vinukonda
Narasaraopet
Depot
6:23 AM
23 Depot
Amaiapuram
Kakinada
Depot
4:23 AM
Driver I, Helper I
4:40 PM
243
Driver II, Helper II
5:04 PM
315
Driver I, Helper I
5:51 PM
236
7:49 PM
318
Driver I, Helper I
7:52 PM
420
Open
Guntur
Rajahmundry
59
60
Route
no
1
2
Route
Run Stop
Brk Stem
time, time, time, time, time,
Start Return No of
Route
hr
hr
hr
hr
hr
time
time stops dist,Mi
14.6 10.1
2.5
2.0
7.8 04:27AM 07:04PM
3
405
19.6 13.1
4.5
2.0
9.4 03:44AM 11:21PM
4
525
Route
cost,$
1215.00
1575.00
13.6
7.6
4.0
2.0
306
918.00
7
8
12.8
13.2
5.8
7.7
5.0
3.5
2.0
2.0
4
2
232
307
696.00
921.00
11
10.0
6.0
3.0
1.0
240
720.00
13
12.3
6.3
4.0
2.0
253
759.00
19
13.7
8.7
3.0
2.0
349
1047.00
24
25
8.2
12.0
4.8
8.5
2.5
2.5
1.0
1.0
2
2
190
339
570.00
1017.00
VEHICLE INFORMATION
Route Veh Weight Delvry Pickup Weight
Cube Delvry Pickup
no typ capcty weight weight
util capcty
cube
cube
1
1
350
180
0 51.4%
9999
0
0
2
1
350
134
0 38.3%
9999
0
0
Cube
util
.0%
.0%
Vehicle
description
T310
T310
350
342
97.7%
9999
.0% T310
7
8
1
1
350
350
337
350
0 96.3%
0 100.0%
9999
9999
0
0
0
0
.0% T310
.0% T310
11
350
350
0 100.0%
9999
.0% T310
13
350
326
93.1%
9999
.0% T310
19
350
330
94.3%
9999
.0% T310
24
25
1
1
350
350
266
296
0
0
76.0%
84.6%
9999
9999
0
0
0
0
.0% T310
.0% T310
Stop
No description
Stop
Drive Distance Time
Arrive
Depart
time to stop to stop wind
time Day
time Day Min
Min
Miles met?
-->Break 30 minutes
61
2 Podili
6 Kondulur
09:00AM
1 09:30AM
11:28AM
1 12:28PM
-->Break 30 minutes
01:19PM
1 02:19PM
-->Break 60 minutes
07:04PM
1 -------
1 Tanguturu
Depot
Stop
No description
Stop volume
Weight
Cube
2 Podili
6 Kondulur
1 Tanguturu
Totals Weight: Del = 180
24
90
66
Pickups
Route time:
Driving
Load/unload
Break
Total
10.1 hr
2.5
2.0
14.6 hr
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
0
0
0
=
1
1
30
60
243.0
118.5
162
79
YES
YES
60
21.0
14
YES
--
---
225.0
150
Capacity in use
Weight
Cube
51.4%
.0%
183.00
7.6
44.6
.0
51.00
.6
18.9
.0
-48.00
-.7
.0
.0
0 Cube: Del = 0 Pickups = 0
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
162 mi
150
93
405 mi
9999 mi
$.00
.00
1215.00
.00
$1215.00
Stop
No description
5 Kani Giri
9 Bestavaipetta
7 Giddalur
4 Markapur
Depot
Stop
No description
Arrive
time
-->Break
09:00AM
-->Break
01:21PM
02:41PM
05:12PM
-->Break
11:21PM
Stop
Drive Distance Time
Depart
time to stop to stop wind
Day
time Day Min
Min
Miles met?
30 minutes
1 11:30AM
1 150
286.5
191 YES
30 minutes
1 01:51PM
1
30
81.0
54 YES
1 03:41PM
1
60
49.5
33 YES
1 05:42PM
1
30
91.5
61 YES
60 minutes
1 ------- -- --279.0
186
Stop volume
Weight
Cube
5 Kani Giri
9 Bestavaipetta
7 Giddalur
4 Markapur
Totals Weight: Del = 134
24
25
25
60
Pickups
Route time:
Driving
Load/unload
Break
Total
13.1 hr
4.5
2.0
19.6 hr
Max allowed
24.0 hr
0
0
0
0
=
Capacity in use
Weight
Cube
38.3%
.0%
87.00
3.6
31.4
.0
15.00
.6
24.3
.0
210.00
8.4
17.1
.0
-21.00
-.4
.0
.0
0 Cube: Del = 0 Pickups = 0
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
191 mi
186
148
525 mi
9999 mi
62
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
$.00
.00
1575.00
.00
$1575.00
Stop
No description
Arrive
time
-->Break
11 Chilakalurupet 09:00AM
12 Narasarapet
10:31AM
-->Break
42 Macheria
01:46PM
-->Break
Depot
08:18PM
Stop
No description
Stop
Drive Distance Time
Depart
time to stop to stop wind
Day
time Day Min
Min
Miles met?
30 minutes
1 10:00AM
1
60
111.0
74 YES
1 11:31AM
1
60
31.5
21 YES
30 minutes
1 03:46PM
1 120
105.0
70 YES
60 minutes
1 ------- -- --211.5
141
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
97.7%
.0%
57.00
.6
71.4
.0
.00
.0
42.9
.0
405.00
2.7
.0
.0
0 Cube: Del = 0 Pickups = 0
11 Chilakalurupet
92
12 Narasarapet
100
42 Macheria
150
Totals Weight: Del = 342 Pickups
0
0
0
=
Route time:
Driving
Load/unload
Break
Total
7.6 hr
4.0
2.0
13.6 hr
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Max allowed
74 mi
141
91
306 mi
9999 mi
$.00
.00
918.00
.00
$918.00
Stop
No description
14 Tadikonda
19 Tenali
8 Chirala
18 Vuyyuru
Depot
Stop
No description
Arrive
time
-->Break
09:00AM
10:58AM
-->Break
02:00PM
05:57PM
-->Break
08:47PM
Stop
Drive Distance Time
Depart
time to stop to stop wind
Day
time Day Min
Min
Miles met?
30 minutes
1 10:00AM
1
60
31.5
21 YES
1 11:58AM
1
60
58.5
39 YES
30 minutes
1 04:00PM
1 120
91.5
61 YES
1 06:57PM
1
60
117.0
78 YES
60 minutes
1 ------- -- --49.5
33
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
63
14 Tadikonda
60
19 Tenali
140
8 Chirala
98
18 Vuyyuru
39
Totals Weight: Del = 337 Pickups
96.3%
0
72.00
1.2
79.1
0
33.00
.2
39.1
0
219.00
2.2
11.1
0
66.00
1.7
.0
= 0 Cube: Del = 0 Pickups = 0
Route time:
Driving
Load/unload
Break
Total
5.8 hr
5.0
2.0
12.8 hr
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Max allowed
.0%
.0
.0
.0
.0
21 mi
33
178
232 mi
9999 mi
$.00
.00
696.00
.00
$696.00
Stop
No description
15 Sattenapalle
3 Ongole
Depot
Arrive
time
-->Break
09:00AM
-->Break
01:13PM
-->Break
08:12PM
Stop
No description
Stop
Drive Distance Time
Depart
time to stop to stop wind
Day
time Day Min
Min
Miles met?
30 minutes
1 10:00AM
1
60
88.5
59 YES
30 minutes
1 03:43PM
1 150
163.5
109 YES
60 minutes
1 ------- -- --208.5
139
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
100.0%
.0%
15 Sattenapalle
45
0
87.00
1.9
87.1
.0
3 Ongole
305
0
567.00
1.9
.0
.0
Totals Weight: Del = 350 Pickups = 0 Cube: Del = 0 Pickups = 0
Route time:
Driving
Load/unload
Break
Total
7.7 hr
3.5
2.0
13.2 hr
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
59 mi
139
109
307 mi
9999 mi
$.00
.00
921.00
.00
$921.00
64
Stop
No description
20 Pamarru
22 Machilipatnam
38 Palakolu
Depot
Arrive
Depart
time to stop
time Day
time Day Min
Min
-->Break 30 minutes
09:00AM
1 10:00AM
1
60
70.5
10:36AM
1 11:36AM
1
60
36.0
-->Break 30 minutes
02:12PM
1 03:12PM
1
60
126.0
05:19PM
1 ------- -- --127.5
Stop
No description
Stop volume
Weight
Cube
47
24
YES
YES
84
85
YES
Capacity in use
Weight
Cube
100.0%
.0%
-9.00
-.1
82.3
.0
126.00
1.2
51.4
.0
285.00
1.6
.0
.0
0 Cube: Del = 0 Pickups = 0
20 Pamarru
62
22 Machilipatnam
108
38 Palakolu
180
Totals Weight: Del = 350 Pickups
0
0
0
=
Route time:
Driving
Load/unload
Break
Total
6.0 hr
3.0
1.0
10.0 hr
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
to stop wind
Miles met?
Max allowed
47 mi
85
108
240 mi
9999 mi
$.00
.00
720.00
.00
$720.00
Stop
No description
Arrive
time
-->Break
23 Kaikalur
09:00AM
39 Bhimavaram
10:54AM
-->Break
36 Tadepallegudem 01:48PM
-->Break
Depot
07:07PM
Stop
No description
Stop
Drive Distance Time
Depart
time to stop to stop wind
Day
time Day Min
Min
Miles met?
30 minutes
1 10:00AM
1
60
102.0
68 YES
1 12:24PM
1
90
54.0
36 YES
30 minutes
1 03:18PM
1
90
54.0
36 YES
60 minutes
1 ------- -- --169.5
113
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
93.1%
.0%
-12.00
-.2
79.4
.0
-39.00
-.3
37.1
.0
123.00
.9
.0
.0
0 Cube: Del = 0 Pickups = 0
23 Kaikalur
48
39 Bhimavaram
148
36 Tadepallegudem
130
Totals Weight: Del = 326 Pickups
0
0
0
=
Route time:
Driving
Load/unload
Break
Total
6.3 hr
4.0
2.0
12.3 hr
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
21.0 hr
Max allowed
68 mi
113
72
253 mi
9999 mi
65
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
$.00
.00
759.00
.00
$759.00
Stop
No description
30 Narasapur
29 Mandapeta
Depot
Arrive
time
-->Break
09:00AM
11:30AM
-->Break
-->Break
07:03PM
Stop
No description
Stop
Drive Distance Time
Depart
time to stop to stop wind
Day
time Day Min
Min
Miles met?
30 minutes
1 10:00AM
1
60
190.5
127 YES
1 01:30PM
1 120
90.0
60 YES
30 minutes
60 minutes
1 ------- -- --243.0
162
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
94.3%
.0%
30 Narasapur
160
0
75.00
.5
48.6
.0
29 Mandapeta
170
0
285.00
1.7
.0
.0
Totals Weight: Del = 330 Pickups = 0 Cube: Del = 0 Pickups = 0
Route time:
Driving
Load/unload
Break
Total
8.7 hr
3.0
2.0
13.7 hr
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
127 mi
162
60
349 mi
9999 mi
$.00
.00
1047.00
.00
$1047.00
Stop
No description
37 Eluru
41 Chintalapuidi
Depot
Stop
No description
Stop
Drive Distance Time
Arrive
Depart
time to stop to stop wind
time Day
time Day Min
Min
Miles met?
-->Break 30 minutes
09:00AM
1 11:00AM
1 120
94.5
63 YES
-->Break 30 minutes
12:41PM
1 01:11PM
1
30
70.5
47 YES
03:11PM
1 ------- -- --120.0
80
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
76.0%
.0%
37 Eluru
198
0
90.00
.5
19.4
.0
41 Chintalapuidi
68
0
192.00
2.8
.0
.0
Totals Weight: Del = 266 Pickups = 0 Cube: Del = 0 Pickups = 0
66
Route time:
Driving
Load/unload
Break
Total
4.8 hr
2.5
1.0
8.2 hr
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
63 mi
80
47
190 mi
9999 mi
$.00
.00
570.00
.00
$570.00
Stop
No description
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
84.6%
.0%
40 Jangareddygudem
68
0
-183.00
-2.7
65.1
.0
32 Kakinada
228
0
363.00
1.6
.0
.0
Totals Weight: Del = 296 Pickups = 0 Cube: Del = 0 Pickups = 0
Route time:
Driving
Load/unload
Break
Total
8.5 hr
2.5
1.0
12.0 hr
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
109 mi
200
30
339 mi
9999 mi
$.00
.00
1017.00
.00
$1017.00
67
Route
no
3
Route
Run Stop
Brk Stem
time, time, time, time, time,
Start Return No of
Route
hr
hr
hr
hr
hr
time
time stops dist,Mi
10.4
7.0
2.5
1.0
7.0 05:01AM 03:28PM
1
278
Route
cost,$
834.00
13
14
7.8
9.1
3.8
6.1
3.0
2.0
1.0
1.0
3
2
153
243
459.00
729.00
16
10.4
5.9
3.5
1.0
236
708.00
18
19
4.8
11.4
3.8
4.4
.5
5.0
.5
2.0
2
4
276
175
828.00
525.00
23
24
15.5
11.9
10.5
7.9
3.0
3.0
2.0
1.0
2
3
420
315
1260.00
945.00
29
13.4
8.0
3.5
2.0
318
954.00
VEHICLE INFORMATION
Route Veh Weight Delvry Pickup Weight
Cube Delvry Pickup
no typ capcty weight weight
util capcty
cube
cube
Cube Vehicle
util description
350
305
87.1%
9999
.0% T310
13
14
1
1
350
350
327
320
0
0
93.4%
91.4%
9999
9999
0
0
0
0
.0% T310
.0% T310
16
350
315
90.0%
9999
.0% T310
18
19
1
1
350
350
235
280
0
0
67.1%
80.0%
9999
9999
0
0
0
0
.0% T310
.0% T310
23
24
1
1
350
350
318
229
0
0
90.9%
65.4%
9999
9999
0
0
0
0
.0% T310
.0% T310
29
350
305
87.1%
9999
.0% T310
Stop
No description
3 Ongole
Depot
Stop
No description
Stop
Drive Distance Time
Arrive
Depart
time to stop to stop wind
time Day
time Day Min
Min
Miles met?
-->Break 30 minutes
09:00AM
1 11:30AM
1 150
208.5
139 YES
-->Break 30 minutes
03:28PM
1 ------- -- --208.5
139
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
87.1%
.0%
68
3 Ongole
305
0
834.00
2.7
Totals Weight: Del = 305 Pickups = 0 Cube: Del = 0 Pickups = 0
Route time:
Driving
Load/unload
Break
Total
7.0 hr
2.5
1.0
10.4 hr
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
.0
.0
139 mi
139
0
278 mi
9999 mi
$.00
.00
834.00
.00
$834.00
Stop
No description
18 Vuyyuru
22 Machilipatnam
26 Gudivada
Depot
Stop
Drive Distance Time
Arrive
Depart
time to stop to stop wind
time Day
time Day Min
Min
Miles met?
-->Break 30 minutes
09:00AM
1 10:00AM
1
60
49.5
33 YES
10:58AM
1 11:58AM
1
60
58.5
39 YES
-->Break 30 minutes
01:20PM
1 02:20PM
1
60
51.0
34 YES
03:30PM
1 ------- -- --70.5
47
Stop
No description
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
93.4%
.0%
-6.00
-.2
82.3
.0
153.00
1.4
51.4
.0
21.00
.1
.0
.0
0 Cube: Del = 0 Pickups = 0
18 Vuyyuru
39
22 Machilipatnam
108
26 Gudivada
180
Totals Weight: Del = 327 Pickups
0
0
0
=
Route time:
Driving
Load/unload
Break
Total
3.8 hr
3.0
1.0
7.8 hr
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Max allowed
33 mi
47
73
153 mi
9999 mi
$.00
.00
459.00
.00
$459.00
Stop
No description
Stop
Drive Distance Time
Arrive
Depart
time to stop to stop wind
time Day
time Day Min
Min
Miles met?
-->Break 30 minutes
69
19 Tenali
09:00AM
1 10:00AM
-->Break 30 minutes
01:33PM
1 02:33PM
04:40PM
1 -------
38 Palakolu
Depot
Stop
No description
Stop volume
Weight
Cube
60
54.0
36
YES
1
--
60
---
183.0
127.5
122
85
YES
Capacity in use
Weight
Cube
91.4%
.0%
19 Tenali
140
0
219.00
1.6
51.4
.0
38 Palakolu
180
0
513.00
2.8
.0
.0
Totals Weight: Del = 320 Pickups = 0 Cube: Del = 0 Pickups = 0
Route time:
Driving
Load/unload
Break
Total
6.1 hr
2.0
1.0
9.1 hr
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
36 mi
85
122
243 mi
9999 mi
$.00
.00
729.00
.00
$729.00
Stop
No description
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
90.0%
.0%
-63.00
-1.7
79.4
.0
-66.00
-.5
42.3
.0
93.00
.6
.0
.0
0 Cube: Del = 0 Pickups = 0
21 Nuzvid
37
36 Tadepallegudem
130
39 Bhimavaram
148
Totals Weight: Del = 315 Pickups
0
0
0
=
Route time:
Driving
Load/unload
Break
Total
5.9 hr
3.5
1.0
10.4 hr
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Max allowed
42 mi
108
86
236 mi
9999 mi
$.00
.00
708.00
.00
$708.00
70
Stop
No description
24 Jaggayyapeta
37 Eluru
Depot
Stop
Drive Distance Time
Arrive
Depart
time to stop to stop wind
time Day
time Day Min
Min
Miles met?
