IENG332 Production Planning I

Prepared by: Asst. Prof. Dr. Orhan KORHAN

Lecture15

2. Quantity Decisions (how much to order)

Quantity decision has a major impact on the level of inventory maintained, and it
directly influences inventory costs.
Models for quantity decisions are usually called lot sizing models. There are many of
them, and we group them under two headings:

Static lot sizing models are used for uniform (constant) demand over the
planning horizon
Dynamic lot sizing models are used for changing demand over the planning
horizon. We assume demand is known with certainty, which is sometimes
called lumpy demand.

2.1. Static Lot Sizing Models

A constant and uniform demand environment is not common in the real world.
However, it is a convenient starting point for developing inventory models. There are
four models:
Economic Order Quantity (EOQ)
Economic Production Quantity (EPQ)
Resource Constraints
Fixed Order Quantity
2.1.1. Economic Order Quantity (EOQ)
This model is the most fundamental of all inventory models and it is suitable for
retail.
We assume the following decision environment:
There is a single inventory system.
Demand is uniform and deterministic amounts to D units per unit time day,
week, month, or year. We will use annual demand, but any other unit of time
can be used.
No shortages are allowed.
There is no order lead time (time from ordering to receipt).
All the quantity ordered arrives at the same time, which is called infinite
replenishment rate.
Instantaneous receipt of all quantities ordered.
The model is suitable for raw material purchase in production or for a retail
environment. The decision variable for this model is Q (number of units to be
ordered).
Parameters:
c = unit cost ($/unit)
i = total annual inventory holding cost (% per year)
h = ic = total annual inventory handling cost (dollars per unit per year)
94
94

A = ordering cost ($/order)

D = demand per unit time
T = cycle length, the length of time between placement (or receipt) of replenishment
K(Q) = total average annual cost as a function of the lot size Q
It = on-hand inventory at time t (quantity of material actually in stock)

Objective of EOQ: to create a balance between two opposing costs (ordering costs
and holding costs)
Ordering cost is a fixed cost; the more we order, the less the cost per unit will be.
Holding cost is a variable cost that is lower the less inventory we have. This balance
is achieved through minimizing K(Q).
Sawtooth Model
It

Depletion Line
Slope -D

Q/2

Time t
T
Figure 1. EOQ inventory geometry
We assume inventory level is Q at time zero. As time moves, the inventory is depleted
at a rate of D units per year (i.e. the slope of the inventory line is D). When the
inventory level reaches zero, we order Q units. Because we assume that lead time is
zero and the replenishment rate is infinite, the inventory level will immediately rise to
Q, and the process will repeat.
T

Q
D

I = average inventory
I

Area under the inventory curve Inventory triangle area 1 QT

Q

T
T
T 2
2

The maximum inventory level is;

I max Q

For each cycle;

purchasing cost = ordering cost
Q
Q
hT = average inventory holding cost
2
2
The average cost per cycle;
icT

cQ A hT
K (Q)
Q

cQ

Q
2

K (Q) cD

AD

We wish to find the value of decision variable Q that minimizes K(Q). This is
achieved by solving the equation
K (Q)

dK (Q )
AD h
2 0
dQ
2
Q

Because the second derivate of K(Q) is positive, K(Q) is a convex function and
achieves its minimum at the point where the derivative is zero.
K (Q)

AD
3 0
2Q

alwaysis positive,
K(Q)
a convex function,
minimum is achieved at K (Q) 0

Q* is known as economic order quantity (EOQ).

2 AD
Q*
h
Minimum total average annual cost;
K (Q*) cD 2 ADh
Annual ordering (set-up) cost = AD
Q*
Q
Annual holding cost = h
2
Sensitivity of K(Q*) is measured by
K (Q )

1 Q* Q

K (Q*) 2 Q
Q *

K (Q )
K (Q*)
Q

1 ) will
Q*
cost less than an order smaller by the same amount
placing and order larger than Q* (i.e.

