12.

Inventory Management

BIA 674 - Supply Chain Analytics

 Outline

The Importance of Inventory

Inventory Costs

ABC Analysis

EOQ Models

Probabilistic Models and Safety Stock

Inventory Control Systems

Single-Period Model

Using Simulation for Inventory Management

What is Inventory?
Stock of items kept to meet future demand for
internalcustomers
external customers

Purpose of inventory management

ORDERING POLICY: When and how many units to order
of each material when orders are placed with either
outside suppliers or production departments within
organizations?
ISSUING POLICY: how to issue units from inventory?
(FIFO, LIFO, random?)
Importance of Inventory
Inventories are important to all types of firms:
They have to be counted, paid for, used in operations,
used to satisfy customers, and managed
Too much inventory reduces profitability
Too little inventory damages customer confidence
It is one of the most expensive assets of many
companies representing as much as 50% of total
invested capital
It is one of the 3 most common reasons for SME
bankruptcy
Need to balance inventory investment and
customer service
Why Do We Want to Hold Inventory
Improve customer service
Safe-guard to hazards in demand, supply, and delivery
that might cause stock-out
Take advantage of economies of scale, & reduce:
ordering costs
Stock-out costs
acquisition costs
Fixed costs (e.g. fixed ordering costs)
Contribute to the efficient and effective operation of
the production system, e.g.,
Reduces the number of costly set-ups and reschedulings
Smoothing and stabilizing resource utilization
Why We Do Not Want to Hold Inventory
Certain costs increase such as
Storage costs
insurance costs
outdate costs
large-lot quality cost
cost of production problems

Ties capital for which the company pays interest

Hides productivity and quality problems

Risk of getting stuck with unsalable goods

Types of Inventory
Raw material
Purchased but not processed
Work-in-process (WIP)
Undergone some change but not completed
A function of cycle time for a product (e.g. items being
transported)
Maintenance/repair/operating (MRO)
Necessary to keep machinery and processes productive
Finished goods
Completed product awaiting shipment
The Material Flow Cycle

Cycle time

95% 5%

Input Wait for Wait to Move Wait in queue Setup Run Output
inspection be moved time for operator time time
Inventory and Service Quality

Customers usually perceive quality service as

availability of goods they want when they want
them

Inventory must be sufficient to provide high-quality

customer service
Inventory Costs
Inventory-Related Costs
Ordering costs (unit variable costs & fixed ordering costs)
costs of replenishing inventory, placing orders, receiving goods
costs for to prepare a machine or process for manufacturing an order
Holding or Inventory carrying costs
cost of holding an item in inventory over time
Shortage or Stock-out / penalty costs
How do you handle shortages?
Lost sales vs. backlogging
Watch out for service level

Outdate costs (for perishable products)

Opportunity costs
Holding Costs
Determining Inventory Holding Costs
COST (AND RANGE) AS
A PERCENT OF
CATEGORY INVENTORY VALUE
Housing costs (building rent or depreciation, 6% (3 - 10%)
operating costs, taxes, insurance)
Material handling costs (equipment lease or 3% (1 - 3.5%)
depreciation, power, operating cost)
Labor cost (receiving, warehousing, security) 3% (3 - 5%)
Investment costs (borrowing costs, taxes, and 11% (6 - 24%)
insurance on inventory)
Pilferage, space, and obsolescence (much higher in 3% (2 - 5%)
industries undergoing rapid change like PCs and cell
phones)
Overall carrying cost 26%
Holding Costs
Determining Inventory Holding Costs
COST (AND RANGE) AS
d e p e n d on
ing OF
c o n s id e ra b ly A PERCENT
CATEGORY g cost s v a r y rates. VALUE
stINVENTORY
H oldin tio n , a n d in tere
e b s in e s
Housing costsu(buildings , locoradepreciation, some
rent high6% h - 10%)
tec(3
th costs, taxes, insurance) t h a n 1 5 % ,
operating
n e rally g re a te r
g co s ts g r e ater
G e s ave holdin
h lease or
Material handling costs
a s h io n ite m
(equipment 3% (1 - 3.5%)
and fpower, operating cost)
depreciation,
Labor cost n 40%. warehousing, security)
tha(receiving, 3% (3 - 5%)
Investment costs (borrowing costs, taxes, and 11% (6 - 24%)
insurance on inventory)
Pilferage, space, and obsolescence (much higher in 3% (2 - 5%)
industries undergoing rapid change like PCs and cell
phones)
Overall carrying cost 26%
ABC Analysis
ABC Analysis
Pay attention to your more critical products!
Divides inventory into three classes based on
annual dollar volume
Class A - high annual dollar volume
Class B - medium annual dollar volume
Class C - low annual dollar volume

Used to establish policies that focus on the few

critical parts and not the many trivial ones
ABC Analysis
Concept: All items do not deserve the same
attention in terms of inventory management
Focus on items that have the highest monetary value

Step 1. Start with the inventoried items ranked by

dollar value in inventory in descending order
Step 2. Plot the cumulative dollar/euro value in

inventory versus the cumulative items in inventory

 ABC Analysis
Class C Typical Chart Using ABC Analysis
100 Class B
90 Class A
Class A
Percentage of dollar value

80 5 15 % of units
70 70 80 % of value
60 Class B
50 30 % of units
40 15 % of value

