Ballou 09

Inventory Policy
Decisions
Every management mistake ends up in inventory.
Michael C.
Bergerac
Former Chief Executive
Revlon,
Inc.
Chapter 9
CR (2004) Prentice Hall, Inc.
9-1
CONTROLLING
Customer
service goals
The product
Logistics service
Ord. proc. & info. sys.
Transport Strategy
Transport fundamentals
Transport decisions
PLANNING
Inventory Strategy
Forecasting
Inventory decisions
Purchasing and supply
scheduling decisions
Storage fundamentals
Storage decisions
ORGANIZING
Inventory Decisions in
Strategy
Location Strategy
Location decisions
The network planning process
9-2
What are Inventories?
Finished product held for sale

Goods in warehouses
Work in process
Goods in transit
Staff hired to meet service needs
Any owned or financially controlled
raw material, work in process, and/or

finished good or service held in
anticipation of a sale but not yet sold
9-3
Where are Inventories?

Inbound
transportation
Production
Outbound
transportation
Finished goods
warehousing
Customers
Receiving
Material
sources
Production
materials
Finished goods
Shipping
Inventories
in-process
Inventory
locations
9-4
Reasons for Inventories

Improve customer service
-Provides immediacy in product availability
Encourage production, purchase, and transportation
economies
-Allows for long production runs
-Takes advantage of price-quantity discounts
-Allows for transport economies from larger shipment sizes
Act as a hedge against price changes
-Allows purchasing to take place under most favorable price
terms
Protect against uncertainties in demand and lead times
-Provides a measure of safety to keep operations
running when demand levels and lead times cannot be known
for sure
Act as a hedge against contingencies
-Buffers against such events as strikes, fires, and
disruptions in supply
9-5
Reasons Against Inventories

They consume capital resources that might be put to
better use elsewhere in the firm
They too often mask quality problems that would more

immediately be solved without their presence
They divert managements attention away from careful

planning and control of the supply and distribution
channels by promoting an insular attitude about
channel management
9-6
Types of Inventories
Pipeline
-Inventories in transit
Speculative
-Goods purchased in anticipation of price increases
Regular/Cyclical/Seasonal
-Inventories held to meet normal operating needs
Safety
-Extra stocks held in anticipation of demand and
lead time uncertainties

Obsolete/Dead Stock
-Inventories that are of little or no value due to being
out of date, spoiled, damaged, etc.
9-7
Nature of Demand
Perpetual demand
-Continues well into the foreseeable future
Seasonal demand
-Varies with regular peaks and valleys throughout
the year
Accurately forecasting
demand is singly the
Lumpy demand
most important factor
-Highly variable (3 Mean)
in good inventory
Regular
demand
management
-Not highly variable (3 < Mean)
Terminating demand
-Demand goes to 0 in foreseeable future
Derived demand
-Demand is determined from the demand of another
item of which it is a part
9-8
Inventory Management
Philosophies
Pull
-Draws inventory into the stocking location
-Each stocking location is considered independent
-Maximizes local control of inventories
Push
-Allocates production to stocking locations based on
overall demand
-Encourages economies of scale in production
Just-in-time
-Attempts to synchronize stock flows so as to just
meet demand as it occurs
-Minimizes the need for inventory
9-9
Inventory Management
Philosophies (Contd)
Supply-Driven
-Supply quantities and timing are unknown
-All supply must be accepted and processed
-Inventories are controlled through demand
Aggregate Control
-Classification of items:
Groups items according to their sales level
based on the 80-20 principle

Allows different control policies for 3 or more
broad product groups
9-10
Pull vs. Push Inventory Philosophies

PUSH - Allocate supply to each
warehouse based on the forecast
for each warehouse
PULL - Replenish inventory with

order sizes based on specific needs
of each warehouse
Demand
forecast
Q1
Warehouse #1
A1
A2
Plant
Q2
Warehouse #2
A3
Demand
forecast
Q3
A = Allocation quantity to each warehouse
Q = Requested replenishment quantity
by each warehouse
Warehouse #3
Demand
forecast
9-11
Costs Relevant to Inventory

Management
Carrying costs
-Cost for holding the inventory over time
-The primary cost is the cost of money tied up in
inventory, but also includes obsolescence,

insurance, personal property taxes, and storage
costs
-Typically, costs range from the cost of short term

capital to about 40%/year. The average is about
25%/year of the item value in inventory.
9-12
Relevant Costs (Contd)

Procurement costs
-Cost of preparing the order
-Cost of order transmission
-Cost of production setup if appropriate
-Cost of materials handling or processing at the
receiving dock
-Price of the goods

9-13
Relevant Costs (Contd)

Out-of-stock costs
-Lost sales cost
Profit immediately foregone
Future profits foregone through loss of goodwill
-Backorder cost
Costs of extra order handling
Additional transportation and handling costs
Possibly additional setup costs
9-14
Inventory Management Objectives

Good inventory management is a careful balancing act
between stock availability and the cost of holding
inventory.
Customer Service,
i.e., Stock Availability
Inventory Holding costs
Service objectives
-Setting stocking levels so that there is only a
specified probability of running out of stock

Cost objectives
-Balancing conflicting costs to find the most
economical replenishment quantities and timing
9-15
Inventorys Conflicting Cost Patterns
Minimum cost
reorder quantity
Cost
Total cost
yin
r
r
a
os
c
g
Procurement cost
Stockout cost
Replenishment quantity
9-16
Glossary of Terms
D average annual demand, units
d average period demand, units
S procurement cost per order, $/order
I carrying costs as a percent of product value, % per year
C product value, $ per unit
sd standard deviation of demand (d), units
k out - of - stock cost, $ per unit
p purchase price
s ' standard deviation of compound demand distribution
E( z ) partial expectation or unit normal loss integral
P probability of being in - stock during lead time (Q - system)
or during lead time plus order cycle time (P - system
Q order quantity
ROP reorder point quantity, units
T order interval, e.g., days
MAX target inventory level, units
z normal deviate or number of standard deviations from mean
on compound demand distribution
r safety stock, or z x s' , units
TC total relevant cost, $
SL service level as a percent of total annual demand
LT , sLT average and standard deviation of lead time
CPn probability of n units being sold
9-17
Single Order Purchasing