-->Break 30 minutes
09:00AM
1 09:30AM
1
30
115.5
77 YES
-->Break 30 minutes
13:24AM
1 3:24PM
1 120
94.5
63 YES
5:00PM
1 ------- -- --115.5
77
Stop
No description
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
10.6%
.0%
24 Jaggayyapeta
37
0
462.00
12.5
.0
.0
Totals Weight: Del = 37 Pickups = 0 Cube: Del = 0 Pickups = 0
Route time:
Driving
Load/unload
Break
Total
3.8 hr
.5
.5
4.8 hr
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
77 mi
63
136
276 mi
9999 mi
$.00
.00
828.00
.00
$828.00
Stop
No description
25 Hanuman
8 Chirala
27 Bapatia
16 Repalie
Depot
Arrive
time
-->Break
Junctio 09:00AM
11:11AM
-->Break
02:01PM
04:08PM
-->Break
07:49PM
Stop
No description
Stop volume
Weight
Cube
25 Hanuman Junctio
8 Chirala
27 Bapatia
16 Repalie
Totals Weight: Del = 280
Route time:
Driving
Load/unload
Stop
Drive Distance Time
Depart
time to stop to stop wind
Day
time Day Min
Min
Miles met?
30 minutes
1 10:00AM
1
60
3.0
2 YES
1 01:11PM
1 120
72.0
48 YES
30 minutes
1 03:01PM
1
60
19.5
13 YES
1 05:08PM
1
60
67.5
45 YES
60 minutes
1 ------- -- --100.5
67
50
98
82
50
Pickups
4.4 hr
5.0
0
0
0
0
=
Capacity in use
Weight
Cube
80.0%
.0%
-117.00
-2.3
65.7
.0
-57.00
-.6
37.7
.0
.00
.0
14.3
.0
108.00
2.2
.0
.0
0 Cube: Del = 0 Pickups = 0
Distance:
To 1st stop
From last stop
2 mi
67
71
Break
Total
2.0
11.4 hr
On route
Total
Max allowed
21.0 hr
Max allowed
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
106
175 mi
9999 mi
$.00
.00
525.00
.00
$525.00
Stop
No description
31 Amaiapuram
32 Kakinada
Depot
Arrive
time
-->Break
09:00AM
11:22AM
-->Break
-->Break
07:52PM
Stop
No description
Stop
Drive Distance Time
Depart
time to stop to stop wind
Day
time Day Min
Min
Miles met?
30 minutes
1 10:00AM
1
60
247.5
165 YES
1 01:22PM
1 120
82.5
55 YES
30 minutes
60 minutes
1 ------- -- --300.0
200
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
90.9%
.0%
31 Amaiapuram
90
0
60.00
.7
65.1
.0
32 Kakinada
228
0
270.00
1.2
.0
.0
Totals Weight: Del = 318 Pickups = 0 Cube: Del = 0 Pickups = 0
Route time:
Driving
Load/unload
Break
Total
10.5 hr
3.0
2.0
15.5 hr
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
165 mi
200
55
420 mi
9999 mi
$.00
.00
1260.00
.00
$1260.00
Stop
No description
34 Tanuku
35 Nidadvole
33 Kovvur
Depot
Stop
Drive Distance Time
Arrive
Depart
time to stop to stop wind
time Day
time Day Min
Min
Miles met?
-->Break 30 minutes
09:00AM
1 10:00AM
1
60
198.0
132 YES
10:33AM
1 11:33AM
1
60
33.0
22 YES
11:55AM
1 12:55PM
1
60
22.5
15 YES
-->Break 30 minutes
05:04PM
1 ------- -- --219.0
146
72
Stop
No description
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
65.4%
.0%
60.00
.4
27.1
.0
-9.00
-.2
12.9
.0
81.00
1.8
.0
.0
0 Cube: Del = 0 Pickups = 0
34 Tanuku
134
35 Nidadvole
50
33 Kovvur
45
Totals Weight: Del = 229 Pickups
0
0
0
=
Route time:
Driving
Load/unload
Break
Total
7.9 hr
3.0
1.0
11.9 hr
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Max allowed
132 mi
146
37
315 mi
9999 mi
$.00
.00
945.00
.00
$945.00
Stop
No description
43 Piduguralia
10 Addanki
13 Vinukonda
12 Narasarapet
Depot
Arrive
time
-->Break
09:00AM
11:57AM
-->Break
01:52PM
03:56PM
-->Break
07:49PM
Stop
No description
Stop
Drive Distance Time
Depart
time to stop to stop wind
Day
time Day Min
Min
Miles met?
30 minutes
1 10:00AM
1
60
127.5
85 YES
1 12:27PM
1
30
117.0
78 YES
30 minutes
1 02:52PM
1
60
55.5
37 YES
1 04:56PM
1
60
63.0
42 YES
60 minutes
1 ------- -- --114.0
76
Stop volume
Weight
Cube
Capacity in use
Weight
Cube
87.1%
.0%
147.00
1.8
64.3
.0
150.00
2.5
47.1
.0
54.00
.8
28.6
.0
.00
.0
.0
.0
0 Cube: Del = 0 Pickups = 0
43 Piduguralia
80
10 Addanki
60
13 Vinukonda
65
12 Narasarapet
100
Totals Weight: Del = 305 Pickups
0
0
0
0
=
Route time:
Driving
Load/unload
Break
Total
8.0 hr
3.5
2.0
13.4 hr
Distance:
To 1st stop
From last stop
On route
Total
Max allowed
21.0 hr
Route costs:
Driver (reg time)
Driver (over time)
Vehicle (mileage)
Fixed
Total
Max allowed
85 mi
76
157
318 mi
9999 mi
$.00
.00
954.00
.00
$954.00
73
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 0.01 to 1.0.
We first calculate a starting forecast by averaging the first four weekly
requirements. That is,
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
=
=
=
=
=
=
=
=
=
=
=
.1(2056)
.1(2349)
.1(1895)
.1(1514)
.1(1194)
.1(2268)
.1(2653)
.1(2039)
.1(2399)
.1(2508)
+
+
+
+
+
+
+
+
+
+
.9(1953.50)
.9(1963.75)
.9(2002.20)
.9(1991.48)
.9(1943.73)
.9(1868.76)
.9(1908.00)
.9(1982.50)
.9(1988.15)
.9(2029.24)
Forecast
Error
1953.50
=
1963.75
=
2002.28
=
1991.48
=
1943.73 -749.73
=
1868.76
399.24
=
1908.00
745.00
=
1982.50
56.50
=
1988.15
410.85
=
2029.24
478.76
=
2077.12
Total squared error
Squared
error
562,095.07
159,392.58
555,025.00
3,192.25
168,797.72
229,211.14
1,677,713.76
528.79
6
N
74
.01
.05
.1
.2
.5
1.0
SF
528.72
528.42
528.46
528.89
535.55
566.07
(c) Assuming that the forecast errors are normally distributed around F11, a 95%
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% 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
6
75
Sales, S
27,000
70,000
41,000
13,000
30,000
73,000
48,000
15,000
34,000
82,000
51,000
16,000
500,000
t
1
2
3
4
5
6
7
8
9
10
11
12
78
St
27,000
140,000
123,000
52,000
150,000
438,000
336,000
120,000
306,000
820,000
561,000
192,000
3,265,000
t2
1
4
9
16
25
36
49
64
81
100
121
144
650
Trend
value,a
St
41,087
41,192
41,298
41,403
41,508
41,613
41,719
41,824
41,929
42,035
42,140
42,245
Seasonal
indexb
0.66
1.70
0.99
0.32
0.72
1.75
1.15
0.36
0.81
1.95
1.21
0.38
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:
b
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.
(c) The forecasts using the seasonal factors from the last 4 quarters are as follows.
Seasonal
76
t
13
14
15
16
St
42,351
42,456
42,561
42,666
factors
0.81
1.95
1.21
0.38
Forecast
34,304
82,789
51,499
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.
Forecast, F13
Forecast error, SE
Region 1
219.73
26.89
Region 2
407.04
25.50
Region 3
303.30
17.54
Combined
938.26
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 ST2
might be estimated as the weighted average of the variances as follows.
ST2
F
F1 2 F2 2
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
T
Therefore,
219.73
407.04
30330
.
26.89 2
25.50 2
17.54 2
930.07
930.07
930.07
0.236 723.07 0.438 650.25 0.326 307.65
555.74
S T2
Then,
77
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.
300
250
200
150
100
50
Jly
Oct
Apr
Jan
Oct
Jly
Apr
Jan
Oct
Jly
Apr
Jan
Oct
Jly
Apr
Jan
Oct
Jly
Apr
Jan
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.
78
Prices,
Pt
211
210
214
208
276
269
265
253
244
202
221
210
215
225
230
214
276
261
250
248
229
221
209
214
5,575
Time,
t
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
300
Pt
211
420
642
832
1380
1614
1855
2024
2196
2020
2431
2520
2795
3150
3450
3424
4692
4698
4750
4960
4809
4862
4807
5136
69,678
t2
1
4
9
16
25
36
49
64
81
100
121
144
169
196
225
256
289
324
361
400
441
484
529
576
4,900
Trend,a
Tt
232.4
232.4
232.4
232.4
232.2
232.3
232.3
232.3
232.3
232.3
232.3
232.2
232.3
232.2
232.3
232.3
232.2
232.2
232.2
232.2
232.2
232.2
232.2
232.2
Seasonal
indexb
0.91
0.90
0.92
0.90
1.19
1.16
1.14
1.09
1.05
0.87
0.95
0.90
0.93
0.97
0.99
0.92
1.19
1.12
1.08
1.07
0.99
0.95
0.90
0.92
St
0.92
0.93
0.96
0.91
1.19
1.14
1.11
1.08
1.02
0.91
0.92
0.91
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.
and
a P t b t 232.29 ( 0.008)(12.5) 232.39
Therefore, the trend equation is:
79
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 month, we have:
t
25
26
27
28
29
30
31
32
33
34
35
36
a
Actual
price
210
223
204
244
274
246
237
267
212
211
188
188
Forecast Forecast
price
error
213.6
- 3.6
215.6
7.1
222.9
-18.9
211.3
32.7
276.3
- 2.3
264.6
-18.6
257.7
-20.7
250.7
16.3
236.8
-24.8
211.2
- 0.2
213.5
-25.5
211.2
-23.2
Total squared error
Squared
error
13.0
50.4
357.2
1069.3
5.3
345.9
428.5
265.7
615.0
0.0
51.0
538.2
3,739.5
Revised
seasonala
0.91
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.
80
(1)
Model type
Regression
Exp. smooth.
Total
Forecast error
19.34
17.27
36.61
(2) = (1)/36.61
Proportion of
total error
0.528
0.472
1.000
(3)=1/(2)
Inverse of error
proportion
1.894
2.119
4.013
(4)=(3)/4.013
Model weights
0.472
0.528
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)
Model type
Regression
Exp. smooth.
(2)
(3)=(1)(2)
Weighted
proportion
99.69
106.27
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% confidence band using the combined results might be constructed as:
Y = 205.96 z18.28
where z is 1.96 for 95% of the area under the normal distribution.
Y = 205.96 1.9618.28
Hence, we can be 95% 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
81
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.
30000
25000
20000
15000
10000
Oct
Jly
Apr
Jan
Oct
Jly
Apr
Jan
Oct
Jly
Apr
Jan
Oct
Jly
Apr
Jan
Oct
Jly
Apr
Jan
5000
Time, months
FIGURE 8-2 Plot of Time Series Data for Hudson Paper Company
(c) Assuming that the forecast errors are normally distributed around the forecast, a 95%
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% of the area under the normal distribution curve. Therefore,
we can be 95% 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.
82
(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%) 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.
83
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
constants
Model type
Level only....
Level-trend...
Level-seasonal
Level-trendseasonal......
TS decomp.....
.4
.2
.3
.5
.01 .2
1.0
MAD
37.82
45.85
38.97
BIAS
-5.27
7.13
11.30
RMSE
67.61
67.80
45.71
.4
30.27
59.46
-6.05
37.18
44.17
71.85
Forecast
week 6 of this
year
817.35
860.26
648.75
770.74
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
84
smoothing model seems to do this best. Both MAD and RMSE are the lowest for this
model type among the alternatives.
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
(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.
85
(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% 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% of the area under a normal distribution. FORECAST
computes the root mean squared error (RMSE) as:
N
(A
RMSE =
Ft ) 2
t =1
86
0%
50%
100%
+----+----+----+----+----+----+----+----+----+----+
|
|
|******
|
|***
|
|********
|
|********
|
|*****
|
|********
|
|******
|
|*
|
|*
|
|
|
|
|
+----+----+----+----+----+----+----+----+----+----+
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 ):
CF =
N
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
30 3
. 1054
.
4417
.
4417
46.56
Now, with z@95% = 1.96 from a normal distribution table, we can be 95% 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
87
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:
Month
Cases
1
43
2
36
9
59
10
51
11
77
12
50
88
Decision Worksheet
Month
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
Month
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
Cases
ordered
Actual
demand
Over @
$1/case
Short @
$2/case
Cost, $
Cases
ordered
Actual
demand
Over @
$1/case
Short @
$2/case
Cost, $
89
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
Demand
47 70 55 38 90 24 65 65
21
23
22
55
23
85
24
66
Period
Demand
33
45
34
70
35
50
36
56
25
53
26
64
27
61
28
63
29
65
30
38
31
80
32
88
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 3-month
initialization period and a 3-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
90
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
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).
91
TABLE 3
Period
13
14
15
16
17
18
19
20
21
22
23
24
*
Cost, $
22
6
15
31
50
48
0
2
46
7
44
0
271
Cost, $
295
290
285
280
275
270
265
0
0.2
0.4
0.6
0.8
1.2
1.4
92
100
Order
quantity
90
80
Cases
70
60
50
40
Demand
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
FIGURE 2 Plot of Forecast and Purchase Order Quantity on Product Demand
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.
93
CHAPTER 9
INVENTORY POLICY DECISIONS
1
The probability of finding all items in stock is the product of the individual probabilities.
That is,
(1)
(2)
Frequency
Order Item mix probabilities
of order
0.20
1
.95.95.95.90.90 = .69
2
0.15
.95.95.95 = .86
3
0.05
.95.95.90.90 = .73
4
0.15
.95.95.95.95.95.90.90 = .62
5
0.30
.95.95.90.90.90.90 = .59
6
0.15
.95.95.95.95.95 = .77
Order fill rate
(3)=(1)(2)
Marginal
probability
0.139
0.129
0.037
0.094
0.178
0.116
0.693
Since 69.3% < 92%, the target order fill rate is not met.
(b) The item service levels that will give an order fill rate of 92% 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% for items A, B, C, D, E, and
F, and 97% to 98% for the remaining items would be about right. The order fill rates
can be found as follows.
(1)
Order
1
2
3
4
5
6
(2)
Frequency
of order
0.20
0.15
0.05
0.15
0.30
0.15
Order fill rate
(3)=(1) (2)
Marginal
probability
0.184
0.146
0.047
0.136
0.266
0.143
0.922
94
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:
Warehouse
1
2
3
4
Total
(1)
Demand
forecast, sets
10,000
15,000
35,000
25,000
85,000
(2)
Forecast
error, sets
1,000
1,200
2,000
3,000
(3)
Values
for z
1.28
1.04
1.18
1.41
(4)=(1)+(2)(3)
Total requirements, sets
11,280
16,248
37,360
29,230
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.
Warehouse
1
2
3
4
(1)
Total
requirements
11,280
16,248
37,360
29,230
94,118
(2)
(3)=(1)(2)
(4)
(5)=(3)(4)
On hand
quantity
700
0
2,500
1,800
Net requirements
10,580
16,248
34,860
27,430
89,118
Proration of
excess
3,633
5,450
12,716
9,083
30,882
Allocation
14,213
21,698
47,576
36,513
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
95
. 92 units
ROP d LT (3,200 / 52) 15
(b) The total annual relevant cost of this design is:
TC D S / Q I C Q * / 2
. )(55)(164.78 ) / 2
(3,200)(35) / 164.78 ( 015
679.69 679.97
$1,359.66
(c) The revised reorder point quantity would be:
(b) The number of times that the course should be offered is:
. or about 4 times per year
N * D / Q * 300 / 77.5 39,
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,
CPn
100
0.667
100 50
96
Quantity
50
55
60
65
70
75
Frequency
0.10
0.20
0.20
0.30
0.15
0.05
1.00
Cumulative
frequency
0.10
0.30
0.50
0.80 Q*
0.95
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:
0.00945
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 2 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,
97
CPn
400 320
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:
Sales
500
750
1,000
1,250
1,500
Frequency
0.2
0.2
0.3
0.2
0.1
1.0
Cumulative
frequency
0.2
0.4
0.7
0.9 Q*
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,
CPn
80
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
98
Now,
+ .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:
SL 1
sd' E( z )
Q
751( 01120
.
)
1 015
.
556
814( 01120
.
)
1 016
.
556
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.
100
Step
1
2
3
4
5
6
Q
778.4
860.0
889.9
899.6
902.8
902.8
P
0.9856
0.9799
0.9778
0.9777
0.9767
0.9767
z
2.19
2.06
2.01
2.00
1.99
1.99
E(z)
0.0050
0.0072
0.0083
0.0085
0.0087
0.0087
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 4 significant digits used in the above calculations.
10
Refer to the solution of problem 10-9 for the general approach.
