You are on page 1of 72

1

Inventory Management
2

OBJECTIVES

Basics of Inventory
Purposes of inventory, inventory-related costs
ABC Analysis
Physical Inventory and Cycle Counting

Economic Order Quantity

Inventory Control Systems


Continuous Review System
Periodic Review System
Single-Period Inventory Decision
3

Inventory

An inventory is a stock of any item or


resource used in an organization

Raw materials Semi-finished goods Finished goods


or work-in-process
4

Inventories in a Bread-Making
Process
Work-in-process Work-in-process
Raw Semi-finished
materials Bread- goods
Packing
making

Flour Bread
Ingredients before packing

Finished
goods

Packed bread
ready for shipment
5

Purposes of Inventory

To meet customer demands from stock


To smooth production requirements
e.g.) Seasonal products, fluctuating customer orders
To protect against stockouts
Demand uncertainty, supply uncertainty
To take advantage of economic lot size
Fixed costs spread over more units, quantity discounts
To protect against loss from price increases
Purchase more than needed in anticipation of price
increases
6

Inventory-related Costs

Acquisition costs
Purchase costs, variable production costs
Inventory-holding (or -carrying) costs
Costs of capital, physical storage costs
Production setup costs
Costs of setting up production tools and equipment
Ordering costs
Costs of placing and receiving an order
Shortage costs
Lost profit, loss of goodwill
7

Terminology

On-hand inventory: Physical inventory held in stock


Inventory level
= On-hand inventory Backorders
Backorders: customer orders that have been received
but not yet shipped because of stockouts
Inventory position
= Inventory level + Scheduled receipts
= On-hand inventory Backorders + Scheduled receipts
Scheduled receipts: purchase or production orders that
have been placed but not yet received
8

Terminology (Contd)

Stock keeping unit (SKU): Individual inventory


item that has an identifying code
Cycle-service level (or service level): Probability
of not running out of stock in any one ordering
cycle
An ordering cycle begins at the time an order is
received and ends right before the next order is
received
Safety stock: Inventory held in excess of expected
demand
To protect against demand and supply uncertainty
9

Terminology (Contd)

Lot size (or order quantity): Quantity ordered or


produced
Cycle inventory: Portion of total inventory that
varies directly with lot size
Inventory
level Cycle inventory

Lot size

Safety stock inventory


Time
10

ABC Analysis

ABC analysis is the process of dividing


SKUs into three classes according to their
dollar usages, so that managers can focus
on items that have the highest dollar usages.
11

ABC Analysis (Contd)


Percentage of dollar volume

100
90
80
70
60
50
40 A B C
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
Percentage of inventory items, SKUs
12

ABC Analysis and Inventory Control

Classification of items
List up all the inventory items according to their
dollar usages
Classify roughly top 20% of the items into A
class, the next 30% into B class, and the
remaining 50% into C class
Use of different inventory control methods
for each class
e.g.) frequent inventory review and highly accurate
inventory records for A class items
13

Example of ABC Analysis

Suppose we have 10 items whose unit costs and


monthly sales are as follows:
Monthly
Unit cost Sales
Inventory Item ($) (units)

Cameras 200 40
CDs 10 400
Computers 2500 30
Displays 250 40
Home Theater 5000 30
Refrigerators 1000 15
Software 50 100
Speakers 150 60
Television sets 400 60
Thumb drives 5 1000
14

Example of ABC Analysis (Contd)

Calculate the dollar usage per month of each item:


Monthly Dollar Usage
Unit cost Sales Per Month
Inventory Item ($) (units) ($)

Cameras 200 40 8,000


CDs 10 400 4,000
Computers 2500 30 75,000
Displays 250 40 10,000
Home Theater 5000 30 150,000
Refrigerators 1000 15 15,000
Software 50 100 5,000
Speakers 150 60 9,000
Television sets 400 60 24,000
Thumb drives 5 1000 5,000
15

Example of ABC Analysis (Contd)