30 Class C
20
50 60 % of units
5 10 % of value
10

0
10 20 30 40 50 60 70 80 90 100
Percentage of SKUs
ABC Analysis Example
ABC Calculation
(1) (2) (3) (4) (5) (6) (7)
PERCENT
OF PERCENT
ITEM NUMBER ANNUAL ANNUAL OF ANNUAL
STOCK OF ITEMS VOLUME UNIT DOLLAR DOLLAR
NUMBER STOCKED (UNITS) x COST = VOLUME VOLUME CLASS
#10286 20% 1,000 $ 90.00 $ 90,000 38.8% A
72%
#11526 500 154.00 77,000 33.2% A
#12760 1,550 17.00 26,350 11.3% B
#10867 30% 350 42.86 15,001 6.4% 23% B
#10500 1,000 12.50 12,500 5.4% B
#12572 600 $ 14.17 $ 8,502 3.7% C
#14075 2,000 .60 1,200 .5% C
#01036 50% 100 8.50 850 .4% 5% C
#01307 1,200 .42 504 .2% C
#10572 250 .60 150 .1% C
8,550 $232,057 100.0%
ABC Analysis
Percentage of annual dollar usage

80 A Items
70
60
50
40 B Items
30
20
C Items
10
0 | | | | | | | | | |

10 20 30 40 50 60 70 80 90 100
Percentage of inventory items
ABC Analysis
Other criteria than annual dollar volume may
be used
High shortage or holding cost
Anticipated engineering changes
Delivery problems
Quality problems
ABC Analysis
Policies employed may include
1. More emphasis on supplier development for A
items
2. Tighter physical inventory control for A items
3. More care in forecasting A items
EOQ Models
Ordering Policy under constant
demand
Simple case
1. Demand rate is constant and known with certainty
2. Unit ordering cost = C
3. Every time an order is placed, there is a fixed cost = S
4. There is a unit holding cost = H
5. No constraints are placed on the size of each order
6. The lead time is zero
 What is wrong with this management?
Company with steady rate of demand D = 100 tons/month
Total annual demand = 1200 tons
Purchase price C = $250/ton
Delivery costs S = $50 (each time)
Holding costs (storage, insurance, ...) H = $4/ton/month

Inventory
Level

Time
t
Get rid of pre-conceived ideas

Irrational Ordering (time, quantity) ... why?

Safety stock ... why?

t
T 2T 3T
 Inventory Usage Over Time

Total order received

Average
Order Usage rate inventory on
Inventory level

quantity = Q hand
(maximum QT
inventory 2
level)

Minimum
inventory 0
T 2T 3T
Time
Determining the optimal cycle
Objective is to minimize Total Annual Cost

Total cost of
ordering + holding
inventory

Minimum
total cost
Annual cost

Holding cost

Ordering (Setup)
cost
Optimal order Order quantity
quantity (Q*)
Minimizing Costs
By minimizing the sum of setup (or ordering)
and holding costs, total costs are minimized
Optimal order size Q* will minimize total cost

Optimal order quantity occurs when:

The derivative of the Total Cost with respect to the
order quantity is equal to zero
The holding cost and setup cost are equal
Calculating the Annual Costs

Annual holding cost

Annual holding cost = (Average cycle inventory)
(Unit holding cost)
No of orders placed / year
Annual ordering cost
Annual ordering cost = (Ordering cost / order)
No of orders placed / year

Total annual cycle-inventory cost

Total Annual costs = Annual holding cost
+ Annual ordering cost
Calculating all the costs
Holding Cost / period
The cost of holding one unit in inventory for one cycle
= H (QT)/2
Ordering Cost / period
It is the cost of ordering one lot with Q units
= CQ + S
No. of orders / year
= Annual Demand / Oder Size = 12D/Q

Total Cost (C)

It is the sum of annual holding and annual setup cost
 Calculating the EOQ Variable
ordering cost

Total annual cycle-inventory cost

QT
Fixed ordering TC = N (S + CQ) + H
cost 2
Holding cost

Where
TC = total annual cost
C = unit ordering annual cycle-inventory cost
Q = lot size
H = holding cost per unit per period
D = demand per period
S = fixed ordering or setup costs per lot
T = re-order period
Calculating the EOQ

TC = N[(S + CQ) + H(QT/2)]=

= (12DS/Q) + (12D/Q)CQ + (12D/Q)(HQ2/2D)=
=12DS/Q + 12DC + 6HQ
To find the optimal Quantity Q: Set derivative w.r.t Q = 0
Therefore,
-(12DS/Q2)+6H = 0 The optimal - order- quantity

Q* = 2SD/H = 50 tons
T = 0,5 month
An EOQ Example
Determine the optimal number of units to order
D = 1,000 units per year
S = $10 per order
H = $.50 per unit per year

2DS
Q* =
H

* 2(1,000)(10)
Q = = 40,000 = 200 units
0.50
An EOQ Example
Determine expected number of orders
D = 1,000 units Q* = 200 units
S = $10 per order
H = $.50 per unit per year

Expected Demand D
number of = N = = *
orders Order quantity Q

1,000
N= = 5 orders per year
200
An EOQ Example
Determine optimal time between orders
D = 1,000 units Q* = 200 units
S = $10 per order N = 5 orders/year
H = $.50 per unit per year