101
2
sd' (T * LT )sd2 d 2 sLT
So,
Max d (T * LT ) z sd'
1,250( 0.44 2.5) 0.84(1,027)
4,537 cases
SL = 79.27%
P
11
(a) The production run quantity is:
Q *p
2 DS
p
IC
pd
1,000 units
0.25(75)
300 100
102
12
(a) The order quantity is:
Q * 2 DS / IC 2( 2,000 )( 250 )(100
. ) / ( 0.30)(35) 309 valves
ROP d LT z sd LT
but sd 0 . Therefore,
ROP d LT z sd LT
where z = 1.28 from Appendix A for an area under the curve equal to 0.90.
Therefore,
103
and
ROP d LT z sd'
where
z = 1.04 (see Appendix A) for the area under the curve equal to 0.85 and
2
sd' sd2 LT d 2 sLT
152 ( 7) (50 2 )( 2 2 ) 107.6 units
Therefore,
MAX d (T * LT ) z sd'
where z = 1.04 and sd' is approximated as:
2
sd' (T * LT )( sd2 ) d 2 ( sLT
)
( 7.35 7)(152 ) 50 2 ( 2 2 )
115.0 units
104
Therefore,
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
M B* 500( 2.5 15
. ) 0.842( 70) 2.5 15
. 2,118 units
where zB = 0.842 for P = 0.80.
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
(includes on-order) and 8,256 units. The reorder size for B is the difference between
the amount on hand (includes on-order) and 2,118 units.
(b) The average amount in inventory is expected to be:
105
AIL d T * / 2 z sd T * LT
For A:
AILA 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:
SL A 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
AILA = 4,301
AILB = 1,138
SLA = 0.999
SLB =0 .991
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:
106
Cost type
Transportation
Formula
RD
In-transit
inventory
ICDT/365
Rail
Truck
6(40,000)(1.25)
11(40,000)(1.25)
= $300,000
= $550,000
0.25( 250)( 40,000)( 21) 0.25( 250)( 40,000)(7)
365
365
= $143,836
= $47,945
$597,945
$443,836
Total
Select rail.
17
The two transport options from the consolidation point are diagrammed in Figure 9-1.
Whether to choose one mode other 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 inICDt
transit inventory cost is estimated from
. The following design matrix can be
365
developed.
Cost type
Transportation
In-transit inventory
*
Method
RD
ICDt/365
Total
Air
$180,800
3,447*
$184,247
Ocean
$98,800
34,467
$133,267
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.
FIGURE 9-1 The
Consolidation
Operation for a
Hydraulic
Equipment
Manufacturer
Consolidation
point
Baltimore
20 days
2 days
Sao
Paolo
107
18
The demand pattern is definitely lumpy, since s = 327 > d = 169. To develop the minmax system of inventory control, we first find Q*. That is,
d
The ROP is
ROP d LT z sd' ED
where
z = 1.04 from Appendix A,
ED = 8 unitsthe average daily demand rate,
and
2
sd' sd2 LT d 2 sLT
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
108
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:
I T 1,000,000 9 3,000,000
In each of 3 warehouses, we would have:
I 3,000,000 / 3 $1,732,051
and in all 3 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.
Warehouse
21
24
20
13
2
11
4
1
23
9
18
12
15
14
6
7
22
8
17
16
Annual
warehouse
thruput
2,586,217
4,230,491
6,403,349
6,812,207
16,174,988
16,483,970
17,102,486
21,136,032
22,617,380
24,745,328
25,832,337
26,368,290
28,356,369
28,368,270
40,884,400
43,105,917
44,503,623
47,136,632
47,412,142
48,697,015
Average
inventory
level
504,355
796,669
1,009,402
1,241,921
2,196,364
1,991,016
2,085,246
2,217,790
3,001,390
2,641,138
3,599,421
2,719,330
4,166,288
3,473,799
5,293,539
6,542,079
2,580,183
5,722,640
5,412,573
5,449,058
Turnover
ratio
5.13
5.31 Avg. = 5.59
6.34
5.49
7.36
8.28
8.20
9.53
7.54
9.37
7.18
9.70
6.81
8.17
7.72
6.59
17.25
8.24
8.76
8.94
109
10
19
3
5
57,789,509
75,266,622
78,559,012
88,226,672
818,799,258
6,403,076
7,523,846
9,510,027
11,443,489
97,524,639
9.03
10.00
8.26
7.71
8.40
Avg. = 8.66
The overall turnover ratio is 8.40. Ranking the warehouses by throughput and
averaging turnover ratios for the top 3 and the bottom 3 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.
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 10-1. A linear regression
line is also shown fitted to the data. The equation for this line is:
Inventory = 200,168 + 0.1132Throughput
12
10
8
Estimating line
0
0
20
40
60
80
100
FIGURE 10-1 Plot of Inventory and Warehouse Thruput for California Fruit
Growers Association
110
Throughput
$21,136,032
26,368,290
22,617,380
$70,121,702
Using this total volume and reading the inventory level from Fig. 10-1 or using the
regression equation, we have:
Inventory = 200,168 + .01132(70,121,702)
= $8,137,945
(d) Warehouse 5 has a throughput of $88,226,672. Splitting this throughput by 30% and
70%, 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(3.25) 784 units
QC* 2(9,000)(10 ) / 0.25( 2.50) 537 units
111
For product A:
Q A* 2(51,000)(10) / [175
. ( 0.25 )]
For product B:
QB* 2( 25,000 )(10) / [3.25( 0.25 )]
For product C:
QC* 2(9,000)(10 ) / [2.50( 0.25 )]
3,000 C A ( QA / 2) CB ( QB / 2) CC ( QC / 2)
Expanding we have:
3,000 175
.
2(51,000 )(10) / [175
. ( 0.25 )]
3.25 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.
112
Investment in
A
1,262.44
1,240.48
1,229.92
1,221.67
1,219.63
1,129.16
0.03
0.04
0.045
0.049
0.05
0.10
B
1,204.53
1,183.58
1,173.51
1,165.63
1,163.69
1,077.36
C
633.87
622.84
617.54
613.40
612.37
566.95
Total
inventory
value, $
3,100.84
3,046.90
3,020.97
3,000.70
2,995.69
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:
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 10-22, or:
2( O S i )
T*
I Ci Di
2( 60 0 )
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
wi Truck capacity
113
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 10-30, the increment to add to I
can be found. That is,
2O
Truck capacity
Di wi
C D
i
2( 60 )
2
30,000
30,000
(988,000)
2,340,000
0.25
2( O S i )
( I ) Ci Di
2( 60 0)
( 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 5 days.
23
The average inventory for each item is given by:
Q*
AIL
z sd'
2
2 DS
. z@ 95% = 1.65 from the normal
IC
distribution in Appendix A. The results of these computations can be tabulated.
114
0.25
sd'
Q*
AIL
A
7.75
188.38
106.98
B
15.49
238.28
144.70
C
19.36
421.23
242.56
D
11.62
361.98
200.16
E
27.11
565.14
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
2 DS
IC
2(123 52)(125
. )
255.42 boxes
019
. (129
. )
Checking to see if the shelf space limit will be exceeded by this order quantity
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 low 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 = 0.012 for the last data point in the plot.
115
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.)
Terminal
1
2
3
4
5
6
7
8
9
Totals
Actual
Inventory, $
2,000
1,950
2,000
2,050
3,900
6,000
4,500
4,300
5,500
32,200
Shipments, $
150,000
195,000
200,000
200,000
320,000
330,000
390,000
410,000
500,000
2,695,000
Estimated inventory, $
I 0.012 D
1,800
2,340
2,400
2,400
3,840
3,960
4,680
4,920
6,000
32,340
Revised inventory, $
I 0.419 D 0.7
1,760
2,115
2,152
2,152
2,991
3,056
3,435
3,558
4,088
25,307
Revising the inventory control policy has the potential of reducing inventory from the
32,340 25,307
x100 21.7% .
linear policy by
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% and an = 0.7. First, determine K for an inventory
throughput curve for the item, which is
1.466
6
TO
Next,
116
tD10.7
12(117 x12) 0.3
0.90
ICK
0.20( 400)(1.466)
zs LT
1.96(15) 2
0.18
a
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.
Product A
Product B
2dS
Q
RS IC
2
2
2(3,000)( 25)
0.02(15)
RS A1
354 units
2
2(5,000)( 25)
0.02(15)
RS A2
457 units
2
RS B1
RS B 2
Product C
RS C1
RS C 2
2(8,000)( 25)
0.02(30)
408 units
2
2(9,500)( 25)
0.02(30)
445 units
2
2(12,500)( 25)
0.02( 25)
559 units
2
2(15,000)( 25)
0.02( 25)
612 units
2
Regular system inventory for two warehouses is RS2W = 354 + 457 + 408 + 445 + 559 +
612 = 2,835.
117
RS A
RS B
RS C
2(8,000)( 25)
0.02(15)
577 units
2
2(17,500)( 25)
0.02(30)
604 units
2
2( 27,500)( 25)
0.02( 25)
829 units
2
Total central warehouse regular stock is RS1W =577 + 604 + 828 = 2,009 units.
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
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
118
SS zs LT
SS A 1.65(860) .75 1,229 units
SS B 1.65( 418) .75 597 units
SS C 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.
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
Q1
2( 425 x12)(50)
270 units
0.2(35)
AIL1
Q1
270
zsd1 LT1
1.65(65) 0.5 211 units
2
2
2(333x12)(50)
239 units
0.2(35)
AIL2
Q2
239
zsd 2 LT2
1.65(52) 0.5 180 units
2
2
For customer 3
Q3
2( 276 x12)(50)
218 units
0.2(35)
AIL3
Q3
218
zs d 3 LT3
1.65( 43) 0.5 159 units
2
2
119
Total customer echelon inventory is AILC = 211 + 180 + 159 = 550 units
For the distributors echelon
QD 2,000 units as given
AILD
QD
2,000
zsd D LTD
1.28(94) 1.0 1,120 units
2
2
120
$556,912
0
4,425
4,425
0
2,529
$568,291
$27,801
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 2 days, or (2/5) = 0.40 weeks and increasing order processing costs by
$2, 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
121
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 5% 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% 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:
ROI
12,700
0.25, or 25% / year
50,000
This would appear to be attractive since carrying costs are 25% per year and the
company's return on investment probably makes up about 80% 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%. 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.
122
Item
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Base
case
99.88%
99.92
99.96
99.98
99.98
99.96
99.97
99.96
99.92
99.98
99.99
99.99
99.92
99.98
99.96
Revised
96.26%
98.02
98.54
99.15
99.45
98.60
98.84
98.61
97.29
99.26
99.70
99.43
97.30
99.14
98.84
Item
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Base
case
99.98%
99.90
99.95
99.89
99.97
99.69
99.97
99.97
99.96
99.92
99.97
99.93
99.89
99.97
99.91
Revised
99.56%
97.57
97.81
95.96
98.15
89.53
98.96
98.96
97.58
99.33
96.68
97.45
98.78
96.92
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% 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.
123
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.5 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.
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.
124
FIGURE 1 Plot of
MDC average
inventory vs. annual
throughput.
1,373,545
100 17.8%
7,713,229
in Consumer inventory levels, but only a 6% reduction in overall inventory levels. The
20% 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%? 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 =
125
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% 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
Option
1
MDC combinations
LA/Seattle
Ravenna/Htown/Chicago
Total reduction
Inventory
reduction,
lamps
1,217,481
3,423,196
4,640,677
Total
inventory
reduction
LA/Seattle
Kansas City/Hagerstown
Ravenna/Hagerstown
Total reduction
1,217,481
1,224,721
1,715,602
4,157,804
18.0%
LA/Seattle
Ravenna/Hagerstown
Atlanta/Dallas/K City
Total reduction
1,217,481
1,715,602
2,293,566
5,226,649
22.6%
20.1%
Options 1 and 3 achieve the 20% 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%. 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 replenishment shipments are
126
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
demand,
Location
lamps
Seattle
4,922,000
Los Angeles
21,470,000
Ravenna
25,853,000
Hagerstown
38,193,000
Chicago
23,321,000
Total
113,759,000
a
(4,922,000/35,000)1800 = 253,131
TL rate,
$/TL
1800
1800
250
475
350
Transport
cost, $
253,131a
1,104,171
184,664
518,334
233,210
2,293,510
Combined
annual
demand, lamps
Transport
cost, $
26,392,000
1,357,302
87,367,000
1,185,695
113,759,000
2,542,997
Cost savings
0.200.8824,640,677 = $818,615
0.104,640,677 = $464,068
Unknown, but may be included in warehouse cost
Unknowndata not given
($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.
(4) How might customer service be affected by the proposed inventory reduction?
127
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 leadtime 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
128
Shipments
26,070,000
13,244,000
87,367,000
15,950,000
26,392,000
169,023,000
Inbound
lead time,
days
2
3
1
2
5
Weighted
lead time,
days
0.308
0.235
0.517
0.094
0.781
1.935
Consolidation is assumed to take place at the MDC with the largest number of current shipments.
129
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
130
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 8 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.
We do expect that demand is price elastic, since hospitals do shop around for blood
131
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.
132
(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.
133
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
Sales forecast
Receipts
Qty on hand 0
Releases to prod.
1
150
200
50
300
2
150
300
200
3
200
1
60
2
60
100
60
3
60
2
120
100
80
100
3
100
100
80
0
300
Week
4
200
300
100
300
5
150
300
250
6
200
7
200
300
150
8
150
6
100
100
40
100
7
80
100
60
8
60
6
60
100
60
7
60
8
80
100
60
50
300
Desk style B
Sales forecast
Receipts
Qty on hand 80
Releases to prod.
20
100
0
100
Week
4
5
80
80
100 100
20
40
100 100
Desk style C
Sales forecast
Receipts
Qty on hand 200
Releases to prod.
1
100
100
100
Week
4
5
80
80
100
0
20
100 100
0
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 requirements
500 100
400 500 200
400 100
1500 300 1200 1500 600 1200 300 0
Plywood sheetsa
a
Desk requirements times 3
Now, find the purchase order releases for the plywood sheets.
Sales forecast
Receipts
Qty on hand 2400
Releases to prod.
1
2
1500 300
600
900 1200
1000 1000
3
1200
1000
1000
1000
Week
4
5
6
7
1500 600 1200 300
1000 1000 1000
500 900 700 400
1000
8
0
400
Therefore, purchase orders should be placed in weeks 1, 2, 3, and 4 for 1000 sheets
each.
134
(b) Using Equation 10-2 in the text, the probability of not having the plywood sheets at
the time needed would be
Pr 1
Pc
5
1
0.02
01
. 5
Cc Pc
Pc
150
0.9999
C c Pc (0.2 35 / 365) 150
1
2
3
4
5
6
7
1200 1200 1200 1200 1200 1200 1200
7500
500 6800 5600 4400 3200 2000 800
7500
7500
8
9
10
1200 1200 1200
7500
7100 5900 4700
135
Whse 2
Requirements
Schd receipts
On-hand qty 3300
Releases
1
2
3
2300 2300 2300
7500
1000 6200 3900
7500
Whse 3
Requirements
Schd receipts
On-hand qty 3400
Releases
1
2
3
4
2700 2700 2700 2700
7500
700 5500 2800 100
7500
7500
Regnl whse A
Requirements
Schd receipts
On-hand qty
52300
Releases to
plant
1
2
22500
3
0
4
5
6
2300 2300 2300
7500
1600 6800 4500
7500
5
6
2700 2700
7500
4900 2200
7500
4
5
0 15000
7
8
9
10
2300 2300 2300 2300
7500
2200 7400 5100 2800
7500
7
8
9
2700 2700 2700
7500
7000 4300 1600
7500
6
7
8
0 7500 15000 0
15000
10
2700
7500
6400
9
7500
15000
7300 7300
10
0
1300
1300
15000
Whse 4
Requirements
Schd receipts
On-hand qty 5700
Releases
1
2
3
4100 4100 4100
7500
1600 5000 900
7500
7500
Whse 5
Requirements
Schd receipts
On-hand qty 2300
Releases
1
2
3
4
1700 1700 1700 1700
7500
600 6400 4700 3000
7500
Whse 6
Requirements
Schd receipts
On-hand qty 1200
Releases
1
2
3
4
5
6
7
8
9
10
900 900 900 900 900 900 900 900 900 900
7500
7500
300 6900 6000 5100 4200 3300 2400 1500 600 7200
7500
7500
Regnl whse B
Requirements
Schd receipts
On-hand qty
31700
Releases to
plant
1
2
3
4
22500
0 7500
15000
Plant
Requirements
Schd receipts
On-hand qty 0
Releases-matls
5
6
7
0 15000 7500
15000
15000
2
3
0 0
0
0 0
20000 20000
7
8
4100 4100
7500
7000 2900
7500
9
10
4100 4100
7500
6300 2200
5
6
7
8
9
10
1700 1700 1700 1700 1700 1700
7500
1300 7100 5400 3700 2000 300
7500
6
4100
7500
3600
7500
4
15000
20000
5000
20000
1700
8
9
10
0 7500 7500
0
15000
9200 9200
15000
5
6
7
15000
0 30000
20000
20000
10000 10000
0
8
0
9
0
10
0
Summing the releases to the plant shows that the plant should place into production
15,000 cases in weeks 4 and 5 and 30,000 cases in week 7. Orders for materials should
be placed in weeks 1, 2, and 4 in an order size to make 20,000 units.
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,
136
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 to $5,000,000.