List up the items according to their dollar usages


and classify them into A, B, and C classes:
Monthly Dollar Usage
Unit cost Sales Per Month Percent of
Inventory Item ($) (units) ($) SKUs Class

Home Theater 5000 30 150,000 20 A


Computers 2500 30 75,000
Television sets 400 60 24,000
Refrigerators 1000 15 15,000 30 B
Displays 250 40 10,000
Speakers 150 60 9,000
Cameras 200 40 8,000
Software 50 100 5,000 50 C
Thumb drives 5 1000 5,000
CDs 10 400 4,000
16

Physical Inventory and Cycle Counting

Inventory accuracy refers to how well the


inventory records agree with physical inventories
Cycle Counting is a physical inventory-counting
technique in which inventory is counted on a
frequent basis rather than once or twice a year
Procedure of cycle counting
1. Set rules for choosing items to be counted
2. Conduct a cycle inventory count on a cyclic schedule
1) Run the cycle counting program to get the list of items
to be counted
2) Count inventory for the items in the list and update their
inventory records
17

Physical Inventory and Cycle Counting

Examples of rules for choosing items to be


counted
Count A class items every two months, B class
items every 4 months, and C class items every 6
months
Choose items whose inventory is lower than 10 units
in the inventory records
Choose items that show positive inventory records but
also have backorders
18

Inventory Control

How often to review the inventory status?


When to order?

How much to order?

18-18
19

Inventory Models and Control


Systems

Inventory models
EOQ (Economic Order Quantity) model
Single-period inventory model
Inventory control systems
Continuous review system
Also referred to as Q system or (Q, R) system
Periodic review system
Also referred to as P system
20

EOQ Model: Assumptions

Inventory is monitored continuously and orders


can be placed at any time
Demand for the product is known and occurs at a
constant rate
Leadtime (time from order to receipt) is constant
Ordering costs and inventory holding costs are
incurred
No shortage is allowed
The objective is to minimize annual total cost
21

EOQ Model: Form of Optimal


Ordering Policy

Form of optimal ordering policy


Place an order of fixed quantity whenever
the inventory position drops to a fixed reorder
point
Two decision variables
Reorder point (R): When to order
Order quantity (Q): How much to order
22

EOQ Model:
Optimal Reorder Point

R = demand during leadtime = d L

_
d = Average demand per unit time (e.g., day, week)
L = Lead time in unit time
23

EOQ Model:
Example of Optimal Reorder Point

Suppose that daily demand is 10 units, and


leadtime is 5 days.

Demand during leadtime = 105 = 50 units


Optimal reorder point is 50 units.

What will happen if R < 50 units?


What will happen if R > 50 units?
24
Behavior of Inventory Level
in EOQ Model
1. You receive an order of size Q. 4. The cycle then repeats.

Inventory
level

Q Q Q

R
Time
L L
2. You start using
them up over time. 3. When inventory reaches
down to reorder point, R,
R = Reorder point you place your next order
Q = Order quantity of size Q.
L = Leadtime
25

Impact of Order Quantity on


Average Inventory and Order Frequency

Large Q high inventory, less frequent orderings


Q

Average
Inventory

Small Q low inventory, more frequent orderings


Q
Average
Inventory
26

Annual Inventory-Holding Costs

Q
Average inventory level =
2
Q
Annual inventory-holding costs = H
2
where H = annual inventory-holding cost per unit

E.g.) If Q=100 units and H=$3 per unit per year,


100
Average inventory level = = 50 units
2
100
Annual inventory-holding costs = 3 = $150
2
27

Annual Ordering Costs


D
Number of orders per year =
Q
D
Annual ordering costs = S
Q
where D = annual demand
S = ordering cost per order

E.g.) If D=1,200 units, Q=100 units, and S=$30 per order,


1,200
Number of orders per year = = 12
100
1,200
Annual ordering costs = 30 = $360
100
28