Make a one-time purchase of an item. How much to order?
Procedure: Balance incremental profit against incremental loss.
Profit = Price per unit Cost per unit
Loss = Cost per unit Salvage value per unit
If CPn is probability of n units being sold, then
CPn x Loss = (1 CPn) x Profit
or
CPn = Profit/(Profit + Loss)
Daily stocking of
newspapers in
vending
machines is a
good example
Now, increase order quantity until CPn just matches cumulative

probability
of selling additional units.
9-18
Single Order Purchasing (Contd)

Example A clothing item is purchased for a seasonal sale. It
costs $35, but it has a sale price of $50. After the season is
over, it is marked down by 50% to clear the merchandise.
The estimated quantities to be sold are:
Probability of
Number of selling exactly n
items, n
items
10
0.15
15
0.20
20
0.30
25
0.20
30
0.10
35
0.05
1.00
Cumulative
probability
0.15
0.35
0.65
0.85
0.95
1.00
9-19
Single Order Purchasing (Contd)
Solution
Profit = $50 35 = $15
Loss = $35 (0.5)(50) = $10
CPn = 15/(15 + 10) = 0.60
CPn is between 15 and 20 items, round up and order 20
items.
9-20
Simple Two-Bin Pull Method

Given:
d = 50 units/week
I = 10%/year
S = $10/order
C = $5/unit
LT = 3 weeks
Note: No uncertainty
in demand or lead
timemanage
regular (cycle) stock
only
Develop a simple control system by finding the

replenishment quantity (Q) and the reorder point
(ROP).
The relevant total cost is:
TC ordering cost carrying costs
DS ICQ
2
Q
9-21
Reorder Point Method Under Certainty

for a Single Item
Quantity on-hand
plus on-order
Inventory
Level
Q
Reorder
point, R
0
Lead
time
Order
Order
Placed Received
Lead
Time
time
Order
Order
Placed
Received
9-22
Two-Bin Method (Contd)

Using differential calculus, the optimal value for Q will
be:
Q* 2DS/IC 2(50x52)(10)/(0.10x5) 322 units
The reorder point is:

ROP = d(LT) = 3(50) = 150 units
Famous EOQ
formula
Rule
Rule When
Whenthe
theinventory
inventorylevel
leveldrops
drops to
to150
150units
units
(ROP)
(ROP) then
then reorder
reorder 322
322 units
units(Q*).
(Q*).
9-23
Reorder Point Control with

Demand Uncertainty
Given:
d = 50 units/week
sd = 10 units/week
I = 10%/year
S = $10/order
C = $5/unit
LT = 3 weeks
P = 99% during lead time
Find Q* and ROP

From the EOQ formula
Q* 2(50x52)(10) 322 units
0.10(5)
Good method for products:

1. Of high value
2. That are purchased from
one vendor or plant
3. Having few economies of
scale in production,
purchasing, or
transportation
9-24
Quantity on hand
Reorder Point Control for a Single Item
Q
Place
order
DDLT
ROP
Receive
order
Stockout
LT
LT
Time
9-25
Reorder Point Control for a Single Item

Quantity on hand
+on order
backorders
Inventory level
Quantity for
control
Actual
on hand
ROP
Safety stock
0
LT
Time
LT
9-26
Reorder Point Control (Contd)

Finding the reorder point requires an understanding of
the demand-during-lead-time distribution
DDLT
P
Week 1
Week 2
+
Week 3
=
sd=10
sd=10
sd=10
d =100
d =100
d =100
Weekly demand is normally distributed

with a mean of d = 100 and a standard
deviation of sd = 10
Lead time is 3 weeks
S=17.3
X = 300 ROP
X d LT 100(3) 300
s ' sd LT 10 3 17.3
9-27

Now,
X d(LT ) 50(3) 150 units
s' sd LT 10 3 17.32 units
Hence,
ROP X zs' X r
150 2.33(17.32) 190 units
where 2.33 is the normal deviate at a probability of
0.01 taken from a normal distribution table.
9-28

Total relevant cost
The total relevant cost equation is now extended to
include the costs of safety stock as well as out-of-stock.
The out-of-stock cost (k) is $2/unit. The price term is
dropped. Hence,
TC DS IC Q ICr k D s'E(z)
2
Q
Q
2,600(10) (0.1)(5) 322 (0.1)(5)(40)
322
2
2 2,600(17.32)(0.0034)
322
$182.20
where E(z) = 0.0034 from a unit normal loss table at a
z value of 2.33
9-29

With known stockout costs k
Setting Q involves balancing both costs and service
level at optimum. Since P and Q are interrelated,
an iterative approach is required.
1 Solve initially for Q
Q 2DS
IC
2 Using Q, find
P 1 QIC If backordering is allowed
Dk
or
P 1 QIC
Dk QIC If sales are lost
9-30

3 Using P, find revised Q
2D[S ks' E ]
d (z)
Q
IC
4 Repeat steps 2 and 3 until no further change

5 Compute ROP and other statistics
9-31

Example Given:
Monthly demand forecast, d
Std. error of forecast., sd
Replenishment lead-time, LT
Item value, C
Cost for processing
vendor order, S
Carrying cost, I
Stockout cost, k
Backordering is allowed
11,107 units
3,099 units
1.5 months
$0.11/unit
$10/order
20%/year
$0.01/unit
Find optimal Q and P

9-32

Solution
Estimate Q
Q 2DS 2(11,107)( 12)(10) 11,008 units
0.20(0.11)
IC
Estimate P
P 1 QIC 1 11,008(0.2 0)(0.11) 0.82
Dk
11,107(12) (0.01)
Revise Q
Find App A, z@0.82=0.92 and from App B,
E(0.92)=0.0968
For these data, s'd was previously calculated as
3,795 units
9-33
2D S ks' E
IC
(z)
2(11,107)( 12)[(10 0.01(3,795)(0.068)]

0.20(0.11)
12,872 units
Revise P
P 1 12,872(0.20)(0.11) 0.79
11,107(12) (0.01)
Now, z@0.81=0.81 and E(0.81)=0.1181
9-34