Sales Price L&S
OH
COG
Current (+4%) (1%)
(-3%) (-6%) (-2%)
Sales
$55.0
$57.2a $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
a
Sales - .7727xSales -8 = 5, where L&S is 0.2727 of Sales and COG is
0.5 of Sales. So, Sales = (5 + 8)/(1 0.7727) = 57.2
Due to the magnitude of cost of goods sold, it requires less than a 2 percent
change in COG to increase profits to $5,000,000.
(b) The current ROA as:
Profit margin = (4.5/55)100 = 8.2%
Investment turnover = 55/20 = 2.75
ROA = 2.758.2 = 22.6%
Or, ROA =
Profit/Assets
137
138
TABLE 10-1 A Comparison of Various Forward Buying Strategies with Hand-to-Mouth Buying
Hand-to-mouth buy
Jan
Feb
Mar
Apr
May
Jun
Jly
Aug
Sep
Oct
Nov
Dec
Price, Quantity,
$/unit
units
4.00
50,000
4.30
50,000
4.70
50,000
5.00
50,000
5.25
50,000
5.75
50,000
6.00
50,000
5.60
50,000
5.40
50,000
5.00
50,000
4.50
50,000
4.25
50,000
Subtotals
Inventory costs
Totals
Average price/unit
Total
$200,000
215,000
235,000
250,000
262,000
287,500
300,000
280,000
270,000
250,000
225,000
212,000
$2,987,500
37,350
$3,024,850
$4.98
Price,
$/unit
4.00
Quantity,
units
100,000
Total
$400,000
4.70
100,000
470,000
5.25
100,000
525,000
6.00
5.60
5.40
5.00
4.50
4.25
50,000
50,000
50,000
50,000
50,000
50,000
300,000
280,000
270,000
250,000
225,000
212,000
$2,932,500
54,900
$2,987,400
$4.88
Price,
$/unit
4.00
Quantity,
units
150,000
Total
$600,000
5.00
150,000
750,000
6.00
5.60
5.40
5.00
4.50
4.25
50,000
50,000
50,000
50,000
50,000
50,000
300,000
280,000
270,000
250,000
225,000
212,500
$2,887,500
72,150
$2,959,650
$4.81
Price,
$/unit
4.00
Quantity,
units
300,000
6.00
5.60
5.40
5.00
4.50
4.25
50,000
50,000
50,000
50,000
50,000
50,000
Total
$1,200,000
300,000
280,000
270,000
250,000
225,000
212,500
$2,737,500
119,700
$2,857,200
$4.56
139
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)
Price, No. of
Month $/lb. lb.
1
1.32
20,833
2
1.05
26,190
3
1.10
25,000
4
0.95
28,947
100,970
a
50,486/4 = 12,622
(3)=(1)(2)
Total
cost,$
27,500
27,500
27,500
27,500
$110,000
(4)=(2)/2
Average
inventory, lb.
10,417
13,095
12,500
14,474
12,622a
(3)=(1)(2)
Total
cost,$
33,000
26,250
27,500
23,750
$110,500
(4)=(2)/2
Average
inventory, lb.
12,500
12,500
12,500
12,500
12,500a
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
140
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.
141
Q
20
50
100
200
300
400
500
550
600
Price
795
795
795
795
200795+100750
300
200795+200750
400
200795+200750
+100725
500
200795+200750
+150725
550
200795+200750
+200725
600
P D
+DS/Q
+ICQ/2
= Total cost
1,113,000.00 5,250.00
1,987.50 $1,120,237.50
1,113,000.00 2,100.00
4,968.75
1,120,068.75
1,113,000.00 1,050.00
9,937.50
1,123,987.50
1,113,000.00
525.00 19,875.00
1,133,400.00
1,092,000.00
350.00 29,250.00
1,121,600.00
1,081,500.00
262.50
38,625.00
1,120,387.50
1,068,200.00
210.00
47,687.50
1,116,097.50
1,063,363.64
190.91
52,218.75
1,115,773.27
1,059,333.33
175.00
56,750.00
1,116,258.33
Destination
Cincinnati
Baltimore
Dallas
Los Angeles
Price
3.40
3.40
3.45
3.25
Transport
0.05
0.15
0.08
0.24
Volume Cost
5,000 $17,250
1,000
3,550
2,500
8,825
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.
Cincinnati
3.40
Dallas
3.44
Minneapolis
3.55
3.53
Los
Angeles
3.49
1200
3.65
Baltimore
3.46
1200
3.63
Kansas City
4800
3.52
3.45
Dayton
Requirements
Capacity
5000
5000
3.67
2500
2500
3.55
1200
1000
1000
9999
The total cost for this solution is $33,788 or a savings of $25 over the current sourcing.
142
Q*
2 DS
IC
2(120,000)( 40)
566 units
0.30(100)
Next, find the adjusted order quantity after the discount has been applied.
Q
dD
pQ *
5(120,000)
100(566)
21,648 units
(p d)I p d (100 5)(0.30) (100 5)
143
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.
TC = PD + RD +
SD
ICQ
+
+ 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:
144
P250 =
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
Q* =
2 DS
2(1500)( 25)
=
= 18.3, or 18 units
0.3(700 + 45)
IC
where C is the $700 price per unit at Baltimore plus the $45 transport cost from
Baltimore, as determined by an LTL shipment (18 units 250 lb. = 4,500 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
Quantity
price
cost
cost
18a
$700.00
$1,050,000
$67,500
50
700.00
1,050,000
67,500
100
700.00
1,050,000
67,500
693.33
1,039,995
67,500
150
692.50
1,038,750
45,000
160
200
690.00
1,035,000
45,000
686.00
1,029,000
45,000
250
300
683.33
1,024,995
45,000
400
680.00
1,020,000
45,000
a
EOQ at a price of ($700 + 45+ 0.625) per unit.
b
First price break.
c
Transport rate break.
d
Second price break.
Ordering
cost
$2,083
750
375
250
234
188
150
125
94
Carrying
cost
$2,013
5,592
11,184
16,619
17,355
21,619
26,873
32,128
42,638
Warehouse
cost
$0
0
0
0
0
0
0
0
1,000
Total cost
$1,121,596
1,123,842
1,129,059
1,124,364
1,101,339
1,101,807
1,101,023Opt.
1,102,248
1,108,732
145
Carrying
cost
$2,013
10,993
17,175
21,124
Warehouse
cost
$0
0
0
0
Total cost
$1,121596
Infeasible
Infeasible
1,098,864
1,082,409
1,071,311Opt.
146
CHAPTER 11
THE STORAGE AND HANDLING SYSTEM
All questions in this chapter require individual judgment and response. No answers are
offered.
147
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.
148
Privately-operated
Rented
Space
equirePrivate
Monthly
Monthly
Rented
Monthly
Monthly
ments,
allofixed cost variable
allostorage
handling
Monthly
sq. ft.b
cation
cost
cation
cost
cost
total cost
62,500
0%
$0
$0
100%
$30,000c
$50,000d
$80,000
50,000
0
0
0
100
24,000
40,000
64,000
37,500
0
0
0
100
18,000
30,000
48,000
25,000
0
0
0
100
12,000
20,000
32,000
12,500
0
0
0
100
6,000
10,000
16,000
3,125
0
0
0
100
1,500
2,500
4,000
15,625
0
0
0
100
7,500
12,500
20,000
28,125
0
0
0
100
13,500
22,500
36,000
37,500
0
0
0
100
18,000
30,000
48,000
43,750
0
0
0
100
21,000
35,000
56,000
50,000
0
0
0
100
24,000
40,000
64,000
56,250
0
0
0
100
27,000
45,000
72,000
21,875
$0
$0
$202,500
$337,500
$540,000
($5/lb.)
t.) = Thruput (lb.) (1/ 2 turns) (1/0.40 storage space ratio) (0.1 cu. ft./$) (1/10 ft.) (5 $/lb.)
2 and 100% of the demand through the rented warehouse, then 1,000,000 (1.00/2) 0.06 = $30,000
= $50,000
TABLE 12-2 Costs for a Mixed Warehouse Strategy Using a 10,000 Square Foot PrivatelyOperated Warehouse
Privately-operated
Rented
Space
WarerequirePrivate
Monthly
house
Monthly
Rented
Monthly
Monthly
ments,
allofixed cost variable
thruput,
allostorage
handling
Monthly
sq. ft.b
cation
Month
lb.a
cost
cation
cost
cost
total cost
84%
$25,200f
$42,000g
$80,192
Jan.
1,000,000
62,500
16%
$9,792d $3,200e
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
56,250
18
9,792
3,200
82
22,140
36,900
72,032
Dec.
900,000
Totals
6,750,000
421,875
$117,504 $36,200
$147,930
$246,550
$548,184
a
Thruput (lb.) = Sales ($)/($5/lb.)
b
Space requirements (sq. ft.) = Thruput (lb.) (1/0.40) (0.1/10) 5 = Thruput (lb.) 0.0625
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
Given a turnover ratio of 2 and 84% of the demand through the rented warehouse, then 1,000,000 (0.84/2) 0.06 = $25,200
d
1,000,000 0.84 0.05 = $42,000
150
TABLE 12-3 Costs for a Mixed Warehouse Size Strategy Using a 30,000 Square Foot PrivatelyOperated Warehouse
Privately-operated
Rented
Space
WarerequirePrivate
Monthly
house
Monthly
Rented
Monthly
Monthly
ments,
allofixed cost variable
thruput,
allostorage
handling
Monthly
sq. ft.b
cation
Month
lb.a
cost
cation
cost
cost
total cost
$29,375d $9,600e
52%
$15,600f
$26,900g
$80,575
Jan.
1,000,000
62,500
48%c
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
56,250
53
29,375
9,600
47
12,690
21,150
72,815
Dec.
900,000
Totals
6,750,000
421,875
$352,500 $94,200
$61,200
$104,170
$612,070
a
Thruput (lb.) = Sales ($)/($5/lb.)
b
Space requirements (sq. ft.) = Thruput (lb.) 1/0.40 (0.1/10) 5 = Thruput (lb.) 0.0625
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
Given a turnover ratio of 2 and 52% of the demand through the rented warehouse, then 1,000,000 (0.52/2) 0.06 = $15,600
d
1,000,000 0.52 0.05 = $26,000
151
TABLE 12-3 Costs for a Mixed Warehouse Size Strategy Using a 40,000 Square Foot PrivatelyOperated Warehouse
Privately-operated
Rented
Space
WarerequirePrivate
Monthly
house
Monthly
Rented
Monthly
Monthly
ments,
allofixed cost variable
thruput,
allostorage
handling
Monthly
sq. ft.b
cation
Month
lb.a
cost
cation
cost
cost
total cost
$39,167 $12,800e
36%
$10,800f
$18,000g
$80,767
Jan.
1,000,000
62,500
64%c
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
56,250
71
39,167
12,800
29
7,830
13,050
72,847
Dec.
900,000
Totals
6,750,000
421,875
$470,004 $115,000
$30,120
$50,200
$665,324
a
Thruput (lb.) = Sales ($)/($5/lb.)
b
Space requirements (sq. ft.) = Thruput (lb.) 1/0.40 ( 0.1/10) 5 = Thruput (lb.) 0.0625
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
Given a turnover ratio of 2 and 52% of the demand through the rented warehouse, then 1,000,000 (0.36/2) 0.06 = $10,800
d
1,000,000 0.36 0.05 = $18,000
152
FIGURE 12-1
Total Annual
Costs for a
Combined
Warehouse
Size Using
Private and
Public
Warehouse
Space
670
650
630
610
590
570
550
530
0
10,000
30,000
40,000
3
The annual cost of public warehousing is:
Handling
Storage
Total
$ 600,000
300,000
$ 900,000
153
TABLE 12-5 Ten-Year Cash Flow Stream for Public vs. Leased Warehouse Comparison
Savings
DepreSavings
less
Savings: Pre-tax ciation
less
depreAfter-tax
Lease vs. net cash schedepreTaxes
ciation Savings
net cash
Year public
flow
dule
ciation
(35%)
& tax
less tax
flow
0
$0
(3,050)a
$0
0
0
0
0
($3,050)
593
208
385
442c
442
1
650
650
57b
2
650
650
57
593
208
385
442
442
3
650
650
57
593
208
385
442
442
4
650
650
57
593
208
385
442
442
5
650
650
57
593
208
385
442
442
6
650
650
57
593
208
385
442
442
7
650
650
58
592
207
385
443
443
8
650
650
0
650
228
422
442
442
9
650
650
0
650
228
422
442
442
650
0
650
228
422
442
442
10
650
$6,500
$3,450
$400
$6,100
$2,139 $3,961
$4,361
$1,311
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
Discount
factor
1/(1+i)j
0.9009
0.8116
0.7312
0.6587
0.5935
0.5346
0.4817
0.4339
0.3909
0.3522
NPV =
Discounted
cash flow
($3,050)
398
359
323
291
262
236
213
183
165
149
($471)
154
k = $210/sq ft.
S = 100,000 sq. ft.
C = $0.01/ft.10,000 = $100/ft.
The width is:
C 8k
S
2C 8k
W*
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 Fig. 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:
m1*
1
L
dCh 2aCs 2C p K ( w a ) L
2h
2( dCh C p )
4
2( 400,000)( 0.001) 3.00
2( 4)
120.48, or 121
155
1
wa
2( dCh C p ) K ( w a ) L
2h
dCh 2aCs 2C p
50,000(8 10 )( 4)
1
2[( 400,000)( 0.001) 3.00]
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
CS
Therefore,
N = (7512,000) 3/(312,000) 8 = 9.37, or 10 doors
7
Summarizing the given information as follows:
Initial investment
Useful life
Salvage value @15%
of initial cost
Annual operating
expenses
Return on investment
before tax
Three
Five
Seven
type 1
type 2
type 3
units
units
units
$60,000
$50,000
$35,000
10 yr.
10 yr.
10 yr.
$ 9,000
$ 7,500
$ 5,250
$ 6,000
$12,500
$21,000
20%
20%
20%
An initial solution to this problem can be found through a discounted cash flow
analysis. Three alternatives are to be evaluated.
156
(1 0.2)10 1
1
9,000
PV1 60,000 6,000
10
10
(1 0.2)
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:
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 5year replacement cycle is also attractive.
157
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 used as follows.
Inbound
H, F, A
A, E
J, C, G
H, J
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:
158
Inbound
F, D
H, D
H, G, B
D, J, C
E, A
E, I, C
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)
Space
required Daily
Product
cu. ft.
orders
A
5,000a
56
B
30,000
103
C
15,000
27
D
17,000
15
E
55,000
84
F
11,000
55
G
7,000
26
H
28,000
45
I
13,000
94
J
9,000
35
a
500 sq. ft. stacked 10 ft. high
(3)=(1)/(2)
CPO
index
89
291
556
1,133
655
200
269
622
138
257
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
159
Inbound
E, F
H, C
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 is an extra challenging problem that 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
160
Solving the linear programming problem by means of any standard transportation code of
linear programming, such as LNPROG in LOGWARE, yields:
X12
X21
X22
X31
X43
X51
X53
X63
=
=
=
=
=
=
=
=
1.610
1.020
2.390
5.000
5.988
4.980
0.024
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
Product
1
2
3
% of bay
capacity
3
5,000
1,610
1,020
2,390
53.7
100.0
100.0
4,980
5,988
24
5,988
100.0
100.0
100.0
Require
-ments
11,000
4,000
12,000
161
CHAPTER 13
FACILITY LOCATION DECISIONS
1
(a) The center-of-gravity method involves finding the X,Y coordinates according to the
formulas:
V R X
X
V R
i
and
V R Y
Y
V R
i
i i
X Y
3 8
8 2
2 5
6 4
8 8
Totals
Vi
5,000
7,000
3,500
3,000
5,500
Ri
.040
.040
.095
.095
.095
ViRi
200.0
280.0
332.5
285.0
522.5
1620.0
ViRiXi
600.0
2240.0
665.0
1710.0
4180.0
9395.0
ViRiYi
1600.0
560.0
1662.5
1140.0
4180.0
9142.5
Now,
X
9,395.0
5.8
1,620.0
and
Y
9,142.5
5.64
1,620.0
This solution has a total cost for transportation of $53,614.91. This problem may also
be solved using the COG module in LOGWARE.
(b) Solving for the exact center-of-gravity method requires numerous computations. We
now use the COG module of LOGWARE to assist us. A table of partial results is
shown below.
162
Iteration
number
0
1
2
3
.
.
.
50
X coord
5.799383
5.901199
5.933341
5.941554
.
.
.
5.939314
Y coord
5.643518
5.518863
5.446919
5.402429
.
.
.
5.317043
Total
cost
53,614.91 COG
53,510.85
53,483.97
53,474.60
.
.
.
53,467.71
After 50 iterations, there is no further change in total cost. The revised coordinates
are X = 5.94 and Y = 5.32 for a total cost of $53,467.71.
(c) The center-of-gravity solution can be one that is close to optimum when there are
many points in the problem and no one point has a dominant volume, that is, has a
larger volume relative to the others. Otherwise, the best single location can be at a
dominant location. The exact center-of-gravity approach has the capability to find
the minimal cost location.
Although the COG model only considers transportation costs that are constant per
mile, the transportation cost can be the major consideration in single facility location.
However, other costs such as labor, real estate, and taxes can also be important in
selecting one location over another. These are not directly considered by the model.
Although the COG model may seem of limited capability, it is a useful tool for
locating facilities where transportation costs are dominant. Location of oil wells in
the Gulf, truck terminals, and single warehouses are examples of application. It also
can be quite useful to provide a starting solution to more complex location models.