EOQ Model:
Annual Total Cost

Annual Annual Annual Annual C=Annual total cost


Total = Acquisition + Ordering + Holding D =Annual demand
Cost Cost Cost Cost A =Acquisition cost
per unit
S =Ordering cost
H =Annual holding
cost per unit
D Q
C=DA+ S+ H
Q 2
29

Minimizing Annual Total Cost


D Q
C=DA+ S+ H
Q 2

Cost Annual Total Cost

Annual
Holding Cost
Annual
Acquisition Cost
Annual
Ordering Cost

EOQ Order Quantity (Q)


30

EOQ Formula

2DS 2(Annual Demand)(Ordering Cost)


EOQ = =
H Annual Holding Cost Per Unit
31

EOQ Example 1

Given the information below, what is the economic


order quantity?

Annual Demand = 1,000 units


Cost to place an order = $10
Holding cost per unit per year = $2.50

2DS 2 1,000 10
EOQ = =
H 2.50
= 89.443 or 89 units
32

EOQ Example 2

Nelsons Hardware Store stocks a 19.2 volt


cordless drill. Annual demand is 5,000 units, the
ordering cost is $15 per order, and the annual
inventory holding cost is $4 per unit. The
purchase cost of this inventory item is $20 per
unit. What is the economic order quantity? What
is the annual total cost for this item given the
EOQ?
33

EOQ Example 2: Solution

2DS 2 5,000 15
EOQ = =
H 4
= 193.65 or 194 units

D Q
C=DA+ S+ H
Q 2
5,000 194
= 5,000 20 + 15 + 4
194 2
= $100,774.60
34

Sensitivity Analysis of the EOQ

2DS 2(Annual Demand)(Ordering Cost)


EOQ = =
H Annual Holding Cost Per Unit

Parameter EOQ
Parameter
Change Change

Demand (D)

Ordering cost (S)

Holding cost (H)


35

TBO (Time Between Orders)

TBO (Time Between Orders) is the average


time between receiving (or placing) orders

Inventory Q
TBO =
level D
Q Q Q

Time
TBO
36

TBO (Time Between Orders) (Contd)

If the order quantity is Q, then


Q
TBO =
D
e.g.) Q=100 units, D=20 units per day TBO=5 days

Optimal time between orders


EOQ
Optimal TBO =
D
37

TBO Example 1

Suppose that the weekly demand for a product is


20 units. If Q = 100 units is used, what will be the
time between orders?

100
TBO 5 weeks
20
38

TBO Example 2 (Optimal TBO)

Suppose that the annual demand for a product is


1000 units. If EOQ = 100 units, what is the optimal
time between orders? Assume 52 weeks per year.

EOQ
Optimal TBO
D
100
0.1 year
1000
0.1 year 52 weeks/year
5.2 weeks
39

Continuous Review System

Review the inventory status continuously


Place an order of fixed quantity whenever
the inventory position drops to a fixed
reorder point

* This system can be used when demand for the


product is stationary over time
40

Visual Version of
Continuous Review System

Two-Bin System

As soon as
Bin 2 is empty,
place an order
Bin 1 Bin 2 for Bin 2
41

Continuous Review System:


Decision Variables

Two decision variables


Reorder point (R): When to order
Order quantity (Q): How much to order
Assumptions
Demand for the product is stationary and
independent over time
We have a target cycle-service level, SL
Leadtime is constant
42

Continuous Review System:


Behavior of Inventory Position/Level

Q Q Q

|L| |L| |L|

Reorder point, R, should cover DDLT (Demand During


Leadtime)
43

Continuous Review System:


Reorder Point

Given a target service level SL,


reorder point R should be chosen such that
P(DDLT R) = SL

Distribution of
DDLT

SL

R
44

Distribution of DDLT
when Daily Demand is Normally Distributed

Suppose that:
Demand per unit time (e.g., daily or weekly
demand) follows a normal distribution with
mean d and standard deviation d
Leadtime is L unit times (e.g., days)

Then, DDLT follows a normal distribution with


and standard deviation of L , where
mean dL
L = d L
45

Demand During Leadtime: Example

Suppose that weekly demand follows a normal


distribution with mean 100 units and standard
deviation 15 units.
Suppose that leadtime is 3 weeks.