Expected time Number of working days per year

between =T =
orders Expected number of orders

250
T= = 50 days between orders
5
 An EOQ Example
Determine the total annual cost
D = 1,000 units Q* = 200 units
S = $10 per order N = 5 orders/year
H = $.50 per unit per year T = 50 days
Total annual cost = Setup cost + Holding cost
D Q
TC = S+ H
Q 2
1,000 200
= ($10) + ($.50)
200 2
= (5)($10) + (100)($.50)
= $50 + $50 = $100
Note: the cost of materials is not included, as it is assumed that the demand will
be satisfied and therefore it is a fixed cost
Calculating EOQ
EXAMPLE 1
A museum of natural history opened a gift shop which operates
52 weeks per year. Managing inventories has become a
problem. Top-selling SKU is a bird feeder. Sales are 18 units
per week, the supplier charges $60 per unit. Ordering cost
is $45. Annual holding cost is 25 percent of a feeders value.
Management chose a 390-unit lot size.

What is the annual cycle-inventory cost of the current policy of

using a 390-unit lot size?

Would a lot size of 468 be better?

Calculating EOQ
SOLUTION
We begin by computing the annual demand and holding cost as
D = (18 units/week)(52 weeks/year) = 936 units
H = 0.25($60/unit) = $15
The total annual cycle-inventory cost for the current policy is
Q D 390 936
C = 2 (H) + Q (S) = ($15) + ($45)
2 390
= $2,925 + $108 = $3,033

The total annual cycle-inventory cost for the alternative lot size is

468 936
C= ($15) + ($45) = $3,510 + $90 = $3,600
2 468
Calculating the EOQ
Current Total Annual Cycle-Inventory Cost Function for
cost the Bird Feeder

3000
Total Q D
Annual cost (dollars)

cost = (H) + (S)

2 Q

2000
Q
Holding cost = (H)
2

1000
D
Ordering cost = (S)
Lowest Q
cost
0 | | | | | | | |
50 100 150 200 250 300 350 400
Lot Size (Q)
Best Q Current
(EOQ) Q
Finding the EOQ, Total Cost, TBO
EXAMPLE 2
For the bird feeders in Example 1, calculate the EOQ and its
total annual cycle-inventory cost. How frequently will orders be
placed if the EOQ is used?

SOLUTION
Using the formulas for EOQ and annual cost, we get

2DS 2(936)(45)
EOQ = = = 74.94 or 75 units
H 15
Finding the EOQ, Total Cost, TBO

The total annual cost is much less than the $3,033 cost of the
current policy of placing 390-unit orders.
Finding the EOQ, Total Cost, TBO
When the EOQ is used, the TBO can be expressed in various
ways for the same time period.

EOQ 75
TBOEOQ = D = = 0.080 year
936

EOQ 75
TBOEOQ = D (12 months/year) = (12) = 0.96 month
936

EOQ 75
TBOEOQ = D (52 weeks/year) = (52) = 4.17 weeks
936

EOQ 75
TBOEOQ = (365 days/year) = (365) = 29.25 days
D 936
Finding the EOQ, Total Cost, TBO
EXAMPLE 3
Suppose that you are reviewing the inventory policies on an $80
item stocked at a hardware store. The current policy is to replenish
inventory by ordering in lots of 360 units. Additional information is:

D = 60 units per week, or 3,120 units per year

S = $30 per order
H = 25% of selling price, or $20 per unit per year

What is the EOQ?

SOLUTION
2DS 2(3,120)(30)
EOQ = = = 97 units
H 20
Finding the EOQ, Total Cost, TBO
What is the total annual cost of the current policy (Q = 360),
and how does it compare with the cost with using the EOQ?
Current Policy EOQ Policy

Q = 360 units Q = 97 units

C = (360/2)(20) + (3,120/360)(30) C = (97/2)(20) + (3,120/97)(30)
C = 3,600 + 260 C = 970 + 965
C = $3,860 C = $1,935
Finding the EOQ, Total Cost, TBO

What is the time between orders (TBO) for the current policy
and the EOQ policy, expressed in weeks?

360
TBO360 = (52 weeks per year) = 6 weeks
3,120

97
TBOEOQ = (52 weeks per year) = 1.6 weeks
3,120
 Managerial Insights

SENSITIVITY ANALYSIS OF THE EOQ

Parameter EOQ Parameter EOQ Comments
Change Change

Increase in lot size is in proportion

Demand 2DS
H
to the square root of D.
Weeks of supply decreases and
Order/Setup
Costs
2DS
H
inventory turnover increases because
the lot size decreases.

Holding Larger lots are justified when holding

Costs
2DS
H
costs decrease.
Robustness
The EOQ model is robust
It works even if all parameters and
assumptions are not met
The total cost curve is relatively flat in the
area of the EOQ
 Introducing delivery lag
EOQ answers the how much question
The reorder point (ROP) tells when to order

Lead time (L) is the time between placing and

receiving an order
Demand Lead time for a new
ROP = per day order in days

=dxL
D
d=
Number of working days in a year
Reorder Point Curve

Q*
Resupply takes place as order arrives
Inventory level (units)

Slope = units/day = d

ROP
(units)

Time (days)
Lead time = L
Reorder Point Example
Demand = 8,000 iPods per year
250 working day year
Lead time for orders is 3 working days, but it may also
take 4 days
D
d=
Number of working days in a year
= 8,000/250 = 32 units