(d) Finding multiple locations by means of the center-of-gravity approach requires
assigned supply and demand volumes to specific facilities and then solving for the
center of gravity for each. In this problem, there are 3 market combinations that need
to be considered. This creates 3 scenarios that need to be evaluated. They can be
summarized as follows. Point volumes appear in the body of the table.
Scenario
Whse
1
P1
P2
1458a 2042
3542
4958
2708
3792
2292
3208
3750
5250
M1
3500
M2
M3
I
3000
3500
5500
3000
II
5500
3500
5500
III
2
1250 1750
3000
Allocated as a proportion of the volume to be served through the
warehouse. That is, 50003500/(3500 + 3000 + 5500) = 1458. The
volumes associated with other supply points are computed similarly.
a
163
The COG module in LOGWARE was used to find the exact centers of gravity for
each warehouse in each scenario. The computational results are:
Scenario
I
II
III
Warehouse 1
X
Y
2.00 5.00
5.84 4.04
7.06 7.28
Warehouse 2
X
Y
7.88 7.80
8.00 8.00
6.00 4.00
Total
cost
$39,050
35,699
46,568
X Y
50 0
10 10
30 15
40 20
10 25
40 30
0 35
5 45
40 45
20 50
Totals
Vi
9,000
1,600
3,000
700
2,000
400
500
8,000
1,500
4,000
Ri
.75
.75
.75
.75
.75
.75
.75
.75
.75
.75
ViRi
6,750
1,200
2,250
525
1,500
300
375
6,000
1,125
3,000
23,025
ViRiXi
337,500
12,000
67,500
21,000
15,000
12,000
0
30,000
45,000
60,000
600,000
ViRiYi
0
12,000
33,750
10,500
37,500
9,000
13,125
270,000
50,625
150,000
586,500
Now,
X
600,000
.
261
23,025
and
Y
586,500
25.5
23,025
The total cost of this location is $609,765. The exact center-of-gravity coordinates
are:
X 2351
. , Y 26.98
with a total cost of $608,478.
(b) The number of points, even in this small problem, requires us to apply some
heuristics to find which patient clusters should be assigned to which warehouses. We
will use a clustering technique whereby patient clusters are grouped by proximity
until two clusters are found. The procedure works as follows.
164
There are as many clusters as there are points, which is 10 in this case.
The closest points are found and replaced with a single point with the combined
volume located at the center of gravity point. There is now one less cluster.
The next closest two points/clusters are found, and they are further combined and
located at their center of gravity.
The process continues until only two clusters remain. The centers of gravity for
these two clusters will be the desired clinic locations.
Applying the clustering technique, we start by combining points D and F into
cluster DF.
X
and
Y
Continuing this process, we would form two clusters containing A, C, D, and F, and
B, E, G, H, I, and J. The centers of gravity would be:
Cluster 1 - ACDF
X 50.00, X 0.00
Cluster 2 - BEGHIJ
X 5.00, Y 45.00
for a total cost of $241,828.
These are the same results obtained from the MULTICOG module in
LOGWARE.
(c) The second clinic can save $608,278 241,828 = $366,458 in direct costs annually.
This savings does not exceed the annual fixed costs of $500,000 required to maintain
a second clinic. On economic grounds, it should not be built.
3
(a) The center-of-gravity location can be determined by forming the following table or
by using the COG module in LOGWARE. The coordinates for each location must be
approximated.
165
Point
A
B
C
D
E
F
G
H
I
X
Y
1.5 6.6
4.7 7.3
8.0 7.1
1.5 4.0
5.0 4.9
8.5 5.1
1.5 1.3
4.4 1.8
7.8 1.8
Totals
Vi
10,000
5,000
70,000
30,000
40,000
12,000
90,000
7,000
10,000
Ri
.10
.10
.10
.10
.10
.10
.10
.10
.10
ViRi
1,000
500
7,000
3,000
4,000
1,200
9,000
700
1,000
27,400
ViRiXi
1,500
2,350
56,000
4,500
20,000
10,200
13,500
3,080
7,800
118,930
ViRiYi
6,600
3,650
49,700
12,000
19,600
6,120
11,700
1,260
1,800
112,430
118,930
4.34
27,400
and
Y
112,430
.
410
27,400
166
In fact, the cell costs are identical to those in text Fig. 13-11, except that there is no fixed
cost element.
Using the transportation method of linear programming (e.g., the TRANLP module in
LOGWARE), the cell cost and solution matrix for iteration 1 is shown in Fig. 13-1. The
solution shows that only W2 remains, and the solution process can be terminated.
A summary of the costs is shown in Table 13-1. The total cost is $2,213,714, and the
product is produced in plant P2 and stocked in warehouse W2. No further iterations are
needed since only one warehouse is used and no further dropping of warehouses is
possible.
TABLE 13-1 Summary Information for
Solution to Problem 4
Whse 1
Whse throughput
Costs:
Transportation
Inbound
Outbound
Inventory
Warehousing
Fixed
Production
Total
Whse 2
200,000
$0
0
0
0
0
0
$400,000
300,000
513,714
200,000
0
800,000
$2,213,714
Iteration 2
A repeat of the iteration 1 solution.
Plants
P1
P2
Whses
W1
W2
Capacity/
Reqmts
a
Warehouses
W1
W2
9
4
0
0
6
8
0
200,000
99a
0
60,000
99
0
799,999
60,000
999,999
Stop iterating.
Customers
C2
C3
99
99
C1
a
99
Capacity
60,000
99
99
99
999,999
9.2
8.2
10.2
60,000
6.2
50,000
5.2
100,000
6.2
50,000
999,999
50,000
100,000
50,000
FIGURE 13-1 Cell Cost and Solution Matrix for Iteration 1 of Problem 4
5
We begin by forming the cell cost matrix of a 3-dimensional transportation problem, as
shown in Figure 13-2. It is similar to the text Figure 13-1 except that the capacity for
warehouse 1 is set at 75,000. Solving the problem by means of the transportation method
shows the solution given in Figure 13-2.
167
Plants
P1
P2
Whses
W1
Warehouses
W1
W2
9
4
60,000
6
8
40,000
100,000
100
0
899,999
100
0
100
Customers
C2
C3
100
100
100
100
100
9.7
8.7
100,000
7.2
10.7
C1
Capacity
60,000
999,999
8.2
50,000
999,999
8.2
50,000
W2
100,000
Capacity/
Reqmts
999,999
100,000
50,000
100,000
50,000
FIGURE 13-2 Cell Cost and Solution Matrix for Iteration 1 of Problem 5
Given the solution from iteration 1, the per-unit inventory and fixed costs are revised.
Inventory
W1
$100(100,000) 0.7
$3.16 / unit
100,000 units
W2
$100(100,000) 0.7
$3.16 / unit
100,000 units
Fixed
W1
W2
$100,000/100,000 = $1.00/unit
$400,000/100,000 = $4.00/unit
Adding outbound transportation and warehouse handling to per-unit inventory and fixed
costs gives the following cell costs.
W1
W2
C1
10.2
10.2
C2
9.2
9.2
C3
11.2
10.2
Revising the warehouse-customer cell costs and solving gives the same warehouse
throughputs, so cell costs will no longer change. A stopping points is reached. The
solution is the same as that in Figure 13-2.
A summary of the costs is:
168
Cost type
Production
Inbound transportation
Outbound transportation
Fixed
Inventory carrying
Handling
Subtotal
Total
Warehouse 1
100,000 cwt.
60,0004 =
$240,000
40,0004 =
160,000
60,0000 =
0
40,0004 =
160,000
100,0003 =
300,000
Warehouse 2
100,000 cwt.
100,0004 =
$400,000
100,0002 =
200,000
50,0002 =
50,0002 =
100,000
100,000
100,000
400,000
100(100,000)0.7 = 316,228 100(100,000)0.7 = 316,228
100,000
100,0002 =
200,000 100,0001 =
$1,476,228
$1,616,228
$3,092,456
Compared with the costs from the text example, the cost difference is $3,092,456
2,613,714 = $478,742. This is the penalty for restricting a warehouse with economic
benefit to the network.
6
Prepare a matrix for a 3-dimensional transportation problem like that in text Figure 1311, except that the per-unit cell costs for warehouse 2 to customer are reduced by $1/unit
to reflect the reduction in that warehouses fixed costs. That is, $200,000/200,000 =
$1/unit instead of $400,000/200,000 = $2/unit. The matrix setup and first iteration
solution are shown in Figure 13-3.
Customers
Warehouses
W1
4a
Plants
C1
99b
C2
99
C3
99
P1
60,000
8
P2
Warehouses
W2
9
Plant &
warehouse
capacities
W1
W2
Warehouse
capacity &
customer
demand
0
60,000
99b
60,000
6
200,000
99
99
9.7
0
799,999
999,999c
99
99
999,999c
8.7
10.7
7.2e
50,000
6.2
100,000
7.2
50,000
50,000
100,000
50,000
60,000
999,999c
FIGURE 13-2
169
The results show that one warehouse is to be used. Further computations are not
needed, as further warehouse consolidation is not possible. The total network costs are
the same as those in the text example minus the $200,000 reduction in fixed costs for a
total coat of $2,413,714.
8
This problem requires us to rework the dynamic programming solution to the example
problem given in the text. The only change is that the cost of moving from one location
to another is now $300,000 instead of $100,000. We begin with the last year and
determine the best action based on the highest net profits. The action will be to move
(M) or to stay (S). For example, given the discounted moving cost of 300,000/(1 +
0.20)(4) = $144,676, we evaluate each course of action, assuming that we are in location
alternative A at the end of year 4. From the location profits of text Table 13-6, we
generate the following table for location A.
Alternative (x)
A
B
P5(A) = max C
D
E
Location
profit
$1,336,000
1,398,200
1,457,600
1,486,600
1,526,000
Moving
cost
-
0
144,676
144,676
144,676
144,676
Net
profit
= $1,336,000
= 1,253,524
= 1,312,924
= 1,341,924
= 1,381,324
The best action in the beginning of the 5th year, if we are already in location A, is to
move to location E. This is an entry in Table 13-2.
Once each of the five alternatives is evaluated for the 5th year, then the 4th year
alternatives are evaluated. The moving cost is 300,000/(1 + 0.2)(3) = $173,611. We now
include the profits for the subsequent years in our calculations.
After Table 13-2 is completed, we search the first column for the highest cumulative
profit. This is initially to locate in location D and remain there throughout the subsequent
years.
9
(a) Using PMED software in LOGWARE and the PMED02.DAT database, solve for the
number of locations from 1 to 9. The best locations for each number of sites are
given in the table below.
170
TABLE 13-2 Location-Relocation Strategies Over a Five-Year Planning Horizon with Cumulative Profits Shown from Year j
to Year 5 for Problem 8
Year from present date j
Warehouse
5th
1st
4th
3rd
2nd
location
StraStraStraStraaltertegya
P2(x)
tegya
P3(x)
tegya
P4(x)
tegya
P5(x)
P1(x)
natives (x)
A
$3,557,767
SA
$3,363,767
SA
$3,007,667
MD
$2,268,289
MD
$1,381,324
3,379,667
SB
3,007,667
MD
2,268,289
MD
1,398,000
B
3,556,167
SB
3,500,900
SC
3,156,200
SC
2,319,800
SC
1,457,600
C
3,673,200
SC
3,553,600
SD
3,216,000
SD
2,459,900
SD
1,486,600
D b
3,720,300
SD
3,374,700
SE
3,071,300
SE
2,355,800
SE
1,526,000
E
3,534,100
SE
a
Strategy symbol refers to staying (S) in the designated location or moving (M) to a new location as indicated.
b
Arrows indicate maximum profit location plan when warehouse is initially located at D.
Strategya
ME
SB
SC
SD
SE
171
Number
of sites
1
2
3
4
5
6
7
8
9
Total cost
$11,694,821
7,634,242
7,596,604
8,381,775
9,455,339
11,536,669
13,909,997
16,581,348
20,000,338
The optimal number of sites is 3 and they are to be located at Cincinnati, Phoenix,
and Denver.
(b) The optimal cost for four sites is $8,381,775, as found in part a. The company
operates the same sites as found in the optimal solution for four sites. Therefore, the
cost savings comes from a reassignment of customers to the sites. The savings is
$35,000,000 8,381,775 = $26,618,225 without any major investment. All of this
may not be recovered since there may be other operating costs included that are not
directly associated with location.
The savings between the optimal 4 sites and the optimal 3 sites is $8,381,775
7,596,604 = $785,171. There is also the reclamation of the salvage value of two
incinerators (close Chicago and Atlanta) and the construction of a new one
(Cincinnati). If the real estate recovery cost exceeds the new construction cost, that
would add to the savings, otherwise a ROI estimation is needed to see there is an
adequate return from the savings on the net investment.
Reallocation of customers among the existing incinerators is certainly attractive
and the reduction of the number of incinerators from 4 to 3 is also attractive as long
as there is no net investment required.
(c) Increase the annual volume for Los Angeles and Seattle markets by a factor of 10 and
re-solve the problem as in part a. Selected results are as follows:
Number
of sites
2
3
4
5
Total cost
$10,506,286
9,831,846
9,810,216
10,595,387
The optimal number of sites is four. An additional site at Seattle is needed, compared
with the three locations found in part a.
172
10
(a) We can apply Huff's model of retail gravitation to this problem. The solution table
(Table 13-3) can be developed. Summarizing, branch A can be expected to attract
11,735/(11,735 + 11,765) = 49.9% of the customers, and branch B should attract the
remaining 50.1%.
TABLE 13-3 Estimate of the Number of Customers Attracted to Each Branch
Bank
Pij
S j / Tij2
Time to ja
Customer i
1
2
3
4
5
6
7
8
9
a
A
0.28
0.10
0.28
0.40
0.20
0.50
0.45
0.57
0.67
B
0.72
0.50
0.45
0.20
0.40
0.50
0.28
0.20
0.54
Tij2
A
0.08
0.01
0.08
0.16
0.04
0.25
0.20
0.32
0.45
S j / Tij2
B
0.52
0.25
0.20
0.04
0.16
0.25
0.08
0.04
0.29
A
12.5
100.0
12.5
6.3
25.0
4.0
5.0
3.1
2.2
B
1.4
2.8
3.5
17.5
4.4
2.8
8.8
17.5
2.4
S
j
A
0.90
0.97
0.78
0.26
0.85
0.59
0.36
0.15
0.48
E ij Pij Ci
/ Tij2
B
0.10
0.03
0.22
0.74
0.15
0.41
0.64
0.85
0.52
A
900
1,940
3,120
1,820
850
885
1,440
300
480
11,735
B
100
60
880
5,180
150
625
2,560
1,700
520
11,765
Time = ( X i X )2 (Yi Y )2 / 50
173
11
This problem can be solved as an integer linear programming problem similar to the Ohio
Trust Company example in the text. First, we create a table showing the counties that are
adjacent to each county. That is,
Counties under
consideration
1. Williams
2. Fulton
3. Lucas
4. Ottawa
5. Defiance
6. Henry
7. Wood
8. Sandusky
9. Paulding
10. Putnam
11. Hancock
12. Seneca
13. Van Wert
14. Allen
15. Hardin
16. Wyandot
17. Mercer
18. Auglaize
19. Marion
20. Shelby
21. Logan
Adjacent counties
by number
2,5,6
1,3,6
2,4,6,7
3,7,8
1,6,9,10
1,2,3,5,7,10,11
3,4,6,8,10,11,12
4,7,12
5,10,13
5,6,7,9,11,13,14
6,7,10,12,14,15,16
7,8,11,16
9,10,14,17,18
10,11,13,15,18
11,14,16,18,19,21
11,12,15,19
13,18
13,14,15,17,20,21
15,16
18,21
15,18,20
Next, according to the problem formulation given in the Ohio Trust Company example,
we can build the matrix as given in the prepared database called ILP03.DAT. The
problem formulation is shown in Table 13-4. This matrix can be solved by the integerprogramming module (MIPROG) in LOGWARE for solution. Note that all coefficients
are 1s or 0s. A coefficient of 1 is given to each county and its adjacent counties. The
sum of all constraints must be 1 or greater. The Xs take on the values of 0 or 1. An X of
1 means that the branch is located in the county.
Solving this problem using the integer-programming module in LOGWARE shows
that a minimum of 5 principal places of business are needed. They should be located in
Henry, Wood, Putnam, Hardin, and Auglaize counties.
174
X1
1
X2
1
1
1
1
1
1
1
1
X3
1
1
1
1
1
1
X4
1
X5
1
X6
1
1
1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
1
1
1 1
X7
1
1
1
X8
1
X9
1
1
1
1
1
1
X10 X11 X12 X13 X14 X15 X16 X17 X18 X19 X20 X21
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
RHS
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
175
12
This problem can be solved with the aid of the PMED program in LOGWARE. A
database has been prepared for it called PMED04.DAT. The database shows the fixed
costs for a single site. When other numbers are to be evaluated, the FOC must be
recalculated and entered into the database. The scaling factor is set at 1 for this problem.
The fixed operating cost must be calculated for each possible number of locations. Using
PMED in LOGWARE gives the following results based on recalculated FOC values, an
estimation of vendor to laboratories transportation cost, and an enumerative search.
No. of
locations
1
2
Volume,
Sites
lb.
Chicago
680,000
Cleveland
515,000
L. Angeles
165,000
Total
680,000
3
New York
235,000
Chicago
280,000
L. Angeles
165,000
Total
680,000
4
New York
200,000
Atlanta
145,000
Chicago
170,000
L. Angeles
165,000
Total
680,000
5
New York
200,000
Atlanta
100,000
Chicago
170,000
Dallas
60,000
L. Angeles
150,000
Total
680,000
6
New York
200,000
Atlanta
65,000
Miami
35,000
Chicago
170,000
Dallas
60,000
L. Angeles
150,000
Total
680,000
1
515,000x310x0.02=3,193,000
2
Outbound
cost, $
28,350,000
Inbound
distance,
mi.