Average demand during leadtime


= d L = 100 3 = 300 units
Standard deviation of demand during leadtime
L = d L = 15 3 = 15 1.73 = 25.95 units

DDLT is normally distributed with mean 300 units and


standard deviation 25.95 units.
46

Demand During Leadtime: Example


= 15 = 15 = 15

+ + =
100 100 100
Demand for week 1 Demand for week 2 Demand for week 3

= = .

300
Demand for 3-week leadtime
47

Reorder Point
when DDLT is Normally Distributed


If DDLT is normally distributed with mean dL
and standard deviation L ,

= +

where zSL is the z value such that P z zSL = SL

Note that:

dL = average demand during leadtime
zSL L = safety stock
48

Finding
Standard normal distribution

P(z zSL ) = SL

SL

0 zSL
zSL can be found from a standard normal table or by
using Excel function, NORM.S.INV
Example: z.95 = 1.65
49

What if Daily Demand is Not Normally


Distributed?

DDLT is the sum of demands per unit time over


the leadtime. So, according to the central limit
theorem, DDLT approximately follows a normal
distribution

Even if demand per unit time is not normally


distributed, R = d L + zSL L is a good solution
50

Continuous Review System:


Order Quantity

We use EOQ for order quantity as a heuristic

2DS
EOQ =
H
2(Average Annual Demand)(Ordering Cost)
=
Annual Holding Cost Per Unit
51

Continuous Review System:


Determination of Q and R

Use EOQ for order quantity Q


R = average demand during leadtime + safety stock
= d L + zSL L
where:
d = average demand per unit time (e.g., daily demand)
L = leadtime in unit time
zSL = z value corresponding to service level, SL
L = standard deviation of demand during leadtime
52

Example of Continuous Review


System
Daily demand for a certain product is normally
distributed with mean 60 and standard deviation 7.
The source of supply is reliable and maintains a
constant leadtime of 6 days. The cost of placing
an order is $10 and annual holding cost is $0.5
per unit. Assume sales occur over the entire 365
days of the year. Management wants to satisfy a
95% probability of not running out of stock in any
one ordering cycle.
What should be the order quantity (Q) and
reorder point (R)?
53

Example of Continuous Review


System: Order Quantity

D = average annual demand = 60 365 units


S = cost of placing an cost = $10
H = annual holding cost per unit = $0.5

2DS
Q=
H
2 60 365 10
936 units
0.5
54
Example of Continuous Review
System: Reorder Point
d = average daily demand = 60 units
d = standard deviation of daily demand = 7 units
L = leadtime = 6 days
Target service level = 95%
z 0.95 = 1.65
L = d L = 7 6 = 17.15 units
R = d L + z 0.95 L
= 60 6 + 1.65 17.15 = 388 units
So, we order 936 units whenever the inventory
position drops to 388 units.
55

Periodic Review System

Review the inventory status and place


orders periodically (e.g., weekly, monthly)
Order up to a fixed level T at each review
time

* This system can be used when demand for the


product is stationary over time
56

Visual Version of
Periodic Review System

One-Bin System

Check the bin


periodically and
order enough to
refill the bin
57

Periodic Review System:


Decision Variable

Decision variable
Target inventory level (T)
Assumptions
Demand for the product is stationary and
independent over time
We have a target cycle-service level, SL
Leadtime is constant

Note that in this system,


When to order: At each review time
How much to order: T current inventory position
58

Periodic Review System:


Behavior of Inventory Position/Level

L L L

P P

P = review period = time between reviews


Target inventory level, T, should cover demand during
P+L periods
59

Periodic Review System:


Determination of T

T = average demand during (P+L) periods + safety stock


= d (P+L) + zSL P+L
where:
d = average demand per period
P = number of periods between reviews
L = leadtime in periods
zSL = z value corresponding to service level, SL
P+L = standard deviation of demand during P + L periods
60

Example of Periodic Review System

Daily demand for a certain product is normally


distributed with mean 20 and standard deviation 4.
Inventory is reviewed every 30 days, and lead
time is 10 days. Management wants to satisfy a
95% probability of not running out of stock in any
one ordering cycle.