ROP = d x L
= 32 units per day x 3 days = 96 units
= 32 units per day x 4 days = 128 units
 Introducing volume discounts
A company buys re-writable DVDs (10 disks / box)
from a large mail-order distributor
The company uses approximately 5,000 boxes / year
at a fairly constant rate
The distributor offers the following quantity discount

schedule:
If <500 boxes are ordered, then cost = $10/box
If >500 but <800 boxes are ordered, then cost = $9.50
If >800 boxes are ordered, then cost = $9.25
Fixed cost of purchasing = $25, and the cost of
capital = 12% per year. There is no storage cost.
Introducing volume discounts
Solve 3 EOQ models
Each one will hold for the corresponding region; if it does not
correspond, choose the lowest one that does
Select the one with the lowest cost

Steps in analyzing a quantity discount

1. For each discount, calculate Q*
2. If Q* for a discount doesnt qualify, choose the lowest possible
quantity to get the discount
3. Compute the total cost for each Q* or adjusted value from
Step 2
4. Select the Q* that gives the lowest total cost
Quantity Discount Models
Total cost curve for discount 2
Total cost
curve for
discount 1
Total cost $

Total cost curve for discount 3

b
a Q* for discount 2 is below the allowable range at point a and
must be adjusted upward to 500 units at point b

1st price 2nd price

break break

0 500 800
Order quantity
 Allowing shortages
A company is a mail-order distributor of audio CDs
They sell about 50,000 CDs / year
Each CD is packaged in a jewel box they buy from a supplier
Fixed cost for an order of boxes = $100; variable cost = $0.50,
storage cost = $0.50/unit/year, and cost of money is 10%
The company assumes that shortages are allowed, and lost demand
is backlogged it just gets to the customer a little later (!)
The company assigns a penalty of $1 for every week that a box
is delivered late, so annual shortage cost (penalty) p = $52/unit.
 Allowing shortages
Inventory Level
Q-b

(Q-b)/D b/D Time

b
Q/D

Allow shortages up to b units Reorder period = Q/D

Order quantity Q Period with I>0 = (Q-b)/D
Order-up-to inventory = Q-b Period with I<0 = b/D
 Allowing shortages
Total Annual Cost = Ordering + Shortage + Holding costs
Ordering cost = N [(S+CQ) + (pb(b/D)/2) + H(Q-b)((Q-b)/D)/2]
Since N = D/Q, we have
Total Annual Cost = DS/Q + CD + (pb2/2Q) +H(Q-b)2/2D

To minimize, take derivative = 0, and solve

hQ2 HbQ (DS+pb2) = 0
 Production Order Quantity Model
1. Used when inventory builds up over a period
of time after an order is placed
2. Used when units are produced and sold
simultaneously

Part of inventory cycle during which

Inventory level

production (and usage) is taking place

Demand part of cycle with
no production (only usage)
Maximum
inventory

t Time
Production Order Quantity Model

Q = Number of pieces per order p = Daily production rate

H = Holding cost per unit per year d = Daily demand/usage rate
t = Length of the production run in days

Annual inventory = (Average inventory level) x Holding cost

holding cost per unit per year

Average = (Maximum inventory level)/2

inventory level

Maximum = Total items produced Total items used

inventory level during the production during the
run production run
= pt dt
Production Order Quantity Model
Q = Number of pieces per order p = Daily production rate
H = Holding cost per unit per year d = Daily demand/usage rate
t = Length of the production run in days

Maximum = Total produced during Total used during

inventory level the production run the production run

= pt dt
However, Q = total produced = pt ; thus t = Q/p
Maximum Q Q d
inventory level = p p d
p
=Q 1
p

Maximum inventory level Q d

Holding cost = (H) = 1 H
2 2 p
Production Order Quantity Model
Q = Number of pieces per order p = Daily production rate
H = Holding cost per unit per year d = Daily demand/usage rate
t = Length of the production run in days

Setup cost = (D / Q)S

Holding cost = 21 HQ "#1 d p $% ( )
D
S = 21 HQ "#1 d p $%
( )
Q Remember,
with no
2DS
Q2 = production
H "#1 d p $%
( ) taking place

2DS
Q *p =
H "#1 d p $%
( ) 2DS
Q* =
H
Production Order Quantity Example

D =1,000 units p = 8 units per day

S =$10 d = 4 units per day
H =$0.50 per unit per year

2DS
Q*p =
H "#1 ( d p)$%

* 2(1, 000)(10)
Q =
p
0.50 [1 (4 8)]
20, 000
= = 80, 000
0.50(1 2)
= 282.8 units, or 283 units
Production Order Quantity Model
Note:
D 1,000
d=4= =
Number of days the plant is in operation 250

When annual data are used the equation becomes

* 2DS
Q =
p " Annual demand rate %
H $1 '
# Annual production rate &
Probabilistic Models and Safety Stock
 Probabilistic Models and
Safety Stock
Demand is often UNCERTAIN
The problem appears when there is LEAD TIME, L
We have to set two parameters that define our ordering
policy: Reorder Point (ROP) and Safety Stock (ss)
You reorder when your inventory falls on or below ROP
Use safety stock to achieve a desired service level and
avoid stockouts
ROP = d x L + ss
Expected Annual stockout costs = (expected units short/ cycle)
x the stockout cost/unit x the number of orders per year
 Safety Stock Example
Current policy:
ROP = 50 units Stockout cost = $40 / unit
Orders per year = 6 Carrying cost = $5 / unit / year
Probability distribution for inventory demand during lead time
NUMBER OF UNITS
PROBABILITY
(d X L)
30 .2
40 .2
Current ROP 50 .3
60 .2
70 .1
1.0