0
310
1,750
15,083,500
713
0
1,750
10,641,999
713
585
0
1,750
7,515,250
713
585
0
790
1,750
6,079,249
713
585
1180
0
790
1,750
5,029,250
Inbound
cost, $
0
3,193,0001
5,775,000
8,968,000
3,351,100
0
5,775,000
9,126,100
2,852,000
1,696,500
0
5,775,000
10,313,500
2,852,000
1,170,000
0
948,000
5,250,000
10,220,000
2,852,000
760,500
826,000
0
948,000
5,250,000
10,636,500
FOC, $
5,000,000
3,535,5342
3,535,534
7,071,068
2,886,751
2,886,751
2,886,751
8,660,253
2,500,000
2,500,000
2,500,000
2,500,000
10,000,000
2,236,067
2,236,067
2,236,067
2,236,067
2,236,067
11,180,335
2,041,241
2,041,241
2,041,241
2,041,241
2,041,241
2,041,241
12,247,446
Total cost,
$
33,350,000
31,122,568
28,428,352
27,828,750
27,479,584
27,913,196
5,000,000 2 / 2 3,535,534
The PMED program is used to find the best combination of sites for a particular
number of sites to be found. The fixed cost must be adjusted for the number of sites
being evaluated. It should be recognized that the model handles only the outbound leg of
the network (sites to serve laboratories). The vendor to site transportation cost is
included externally, as shown in the previous table. Calculating the distances between
vendor and the selected sites easily can be done by using the MILES module in
LOGWARE with a scaling factor of 1. Then, inbound transport costs are a product of
site volume, distance, and the inbound rate.
176
Searching from 1 to N sites shows that outbound transportation costs decrease while
inbound and fixed costs increase with increasing numbers of sites. Initially, total cost
declines until 5 sites are reached after which total cost increases. We select 5 sites as
economically the best number. Their customer assignments are
Location
number
1
2
3
4
5
Assignments
New York
Atlanta
Chicago
Dallas
Los Angeles
Volume
200,000
100,000
170,000
60,000
150,000
Customers
1, 2, 3, and 12
4 and 5
6, 7, 8, 9, 10, and 11
13, 14, and 16
15, 17, 18, 19, and 20
177
OPTIMAL SOLUTION
Variable
X(1)
=
X(2)
=
X(3)
=
X(4)
=
X(5)
=
X(6)
=
X(7)
=
X(8)
=
X(9)
=
X(10) =
X(11) =
X(12) =
X(13) =
X(14) =
X(15) =
X(16) =
X(17) =
X(18) =
X(19) =
X(20) =
X(21) =
X(22) =
X(23) =
X(24) =
X(25) =
X(26) =
X(27) =
X(28) =
X(29) =
X(30) =
X(31) =
X(32) =
Value
.0000
10000.0000
50000.0000
.0000
.0000
.0000
50000.0000
90000.0000
.0000
.0000
.0000
.0000
.0000
30000.0000
20000.0000
.0000
.0000
.0000
20000.0000
.0000
40000.0000
.0000
.0000
.0000
1.0000
.0000
1.0000
1.0000
1.0000
.0000
.0000
.0000
Rate
8.0000
7.0000
9.0000
11.0000
10.0000
11.0000
12.0000
11.0000
13.0000
8.0000
7.0000
8.0000
6.0000
5.0000
7.0000
11.0000
10.0000
11.0000
9.0000
8.0000
10.0000
7.0000
6.0000
7.0000
100000.0000
500000.0000
140000.0000
260000.0000
220000.0000
70000.0000
130000.0000
110000.0000
Cost
.0000
70000.0000
450000.0000
.0000
.0000
.0000
600000.0000
990000.0000
.0000
.0000
.0000
.0000
.0000
150000.0000
140000.0000
.0000
.0000
.0000
180000.0000
.0000
400000.0000
.0000
.0000
.0000
100000.0000
.0000
140000.0000
260000.0000
220000.0000
.0000
.0000
.0000
Variable label
P1S1W1C1
P1S1W1C2
P1S1W1C3
P1S1W2C1
P1S1W2C2
P1S1W2C3
P1S2W1C1
P1S2W1C2
P1S2W1C3
P1S2W2C1
P1S2W2C2
P1S2W2C3
P2S1W1C1
P2S1W1C2
P2S1W1C3
P2S1W2C1
P2S1W2C2
P2S1W2C3
P2S2W1C1
P2S2W1C2
P2S2W1C3
P2S2W2C1
P2S2W2C2
P2S2W2C3
zW1
zW2
yW1C1
yW1C2
yW1C3
yW2C1
yW2C2
yW2C3
3700000.00
Note that warehouse 2 is no longer used in favor of all products flowing through
warehouse 1.
14
(a) The demand of customer 1 for product 1 increases to 100,000 cwt. In the problem
matrix of ILP02.DAT the following cell values are changed.
Cell
Dem P1W1C1, yW1C1
Dem P1W2C1, yW2C1
Cap-W1, yW1C1
Cap-W2, yW2C1
Obj. coef., yW2C1
Obj. coef., yW2C1
From
-50000
-50000
70000
70000
140000
70000
To
-100000
-100000
120000
120000
240000
120000
The result shows that warehouse 2 is still the only warehouse used, and the
products are sourced from plant 2. However, the costs have increased to $3,500,000.
178
(b) Using the ILP02.DAT file in MIPROG of LOGWARE, the following changes are
made to the following cells.
Cell
From
9
8
10
7
6
7
To
12
11
13
10
9
10
The result shows that warehouse 2 is still the only warehouse used, and the products
are sourced from plant 2. However, the costs have increased to $3,380,000.
(c) Using the ILP02.DAT file in MIPROG of LOGWARE, the changes are made to the
following cells.
Cell
From
70,000a
130,000
110,000
To
280,000
520,000
440,000
The solution for product 1 shows that 50,000 cwt. flows from plant 1 through
warehouse 1 and on to customer 1. The remainder flows from plant 2 through
warehouse 2 and on to customers 2 and 3.
For product 2, plant 1 supplies warehouse 1 and customer 1 with 20,000 cwt. The
remaining 90,000 cwt. flows from plant 2 through warehouse 2 to customers 2 and 3.
The total cost is $3,920,000.
(d) Making some slight revisions in file ILP02.DAT can adjust the capacities on plant 1.
The cell changes to make are:
Cell
Cap-P1S1, RHS
Cap-P1S2, RHS
From
60,000
999,999
To
150,000
90,000
179
Product
1
1
1
1
2
2
2
Plant
1 50,000 cwt.
1 60,000
2 40,000
2 50,000
1 20,000
2 30,000
2 60,000
Warehouse
1 50,000
2 60,000
2 40,000
2 50,000
1 20,000
2 30,000
2 60,000
Customer
1 50,000 cwt.
2 60,000
2 40,000
3 50,000
1 20,000
2 30,000
3 60,000
Using the MIPROG module in LOGWARE, the following matrix can be defined.
180
Cost
1.0000
.0000
.0000
1.0000
.0000
.0000
.0000
.0000
.0000
.0000
Variable label
1
2
3
4
5
6
7
8
9
10
2.00
181
TIME/DISTANCE/COST INFORMATION
Route
no
1
2
3
4
5
6
7
8
9
Total
Route
Run Stop
Brk Stem
time, time, time, time, time,
Start Return No of
Route
hr
hr
hr
hr
hr
time
time stops dist,Mi
9.9
6.4
2.5
1.0
5.0 08:00AM 05:56PM
4
205
3.4
1.6
1.8
.0
.7 08:00AM 11:25AM
3
52
10.0
6.5
2.5
1.0
3.1 08:00AM 05:58PM
4
207
7.4
4.1
2.3
1.0
2.2 08:00AM 03:24PM
3
130
9.6
6.4
2.3
1.0
5.3 08:00AM 05:37PM
3
204
9.4
5.8
2.6
1.0
4.3 08:00AM 05:22PM
4
185
9.6
6.7
1.9
1.0
6.2 08:00AM 05:34PM
2
214
9.6
5.9
2.7
1.0
2.8 08:00AM 05:35PM
4
189
5.9
2.8
2.2
1.0
1.5 08:00AM 01:56PM
3
89
74.8 46.1 20.7
8.0 31.1
30
1476
Route
cost,$
602.45
220.25
608.59
416.05
599.64
552.96
625.79
563.48
311.65
4500.87
VEHICLE INFORMATION
Route Veh Weight Delvry Pickup Weight
Cube Delvry Pickup
no typ capcty weight weight
util capcty
cube
cube
1
1
1000
925
0 92.5%
9999
0
0
2
1
1000
625
0 62.5%
9999
0
0
3
1
1000
900
0 90.0%
9999
0
0
4
1
1000
950
0 95.0%
9999
0
0
5
1
1000
900
0 90.0%
9999
0
0
6
1
1000
950
0 95.0%
9999
0
0
7
1
1000
825
0 82.5%
9999
0
0
8
1
1000
1000
0 100.0%
9999
0
0
9
1
1000
850
0 85.0%
9999
0
0
Total
9000
7925
0 88.1% 89991
0
0
Cube
util
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
Vehicle
description
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Substituting yard location A for the current yard location and solving for the route
design in ROUTER yields Figure 13-4. No routes can be put end-to-end so that one
truck can be used instead of two, so the minimum number of trucks remains at nine. The
total daily cost for this location is 3872.02 +9 x 200 + 480 = P6,152.092.
Yard Location A
182
Route
no
1
2
3
4
5
6
7
8
9
Total
Route
Run Stop
Brk Stem
time, time, time, time, time,
Start Return No of
Route
hr
hr
hr
hr
hr
time
time stops dist,Mi
9.8
7.0
1.8
1.0
5.4 08:00AM 05:47PM
4
224
8.4
5.1
2.3
1.0
3.8 08:00AM 04:26PM
3
163
8.4
5.0
2.4
1.0
3.9 08:00AM 04:26PM
3
161
8.6
4.8
2.9
1.0
2.2 08:00AM 04:37PM
5
152
7.2
4.0
2.2
1.0
2.6 08:00AM 03:12PM
3
128
6.3
2.9
2.4
1.0
1.6 08:00AM 02:18PM
3
93
5.3
2.1
2.2
1.0
1.6 08:00AM 01:15PM
3
66
8.9
5.7
2.1
1.0
3.7 08:00AM 04:51PM
3
183
5.1
1.7
2.3
1.0
.6 08:00AM 01:03PM
3
55
68.0 38.3 20.7
9.0 25.5
30
1225
Route
cost,$
649.88
498.40
491.57
470.04
410.15
321.27
254.22
548.68
227.82
3872.02
VEHICLE INFORMATION
Route Veh Weight Delvry Pickup Weight
Cube Delvry Pickup
no typ capcty weight weight
util capcty
cube
cube
1
1
1000
475
0 47.5%
9999
0
0
2
1
1000
950
0 95.0%
9999
0
0
3
1
1000
1000
0 100.0%
9999
0
0
4
1
1000
975
0 97.5%
9999
0
0
5
1
1000
875
0 87.5%
9999
0
0
6
1
1000
1000
0 100.0%
9999
0
0
7
1
1000
875
0 87.5%
9999
0
0
8
1
1000
825
0 82.5%
9999
0
0
9
1
1000
950
0 95.0%
9999
0
0
Total
9000
7925
0 88.1% 89991
0
0
Cube
util
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
Vehicle
description
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Continuing with location B, nine trucks are required and the total daily cost for the
route design in Figure 3 is 3370.42 + 9 x 200 + 450= P5,620.42.
Yard Location B
183
TIME/DISTANCE/COST INFORMATION
Route
no
1
2
3
4
5
6
7
8
9
Total
Route
Run Stop
Brk Stem
time, time, time, time, time,
Start Return No of
Route
hr
hr
hr
hr
hr
time
time stops dist,Mi
7.1
3.7
2.4
1.0
2.3 08:00AM 03:06PM
4
120
6.5
3.1
2.3
1.0
1.8 08:00AM 02:27PM
3
100
7.9
4.6
2.4
1.0
3.0 08:00AM 03:56PM
4
146
5.9
2.8
2.2
1.0
1.4 08:00AM 01:56PM
3
89
5.9
2.6
2.3
1.0
1.1 08:00AM 01:55PM
3
84
6.7
3.6
2.1
1.0
2.3 08:00AM 02:41PM
3
114
5.8
2.5
2.3
1.0
2.1 08:00AM 01:50PM
3
80
6.1
3.3
1.9
1.0
2.8 08:00AM 02:07PM
2
104
9.7
5.8
2.8
1.0
3.0 08:00AM 05:40PM
5
187
61.7 32.0 20.7
9.0 19.8
30
1024
Route
cost,$
389.30
340.53
454.97
311.84
300.41
375.39
290.22
350.46
557.31
3370.42
VEHICLE INFORMATION
Route Veh Weight Delvry Pickup Weight
Cube Delvry Pickup
no typ capcty weight weight
util capcty
cube
cube
1
1
1000
825
0 82.5%
9999
0
0
2
1
1000
950
0 95.0%
9999
0
0
3
1
1000
825
0 82.5%
9999
0
0
4
1
1000
850
0 85.0%
9999
0
0
5
1
1000
925
0 92.5%
9999
0
0
6
1
1000
825
0 82.5%
9999
0
0
7
1
1000
950
0 95.0%
9999
0
0
8
1
1000
825
0 82.5%
9999
0
0
9
1
1000
950
0 95.0%
9999
0
0
Total
9000
7925
0 88.1% 89991
0
0
Cube
util
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
Vehicle
description
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Finally, the route design from location C is shown in Figure 4. Although 10 routes
are in the design, two of these can be dovetailed so that only nine trucks are needed. The
total daily cost is 4479.99 + 9 x 200 + 420 = P6,699.99.
Yard Location C
184
TIME/DISTANCE/COST INFORMATION
Route
no
1
2
3
4
5
6
7
8
9
10
Total
Route
Run Stop
Brk Stem
time, time, time, time, time,
Start Return No of
Route
hr
hr
hr
hr
hr
time
time stops dist,Mi
1.5
.6
.9
.0
.6 08:00AM 09:30AM
1
20
6.7
3.7
2.0
1.0
1.9 08:00AM 02:41PM
2
120
8.0
4.1
2.9
1.0
1.5 08:00AM 04:00PM
5
132
9.4
5.6
2.8
1.0
3.4 08:00AM 05:25PM
5
180
7.5
4.2
2.3
1.0
2.7 08:00AM 03:29PM
3
134
7.7
4.2
2.4
1.0
1.9 08:00AM 03:39PM
3
136
7.0
3.7
2.3
1.0
2.8 08:00AM 02:59PM
3
117
7.3
4.0
2.3
1.0
3.5 08:00AM 03:18PM
3
127
9.6
7.2
1.4
1.0
5.6 08:00AM 05:35PM
2
229
9.8
7.4
1.4
1.0
6.3 08:00AM 05:47PM
3
236
74.5 44.7 20.7
9.0 30.3
30
1432
Route
cost,$
140.91
389.21
420.20
540.48
425.69
429.10
383.04
408.53
663.40
679.42
4479.99
VEHICLE INFORMATION
Route Veh Weight Delvry Pickup Weight
Cube Delvry Pickup
no typ capcty weight weight
util capcty
cube
cube
1
1
1000
375
0 37.5%
9999
0
0
2
1
1000
875
0 87.5%
9999
0
0
3
1
1000
975
0 97.5%
9999
0
0
4
1
1000
925
0 92.5%
9999
0
0
5
1
1000
925
0 92.5%
9999
0
0
6
1
1000
1000
0 100.0%
9999
0
0
7
1
1000
950
0 95.0%
9999
0
0
8
1
1000
950
0 95.0%
9999
0
0
9
1
1000
550
0 55.0%
9999
0
0
10
1
1000
400
0 40.0%
9999
0
0
Total
10000
7925
0 79.3% 99990
0
0
Cube
util
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
.0%
Vehicle
description
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
Truck #1
From an economic analysis, it appears that yard location B is the best choice.
185
When a single warehouse is to be located, the primary location costs are transportation,
both inbound to the warehouse and outbound from it, and the warehouse lease, which
varies with the location. The current location serves as a benchmark against which the
costs for other locations can be compared. That is, based on information given in the
case, the total relevant cost for the Kansas City location is:
Inbound transportation
Outbound transportation
Lease $2.75/sq. ft. 200,000 sq. ft.
Total relevant cost
2,162,535
4,819,569
550,000
7,532,104
Using the COG module with inbound transport rates from Phoenix set at $16.73/1163 =
$0.014/cwt./mile and from Monterrey set at $9.40/1188 = $0.008/cwt./mile, and the
outbound transport rate from the unknown warehouse location set at $0.0235/cwt./mile,
the coordinates for the best location are X = 7.61 and Y = 4.51, or approximately
Oklahoma City. Total transportation cost for this location would be $6,754,082. The
total relevant cost would be:
Transportation
Lease $3.25/sq. ft. 200,000 sq. ft.
Total
$ 6,754,082
650,000
$ 7,404,082
186
A new benchmark for the 5th year can be computed from the data given in Tables 1
and 2. After adjusting the plant and market volumes according to the changes indicated,
the 5th-year benchmark costs can be computed as follows.
Volume,
Rate,
Transport
Point
cwt.