Revise Q
Q 2(11,107)( 12)[10 0.01(3,795)(0.1181)] 13,246 units
0.20(0.11)
Continue to revise Q and P until no further

change occurs. P=78% and Q=13,395 units.
Note
Note Although
Althoughthe
thein-stock
in-stockprobability
probabilityduring
during the
the
lead
leadtime
timeisis78%,
78%,the
theactual
actualservice
servicelevel
levelisis
SL=96%
SL=96%
9-35
Pull Methods (Contd)

Noninstantaneous resupply
At times, production or supply continues while demand
is depleting inventories. This requires a slight
modification of the EOQ formula. That is,
p
Qp* 2DS
IC p d
Just add
this term
where
p = output or supply rate
d = demand rate
and p > d. ROP remains unchanged.
9-36

Reorder point control with demand and lead time uncertainties
The combined effect of these two uncertainties is particularly
hard to estimate accurately. It is the standard deviation of the
demand-during-lead-time distribution that is the problem,
especially if the level of demand and the length of the lead time
are related to each other. Ideally, we would simply observe the
actual demand occurring over each lead time period. If the
demand and lead time are independent of each other and each
are represented by separate distributions, we may estimate the
standard deviation (s) from
2 )
s' LT (sd2) d 2(sLT
Caution: Can result in very

high safety stock levels when
lead-time variability is high
After computing s, calculation of the ordering policy would

9-37
be identical to that presented previously.

Periodic review control with demand uncertainty
The inventory is reviewed at the time interval (T) to
determine the quantity on hand. The replenishment
quantity (Q) to be ordered is the difference between a
target level called MAX and the quantity on hand. We
need to find MAX and T*.
Good method for
Given:
d = 50 units/week
sd = 10 units/week
I = 10%/year
S = $10/order
C = $5/unit
LT = 3 weeks
P = 0.99
k = $2/unit
products:
1. Of low value
2. That are purchased
from the same vendor
3. Having economies of
scale in production,
purchasing, and
transportation
9-38
Periodic Control for a Single Item

Quantity on hand
Q2
Q1
Stock
level
reviewed
Order
received
LT
M = maximum level
M - q = replenishment quantity
LT = lead time
LT
Time
T
T = review interval
q = quantity on hand
Qi = order quantity
9-39
Periodic Review (Contd)

Estimate Q* from the EOQ formula as if under demand
certainty conditions. Recall that this is Q* = 322 units.
Now,
T* = Q*/d = 322/50 = 6.4 weeks
Construct the demand-during-lead-time-plus-ordercycle-time distribution.
T is order
review time
9-40

DD(T* + LT)
Z(s)
s sd T LT
'
X
*
MAX
= d(T + LT)
9-41

where
X d(T * LT ) 50(6.4 3) 470
s' s T * LT 10 6.4 3 30.66
d
Find MAX
MAX = d(T* + LT) + z(s)
= 50(6.4 + 3) + 2.33(30.66)
= 470 + 71.44 = 541 units
Rule
Rule Review
Review the
theinventory
inventory every
every6.4
6.4weeks
weeksand
and place
place
an
anorder
order for
for the
thedifference
differencebetween
betweenthe
the MAX
MAXlevel
levelof
of 541
541
units
unitsand
andthe
thequantity
quantityon
onhand
hand++quantity
quantityon
onorder
order
backorders.
backorders.
9-42

The total relevant cost for this design is:
TC = DS/Q + ICQ/2 + ICr + ks(D/Q)E(z)
= 2600(10)/322 + (.10)(5)(322/2)
+ (.10)(5)(71) + 2 (30.66)(2600/322)(.0034)
= $198
Note
Note Compare
Comparethis
thiscost
costwith
withthat
that of
of the
thereorder
reorderpoint
point
method
methodto
tosee
seethat
thatperiodic
periodicreview
review control
controlcarries
carriesaa
slight
slight premium
premium in
incost
costdue
due to
tomore
moresafety
safetystock.
stock.
9-43

In/
Date
Customer
10/26
Bal Fwd
10/26
100M
10/30
Progression
10/30
Ogleby
11/2
Mid Ross
11/9
Unt Sply
11/29
Berea Lit
12/1
Dol Fed
12/13
Card Fed
12/14
Belmont
12/15
Shkr Sav
1/8
BFK
1/8
100M
1/8
Card Fed
1/9
Pt of View
1/17
Am Safety
1/23
Foster
1/24
Gib Prtg
1/26
Bel-Gar
1/26
Copies
1/29
Slvr Lake
1/29
100M
2/2
Sagamore
Size
M/Wgt
8x14
12.72
Sales
20000
25000
15000
50000
25000
10000
20000
15000
5000
500
30000
10000
5000
15000
5000
5000
20000
5000
20000
Basis
20
On
hand
Date
80500 2/2
180500 2/5
160500 2/6
135500 2/6
120500 2/6
70500 2/6
45500 2/8
35500 2/14
15500 2/15
500 2/16
500* 2/21
0 2/26
100000 2/27
70000 2/28
60000 2/28
55000 3/1
40000 3/2
35000 3/8
30000 3/8
10000 3/12
5000 3/12
105000 3/12
85000 3/20
Grain
Color
L
White
In/
Customer
Copies
Bel-Gar
Bel-Gar
Superior
Unt Sply
Berea Prtg
Sagamore
100M
50M
Bel-Gar
Bel-Gar
Inkspot
Lcl 25UAW
Ptrs Dvl
Shkr Sav
Copies
Untd Tor
Sagamore
Sagamore
150M
Untd Tor
Preston
Midland
Finish
RmSeal
No stock or insufficient stock to meet demand
On
Sales
hand
50000
35000
5000
30000
15000
15000
25000
0*
15000
0*
15000
0*
5000
0*
100000
150000
5000
145000
15000
130000
5000
125000
50000
75000
2500
72500
25000
47500
35000
12500
10000
2500
2500
0
12500
0*
150000
40000
110000
50000
60000
15000
45000
Grade
Advantage Bond
In/
Date
Customer
Sales
3/30
Sup Meats
25000
3/30
Copies
50
3/30
Ptrs Dvl
5000
3/30
Belmont
10000
4/2
Berea Prtg
4950
4/2
Berea Prtg
15050
4/9
REM
500
4/12
Mid Ross
5000
5/7
Ohio Ost
5000
5/8
Inkspots
5000
5/8
Prts Dvl
2500
5/11
100M
5/14
BVR
5000
5/15
Guswold
10000
5/16
ESB
15000
5/16
Superior
50000
5/16
J Stephen
5000
5/16
Am Aster
15000
5/16
Am Aster
10000
5/22
Sagamore
15000
Coding
21200
M. Base Cost
Date Min
2.64
4/2
Max
Location
Ctn. Skid Cont.
F 14
5M
On
Hand
20000
19950
14950
4950
0
0*
0*
0*
0*
0*
0*
100000
95000
85000
70000
20000
15000
0
0*
0*
125M
250M
Att.
9-44