How much safety stock should we keep and added to 50 (current ROP)?
Safety Stock Example
ROP = 50 units Stockout cost = $40 /unit
Orders /year = 6 Carrying cost = $5 /unit/ year
SAFETY ADDITIONAL TOTAL
STOCK HOLDING COST STOCKOUT COST COST

20 (20)($5) = $100 $0 $100

10 (10)($5) = $ 50 (10)(.1)($40)(6) = $240 $290

0 $ 0 (10)(.2)($40)(6) + (20)(.1)($40)(6) = $960 $960

A safety stock of 20 units gives the lowest total cost

ROP = 50 + 20 = 70 frames
Probabilistic Demand
Use prescribed service levels to set safety stock
when the cost of stockouts cannot be determined

ROP = demand during lead time + ZdLT

Where:
Z = Number of standard deviations
dLT = Standard deviation of demand during lead time
Probabilistic Demand

Probability of Risk of a stockout

no stockout (5% of area of
95% of the time normal curve)

Mean ROP = ? kits Quantity

demand
350
Safety
stock
0 z
Number of
standard deviations
Probabilistic Demand
= Average demand = 350 kits
dLT = Standard deviation of demand during lead time
= 10 kits
Z = 5% stockout policy (service level = 95%)
Using Normal distribution tables, for an area
under the curve of 95%, the Z = 1.65
Safety stock = ZdLT = 1.65(10) = 16.5 kits

Reorder point = Expected demand during lead time +

Safety stock
= 350 kits + 16.5 kits of safety stock
= 366.5 or 367 kits
Probabilistic Demand

Minimum demand during lead time

Inventory level

Maximum demand during lead time

Mean demand during lead time

ROP = 350 + safety stock of 16.5 = 366.5

ROP
Normal distribution probability of
demand during lead time
Expected demand during lead time (350 kits)

Safety stock 16.5 units

0 Place Lead
time Receive Time
order order
Other Probabilistic Models
When data on demand during lead time is not
available, there are other models available
1. When demand is variable and lead time is
constant
2. When lead time is variable and demand is
constant
3. When both demand and lead time are variable
Other Probabilistic Models:
Variable demand, constant lead time
Demand is variable and lead time is constant

ROP = (Average daily demand

x Lead time in days) + ZdLT

where dLT = d Lead time

d = standard deviation of demand per day
Other Probabilistic Models:
Variable demand, constant lead time
Average daily demand (normally distributed) = 15
Lead time in days (constant) = 2
Standard deviation of daily demand = 5
Service level = 90%
Z for 90% = 1.28
From Appendix I

ROP = (15 units x 2 days) + ZdLT

= 30 + 1.28(5)( 2)
= 30 + 9.02 = 39.02 39

Safety stock is about 9 computers

Other Probabilistic Models:
Constant demand, variable lead time

ROP = (Daily demand x Average lead time in days)

+ Z x (Daily demand) x LT

where LT = Standard deviation of lead time in days

Other Probabilistic Models:
Constant demand, variable lead time
Daily demand (constant) = 10
Average lead time = 6 days
Standard deviation of lead time = LT = 1
Service level = 98%, so Z (from Appendix I) = 2.055

ROP = (10 units x 6 days) + 2.055(10 units)(1)

= 60 + 20.55 = 80.55

Reorder point is about 81 cameras

Other Probabilistic Models:
Variable demand, variable lead time

ROP = (Average daily demand x Average lead time)

+ ZdLT

where d = Standard deviation of demand per day

LT = Standard deviation of lead time in days
dLT = (Average lead time x d2)
+ (Average daily demand)22LT
Other Probabilistic Models:
Variable demand, variable lead time
Average daily demand (normally distributed) = 150
Standard deviation = d = 16
Average lead time 5 days (normally distributed)
Standard deviation = LT = 1 day
Service level = 95%, so Z = 1.65 (from Normal tables)

ROP = (150 packs 5 days) +1.65 dLT

dLT = ( ) ( ) (5 256) + (22,500 1)
5 days 162 + 1502 12 =

= (1,280) + (22,500) = 23,780 154

ROP = (150 5) +1.65(154) 750 + 254 = 1,004 packs
Inventory Control Systems
Inventory Control Systems
Continuous review (Q) system
Reorder point system (ROP) and fixed order
quantity system
For independent demand items (i.i.d.)

Tracks inventory position (IP)

Includes scheduled receipts (SR), on-hand inventory

(OH), and back orders (BO)
Inventory position = On-hand inventory + Scheduled receipts
Backorders

IP = OH + SR BO
Selecting the Reorder Point
IP IP IP
Order Order Order Order
received received received received
On-hand inventory

Q Q Q

OH OH OH
R
Order Order Order
placed placed placed

L L L Time
TBO TBO TBO

Q System When Demand and Lead Time Are Constant and Certain
Continuous Review Systems
The on-hand inventory is only 10 units, and the reorder
point R is 100. There are no backorders, but there is
one open order for 200 units. Should a new order be
placed?
SOLUTION
IP = OH + SR BO = 10 + 200 0 = 210
R = 100

Decision: Do not place a new order

Continuous Review Systems
Reorder Point Level:
Assuming that the demand rate per period and the
lead time are constant, the level of inventory at
which a new order is placed (reorder point) can be
calculated as follows:
R = dL

Where
d = demand rate per period
L = lead time

Remember: The order quantity Q is the EOQ!