1
64,575
2
108,540
3
17,850
4
33,600
5
13,125
6
8,550
7
26,550
8
18,900
9
37,170
10
7,740
11
9,630
Totals 346,230
$/cwt.
$16.73
9.40
33.69
30.43
25.75
18.32
25.24
19.66
26.52
26.17
27.98
cost, $
$1,080,340
1,020,276
601,367
1,022,448
337,969
156,636
670,122
371,574
985,748
202,556
269,447
$6,718,483
$ 6,718,483
550,000
$ 7,268,483
Optimizing the location with the 5th-year data gives a location at X = 7.05 and Y =
4.52. The relevant costs for this location are
Transportation
Lease $3.25/sq. ft. 200,000 sq. ft.
Total
$ 6,464,206
650,000
$ 7,114,206
$141,150
100 471%
.
$300,000
Management must now judge whether 47.1% annual return is worth the risk of changing
warehouse locations.
(3) If by year 5 increases are expected of 25% in warehouse outbound transport rates
and 15% in warehouse inbound rates, would your decision change about the warehouse
location?
187
It is assumed here that the 5th-year demand level applies. A revised 5th-year benchmark
can be recomputed by applying the cost growth factors to the overall 5th-year transport
costs. That is,
Inbound 1.15 2,100,616 =
$2,415,708
Outbound 1.254,617,867 =
Subtotal
$8,188,042
Total
$8,738,042
Lease
5,772,334
550,000
Running COG shows that the minimum transport cost location would be at coordinates
X = 7.20 and Y = 4.62, which is near the previous location in question 2. The shift in
location is minimal. The cost for this location is:
Transportation
$
7,939,545
Lease $3.25/sq. ft. 200,000 sq. ft.
650,000
Total
$ 8,589,545
The annualized cost savings would be $8,738,042 8,589,545 = $148,497.
It can be concluded that:
1. Location is similar to the optimized 5th-year location.
2. The increase in possible cost savings further encourages relocation from Kansas City
and toward a site near Oklahoma City, OK.
(4) If the center-of-gravity method is used to analyze the data, what are its benefits and
limitations for locating a warehouse?
The center-of-gravity method locates a facility based on transportation costs alone. This
is reasonable when only one facility is being located and the general location for it is
being sought. Such costs as inventory carrying, production, and warehouse fixed are not
included, but they are not particularly relevant to the problem. However, costs such as
warehouse storage and handling, and other costs that vary by the particular site are not
included but may be relevant in a given situation. Transportation costs are assumed
linear with distance. This may not be strictly true, although distance may be nonlinear.
The obvious benefits of the method are (1) it is a fast solution methodology; (2) it
considers all possible locations (continuous); (3) it is simple to use; (4) its data are
readily available; and (5) it gives precise locations through a coordinate system. Some
potential limitations are (1) coordinates need to be linear; (2) transportation rates on a
per-mile basis are constant; (3) volumes are known and constant for given demand and
source points; and (4) locations may be suggested that are not feasible such as in lakes,
central cities, or restricted lands.
Concluding Comments
188
The analysis in the case seems to suggest a move from Kansas City to a region around
Oklahoma City would be advantageous. A return on investment of 47 percent or higher
is possible, however management must now seek a particular site in the area whose
choice may add or detract from this savings potential. In any case, the COG method has
assisted in the selection of good potential locations and testing their sensitivity to changes
in costs and volumes.
APPENDIX 1
Rate
0.0140
0.0080
0.0235
0.0235
0.0235
0.0235
0.0235
0.0235
0.0235
0.0235
0.0235
189
The nature of the costs and the number of possible alternative network designs make it
impractical to seek an optimal solution. Therefore, a possible approach to the analysis is
outlined as follows.
First, establish a benchmark against which changes to the network can be compared.
Much of the data for this is given in the case write up. The costs can simply be applied to
the size of each bureau and its associated staff. The cost for residents traveling to the
bureaus is not known because the bureau territories are not known. However, an estimate
can be made of travel costs by solving the problem in MULTICOG for eight bureaus.
Since MULTICOG attempts to optimize bureau location, this travel cost is probably
understated. The benchmark costs are summarized in Table 1. The location costs for the
current operation are estimated to be $1,355,706.
Second, what improvements can be made on the existing eight locations? Besides
moving the locations of the bureaus, which results in resizing the facilities and adjusting
the staff numbers, there are no obvious improvements to be made. Therefore, the
benchmark remains the base for comparison.
Third, it is now necessary to estimate the approximate number of bureaus that are
needed to serve the area. Since the costs for a particular network design depend on the
size of each bureau, which cannot be known until the problem is solved, an initial
assumption must be made. It will be assumed that all bureaus are of the same size.
Hence, for 5 bureaus, the average number of residents in each bureaus territory would be
the total number of residents divided by the number of bureaus, or 691,700/5 = 138,340.
Rent, staff salaries, and utility expenses can be derived from this estimate. Table 2 is
developed to show the bureau size and the number of staff for 1 to 10 bureaus. Table 3
extends the average costs from these estimates. A reduced number of bureaus, in the
range of two, is about right.
190
TABLE 1
Bureau
1
2
3
4
5
6
7
8
Totals
191
No. of
bureaus
1
2
3
4
5
6
7
8
9
10
Rent,
$
99,000
132,000
165,000
176,000
220,000
264,000
231,000
264,000
297,000
330,000
Staff salaries,
$
210,000
294,000
378,000
420,000
525,000
630,000
588,000
672,000
756,000
840,000
Utilities,
$
18,000
24,000
30,000
32,000
40,000
48,000
42,000
48,000
54,000
60,000
Resident
travel, $
662,319
430,922
354,239
298,000
278,181
249,287
237,635
220,106
206,496
198,600
Annual total
cost, $
989,319
880,922
927,239
926,965
1,063,181
1,191,287
1,098,636
1,204,106
1,313,496
1,428,600
The cost estimates can now be refined around two bureaus. A comparison with the
benchmark costs and a return of the initial investment (costs related to changing the
network design) are sought. A sample analysis for two bureaus, based on a design
provided by MULTICOG, is shown below.
Bureau
1
2
Totals
Residents
290,200
401,500
691,700
Size,
sq. ft.
2,500
3,500
6,000
Staff
6
8
14
Rent,
$
55,000
77,000
132,000
Staff
salaries, Utilities,
$
$
126,000
10,000
168,000
14,000
294,000
24,000
Resident
travel cost,
$
166,332
264,590
430,922
The total annual variable cost is $132,000 + 294,000 + 24,000 + 430,922 = $880,922.
There are one-time costs due to staff separation and equipment moves. Compared
with the benchmark, 36 14 = 22 staff members will be separated for a cost of 22
$8,000 = $176,000. Equipment movement costs to two bureaus would be 2 10,000 =
$20,000. Total movement costs would be $176,000 + 20,000 = $196,000.
Annual variable cost savings compared with the benchmark would be $1,355,706
880,922 = $474,784. A simple return on investment would be:
ROI
$474,784
100 242%
$196,000
Similar calculations are carried out for various numbers of bureaus. These results are
tabulated in Table 4.
192
TABLE 4 Cost Savings and Return of Investment for Alternate Network Designs
as Produced by MULTICOG
Total
Annual
Return on
No. of
size, sq.
Total
variable cost,
Moving
investSavings,
bureaus
ft.
staff
$
cost, $
ment, %
$
1
4,500
10
989,319
218,000
366,387
168
2
6,000
14
880,922
196,000
242
474,784
3
7,500
18
927,239
176,000
428,467
243
4
8,500
21
960,965
160,000
394,741
246
5
9,500
24
1,029,181
146,000
326,525
223
6
11,500
29
1,157,287
106,000
198,419
187
7
12,000
31
1,200,635
110,000
155,071
141
8
13,000
34
1,272,106
96,000
83,600
87
Benchmark 14,600
36
1,355,706
---------
The maximal annual savings occurs with a network containing two bureaus.
However, the maximal return on investment occurs with four bureaus. ROI is selected as
the appropriate measure on which to base this economic decision. The details for a
design with four bureaus are given in Table 5. The design is shown pictorially in Figure
1 of this note.
TABLE 5 Design Details for a Network with Four Bureaus
Column
grid
Row grid
coordicoordiBureau
nate
nate
Residents Grid box number assignment
1
3.36
2.74
218,200
1,2,3,4,5,6,7,8,9,10,11,12,13,
14,15,16,17,18,19,20,21,22,23,
24,25,26,27,28,29,30,31,32,33
2
7.00
3.00
168,700
34,35,36,37,38,39,40,41,43,44,
45,46,47,50,51,52,53
3
9.74
5.66
195,000
42,48,49,54,55,56,60,61,62,63,
67,68,69,70,74,75,76,77,81,82, 83,84,
4
10.00
2.00
109,800
57,58,59,64,65,66,71,72,73,78, 79,80
193
The effect of bureau location on the resident's perception of service is not as well
known as portrayed in the case. In addition, service may need to be represented by
more than location.
Travel to the bureaus is assumed straight line. However, location in the area is
likely to be influenced by a road network. Time may be more important than
distance to residents.
The fixed costs associated with location are not handled directly by the center-ofgravity approach.
Residents are assumed to travel to the locations within their assigned territories.
They may not strictly do this.
Good facilities may not be available at the indicated location coordinates.
The analysis is particularly weak around the estimate of the resident travel cost.
While an exact cost is not likely to be known, Dan should conduct a sensitivity analysis
around this cost. He may find that the design does not change a great deal over a wide
range of assumed values. If this is the case, he can feel comfortable that his
recommendation is fundamentally sound. If not, he should seek to find a more precise
value.
(3) What concerns besides economic ones should Dan have before suggesting that any
changes be made to the network?
A quantitative approach to location will rarely give the precise locations to be
implemented. Rather, it provides a starting point for further analysis. There are a
number of other factors to be considered before the revised network design can be
implemented. First, there are site selection factors to be taken into account such as the
availability of adequate space near the location coordinates, proximity to good highway
linkages, and reasonable neighborhood reactions to this type of operation.
Second, there are political concerns. Reducing the number of locations will result in
a releasing some of the staff. Dan may experience some political resistance to this.
Since staff is a large expense in the operation, currently about 2/3 of the costs, retention
of a larger number of bureaus may be required. Of course, transferring staff to other
governmental operations may be a way of dealing with this issue. This assumes their
willingness to be relocated, although this is not likely to be a strong issue if relocation
were to occur in the same area.
Third, there may be difficulty in demonstrating the economics of network redesign.
Although others may appreciate the costs of rent, salaries, and utilities, the cost of
resident travel is subject to much interpretation. Those favoring many bureaus may argue
the high cost while those wanting to reduce the number of bureaus may perceive it as not
very significant.
194
SUPPLEMENT Sample Input Data File for MULTICOG in LOGWARE for the
Ohio Auto & Driver's License Bureau Case Study
Title: LICENSE BUREAU
Number of sources: 4
Number of demand points: 84
Scaling factor: 2.5
POINT
X-COORDINATE
Y-COORDINATE
1
1
1
2
1
2
3
1
3
4
1
4
5
1
5
6
1
6
7
1
7
8
2
1
9
2
2
10
2
3
11
2
4
12
2
5
13
2
6
14
2
7
15
3
1
16
3
2
17
3
3
18
3
4
19
3
5
20
3
6
VOLUME
4100
6200
7200
10300
200
0
0
7800
8700
9400
11800
100
0
0
8100
10500
15600
10500
200
0
RATE
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
195
21
.12
SUPPLEMENT (Continued)
POINT
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
X-COORDINATE
4
4
4
4
4
4
4
5
5
5
5
5
5
5
6
6
6
6
6
6
6
7
7
7
7
7
7
7
8
8
8
8
8
8
8
9
9
9
9
9
9
9
10
10
10
10
10
10
10
11
11
11
11
11
11
11
12
Y-COORDINATE
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
VOLUME
10700
12800
13800
15600
400
0
0
11500
13900
14500
13700
600
0
0
9300
14900
13700
10200
1200
0
0
10100
12600
16700
15800
12400
2600
0
8800
13700
15200
14100
10800
17200
500
5300
16700
13800
11900
13500
18600
12000
5100
17400
10300
9800
10300
15500
11700
7700
9200
7500
8500
7800
9900
8700
4300
RATE
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
.12
196
79
80
12
12
2
3
6700
5800
.12
.12
Y-COORDINATE
4
5
6
7
VOLUME
6800
5400
7100
6400
RATE
.12
.12
.12
.12
SUPPLEMENT (Continued)
POINT
81
82
83
84
X-COORDINATE
12
12
12
12
197
SOUTHERN BREWERY
Teaching Note
Strategy
The purpose of this case study is to provide students with the opportunity to design a
distribution network where plant location is at issue. They first should identify the major
costs and alternatives that are important to such a design problem. Second, they should
be encouraged to apply the transportation method of linear programming to assist in the
analysis of alternatives using the TRANLP module in LOGWARE. Finally, they should
consider factors other than those in the analysis that might alter the course of their
recommendation and be sensitive to the limitations and benefits of linear programming as
a solution methodology.
Answers to Questions
(1) If you were Carolyn Carter, would you agree with the proposal to build the new
brewery? If you do, what plan for distribution would you suggest?
If growth is uniform over the next five years, Southern can expect that demand for its
products will exceed the currently available plant capacity. That is, demand is increasing
at the rate of (0 + (595,000 403,000)) 5 = 38,400 barrels per year. Thus, the current
annual capacity of 500,000 barrels will be used up in (500,000 403,000) 38,400 = 2.5
years. A major concern is whether it would be profitable to construct the new plant.
Rough estimates of its profitability can be made, projecting profits with and without the
new plant. We know that 595,000 500,000 = 95,000 barrels of beer would not be sold
annually in the 5th year if additional capacity is not constructed. This represents a
potential average lost revenue of $280/barrel 0.20 [(0 + 95,000) 2.5] = $2,128,000.
(The figure of 2.5 years assumes that the new brewery can be brought on stream at
approximately the time when capacity will be used up in the existing plants.) From the
benchmark1 costs for the current system, as shown in Table 1, Southern is currently
producing and distributing 403,000 barrels at a total cost of $60,015,000. This is an
average cost per barrel of $60,015,000 403,000 = $148.92. The overhead and sales
expense is 27 percent, or $280 0.27 = $75.60 per barrel. Total costs per barrel are
$148.92 + 75.60 = $224.52, which is about 80 percent of the sales dollar. The 20 percent
profit margin seems valid. Therefore, the benefit of serving the potentially lost demand
with a new brewery can be estimated using on a simple return on investment:
ROI
$2,128,000
0.21, or 21%
$10,000,000
If management feels that this is an adequate return for such a project, Carolyn should
proceed with her analysis. Let's assume that she has this approval.
Next, she may wish to explore the opportunities available by improving upon the
existing distribution system without the presence of the new plant. This is an improved
1
A benchmark refers to the costs of producing and distributing demand as currently allocated throughout
the network.
198
RA-
KNOX-
COL-
AT-
SAVAN-
MONT-
TALL
JACK-
From\To
MOND
LEIGH
VILLE
UMBIA
LANTA
NAH
GOMRY
AHAS
SONVL
RICHMD
148.49
150.70
156.38
152.54
155.48
154.64
159.98
164.30
158.84
100
COLMBA
157.54
154.78
157.81
151.96
156.85
154.54
157.93
160.18
157.27
100
MONTGM
156.98
153.35
150.80
149.93
147.20
150.80
143.69
150.65
152.18
300
JACKVL
152.13
149.25
150.48
146.16
148.80
144.54
148.80
144.72
142.68
Demand
56
31
22
44
94
13
79
26
38
Supply
An improved benchmark refers to a reallocation of current demand in an optimal way, respecting plant
capacity restrictions.
199
Adding a plant at Jacksonville with a capacity of 100,000 barrels per year provides
enough capacity to satisfy demand out to the 5th year. If the new plant were constructed
and producing immediately, total costs could be reduced from the improved benchmark
by $59,804,000 59,090,000 = $714,000 per year. (Compare the total costs in Tables 3
and 4.) The Columbia plant would not be needed if the lower-cost Jacksonville plant
were on line.
We do not know the savings for the 5th-year demand level since not all demand can
be served without the presence of the new plant. Therefore, a future-year benchmark
cannot be determined. However, we do know how the new plant should be utilized
within the system (see Table 5) and how demand allocation should be adjusted to
accommodate it. Also, note that the Columbia plant is needed once again although its
capacity is not required until the last one-half year of the 5-year planning horizon. This
suggests that the Columbia plant should not be sold, but perhaps some alternate use could
be made of the facility in the interim, such as subcontracting beer production to a noncompeting company.
The new plant is not likely to be brought on stream immediately nor is it needed for
2.5 years, so Carolyn might suggest a distribution plan similar to that in Table 5. A
careful inspection of this plan shows that only 1 barrel of demand in the Knoxville region
is assigned to Richmond. Splitting demand to this extent is probably not practical and
can be assigned to Columbia where there is excess capacity. Costs will rise only slightly.
An interesting question is whether the Columbia plant, through modernization, could
be made as efficient as the new brewery, and what the implications for distribution might
be. We know that this could potentially save $145 135 = $10 per barrel in production
costs. At a 100,000-barrel capacity, this is $1,000,000 in cost savings. If the
modernization were to cost no more than $5,000,000, this option might be attractive. Of
course, we would need to resolve the linear programming problem with Columbia's per
barrel costs at $135 plus transportation costs. This would tell us how and to what extent
demand would be allocated to Columbia and give a more accurate basis for determining
200
the cost savings. Similarly, it would be interesting to explore what it means to expand
the capacity of an existing brewery at a lower investment cost per barrel than the
construction of a new facility.