Supply chain example
Suppose that inventory is to be maintained on a
distributors shelf for an item whose demand is
forecasted to be d = 100 units per day and sd = 10 units
per day. A reorder point is the method of inventory
control. The supply channel is shown in the diagram.
Determine the average inventory to be held at the
distributor where we have:
I = 10%/year
S = $10/order
C = $5/unit
P = 0.99 during lead time
9-45
Supply Chain Example (Contd)

Supplier
Processing time
1, s 2p 0 .1
Transport time
Inbound transport
4 , si2 1.0
Outbound transport
Pool point
Transport time
X
2 , s o2 0 .25
Distributor
9-46

Solution The reorder point inventory theory applies.
However, determining the statistics of the demandduring-lead-time distribution requires taking the leadtime for the entire channel into account.
Recall,
2
s' LT(s d2 ) d 2(sLT
)
where
2
sLT
sp2 si2 so2
0.1 1.0 0.25
1.35 days
9-47

Average lead time
LT X p X i X o 1 4 2 7 days
Now
s' 7x102 1002 x1.35 14,200 119 .16 days
and
Q* 2(100)(10) 63 units
0.1(5)
*
Q
AIL z(s' ) 63 2.33(199.16) 309 units
2
2
9-48

Joint ordering
Perpetual inventory control for most firms is the problem of
managing items jointly rather than singly. This occurs since
more than one item is typically purchased from the same
vendor. The approach to joint ordering is to find a common
order review interval (T) and then to set separate target levels
(MAX) based on specific item costs and service levels.
A common review time may be specified, or it may be
computed based on appropriate economics.
T*
2(O S )
i
(I C D )
i
where
O = common procurement cost, $/order
*
*
Note:
Note: Q
Q* == TT*xd
xd
9-49
Joint Ordering Example

Given
Item
A
30
8
14
25
30
Average daily demand (d)

Demand std. dev. (sd)
Average lead time (LT)
Annual carrying cost (I)
Procurement cost (S)
with common cost (O)
In-stock probability (P)
80
Product value (C)
170
Out-of-stock cost (k)
25
Selling days per year
365
B
75 units
10 units
14 days
25 %
20 $/order
80 $/order
92 %
200 $/unit
45 $/unit
365 days
9-50
Joint Ordering Example (Contd)

Find common review time
T*
2[80 (30 20)]

4.35 days
[0.25/365][170(30) 200(75)]
Find target quantity (MAX) for item A

s'A sd
T * LT 8 4.35 14 34.3 units

A
then z@80%=0.84
MAX A X z(sA' ) 30(4.35 14) 0.84(34.3) 579 units
9-51
Joint Ordering Example (Contd)

which has an average inventory of
Avg. Inventoryi T *(di / 2) zi (si' )
Avg. InventoryA 4.35(30 / 2) 0.84(34.3) 94.1units
Find target quantity (MAX) for item B
sB' sd
T * LT 10 4.35 14 42.8 units

B
then for z@90%=1.41

MAX B 75(4.35 14) 1.41(42.8) 1437 units
which has an average inventory of
Avg. inventoryB 4.35(75 / 2) 1.41(42.8) 223 units
9-52
The Min-Max variant

This is basically a reorder point system, but the order
quantity is incremented by the amount of the difference
between the reorder point quantity and the quantity on
hand + quantity on order backorders. This takes into
account that demand does not decrement inventory
levels evenly. Therefore, inventory levels may fall
below the reorder point at the time that it is reached.
9-53
Min-Max Inventory Control

Add increment ROPq to order size
Quantity on hand
Q1
Q2
Q*
ROP
q
LT
LT
Time
9-54
The T, R, M variant
This is a combination of the min-max and the periodic
review systems. The stock levels are reviewed
periodically, but control the release of the replenishment
order by whether the reorder point is reached. This
method is useful where demand is low, such that small
quantities might be released under a periodic review
method.
9-55

Inventory not
below R, so dont
place an order
Inventory level
T,R,M variant
Q1
Q2
R
q
LT
LT
Time
T = review time
R = reorder point M Q = replenishment quantity
9-56

Stock to demand (a periodic review method)
This is an important periodic review method, not so much
because of its accuracy but because of its popularity in
practice. The method is synchronized with the period of
the forecast. The target quantity (MAX) is developed as
follows.
An example
Set the period of the forecast, say 4 weeks

Add time for lead time, say 1 week
Add an increment of time for safety stock, say 1 week
9-57
Stock to Demand (Contd)

Therefore, MAX is 6/4 times the monthly forecast. The
replenishment quantity is determined as follows.
At the time (T) of the monthly stock-level review, make a
forecast and determine the MAX level.
MAX = Forecast x 6/4
Plus: Backorders
Less: Quantity on hand*
Less: Quantity on order
Order quantity (Q)
Units
12,500
0
-5,342
-4,000
3,158
*Quantity on hand = actual quantity on hand +

quantity on order backorders
9-58
Multiple item, multiple-location control

The theory that has been discussed previously is
useful when designing inventory control systems for
the practical problem of controlling many items at
many locations. Consider how a specialty chemical
company designed such a practical system. TASO is
the time to accumulate a stock order (truckload) for all
items in warehouse.
9-59
Q3
Q2
Q1
Stock
order
Order
received
0
TASO
LT
TASO
M = maximum level
TASO = time to accumulate stock order
LT
TASO
Time
Qi = order quantity
LT = lead time
Multiple-Item, Multiple-Location Control
Quantity on hand
9-60
Customer Service Level

For individual items
The service level (stock availability) actually achieved by
inventory control methods is not best represented by the
probability (P) of a stockout during the lead time. It is
more accurate to compute it as follows.
SL 1
s 'E D /Q
(z)
s 'E
(z)
Using data from the reorder point under uncertainty

example, the service level would be:
SL 1 17.32(.0034)(2,600 / 322) 0.999
2,600
Note: Higher
than P
9-61
Customer Service Level (Contd)