Continuous Review Systems
EXAMPLE 4
Demand for chicken soup at a supermarket is always 25 cases
a day and the lead time is always 4 days. The shelves were just
restocked with chicken soup, leaving an on-hand inventory of
only 10 cases. No backorders currently exist, but there is one
open order in the pipeline for 200 cases. What is the inventory
position? Should a new order be placed?
SOLUTION
R = Total demand during lead time = (25)(4) = 100 cases
IP = OH + SR BO
= 10 + 200 0 = 210 cases

Decision: Do not place a new order

Continuous Review Systems

Selecting the reorder point with variable demand

and constant lead time
Reorder point = Average demand during lead time
+ Safety stock
= dL + safety stock

where
d = average demand per week (or day or months)
L = constant lead time in weeks (or days or months)
Continuous Review Systems
(uncertain demand)
IP IP IP
Order
Order Order
received
received received
Order
On-hand inventory

received

Q Q Q

R
Order Order Order
placed placed placed

0
L1 L2 L3 Time
TBO1 TBO2 TBO3

Q System When Demand Is Uncertain

How to determine the Reorder Point

1. Choose an appropriate service-level policy

Select service level or cycle service level
Protection interval
2. Determine the demand during lead time
probability distribution
3. Determine the safety stock and reorder point
levels
Demand During Lead Time
Specify mean d and standard deviation d for the demand
(typically these values are given)
Calculate standard deviation of demand during lead time L
dLT = d2L = d L
Then, the safety stock and reorder point are

Safety stock = zdLT

where
z = number of standard deviations needed to achieve the
cycle-service level (found from tables)
dLT = stand deviation of demand during lead time

Reorder point = R = dL + safety stock

Demand During Lead Time

Cycle-service level = 85%

Probability of stockout
(1.0 0.85 = 0.15)
Average
demand
during
lead time R

zdLT

Finding Safety Stock with a Normal Probability Distribution for an 85 Percent

Cycle-Service Level
Reorder Point for Variable Demand
EXAMPLE 5
Let us return to the bird feeder in Example 2.
The EOQ is 75 units.
Suppose that the average demand is 18 units per week with
a standard deviation of 5 units.
The lead time is constant at two weeks.
Determine the safety stock and reorder point if
management wants a 90 percent cycle-service level.
Reorder Point for Variable Demand
SOLUTION
In this case, d = 5, d = 18 units, and L = 2 weeks, so
dLT = d L = 5 2 = 7.07. Consult the body of the table in the
Normal Distribution appendix for 0.9000, which corresponds to
a 90 percent cycle-service level. The closest number is 0.8997,
which corresponds to 1.2 in the row heading and 0.08 in the
column heading. Adding these values gives a z value of 1.28.
With this information, we calculate the safety stock and reorder
point as follows:

Safety stock = zdLT = 1.28(7.07) = 9.05 or 9 units

Reorder point = dL + Safety stock = 2(18) + 9 = 45 units

Reorder Point for Variable Demand
EXAMPLE 6
Suppose that the demand during lead time is normally distributed with an
average of 85 and dLT = 40. Find the safety stock, and reorder point R,
for a 95 and 85 percent cycle-service level.
SOLUTION
Safety stock = zdLT = 1.645(40) = 65.8 or 66 units
R = Average demand during lead time + Safety stock
R = 85 + 66 = 151 units

Find the safety stock, and reorder point R, for an 85 percent

cycle-service level.
Safety stock = zdLT = 1.04(40) = 41.6 or 42 units
R = Average demand during lead time + Safety stock
R = 85 + 42 = 127 units
Reorder Point for Variable Demand &
Variable Lead Time
Often the case that both are variable
The equations are more complicated
Safety stock = zdLT
R =(Average weekly demand Average lead time)
+ Safety stock
= dL + Safety stock
where
d = Average weekly (or daily or monthly) demand
L = Average lead time
d = Standard deviation of weekly (or daily or monthly) demand
LT = Standard deviation of the lead time
dLT = Ld2 + d2LT2
Reorder Point for Variable Demand
& Variable Lead Time
EXAMPLE 7
The Office Supply Shop estimates that the average demand for
a popular ball-point pen is 12,000 pens per week with a
standard deviation of 3,000 pens. The current inventory policy
calls for replenishment orders of 156,000 pens. The average
lead time from the distributor is 5 weeks, with a standard
deviation of 2 weeks. If management wants a 95 percent cycle-
service level, what should the reorder point be?
Reorder Point for Variable Demand
& Variable Lead Time
SOLUTION
We have d = 12,000 pens, d = 3,000 pens, L = 5 weeks,
and LT = 2 weeks

dLT = Ld2 + d2LT2 = (5)(3,000)2 + (12,000)2(2)2

= 24,919.87 pens
From the Normal Distribution appendix for 0.9500, the
appropriate z value = 1.65. We calculate the safety stock and
reorder point as follows:

Safety stock = zdLT = (1.65)(24,919.87)