TABLE 4 Production and Transportation Costs ($000s) for Current Demand with
the Jacksonville Plant
Demand
TransBrewery of
in 000s
Producport
Market area
origin
barrels tion costs
costs
Total costs
1 Richmond
Richmond
56
$7,840
$475
$8,315
2 Raleigh
Richmond
31
4,340
332
4,672
3 Knoxville
Montgomery
22
3,014
304
3,318
4 Columbia
Montgomery
21
2,877
272
3,149
4 Columbia
Jacksonville
23
3,105
257
3,362
5 Atlanta
Montgomery
94
12,878
959
13,837
6 Savannah
Jacksonville
13
1,755
124
1,879
7 Montgomery
Montgomery
79
10,823
550
11,373
8 Tallahassee
Jacksonville
26
3,510
253
3,763
5,130
292
5,422
9 Jacksonville
Jacksonville
38
Total
403
$55,272
$3,818
$59,090
(2) If the new brewery is not to be constructed, what distribution plan would you propose
to top management?
Table 6 shows a linear programming solution where the new plant is not brought on
stream and the demand in the markets is set at the 5-year level. An interesting solution
occurs when demand exceeds capacity. The most costly demand region to serve is not
assigned to any plant. As can be seen in Table 6, portions of the demand in Tallahassee
and Jacksonville should not be served, and essentially the entire Knoxville market should
not be served at all. Top management may wish to adjust this plan for reasons other than
economic ones.
201
TABLE 6 Production and Transportation Costs ($000s) for Projected 5thYear Demand Without the Jacksonville Plant
Expected
Served
demand in demand in
Brewery of
000s
000s
Market area
origin
barrels
barrels
Total costs
1 Richmond
Richmond
64
64
$9,503
2 Raleigh
Richmond
35
35
5,275
3 Knoxville
Richmond
33
1*
156
4 Columbia
Columbia
55
55
8,358
5 Atlanta
Montgomery
141
141
20,755
6 Savannah
Columbia
20
20
3,091
7 Montgomery
Montgomery
119
119
17,131
8 Tallahassee
Montgomery
52
40*
6,026
25*
3,931
9 Jacksonville
Columbia
76
Total
595
500
$74,226
*Indicates market demand is not fully served due to inadequate plant capacity.
(3) What additional considerations should be taken into account before reaching a final
decision?
A number of assumptions have been implied in the analysis shown above. For example,
Demand has been assumed to grow at a constant rate in the markets.
Production is assumed limited to exactly the values given without the possibility for
expansion through overtime, additional shifts, or subcontracting.
202
Per-unit production and transportation costs are assumed to remain unchanged with
the reallocation of demand throughout the network.
Customer service effects are not considered in reallocation of demand.
There is no change in per-unit costs throughout the 5-year planning horizon.
This case might end with a discussion of the appropriateness of using linear
programming as a vehicle for analysis in a problem such as this. Mentioning that linear
programming does not consider such factors as fixed costs, return on investment, or the
many subjective factors (top management's intuition about location, vested interests, etc.)
that are typically a part of such problems means that linear programming, at best, is a
facilitating vehicle for analysis. It does not provide the final answer.
203
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.
X Coordinate
Lansing, MI
Lubbock, TX
El Paso, TX
Atlanta, GA
Boston, MA
Los Angeles, CA
Seattle, WA
Portland, OR
Y Coordinate
924.3
1488.6
1696.3
624.9
374.7
2365.4
2668.8
2674.2
1675.2
2579.4
2769.3
2318.7
1326.6
2763.9
1900.8
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
204
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% 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% 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% of the
variation in the data with distance. Such a simple relationship seems to represent the
rates quite well.
205
20
18
16
14
12
10
Estimating line
8
6
4
2
0
0
500
1000
1500
2000
2500
3000
3500
Distance, miles
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
Warehouse 22 has a much higher inventory turnover ratio than the average of the
other warehouses. This would suggest that the inventory control procedures might be
different from the others. One reason might be that person in control of the inventory in
this warehouse attempts to keep inventories at low level, demand may be high such that
the inventory level has been restored to a normal level, or lead times have been extended
to the point where replenishment has been delayed. The reason should be investigated.
206
12
10
8
6
4
Estimating line
2
0
0
20
40
60
80
100
207
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 productiondistribution 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
Current demand
Net increase
1,908,606 cwt.
1,477,026
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%, the profit per cwt.
would be 20%$67.7/cwt, or $20/1.477, = $13. Thus, 278,60613 = $3.16 million in lost
sales. The weighted profit loss over the five-year period would be:
2/5 (0) + (3/5) ([0 + 3.6])/2) = $1.08m/yr.
208
Benchmark
Production
$30,762
Warehouse operations
1,578
Order processing
369
Inventory carrying
457
Transportation
Inbound
2,050
Outbound
6,896
Total costs
$42,112
Improved
benchmark
Optimum
number
of whses
Optimum
number
of whses
Relaxed
service (1)
Relaxed
service (2)
Maximum
opportunity
$30,678
1,468
354
431
$30,673
1,608
370
508
$30,675
1,572
358
490
$30,678
1,296
349
390
$30,673
1,420
354
445
$30,386
1,529
358
500
1,802
6,991
$41,725
1,976
6,310
$41,447
1,860
6,365
$41,321
1,249
7,238
$41,201
1,178
6,698
$41,043
1,178
6,458
$40,409
Customer
service:
300 mi.
600 mi.
93%
98%
93%
98%
98%
100%
92%
100%
75%
98%
88%
100%
81%
94%
No. of stocking
points
22
21
31
30
19
26
40
No. of plants
Savings vs.
benchmark
$0
$387
$665
$791
$911
$1,069
$1,703
$0
$278
$404
$524
$1,316
Service
to match
benchmark
600 mi constraint on
current
warehouses
600 mi constraint on
opt no. of
warehouses
682
Unlimited
service,
whses, and
plant cap.
209
Add plant
@ Memphis
Add plant
@ Memphis
& Chicago
Memphis
and opt no.
of whses
Memphis
and opt no.
of whses
$33,965
1,496
393
431
$39,517
1,842
462
505
$39,548
1,847
454
497
$39,524
2,028
470
591
$39,522
1,976
460
573
1,647
7,230
$45,164
2,350
9,030
$53,705
2,000
9,036
$53,382
2,614
8,117
$53,342
2,426
8,222
$53,179
Customer
service:
300 mi
600 mi
98%
99%
94%
98%
95%
98%
98%
100%
92%
100%
No. of stocking
points
20
21
20
31
30
No. of plants
Comments:
Not all
demand
met
Cost type
Production
Warehouse operations
Order processing
Inventory carrying
Transportation
Inbound
Outbound
Total costs
Service
at benchmark
210
Based on a simple rate of return on investment, capturing this profit potential would yield
1.08/4 = 27% annually on a $4,000,000 investment for expansion. The return would
increase to 90% 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 5th 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 5th-year savings
would only amount to about 8% (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.
211
Type of
run
Benchmark
Improved
benchmark
Improved
benchmark
Current
yr. whses
5th yr.
whses
Percent
of demand
300 mi.
Year
Plant
capacities
Total
cost
No. of
whses
Current
Current
93
$42,112
22
Current
93
41,725
21
5th yr.
Current
Current
+ Memphis
94
53,705
31
Current
Current
+ Memphis
92
41,321
30
5th
year
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 5th year. The number of
warehouses should be increased from the present 22 in both the current year and the 5th
year.
42
100
99
41.8
98
97
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 5, the savings of the added warehouse would be:
$53,705,000 53,179,000 = $526,000
212
Thruput
limits
Current
year
solution
Future
year
solution
Covington
New York
Arlington
Long Beach
450,000 cwt.
380,000
140,000
180,000
254,471 cwt.
302,043
66,592
95,943
306,478 cwt.
380,523
66,161
117,288
Customer Service
Currently, a high proportion of demand (93%) 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% of the demand is
within 300 miles and 97.5% is within 600 miles. If a 600-mile constraint is applied to the
current network configuration (run 23), 75% of the demand is within 300 miles and 98%
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
213
problem, we can only estimate the worth of the service. That is, can enough sales be
generated to cover the higher service level? If physical distribution costs for the company 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%.
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 5th 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.
214
Cost type
Production
Whse operations
Order processing
Inventory carrying
Transportation
Inbound
Outbound
Total costs ($000s)
Benchmark
Currentyear
design
Change from
benchmark
$30,762
1,578
369
457
$30,675
1,572
358
490
$ -87
- 6
-11
+33
-0.3%
-0.4
-3.0
+7.2
2,050
6,896
$42,112
1,860
6,365
$41,320
-190
-531
$-792
-9.3
-7.7
-1.9%
215
Run Run
no. description
No of
plants
Plant
capacity
Demand
level
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
4
4
4
6
4
4
4
4
4
4
4
4
4
5
5
6
4
4
4
5
4
4
4
4
5
Current
Current
Current
Crnt+1m
Current
Current
Current
Current
Current
Current
Current
Current
Current
Current
Current
Current
Current
See cmt
Current
Current
Current
Current
Current
See cmt
Current
- mi
Current 300
Current 9000
Current 9000
5th yr. 300
Current 300
Current 300
Current 300
Current 300
5th yr. 300
5th yr. 300
5th yr. 300
5th yr. 300
5th yr. 300
5th yr. 300
5th yr. 300
5th yr. 300
Current 300
Current 300
5th yr. 300
Current 600
Current 600
Current 600
Current 375
5th yr. 375
Benchmark
Improved benchmark
No serv constraint
Max opportunity
Future yr-imp bmk
Test 27 whses
Test 32 whses
Test 37 whses
Test 42 whses
Exp Covington
Exp New York
Exp Arlington
Exp Long Beach
Add Memphis
Add Chicago
Add Mem & Chi
No plant expansion
Optimum whses
Optimum whses
Optimum whses
Test cust service
Test cust service
Test cust service
Optimum whses
Optimum whses
No of
whses
Total
costs
Percent of
demand within
300 600 Comments
22
21
18
40
21
26
31
36
40
21
21
21
21
21
20
20
20
31
30
31
31
26
19
30
30
$42,112
41,725
40,896
40,409
53,777
41,744
41,615
41,501
41,486
54,145
53,986
54,709
55,251
53,705
53,781
53,382
45,164
41,447
41,563
53,342
40,996
41,043
41,201
41,321
53,179
93%
93
71
81
93
95
98
99
99
94
93
94
94
94
94
95
98
98
97
98
80
88
75
92
92
98%
98
89
94
98
100
100
100
100
98
98
98
98
98
98
98
100
100
100
100
100
100
98
100
100
216
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 through the use of 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-today 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 multiechelon 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?
217
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
Retailer
Distri
warehouse
Essen
Warewarehouse
house
Factory
Factory
-butor
Retailer
Retailer
218
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 7 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
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
channel. Current performance of the channel members is summarized in Table 1 for 4
simulation runs using different seed numbers.
219
Table 1
Channel member
Run 1
Run 2
Run 3
Run 4
Average
$73,105,904
37,918
$1,928
$72,967,088
37,846
$1,928
$72,616,192
37,664
$1,928
$72,477,376
37,592
$1,928
$72,791,640
37,755
$1,928
6.52
<50%
$5,578,291
$147.02
6.59
<50%
$5,549,202
$146.50
6.61
<50%
$5,521,236
$146.32
6.58
<50%
$5,540,447
$146.65
6.58
<50%
$5,547,294
$146.62
12.95
<50%
$3,873,236
$101.94
13.01
<50%
$3,895,406
$102.15
12.96
<50%
$3,853,890
$102.14
12.89
<50%
$3,912,699
$103.07
12.95
<50%
$3,883,808
$102.33
81.02
<50%
38,017
$2,884,527
$76.19
80.38
53.02%
37,983
$2,768,697
$72.89
80.45
54.09%
37,774
$2,939,464
$77.82
80.60
<50%
37,804
$2,981,607
$78.87
80.61
<50%
37,895
$2,893,574
$76.44
$24,426,593
22.23%
123456
$24,591,053
22.40%
444444
$24,235,498
22.20%
555555
$24,340,563
22.28%
666666
$24,398,426
22.28%
Essens factory
Total cost
Units produced
Cost per unit
Essens
warehouse
TO ratio
Fill rate
Cost
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?
There are a number of actions that generally 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.
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
220
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-todemand 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 9 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.
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
1. The average fill rate has increased for all members of the channel, but the cost effects
221
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
member of the supply channel, Essen undoubtedly suffers from the variability in demand,
which cannot entirely be controlled.
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
Channel member
Run 1
Run 2
Run 3
Run 4
Average
$68,638,738
35,950
$1,909
$74,761,258
39,286
$1,903
$78,418,188
41,268
$1,900
$75,400,666
39,622
$1,903
$74,304,713
39,032
$1,904
1.47
90.65%
$12,060,444
$317.51
1.42
100%
$12,449,170
$326.28
1.51
63.16%
$11,895,743
$311.87
1.46
72.92%
$12,178,581
$319.61
1.47
81.68%
$12,145,984
$318.82
9.24
96.64%
$4,018,143
$105.70
9.02
96.29%
$4,080,160
$107.09
9.14
97.60%
$4,043,973
$106.55
9.28
98.09%
$3,992,791
$105.51
9.17
97.16%
$4,033,767
$106.21
35.44
99.52%
38,017
$727,378
$19.13
35.36
99.53%
37,936
$717,363
$18.91
34.99
99.70%
37,979
$709,772
$18.69
35.14
99.40%
37,730
$748,198
$19.75
35.23
99.54%
37,916
$725,677
$19.12
$24,423,849
22.23%
123456
$17,627,666
16.08%
111111
$14,690,765
13.38%
222222
$17,184,464
15.69%
333333
$18,481,686
16.85%
Essens factory
Total cost
Units produced
Cost per unit
Essens
warehouse
TO ratio
Fill rate
Cost
222
Seed value
Length of simulation, years
Annual price, $/unit
Index
0.75
1.50
2.50
223
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
2220
Item value in inventory, $/unit
20
Retailer order filling cost, $/unit
75
Purchase order processing cost, $/order
25
Inventory carrying cost, %/year
2
Average retailer order fill time, days
0.2
Retailer order fill time standard deviation, days
95
In-stock probability, %
100
Back order cost, $/unit
Forecasting method
Moving average
30
Number of periods
Reorder policy
Stock-to-demand control method
45
Target days of inventory
30
Review time in days
Warehouse/Level 3
Product item data
1710
Item value in inventory, $/unit
15
Distributor order filling cost, $/unit
75
Purchase order processing cost, $/order
20
Inventory carrying cost, %/year
3
Average distributor order filling time, days
0.3
Distributor order fill time, days
95
In-stock probability, %
25
Back order cost, $/unit
Forecasting method
Moving average
360
Number of periods
Reorder policy
Stock-to-demand control method
Target days of inventory
90
Review time in days
30
224
Factory/Source
Product item data
850
10
10
8
2
1000
25
1
0
70
5
1
78
9
3
Appendix B
322,303,000
949,025
2,656,010
2,957,604
9,490,250
26,560,100
29,576,040
38,017
759,890
569,145
380,168
7,598,900
5,691,450
1,638
683
675
16,380
6,825
6,750
Production cost
Transportation costs:
Distributor to retailer
Warehouse to distributor
Factory to warehouse
Sales order handling cost for:
Customer orders
Retailer orders
Distributor orders
Order processing cost for:
Orders to distributor
Orders to warehouses
Orders to factory
Inventory costs
225
260,414
1,627,909
1,988,984
2,604,145
16,279,092
19,889,837
Retailer
Distributor
Warehouse
2,584,458
535,730
363,478
25,844,580
5,357,300
3,634,775
$24,426,593
$244,265,928
Seed value
Length of simulation, years
Annual price, $/unit
Index
0.75
1.50
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
98
In-stock probability, %
670
Back order cost, $/unit
Forecasting method
Moving average
7
Number of periods
Reorder policy
Stock-to-demand control method
14
Target days of inventory
226
Distributor/Level 2
Product item data
2220
Item value in inventory, $/unit
20
Retailer order filling cost, $/unit
75
Purchase order processing cost, $/order
25
Inventory carrying cost, %/year
2
Average retailer order fill time, days
0.2
Retailer order fill time standard deviation, days
95
In-stock probability, %
100
Back order cost, $/unit
Forecasting method
Moving average
7
Number of periods
Reorder policy
Stock-to-demand control method
35
Target days of inventory
7
Review time in days
Warehouse/Level 3
Product item data
1710
Item value in inventory, $/unit
15
Distributor order filling cost, $/unit
75
Purchase order processing cost, $/order
20
Inventory carrying cost, %/year
3
Average distributor order filling time, days
0.3
Distributor order fill time standard deviation, days
95
In-stock probability, %
25
Back order cost, $/unit
Forecasting method
Moving average
7
Number of periods
Reorder policy
Stock-to-demand control method
25
Target days of inventory
7
Review time in days
Factory/Source
Product item data
825
20
10
10
2.1
1000
227
Transportation
Transport between Distributor and Retailer
25
1
0
70
5
1
78
9
3
228
CHAPTER 15
LOGISTICS/SUPPLY CHAIN ORGANIZATION
All questions in this chapter require individual judgment and response. No answers are
offered.
229
CHAPTER 16
LOGISTICS/SUPPLY CHAIN AND CONTROL
All questions in this chapter require individual judgment and response. No answers are
offered.
230