This actual level is higher than P = 0.99 that was
used to set the inventory level. The reason is that
there are periods of time when the stock level is
above the reorder point and there is no risk of being
out of stock.
Methods for defining stock availability include:
Probability of filling all item demand

Probability of filling an order completely
Probability of filling a percent of all item demand
Weighted average of items filled on an order (fill
rate)
9-62
Customer Service Level (Contd)

For multiple items on the same order
If all items on an order have the same service level,
what is the probability of filling the order complete?
The service level for multiple items is the combination
of the individual item service levels as follows:
SL = SL1 x SL2 x SL3 x SLn
Suppose 3 items have the following service levels
0.95, 0.89, and 0.92. The probability of filling the order
complete is:
SL = 0.95 x 0.89 x 0.92 = 0.78
9-63
Push Inventory Control

Example
Three warehouses are used to supply 900 retail
drugstores. Each warehouse serves approximately 300
stores. A large purchase of clock radios is made, where
radios were to be a promotional item in the next forecast
period. The special buy will result in more stock than
needed, but the company expects to sell all stock
eventually. Warehouses are to have a 92% in-stock
probability. All of the purchased radios are to be allocated
to the warehouses based on the anticipated demand
levels at each warehouse. Account is taken of the
inventory already on hand. A total of 5000 radios is
purchased. The next purchase will be made in one
month. Further information is given below.
9-64
Push Inventory Control (Contd)
Warehouse
1
2
3
Current
stock
level,
units
400
350
0
Forecasted
demand,
units
2,300
1,400
900
4,600
Forecast
error (std.
dev.), units
100
55
20
How should the allocation to the warehouses be

made?
9-65

Solution
Warehouse
1
2
3
a
Total
requirements
a
2,428
1,470
926
4,824
2,428 = 2,300 + 1.28(100)
Total requirements = Forecast + z(Forecast error)

where z@90% = 1.28. Therefore,
9-66

(1)
(2)
(3)=
(4)
(5)=
Pro(4)+(3)
(1)(2)
Total
Onration
Net
require- hand requireof
AlloWare- ments, stock, ments, excess, cation,
house units
units
units
units
units
a
b
1
2,428
400 2,028
463
2,491
2
1,470
350 1,120
282
1,402
3
926
0
926
181
1,107
c
Total 4,824
4,074
926
5,000
Total requirements less (quantity on hand + quantity on order backorders)
b
Excess purchase quantity times forecast for warehouse divided by total
forecast quantity. For example, (5,000 4,074) x 2,300/4,600 = 463
c
5,000 4,074 = 926
a
9-67
Multi-Echelon Inventories
Control the entire channel inventory levels, not just a
single echelon.
Warehouse
echelon
R1
Warehouse
lead-time, LTw
S
Supplier
R2
Warehouse
d 1 , sd1
d 2 , sd 2
R3
End customer demand
How much stock here when

retailers also carry stock?
d 3 , sd 3
Retailer
9-68
Multi-Echelon Inventories (Contd)

Example
An item has the following cost characteristics. Item
values are CR=$10/unit and CW=$5/unit. Carrying cost
is I = 20%/year. Ordering costs are SR=$40/order and
SW=$75/order. Lead times are LTR=0.25 month and
LTW=0.5 months. In-stock probability for retailers and
warehouses is 90%. Monthly demand statistics are:
Monthly
avg., units
Std. dev.,
units
Retailer 1
202.5
16.8
Retailer 2
100.5
15.6
Retailer 3
302.5
18.0
Combined
605.5
32.4
9-69

Solution
Based on reorder point inventory control, the retailers
inventory statistics are
Retailer 1
Retailer 2
Retailer 3
Reorder qty, Q
312
220
381
Reorder point, ROP
61
35
87
Avg. inv., AIL
167
120
202
The warehouse echelon order quantity is

2DW SW
QW
2(605.5x12)(75) 1,043.98, or 1,044 units
0.20(5)
ICW
ROPW dW xLTW zsW LTW 605.5(.5) 1.28(32.4) .5
332 units
9-70

The warehouse echelon inventory is
Q
AILW W zsW LTW
2
1,043.98 1.28(32.4) 0.5
2
551.32. or 551units
The average warehouse inventory is the warehouse
echelon inventory less the retailers inventory, or 551
167 120 202 = 62 units.
Rule
Rule When
Whenthe
thetotal
totalwarehouse
warehouseinventory
inventory(sum
(sum of
of
retailers
retailersinventory,
inventory,inventory
inventoryat
at the
thewarehouse
warehouseand
andon
on
order,
order,and
andretailers
retailersorders
ordersless
less any
anyinventory
inventorycommitted
committed
to
tocustomers
customersdrops
dropsbelow
below 332
332units,
units,order
order 1,044
1,044units.
units.
9-71
100
90
Total sales (%)
80
70
60
50
40
30
A items
20
B items
C items
10
0
0
20
40
60
Total items (%)
80
100
Aggregate Inventory Control
Product items can be grouped according to 80-20

curve, each with different stocking policies
9-72
Inventory Consolidation
(Risk Pooling)
Illustration of risk pooling
Suppose there is a product stocked in two warehouses.
The replenishment quantities are determined by the
economic order quantity formula. The replenishment
lead-time is 0.5 months, the cost for a replenishment
order is $50, the inventory carrying cost is 2% per
month, and the item value is $75 per unit. The
probability of an out of stock during the lead-time period
is 5%. The demand is normally distributed with typical
demand over six months as follows.
9-73
Risk Pooling (Contd)
Month
1
2
3
4
5
6
Avg. (D)
Std. Dev. (sd)
Demand
in Whse
A
35
62
46
25
37
43
41.33
11.38
Combined
Demand Demand in a
in Whse
Central
B
Whse
67
102
83
145
71
117
62
87
55
92
66
109
67.33
108.66
8.58
19.07
Estimate the average inventory levels for twowarehouse and one-warehouse supply channels.
9-74
Regular stock
2DS
RS Q IC
2
2
2(41.33)(50)
0.02(75) 52.49 26.25 units
RSA
2
2
2(67.33)(50)
0.02(75) 67.00 33.50 units
RSB
2
2
Regular stock in system is
RSs RSA RSB 26.25 33.50 59.75 units
9-75