= 41,117.79 or 41,118 pens

Reorder point = dL + Safety stock = (12,000)(5) + 41.118

= 101,118 pens
Reorder Point for Variable Demand
& Variable Lead Time
EXAMPLE 8
Grey Wolf lodge is a popular 500-room hotel in the North
Woods. Managers need to keep close tabs on all of the room
service items, including a special pint-scented bar soap. The
daily demand for the soap is 275 bars, with a standard
deviation of 30 bars. Ordering cost is $10 and the inventory
holding cost is $0.30/bar/year. The lead time from the supplier
is 5 days, with a standard deviation of 1 day. The lodge is
open 365 days a year.
What should the reorder point be for the bar of soap if
management wants to have a 99 percent cycle-service?
Reorder Point for Variable Demand
& Variable Lead Time
SOLUTION
d = 275 bars
L = 5 days
d = 30 bars
LT = 1 day
dLT = Ld2 + d2LT2 = 283.06 bars

From the Normal Distribution appendix for 0.9900, z = 2.33.

We calculate the safety stock and reorder point as follows;

Safety stock = zdLT = (2.33)(283.06) = 659.53 or 660 bars

Reorder point + safety stock = dL + safety stock
= (275)(5) + 660 = 2,035 bars
 Periodic Review (or fixed period)
System (P)
Fixed interval reorder system or periodic reorder system
Four of the original EOQ assumptions maintained
No constraints are placed on lot size
Holding and ordering costs
Independent demand
Lead times are certain
Order is placed to bring the inventory position up to the target
inventory level, T, when the predetermined time, P, has elapsed
Only relevant costs are ordering and holding
Lead times are known and constant
Items are independent of one another
Periodic Review Systems
Target quantity (T)

Q4
Q2
On-hand inventory

Q1 P
Q3

Time
Periodic Review Systems
Inventory is only counted at each review
period
May be scheduled at convenient times
Appropriate in routine situations
May result in stockouts between periods
May require increased safety stock
Periodic Review System (P)
T
IP IP IP
Order Order Order
received received received
On-hand inventory

Q1 Q3
OH Q2 OH
IP1

IP3
Order Order
placed placed
IP2

L L L Time
P P
Protection interval

P System When Demand Is Uncertain

 How Much to Order in a P System
EXAMPLE 9
A distribution center has a backorder for five 36-inch color TV sets.
No inventory is currently on hand, and now is the time to review. How
many should be reordered if T = 400 and no receipts are scheduled?
SOLUTION
IP = OH + SR BO
= 0 + 0 5 = 5 sets
T IP = 400 (5) = 405 sets

That is, 405 sets must be ordered to bring the inventory

position up to T sets.
How Much to Order in a P System
EXAMPLE 10
The on-hand inventory is 10 units, and T is 400. There are no
back orders, but one scheduled receipt of 200 units. Now is the
time to review. How much should be reordered?
SOLUTION
IP = OH + SR BO
= 10 + 200 0 = 210

T IP = 400 210 = 190

The decision is to order 190 units

Periodic Review System
Selecting the time between reviews, choosing P and T
Selecting T when demand is variable and lead time is
constant
IP covers demand over a protection interval of P + L

The average demand during the protection interval is

d(P + L), or

T = d(P + L) + safety stock for protection interval

Safety stock = zP + L , where P + L = d P + L

Calculating P and T

EXAMPLE 11
Again, let us return to the bird feeder example. Recall that
demand for the bird feeder is normally distributed with a mean
of 18 units per week and a standard deviation in weekly
demand of 5 units. The lead time is 2 weeks, and the business
operates 52 weeks per year. The Q system developed in
Example 5 called for an EOQ of 75 units and a safety stock of
9 units for a cycle-service level of 90 percent. What is the
equivalent P system? Answers are to be rounded to the nearest
integer.
Calculating P and T
SOLUTION
We first define D and then P. Here, P is the time between
reviews, expressed in weeks because the data are expressed
as demand per week:
D = (18 units/week)(52 weeks/year) = 936 units

EOQ 75
P= (52) = (52) = 4.2 or 4 weeks
D 936

With d = 18 units per week, an alternative approach is to calculate P

by dividing the EOQ by d to get 75/18 = 4.2 or 4 weeks. Either way,
we would review the bird feeder inventory every 4 weeks.
 Calculating P and T

We now find the standard deviation of demand over the protection

interval (P + L) = 6:
P+ L = d P + L = 5 6 = 12.25 units

Before calculating T, we also need a z value. For a 90 percent

cycle-service level z = 1.28. The safety stock becomes

Safety stock = zP + L = 1.28(12.25) = 15.68 or 16 units

We now solve for T:

T = Average demand during the protection interval + Safety stock

= d(P + L) + safety stock
= (18 units/week)(6 weeks) + 16 units = 124 units
Comparative Advantages

Primary advantages of P systems

Convenient

Orders can be combined

Only need to know IP when review is made

Primary advantages of Q systems

Review frequency may be individualized
Fixed lot sizes can result in quantity discounts
Lower safety stocks
Single Period Model
Single-Period Model
Only one order is placed for a product
Units have little or no value at the end of the sales
period
Newsboy problem 1:
Demand = 500 papers/day, = 100 paper
Cost to newsboy = 10c, Sales price = 30c
How many papers should he order?
The newsboy problem
If he sells a paper, he makes a profit = 20c
If he doesnt sell a paper he makes a loss = 10c
If he orders x papers, on the i-th paper he makes
an expected profit:
E(Profit)i = 20p 10(1-p), where p=sale of i-th paper
Breakeven occurs when the Expected profit = 0
So, 20p 10(1-p) = 0, and therefore p* = 1/3
By looking at the Normal tables, Z = .431
p
1-p