Example 1:
A small welding shop uses welding rods at a uniform rate. Marwin, the owner, orders
the rods from a local supplier. Marwin estimates the annual demand is about 1000
pounds. To place each order, he has to spend about $3.60 for the phone call and
paperwork. Marwin pays $2 per pound of rods, and holding costs are based on a 25
percent annual rate. Analyze the system.
Solution:
A = $3.60
D = 1000 punds per year
c = $2 per pound
i = 25% annually
h = 0.25 $2 = $0.50 per pound per year
Q* = EOQ =

2 AD
h

2($3.60)(1000)
0.5

120

It is best for Marwin to place and order for 120 pounds. He should issue an order
120
every T
0.12 year or 1.44 months.
1000
The total average annual cost is
K (Q*) cD 2 ADh ($2)(1000) 2(3.60)(1000)(0.5) $2060
The annual ordering cost is
1000
AD 3.60

$30
Q*
120

Example 2:
Suppose the welding rods are ordered in packages of 75 pounds each. How many
packages should Marwin order?
Solution:
Sensitivity:

Q* = 120
Q = 75, or Q = 150 (two packages)

1. Set Q = 75;
K ( 75)
K (120)

1 120

2 75

75

120

1.1125

2. Set Q = 150;
K (150) 1 120 150

1.025
K (120) 2 150 120
Select the smaller value between the alternatives
Therefore, Marwin will better off by ordering two packages each time.

2.1.1. Economic Production Quantity (EPQ) with Extensions

This extension of the EOQ model relaxes the assumption of infinite replenishment
rate. Instead, there is a finite replenishment rate, which is typical of a manufactured
item in which the lot is typical of a manufactured item in which the lot is delivered
overtime according to the production rate.
We also allow shortages to occur and be backlogged, assuming there is a maximum
level of backlog that management is willing to tolerate. Backlog occur in production
systems because of either lack of material, lack of capacity, or both.
= replenishment production rate, measured in the same units as the demand
Q = size of the production lot
A = set-up cost
c = unit production cost
Bt = shortage (backorder) level at time t
B = average shortage level
b = max Bt
It
Imax

Depletion

Replenishment

Slope (-D)

Imax+b

T1

Slope (-D)

T2

T3

T4

0
TP
-b

TD

Figure 2. Inventory geometry: EPQ with backlog

Q
D

Cycle time:

Time to produce Q units:

TP

TD

Time to deplete maximum inventory:

I max
D

D
I max Q 1 t

Time to recover from the backlog:

Time to generate Imax:

T2

Time to deplete Imax:

T3

T1

b
D

I max
D
I
g max
D

Time to generate backlog of b:

T4

b
D

To get the equation for K(Q, b), we need I (average of inventory) and B (average of
backlog)
I
I
1
T T
2T

I
I

max

max

2T D D
The total shortage cost per cycle is:

b TB

The total annual inventory holding cost is:

1
T

hTI hI

D
h Q 1 b

D
2Q 1

The average annual shortage cost:

1
T

b TB

bbD
Q

D
2Q 1

The total annual cost is:

K (Q, b) cD

AD

D
h Q 1 b

D
2Q 1

bD
Q

2b

D
2Q 1

K(Q,b) = purchasing + ordering + holding + fixed + backlog

cost
cost
cost
storage
To find Q* and b*, solve:
K
and
K
0
0
Q
b

(for 0 )

D h
2 AD

h
h
h 1 D

Q*

b*

hQ * D 1

Example 3:
SuperSauce produces a certain salad dressing. The demand for this dressing is about
400 pounds per month, and SuperSauce can manufacture it at the rate of 2000 pounds
per month. To initiate production, the machines have to be thoroughly checked and
cleaned, and it costs the company $120 per set-up. The cost to produce this dressing is
$3 a pound, and the inventory holding cost is estimated as 20 percent annually. If the
demand for this dressing exceeds the available inventory, it is backlogged.
Management estimates that a backlog accrues two types of cost loss of goodwill and
shortage penalty. The loss of goodwill is estimated to be $0.1 per pound short, and the
shortage penalty is estimated to be $1.2 per pound short per month. Analyze the
problem.
Solution:
A = $120 per set-up
D = 400/ month = 4800/ year
i = 20% annually
= 2000/ month = 24,000/ year
c = $3 per pound
h = 0.2 $3 = $0.6 per pound
= $0.1 per pound
= $1.2 per pound per month = $14.4 per pound per year

100
1001

Economic order quantity

D 2

2 AD
Q*

h 1 D h h

( 2)(120)(4800)

4800
0.6 1

24000

(0.1)(4800)2

0.6 14.4

(0.6)0.6 14.4

14.4

Q* = 1573 pounds/lot
The optimal backorder level is

hQ * D 1

b*

(0.6)(1573) (0.1)(4800) 1

D
h Q 1 b

AD

K (Q*, b*) cD

D
2Q 1

0.6 14.4

4800
24000

24.73 25

bD

b2

D
2Q 1

4800

1
(0.6)
25
(120)(4800)

24000

K
(1573,
25)
(3)(4800)

)
1573
4800

2(1573 1 24000

2
(0.1)(25)(4800)
(14.4)( 25)g

$15,136

24000