Regular stock if item is entirely in one warehouse
2(108.66)(50)
0.02(75) 85.11 42.56 units
RSC
2
2
Safety stock
SS z(sd ) LT
SSA 1.96(11.38) 0.5 15.77 units
SSB 1.96(8.58) 0.5 11.89 units
System safety stock in 2 warehouses
SSA SSB 15.77 11.89 27.66 units
9-76

Safety stock in 1 warehouse
SSc 1.96(19.07) 0.5 26.43 units
Total inventory
AIL = Regular stock + Safety stock
Two
warehouses
AIL = 59.75 + 27.66 = 87.41 units

In a one-warehouse channel
AIL = 42.56 + 26.43 = 68.99 units
Conclusion
Conclusion There
Thereisisaa reduction
reductionin
in the
theaverage
average
inventory
inventorylevel
levelof
ofan
anitem
item as
as the
thenumber
numberof
ofstocking
stocking
points
pointsin
inthe
thesupply
supplychannel
channelisis decreased.
decreased. In
Inthis
this
example,
example,both
bothregular
regular stock
stockand
andsafety
safetystock
stockdecline.
decline.
9-77

System-wide Inventory as a Fraction of the
Demand Divided Between Two Warehouses
Percent of Peak
System-Wide Inventory
100
95
90
85
80
75
70
0
0.2
0.4
0.6
0.8
One warehouse's demand as a fraction

of the total
9-78
Virtual Inventories
Stockouts are filled from other stocking locations in
the distribution network
Customers assigned to a primary stocking location

Backup locations are usually determined by
zoning rules
Expectation is that lower system-wide inventories

can be achieved while maintaining or improving
stock availability levels
Total distribution costs should be lower to support

the cross filling of customer demand
9-79
Cross Filling Among 2

Stocking Locations
Demand 1
Virtual Inventories
Stock
location B
Secondary
assignment
Primary
assignment
Primary
assignment
Stock
location A
Demand 2
9-80
Potential Benefit
of Cross Filling
Suppose that an item is stocked at a fill rate of 80% in
4 stocking locations. If cross filling is used, what is
the effective fill rate for the customer?
Fill rate = [1 (.20)(.20)(.20)(.20)] x 100 = 99.8%
Customer service levels can be quite high

even if the item fill rate is low!
But are inventory costs lower?
Virtual Inventories
9-81
Regular Stock in 2 Locations
Demand 1
Stock
location B
Secondary
assignment
Primary
assignment
Demand dispersion
Stock
location A
Primary
assignment
Meaning of regular stock

How it varies with:
Demand 2
Method of stock control

Fill rate
Virtual Inventories
9-82
Stock Control Methods

and Regular Stock
If control is EOQ-based, average inventory level (AIL) is
EOQ
formula
2S D
AIL Q IC
2
2
0.5
kD0.5
If stock-to-demand control
AIL is a function of
demand with
exponents ranging
from 0.5 to 1.0
AIL kD1.0
Virtual Inventories
9-83
Percent of peak sytem-wide

inventory
Regular Stock as a Percent of Demand Divided

Stock-toBetween Two Warehouses
100
demand
control
95
90
85
80
75
70
0
EOQ-based
control
20
40
60
80
100
One warehouse's demand

as a percent of total demand
Virtual Inventories
9-84
Observation about
Regular Stock
A system of multiple stocking locations
will carry its maximum regular stock
when demand is balanced among them
Virtual Inventories
9-85
Fill Rate and Regular Stock

Cross filling increases regular stock as lower fill rates
are specified
Example
2 locations
Demand is dispersed 50 and 150
Fill rate is 90%
Stocking policy is D with k=1
0.5
Virtual Inventories
9-86
Example (Contd)
No cross filling
Location A
Cross filling
Location B
Location A
Location B
Demand 1
50
45a
5b
Demand 2
150
15
135
Total
50
150
60
140
Regular
stock
7.1
12.2
7.7
11.8
System inv.
a
50x.90=45
[50x(1-0.90)]x0.905
19.3
19.5
Regular stock increases

with cross filling
Virtual Inventories
9-87
% of inventory compared
with no cross-filling
Regular Stock Penalty for Cross Filling Under

Various Stocking Policies
EOQ-based
150
control
140
130
120
110
100
50
Stock-todemand control
60
70
80
90
100
Average fill rate on warehouses
Virtual Inventories
9-88
Safety Stock in 2 Locations

Meaning of safety stock
Safety stock depends on
Demand dispersion (variance is proportional
Secondary
assignment
Demand 1
Primary
assignment
Stock
locationB
location
B
Primary
assignment
Stock
locationA
location
A
Demand 2
to (demand)
Fill rate
Observation
A system of multiple stocking
locations will carry its minimum
safety stock when demand is
balanced among them
Virtual Inventories
9-89
Safety Stock Estimation

Safety stock is estimated by
ss zs * LT
where s* is the demand standard deviation at
location N
When cross filling,
s * FR 2sd2
where sd is the demand standard deviation at the

primary location
At any location N
sN* [FR (1 FR )N 1]2 sd2
Virtual Inventories
9-90
Safety Stock in 2 Locations

Example
2 locations
Weekly demand and std. dev. are (50,5) and
(150,15)
Lead time is 1 week

Fill rate (FR) is 95%
z is 1.65 for 95% stocking level (demand normally
distributed)
Inventory control is EOQ based

Virtual Inventories
9-91
Safety Stock for 2 Locations

No cross filling
Location A
Location B
Cross filling
Location A
Location B
Std. Dev. 1
4.7500
0.2375
Std. Dev. 2
15
0.7125
14.2500
Combined
15
4.8
14.3
8.3
24.8
7.9
23.5
Safety stock
System inv.
33.1
31.4
Safety stock decreases

with cross filling
Virtual Inventories
9-92
Safety Stock Reduction Due to Cross Filling as a

Percent of Demand Divided Between Two Warehouses
Percent reduction
25
FR=70%
20
Lower safety
stocks from
lower fill rates
15
10
FR=90%
No crossfilling
0
0
15 25 35 45 55 65 75 85 95 100
One warehouse's demand as a percent of

the system-wide demand
Virtual Inventories
9-93
Simplifying the Decision

Problem
An item potentially can be cross filled from 1 backup
warehouse. The item has a value of $200/unit, a carrying
cost of 25%/per year, a stocking level of 6-weeks demand, a
replenishment lead-time of 8 weeks, and a target fill rate of
95%. To cross haul the item from the secondary warehouse
incurs an extra $10/ unit in transportation. The stock control
policy is not known. Demand statistics are as follows:
Location
Mean demand,
units
Std. Dev.,
units
300
138
100
80
System
400
160
Should the item be cross filled?