= 500
Optimal stocking level
The newsboy problem

1-p

= 500

Optimal stocking level

Therefore:
Z = .431 = (X*-)/ = (X*-500)/100
X* = 543
 A small variation
Assume that he can return the paper, if unsold for 5c each
E(Profit)i = 20p 5(1-p), where p=sale of i-th paper
Breakeven occurs when the Expected profit = 0
So, 20p 5(1-p) = 0, and therefore p* = 1/5
By looking at the Normal tables, Z = .842
Then, X* = 584
In general, if MR = Marginal Return and ML = Marginal
Loss, then p MR - (1-p) ML = 0
p* = ML/(MR+ML)
Using Simulation for stochastic inventory
management
 What is Simulation ?
Simulation is a model computer code imitating
the operation of a real system in the computer.
It consists of:
a) A set of variables representing the basic features
of the real system and
b) A set of logical commands in the computer that
modify these features as a function of time in
accordance with the rules (logical of physical)
regulating the real system.
Main Features of a Simulation
System
The capacity to "advance time" through the use of a
simulation build-in clock that monitors and the events while
stepping up real time
The capacity of drawing samples through the creation of
artificial observations that behave "like" random events in
the real system
a) Creation of random numbers (independent & uniformly
distributed) by the computer (according to an internal algorithm -
function)
Conversion in the observations distribution
Examples of applications
Very important tool for

l Service management - Analysis of queuing systems

l Business Process Reengineering
l Strategic planning
l Financial planning
l Industrial design (e.g. chemical plants)
l Short term production planning
l Quality and reliability control,
l Training business games, etc
Creating random numbers according
to a probability distribution

Let us suppose that the weekly demand of a

product can take the following values with the
respective probabilities:

Probability of Demand Cumulative Distribution of

Demand (D)
P(D) Demand F(D)
1,000 0.20 0.20
1,500 0.10 0.30
2,000 0.30 0.60
2,500 0.25 0.85
3,000 0.15 1.00
 The cumulative probability
distribution F
1.00

0.85

0.60

R1
.,30

0.20

R2
D2 D1
1.000 1,500 2,000 2,500 3,000
 Create a "demand

Example
series corresponding
to the random numbers
generated

Remember, the
Cumulative
Distribution F Unif. Random Week's
Week
Number Demand
1 32 2,000
2 8 1,000
3 46 2,000
4 92 3,000
Cumulative Distribution of
Demand (D) 5 69 2,500
Demand F(D)
1,000 0.20 6 71 2,500
1,500 0.30
7 29 1,500
2,000 0.60
2,500 0.85 8 46 2,000
3,000 1.00 9 80 2,500
10 14 1,000
Using Simulation to define an
Inventory Policy
l Assume that a company is interested to implement an (s, S)
ordering policy. Determine values of s and S:
l s = Safety stock S = Order-up-to quantity

Inventory Behavior (s = 200, S = 700)

800
700
Inventory levels

600

Demand
500
Series2
400
Series1
300
200
100
0
1 2 3 4 5 6 7 8 9 10 11 12 13
Weeks
 Application to our problem
Probability of Demand Cumulative Distribution of
Demand (D)
P(D) Demand F(D)
1,000 0.20 0.20
1,500 0.10 0.30
2,000 0.30 0.60
2,500 0.25 0.85
3,000 0.15 1.00

Key assumptions:
When < the safety stock s, order up to the reorder point S
When short, make emergency order for quantity short
Normal ordering cost = 200 + 10 quantity
Emergency ordering cost = 500 + 15 quantity
Leftover inventory cost = 3 quantity
 Flow chart
START OF SIMULATION
Define initial conditions
Define strategy parameters s, S

Next week, t = t+1

Simulation complete? STOP

Order amount S-I

Update inventory to S Compute Average
Pay ordering cost Need to order?
Weekly cost
Update total cost Compute other
performance criteria
Create demand D for this week

Update inventory level Place emergency order

Pay holding cost Update inventory level
Is demand satisfied? Pay emergency cost
Update total cost
Update total cost
 Manual Simulation of Inventory
System
l Assume: s = 1,500 and S = 2,500

Starting Need to Size of Available Week's Emergency Size of emerg. Ending Weekly
Week
Inventory order? order inventory Demand order? order Inventory Cost
1 2,000 no 0 2,000 2,000 no 0 0 0
2 0 yes 2,500 2,500 2,000 no 0 500 26,700
3 500 yes 2,000 2,500 3,000 yes 500 0 30,500
4 0 yes 2,500 2,500 1,000 no 0 1,500 29,700
5 1,500 yes 1,000 2,500 2,500 no 0 0 10,200
6 0 yes 2,500 2,500 2,500 no 0 0 25,200
7 0 yes 2,500 2,500 3,000 yes 500 0 38,000
8 0 yes 2,500 2,500 1,500 no 0 1,000 28,200
9 1,000 yes 1,500 2,500 2,000 no 0 500 16,700
10 500 yes 2,000 2,500 2,500 no 0 0 20,200
Average 389 always 2,111 2,500 2,200 20% 500 389 25,044