What should be the target inventory level, T?


At a review time, if there are 200 units of inventory,
how many units should be ordered?
61

Example of Periodic Review System:


Solution

d = average daily demand = 20 units


d = standard deviation of daily demand = 4 units
P = number of days between reviews = 30 days
L = leadtime = 10 days
Target service level = 95%

z 0.95 = 1.65
P+L = d P+L = 4 30+10 = 25.30
62

Example of Periodic Review System:


Solution (Contd)

d = average daily demand = 20 units


P = the number of days between reviews = 30 days
L = leadtime = 10 days
z 0.95 = 1.65
P+L = 25.30 units

T = d (P+L) + z 0.95 P+L


= 20(30+10) + 1.65 25.30 = 842

So, the target inventory level (T) is 842 units.


63

Example of Periodic Review System:


Solution (Contd)

T = target inventory level = 842 units


I = current inventory position = 200 units

Order quanity at this review time


= T - I = 842 - 200 = 642

So, at this review time we should order 642 units.


64

Single-Period Inventory Model:


Assumptions

Selling Period

Choose Q Demand realized Leftovers salvaged


Profit realized
Single selling-period
Demand, D, is uncertain, but we know its
probability distribution
Place an order of Q before the start of the
selling period
65

Single-Period Inventory Model:


Assumptions (Contd)

Revenue and cost parameters


Selling price
Purchase cost (or production cost)
Salvage value
The objective is to choose Q to maximize
the expected profit
66

Overage and Underage Costs

The realized demand can be smaller or


larger than order quantity
Overage cost per unit unsold (Co)
= cost of ordering too much
= purchase cost salvage value
Underage cost per unit short (Cu)
= cost of ordering too less
= selling price purchase cost
Selling price purchase cost is an opportunity cost
67

Overage and Underage Costs:


Example

A newsvendor has to order copies of a weekly


magazine to sell in a week. The magazine sells at
$5 per copy, and its purchase cost is $2 per copy.
Any leftovers can be sold to a waste management
company at $0.50
Overage cost per unit unsold
Co = purchase cost salvage value
= $2 $0.50 = $1.50
Underage cost per unit short
Cu = selling price purchase cost
= $5 $2 = $3
68

Optimal Order Quantity

If we assume continuous demand,


The product is divisible (e.g., oil and gas)
Or, we are dealing with a large quantity of the
product
Then the optimal order quantity, Q*, can
be obtained from:
P(D Q*) = Cu / (Cu + Co)
Cu / (Cu + Co) is the optimal service level
Also called the critical fractile or newsvendor ratio
69

Optimal Order Quantity (Contd)

P(D Q*) = Cu / (Cu + Co)

Demand distribution

SL = Cu / (Cu + Co)
SL

Q*
70

Optimal Order Quantity


when Demand is Normally Distributed

If demand follows a normal distribution with mean


and standard deviation , the optimal order quantity
Q* can be found as follows:
1. Find zSL such that P z zSL SL
2. The optimal order quantity is

Q = + zSL
*
71

Single-Period Model Example 2

Our college basketball team is playing in a


tournament game this weekend. Based on
our past experience we sell on average 2,400
shirts with a standard deviation of 350. The
demand is estimated to follow a normal
distribution. We make $10 on every shirt we
sell at the game, but lose $5 on every shirt
not sold. How many shirts should we make
for the game?
72

Single-Period Model Example 2:


Optimal Order Quantity

Cu = $10
Co = $5
Optimal service level = Cu / (Cu + Co)
= 10 / (10 + 5) = .667
z.667 = .43
Therefore, we need 2,400 + .43(350) =
2,551 shirts

You might also like