Virtual Inventories 9-94
Simplifying the Decision
Answer
Solve for K in AIL=KD. K=D1-/TO, where TO is 52
weeks/6 weeks of demand or 8.67. This assumes a
control policy based on of 0.7. Hence,
K=(400x52)1-.7/8.67=2.28. The demand ratio r between
the 2 locations is 100/400=0.25. Now solve for two
parameters of the cross-filling curve.
1.7
1
10
([
400
x
52
])
tD
X
1.73
ICK 0.25(200)(2.28)
and
1.96(160) 8 0.4
Y zs LT
KD
2.28(400x52)0.7
Now, from the decision curve, dont cross fill this item.
Virtual Inventories 9-95
LEGEND
D = annual system demand
s = system demand std.
dev.
C = item unit cost
I = annual carrying cost
rate (%)
K, = inventory
control parameters
t = transportation rate
FR = item fill rate
r = ratio of minimum
demand to system
demand
LT = lead-time in demand
std. dev. time units
z = normal deviate at FR %
tD 1
X
ICK
Ratio of minimum demand to total

demand, r
Decision Curve for FR=.95 and =0.7
zs LT
KD
Below
decision
curvedont
cross fill
Virtual Inventories
9-96
Square Root Law

of Inventory Consolidation
The amount of inventory (regular stock) at multiple stocking
points can be estimated by the square root law when
Inventory control at each point is based on EOQ

principles
There is an equal amount on inventory at each point

The square root law is:
IT Ii n
where
IT = amount of inventory at one location
Ii = amount of inventory at each of n locations
n = number of stocking points
9-97
Square Root Law (Contd)

Example Suppose that there is $1,000,000 of inventory at 3
stocking points for a total of $3,000,000. If it were all
consolidated into 1 location, we can expect:
IT 1,000,000 3 $1,732,051
If we wish to consolidate from 3 to 2 warehouses, the

level of inventory in each warehouse would be:
I
Ii T $1,732,051 $1,224,745
1.41
2
For a total system inventory of n x I = 2 x $1,224,745 =
$2,449,490.
9-98
Square Root Law (Contd)

More simply
n2
I2 I1 n
1
where
I2 system inventory in n2 locations
I1 system inventory in n1 locations
n2 no. of new locations
n1 no. of previous locations
So,
I2 3,000,000 2 2,449,490
3
9-99
Warehouse average inventory, Ii ($000s)
Inventory-Throughput
Curve
3000
2500
2000
Ii = 1.57Di0.72
R = 0.85
1500
1000
500
0
0
10000
20000
30000
40000
Annual warehouse throughput, Dj ($000s)
50000
9-100
Inventory-Throughput Curve
Example
Suppose two warehouses with 390,000 lb. and
770,000 lb. of throughput respectively are to be
consolidated into a single warehouse with 390,000 +
770,000 = 1,160,000 lb. of annual throughput. How
much inventory is likely to be in the single warehouse?
The inventory-throughput curve developed from the
companys multiple warehouse stocks is shown in the
figure below.
Note Reading from the plot, the inventory in the
consolidated warehouse has dropped to 262,000 lb.
from 132,000 + 203,000 = 335,000 lb. in the two
warehouses.
9-101
Inventory-Throughput Curve
450
Average inventory level, AIL i (000 lb.)
400
350
AILi = 3.12Di0.628
R2 = 0.77
300
262
250
203
200
150
132
100
Current
warehouse
50
Consolidated
warehouse
0
0
200
400
600
800
1000
1200
1400
1600
1800
Annual warehouse throughput,D i (000 lb.)

9-102
Turnover Ratio
Annual sales
Turnover ratio
Average inventory
A fruit grower stocks its dried fruit products in 12 warehouses
around the country. What is the turnover ratio for the
distribution system?
Warehouse
no.
1
2
3
4
5
6
Annual
Average
warehouse
inventory
throughput, $
level, $
21,136,032
2,217,790
16,174,988
2,196,364
78,559,012
9,510,027
17,102,486
2,085,246
88,228,672 11,443,489
40,884,400
5,293,539
TO ratio
$425,295,236
9.7
$43,701,344
Warehouse
no.
7
8
9
10
11
12
Totals
Annual
warehouse
throughput, $
43,105,917
47,136,632
24,745,328
57,789,509
16,483,970
26,368,290
425,295,236
Average
inventory
level, $
6,542,079
5,722,640
2,641,138
6,403,076
1,991,016
2,719,330
43,701,344
$ are at cost
9-103
Areas under
Standardized
Normal
Distribution
A table entry is the proportion of the

area under the curve from a z of 0 to a
positive value of z. To find the area
from a z of 0 to a negative z, subtract the
tabled value from 1.
z
0.0
0.1
0.2
0.3
0.4
.00
0.5000
0.5398
0.5793
0.6179
0.6554
.01
0.5040
0.5438
0.5832
0.6217
0.6591
.02
0.5080
0.5478
0.5871
0.6255
0.6628
.03
0.5120
0.5517
0.5910
0.6293
0.6664
.04
0.5160
0.5557
0.5948
0.6331
0.6700
.05
0.5199
0.5596
0.5987
0.6368
0.6736
.06
0.5239
0.5636
0.6026
0.6406
0.6772
.07
0.5279
0.5675
0.6064
0.6443
0.6808
.08
0.5319
0.5714
0.6103
0.6480
0.6844
.09
0.5359
0.5753
0.6141
0.6517
0.6879
0.5
0.6
0.7
0.8
0.9
0.6915
0.7257
0.7580
0.7881
0.8159
0.6950
0.7291
0.7611
0.7910
0.8186
0.6985
0.7324
0.7642
0.7939
0.8212
0.7019
0.7357
0.7673
0.7967
0.8238
0.7054
0.7389
0.7704
0.7995
0.8264
0.7088
0.7422
0.7734
0.8023
0.8289
0.7123
0.7454
0.7764
0.8051
0.8315
0.7157
0.7486
0.7794
0.8078
0.8340
0.7190
0.7517
0.7823
0.8106
0.8365
0.7224
0.7549
0.7852
0.8133
0.8389
1.0
1.1
1.2
1.3
1.4
0.8413
0.8643
0.8849
0.9032
0.9192
0.8438
0.8665
0.8869
0.9049
0.9207
0.8461
0.8686
0.8888
0.9066
0.9222
0.8485
0.8708
0.8907
0.9082
0.9236
0.8508
0.8729
0.8925
0.9099
0.9251
0.8531
0.8749
0.8944
0.9115
0.9265
0.8554
0.8770
0.8962
0.9131
0.9279
0.8577
0.8790
0.8980
0.9147
0.9292
0.8599
0.8810
0.8997
0.9162
0.9306
0.8621
0.8830
0.9015
0.9177
0.9319
1.5
1.6
1.7
1.8
1.9
0.9332
0.9452
0.9554
0.9641
0.9713
0.9345
0.9463
0.9564
0.9649
0.9719
0.9357
0.9474
0.9573
0.9656
0.9726
0.9370
0.9484
0.9582
0.9664
0.9732
0.9382
0.9495
0.9591
0.9671
0.9738
0.9394
0.9505
0.9599
0.9678
0.9744
0.9406
0.9515
0.9608
0.9686
0.9750
0.9418
0.9525
0.9616
0.9693
0.9756
0.9429
0.9535
0.9625
0.9699
0.9761
0.9441
0.9545
0.9633
0.9706
0.9767
2.0
2.1
2.2
2.3
2.4
0.9772
0.9821
0.9861
0.9893
0.9918
0.9778
0.9826
0.9864
0.9896
0.9920
0.9783
0.9830
0.9868
0.9898
0.9922
0.9788
0.9834
0.9871
0.9901
0.9925
0.9793
0.9838
0.9875
0.9904
0.9927
0.9798
0.9842
0.9878
0.9906
0.9929
0.9803
0.9846
0.9881
0.9909
0.9931
0.9808
0.9850
0.9884
0.9911
0.9932
0.9812
0.9854
0.9887
0.9913
0.9934
0.9817
0.9857
0.9890
0.9916
0.9936
2.5
2.6
2.7
2.8
2.9
0.9938
0.9953
0.9965
0.9974
0.9981
0.9940
0.9955
0.9966
0.9975
0.9982
0.9941
0.9956
0.9967
0.9976
0.9982
0.9943
0.9957
0.9968
0.9977
0.9983
0.9945
0.9959
0.9969
0.9977
0.9984
0.9946
0.9960
0.9970
0.9978
0.9984
0.9948
0.9961
0.9971
0.9979
0.9985
0.9949
0.9962
0.9972
0.9979
0.9985
0.9951
0.9963
0.9973
0.9980
0.9986
0.9952
0.9964
0.9974
0.9981
0.9986
3.0
3.1
3.2
3.3
3.4
0.9987
0.9990
0.9993
0.9995
0.9997
0.9987
0.9991
0.9993
0.9995
0.9997
0.9987
0.9991
0.9994
0.9995
0.9997
0.9988
0.9991
0.9994
0.9996
0.9997
0.9988
0.9992
0.9994
0.9996
0.9997
0.9989
0.9992
0.9994
0.9996
0.9997
0.9989
0.9992
0.9994
0.9996
0.9997
0.9989
0.9992
0.9995
0.9996
0.9997
0.9990
0.9993
0.9995
0.9996
0.9997
0.9990
0.9993
0.9995
0.9997
0.9998
9-104
Unit Normal
Loss
Integrals
9-105

Ballou 09

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Ballou 09

Diunggah oleh

Hak Cipta:

Format Tersedia

Inventory Policy

CR (2004) Prentice Hall, Inc.

What are Inventories?

Finished product held for sale

raw material, work in process, and/or

Where are Inventories?

CR (2004) Prentice Hall, Inc.

Reasons for Inventories

Reasons Against Inventories

They too often mask quality problems that would more

They divert managements attention away from careful

CR (2004) Prentice Hall, Inc.

lead time uncertainties

based on the 80-20 principle

CR (2004) Prentice Hall, Inc.

Pull vs. Push Inventory Philosophies

PULL - Replenish inventory with

Costs Relevant to Inventory

inventory, but also includes obsolescence,

-Typically, costs range from the cost of short term

CR (2004) Prentice Hall, Inc.

Relevant Costs (Contd)

-Price of the goods

Relevant Costs (Contd)

Inventory Management Objectives

Inventory Holding costs

specified probability of running out of stock

Inventorys Conflicting Cost Patterns

Single Order Purchasing

Now, increase order quantity until CPn just matches cumulative

Single Order Purchasing (Contd)

Single Order Purchasing (Contd)

Simple Two-Bin Pull Method

Develop a simple control system by finding the

Reorder Point Method Under Certainty

CR (2004) Prentice Hall, Inc.

Two-Bin Method (Contd)

The reorder point is:

Reorder Point Control with

Find Q* and ROP

Good method for products:

Reorder Point Control for a Single Item

CR (2004) Prentice Hall, Inc.

Reorder Point Control for a Single Item

Reorder Point Control (Contd)

Weekly demand is normally distributed

Reorder Point Control (Contd)

CR (2004) Prentice Hall, Inc.

Reorder Point Control (Contd)

Reorder Point Control (Contd)

Reorder Point Control (Contd)

4 Repeat steps 2 and 3 until no further change

CR (2004) Prentice Hall, Inc.

Reorder Point Control (Contd)

Find optimal Q and P

Reorder Point Control (Contd)

Reorder Point Control (Contd)

2(11,107)( 12)[(10 0.01(3,795)(0.068)]

Now, z@0.81=0.81 and E(0.81)=0.1181

CR (2004) Prentice Hall, Inc.

Reorder Point Control (Contd)

Continue to revise Q and P until no further

CR (2004) Prentice Hall, Inc.

Pull Methods (Contd)